Food Traceability FRIA

FDA - Food Traceability FRIA Final NOV 2022.pdf

Establishment, Maintenance, and Availability of Records; Additional Traceability Records for Certain Foods

Food Traceability FRIA

OMB: 0910-0560

⚠️ Notice: This form may be outdated. More recent filings and information on OMB 0910-0560 can be found here:

2022-12-05 - Revision of a currently approved collection

Document [pdf]

Download: pdf | pdf

DEPARTMENT OF HEALTH AND HUMAN SERVICES
Food and Drug Administration

Requirements for Additional Traceability Records
for Certain Foods

Docket No. FDA- FDA-2014-N-0053

Final Regulatory Impact Analysis
Final Regulatory Flexibility Analysis
Unfunded Mandates Reform Act Analysis

Economics Staff
Office of Economics and Analysis
Office of Policy, Legislation, and International Affairs
Office of the Commissioner

Table of Contents

I. Introduction and Summary .......................................................................................................... 6
A. Introduction ............................................................................................................................ 6
B. Summary of Costs and Benefits ............................................................................................. 7
C. Terminology ......................................................................................................................... 14
D. Comments on the Preliminary Economic Analysis of Impacts and Our Responses ........... 15
Comment 1 (Underestimation of Costs) ................................................................................. 15
Comment 2 (Costs of Reading the Rule) ................................................................................ 16
Comment 3 (Capital Investment Costs) ................................................................................. 16
Comment 4 (Need for Costly New Systems) ........................................................................... 17
Comment 5 (Electronic Records) ........................................................................................... 19
Comment 6 (Training)............................................................................................................ 20
Comment 7 (General Recordkeeping).................................................................................... 21
Comment 8 (Recordkeeping per Lot) ..................................................................................... 21
Comment 9 (Number of Lots) ................................................................................................. 22
Comment 10 (Product Identifier) ........................................................................................... 22
Comment 11 (Effect on Food Prices, Availability, and Product Diversity) ........................... 23
Comment 12 (Costs to State and Local Jurisdictions) ........................................................... 24
Comment 13 (Identifying FTL Foods) ................................................................................... 25
Comment 14 (Additional Employees) .................................................................................... 25
Comment 15 (Benefits and Electronic Systems)..................................................................... 26
Comment 16 (Benefits and Epidemiology)............................................................................. 27
Comment 17 (Benefits Will be Marginal as Compared to Costs of the Rule) ........................ 27
Comment 18 (Benefits Estimates are Flawed) ....................................................................... 28
Comment 19 (Feedback Supporting Benefits Estimates) ....................................................... 28
Comment 20 (Traceback Time) .............................................................................................. 29
Comment 21 (Cost Savings) ................................................................................................... 29
Comment 22 (Baseline and COVID-19) ................................................................................ 29
Comment 23 (PTI and Current Traceability Practices) ........................................................ 30
Comment 24 (Coverage Underestimated).............................................................................. 31
Comment 25 (Spillover on Non-FTL Foods) ......................................................................... 32
Comment 26 (Effects on Global Supply of Seafood) .............................................................. 34

Comment 27 (Mung Bean Sprouts) ........................................................................................ 34
Comment 28 (Disproportionate Effects on Consumers and Small Businesses) ..................... 35
Comment 29 (Gasoline Sales) ................................................................................................ 36
Comment 30 (Small Business Exit) ........................................................................................ 37
Comment 31 (Input from Small Businesses) .......................................................................... 38
Comment 32 (Proposed Option 1 - Very small, Retail) ......................................................... 39
Comment 33 (Number of Foods by FTL Commodity) ............................................................ 39
Comment 34 (General) .......................................................................................................... 40

E. Summary of Changes ........................................................................................................... 41
1. General Changes to the Rule ............................................................................................. 41
2. Baseline Conditions: Coverage and Current Industry Practices....................................... 43
3. Benefits............................................................................................................................... 44
4. Costs................................................................................................................................... 45
5. International Effects........................................................................................................... 46
6. Distributional Effects ......................................................................................................... 46
7. Regulatory Flexibility Analysis .......................................................................................... 46

II. Final Economic Analysis of Impacts ....................................................................................... 47
A. Background .......................................................................................................................... 47
B. Potential Need for Federal Regulatory Action ..................................................................... 49
C. Purpose of the Rule .............................................................................................................. 56
D. Baseline Conditions ............................................................................................................. 57
1. Bioterrorism Act of 2002 and the 2004 BT Final Rule Recordkeeping Requirements ...... 58
2. Coverage of the Rule .......................................................................................................... 59
3. Current Industry Practices................................................................................................. 66

E. Benefits of the Rule .............................................................................................................. 70
1.

Public Health Benefits from Averted Illnesses ............................................................... 72

2. Benefits from Avoiding Overly Broad Recalls Following FDA Issued Public Health Advisories
............................................................................................................................................... 84
3. Other Benefits .................................................................................................................. 109

F. Costs of the Rule ................................................................................................................. 111
1. Main Assumptions of the Cost Analysis ........................................................................... 113
2. Costs of Reading and Understanding the Rule ................................................................ 115
3. Costs of Capital Investment ............................................................................................. 117
4. Costs of Training.............................................................................................................. 127
5. Costs of Recordkeeping.................................................................................................... 133

6. Non-Quantified Costs....................................................................................................... 162
7. Summary of Costs ............................................................................................................ 165

G. Distributional Effects ......................................................................................................... 166
H. International Effects ........................................................................................................... 171
I. Uncertainty and Sensitivity Analysis ............................................................................... 176
1. Coverage .......................................................................................................................... 176
2. Costs................................................................................................................................. 177
3. Benefits from Avoiding Overly Broad Recalls ................................................................. 179

J. Analysis of Regulatory Alternatives to the Rule ................................................................ 179
Alternative a. No Action ....................................................................................................... 180
Alternative b. Broader exemption for retail food establishments and restaurants .............. 181
Alternative c. Reduce compliance date to two years............................................................ 181
Alternative d. Extend compliance date to four years ........................................................... 182

III. Final Small Entity Analysis .................................................................................................. 183
A. Description and Number of Affected Small Entities ......................................................... 183
B. Description of the Potential Impacts of the Rule on Small Entities ................................... 189
C. Alternatives to Minimize the Burden on Small Entities .................................................... 193
IV. References............................................................................................................................. 194
V. Appendices ............................................................................................................................. 198
A. Food Traceability List (FTL) ............................................................................................. 198
B. Methodology Used to Estimate the Number of Illnesses ................................................... 202
C. Outbreak Case Studies Used in Estimation of Public Health Benefits .............................. 207
Definitions ............................................................................................................................ 207
Limitations ........................................................................................................................... 209

D. Estimation of the Number of Covered Entities .................................................................. 212
Produce Farms..................................................................................................................... 212
Shell Egg Farms Estimates .................................................................................................. 213
Aquaculture Farms Estimates .............................................................................................. 214
Manufacturers, Wholesalers, Warehouses, and Retailers ................................................... 215
Difference Between Proposed and Final RIA Estimates...................................................... 216

E. Changes to Cost Estimation from the Preliminary Analysis .............................................. 217
F. Case Studies Considered by Experts in Estimating Reduced Costs from Avoiding Overly
Broad Recalls Following an FDA Issued Public Health Advisory. ........................................ 221
G. Detailed Calculations Used for Estimating Benefits from Avoiding Overly Broad Recalls
Following an FDA Issued Public Health Advisory ................................................................ 226
4

H. Accounting for English Proficiency and Internet Access .................................................. 229

I. Introduction and Summary
A. Introduction
We have examined the impacts of the final rule under Executive Order 12866, Executive
Order 13563, the Regulatory Flexibility Act (5 U.S.C. 601-612), and the Unfunded Mandates
Reform Act of 1995 (Pub. L. 104-4). Executive Orders 12866 and 13563 direct us to assess all
costs and benefits of available regulatory alternatives and, when regulation is necessary, to select
regulatory approaches that maximize net benefits (including potential economic, environmental,
public health and safety, and other advantages; distributive impacts; and equity). The Office of
Information and Regulatory Affairs has designated this final rule as an economically significant
regulatory action as defined by Executive Order 12866.
The Regulatory Flexibility Act requires us to analyze regulatory options that would
minimize any significant impact of a rule on small entities. Because some small firms may incur
annualized costs that exceed one percent of their annual revenue, we find that the final rule will
have a significant economic impact on a substantial number of small entities.
The Unfunded Mandates Reform Act of 1995 (section 202(a)) requires us to prepare a
written statement, which includes an assessment of anticipated costs and benefits, before issuing
“any rule that includes any Federal mandate that may result in the expenditure by State, local,
and tribal governments, in the aggregate, or by the private sector, of $100,000,000 or more
(adjusted annually for inflation) in any one year.” The current threshold after adjustment for
inflation is $165 million, using the most current (2021) Implicit Price Deflator for the Gross
Domestic Product. This final rule would result in an expenditure in at least one year that meets
or exceeds this amount.

B. Summary of Costs and Benefits
This final rule will allow FDA and industry to more rapidly and effectively trace food
products that cause illnesses back through the food supply system to the source and forward to
recipients of the contaminated product. This rule will only apply to foods FDA has designated
for inclusion on the Food Traceability List1 and foods that contain listed foods as ingredients that
remain in the same form (e.g., fresh) in which they appear on the list. By allowing faster
identification of contaminated foods and increasing rates of successful tracing completions, the
rule results in public health benefits if foodborne illnesses directly related to those outbreaks are
averted. This might also lead to more efficient use of FDA and industry resources needed for
outbreak investigations by potentially resulting in more precise recalls and avoidance of overly
broad market withdrawals and advisories for covered foods.
The primary public health benefits of this rule are the value from the reduction of
foodborne illnesses and deaths because records required by the rule are likely to reduce the time
that a violative or contaminated covered food product is distributed in the market. Benefits from
this rule are generated if the following two conditions hold: (1) a foodborne outbreak occurs and
(2) the traceability records required by this rule help FDA to locate a commercially distributed
violative product quickly and accurately and to ensure it is removed from the market.
While the primary benefits from the rule are the value of the reduction of foodborne
illnesses and deaths, we also examine non-health related benefits. Non-health related benefits of
this rule will be from avoiding costs associated with conducting overly broad recalls and market
withdrawals that affect products that otherwise would not need to be withdrawn or recalled.

The list of applicable foods may be updated by publication of a notice in the Federal Register following
consideration of comments on proposed changes. See Appendix A for the list as of this writing.

Although recalls of rightly implicated foods come with necessary costs, overly broad recalls that
involve loosely related or unrelated products can make overall recalls unnecessarily costly. The
costs of a broad recall or market withdrawal include lost revenues from unimplicated products,
plus expenses associated with notifying retailers and consumers, collection, shipping, disposal,
inventory, and legal costs.2 There are no benefits from removing unimplicated products from the
market. Benefits from avoiding overly broad recalls may be realized only when recalls are
initiated in response to an FDA public health advisory.
It is possible, but not certain, that both of these categories of benefits could be
experienced to the extent quantified in this regulatory impact analysis. On the other hand, it is
also possible that a given instance of baseline contamination would lead to a very broad recall
(that could be narrowed by the final rule) or to illnesses (that could be avoided due to the final
rule) but not both.
Additional benefits of the rule may include increased food supply system efficiencies,
such as improvements in supply chain management and inventory control; more expedient
initiation and completion of recalls; avoidance of costs due to unnecessary preventive actions by
consumers; reduction of food waste; and other food supply system efficiencies due to a
standardized approach to traceability, including an increase in transparency and trust and
potential deterrence of fraud (Ref. [1, 2]).
This rule will impose compliance costs on covered entities by increasing the number of
records that are required for covered food products. Entities that manufacture, process, pack, or

2
For example, in an undifferentiated product recall, a single firm’s investment in traceability may be ineffective
when competitors and partners have not instituted a traceability system. This is problematic because, for example, in
the event of an undifferentiated leafy greens outbreak, issuing a broad recall could be unavoidable, at least until the
implicated product is identified and removed from the market. In situations where the recalled products are insured,
targeted recalls will help prevent unnecessary recalls of insured products which may have long term consequence to
retailers from increases in their insurance rates due to imprecise recalls.

hold covered foods will incur costs to establish and maintain a traceability plan and traceability
records. Some firms may also incur initial and recurring capital investment and training costs for
systems that will enable them to keep, maintain, and make available to other supply chain
entities (and to us upon our request) their traceability records. Moreover, firms will incur onetime costs of reading and understanding the rule.3
Table 1a and Table 1b summarize the costs and benefits of the final rule. At a seven
percent discount rate, 20-year annualized costs range from about $63 million to $2.3 billion, with
a primary estimate of $570 million per year. At a three percent discount rate, annualized costs
range from about $53 million to $2.3 billion, with a primary estimate of $551 million per year.
The present value of costs with seven percent discounting over 20 years (not shown in Table 1a)
ranges from about $0.7 billion to $24.6 billion, with a primary estimate of about $6 billion. The
present value of costs with three percent discounting over 20 years (not shown in Table 1a)
ranges from about $0.8 billion to $33.7 billion, with a primary estimate of $8.2 billion.
In section II.E.1 we estimate public health benefits using several case studies of outbreak
tracebacks for four pathogens associated with illnesses caused by covered foods.4 We calculate
these benefits based on an estimated 83 percent reduction of traceback time resulting from the
requirements of this rule. At a seven percent discount rate over twenty years, the annualized
monetized health benefits of the rule range from $59 million to $2.2 billion with a primary

The information flows brought about by the rule may also prompt new protective actions — for example, in
farming, manufacturing, or cooking processes — that could also have costs. We have not quantified these potential
costs, but they would likely correlate with the realization of health and longevity benefits of this rule.
4
This approach has a tendency toward underestimation of the total public health benefits because these four
pathogens do not represent the total burden of all FTL-associated illnesses. However, adjustments made for
undiagnosed and unattributed illnesses may have the opposite tendency of overstating both FTL-associated illnesses
and benefits. We cannot scale up to 100% because our estimates of the percentage of illnesses potentially avoided
with improved traceability depend on data specific to each pathogen. We describe our methods in detail in section
II.E.1 Public Health Benefits from Averted Illnesses. In short, these four pathogens may account for roughly 95%
of the total dollar value of the illnesses for which traceability might be an effective preventive measure.

estimate of $780 million (Table 1a).5 At a three percent discount rate over twenty years, the
annualized monetized health benefits range from $61 million to $2.3 billion with a primary
estimate of $810 million. The present value of health benefits with seven percent discounting
over 20 years (not shown in Table 1a) ranges from about $0.6 billion to $23.7 billion, with a
primary estimate of $8.3 billion. The present value of health benefits with three percent
discounting over 20 years (not shown in Table 1a) ranges from about $0.9 billion to $34.5
billion, with a primary estimate of $12.0 billion.
In section II.E.2 we estimate (non-health) benefits from avoiding overly broad recalls and
market withdrawals. At a seven percent discount rate over twenty years, these annualized
monetized benefits range from $233 million to $1.8 billion with a primary estimate of $575
million (Table 1a). At a three percent discount rate over twenty years, these annualized
monetized benefits range from $242 million to $1.8 billion with a primary estimate of $596
million. The present value of benefits from avoiding overly broad recalls with seven percent
discounting over 20 years (not shown in Table 1a) ranges from about $2.5 billion to $18.8
billion, with a primary estimate of $6.1 billion. The present value of these benefits with three
percent discounting over 20 years (not shown in Table 1a) ranges from about $3.6 billion to
$27.3 billion, with a primary estimate of $8.9 billion.

Table 1a. Summary of Benefits, Costs and Distributional Effects of Final Rule ($Millions)
Category

Benefits

Annualized
Monetized
Millions$/year

Primary
Estimate

Low
Estimate

High
Estimate

$780

$59

$2,238

$810

$61

$2,322

Units
Year
Dollars
2020
2020

Notes

Discount
Rate
7%

Period
Covered
20 years

20 years

Monetized
health
benefits from
an estimated

We examined multiple case studies of tracing success rates for outbreaks. As explained in detail in section
II.E.1.iii, Table 7, and Appendix C, our estimated percentage range of illnesses prevented vary widely.

Category

Primary
Estimate

Low
Estimate

High
Estimate

Units
Year
Dollars

Discount
Rate

Notes
Period
Covered
domestic
consumers.

Annualized
Quantified
Qualitative

Transfers

Effects

Costs of
farming-,
manufacturin
g- or
cookingrelated
actions that,
as a result of
new
information
flows,
address risks
of foodborne
illness.

Federal
Annualized
Monetized
Millions$/year
From/To
From:
To:
Other
Annualized
Monetized
Millions$/year
From/To
From:
To:
State, Local or Tribal Government: No significant effect.
Small Business: Potential impact on small entities that are currently not keeping traceability records
described by the rule.
Wages: N/A
Growth: N/A

Table 1b explores the possibility that baseline costs—of recalls and possibly also FTLassociated foodborne illnesses—are already internalized by market actors. If so, then ruleinduced costs would form an upper bound on rule-induced benefits. Especially in the case of
recall costs, the same entities experiencing baseline costs (or entities with whom they have
business contracts) would incur the costs of the rule. As shown in column (b) in Table 1b, if
these costs are fully internalized in the baseline and if the narrowed-recall benefits estimates in
RIA section II.E.2 are plausible, they would form a lower bound on the cost of the final rule;
alternatively, if the rule-induced cost estimates in sections II.F and II.H are plausible, they would
form an upper bound on the narrowed-recall category of benefits.
12

Table 1b. Summary of Rule-Induced Benefits and Costs, as a Function of Baseline Cost
Internalization *
(a)
Neither adverse health effects nor
recall-associated costs fully
internalized in market transactions
for FTL foods

RIA Section II.E.1

Health Benefits: $780M
(range: $59M to $2.2B)

(b)
Recall-associated costs, but not adverse
health effects, fully internalized in market
transactions for FTL foods

Health Benefits: $780M
(range: $59M to $2.2B)

and/or
Recall-Associated Benefits: $575M
(range: $233M to $1.8B)
RIA Section II.E.2

Recall-Associated Benefits: $575M
(range: $233M to $1.8B)

Direct Compliance Costs > $575M
(range: $233M to $1.8B)
Protective Action Costs (potential): not
quantified
and/or

Protective Action Costs (potential):
not quantified

RIA Sections
II.F and II.H

Direct Compliance Costs (if foreign
passed through to U.S. supply chain
& consumers): $620M
(range: $67M to $2.6B)
Direct Compliance Costs (if foreign
not passed through to U.S. supply
chain & consumers): $570M
(range: $63M to $2.3B)

Recall-Associated Benefits < Costs
Direct Compliance Costs (if foreign passed
through to U.S. supply chain & consumers):
$620M
(range: $67M to $2.6B)
Direct Compliance Costs (if foreign not
passed through to U.S. supply chain &
consumers): $570M
(range: $63M to $2.3B)
Protective Action Costs (potential): not
quantified

* Primary estimates presented in this table are calculated with a 7 percent discount rate; primary estimates
discounted at 3 percent differ only slightly. All estimates are expressed in 2020 dollars and annualized over 20
years. Abbreviations: M=million, B=billion.

C. Terminology
In Table 2, we describe the key terms we use in this document. We note that these
definitions only apply to this document.
Table 2. Key Terms in the Regulatory Impact Analysis
Term
Description
BT Act
Bioterrorism Act of 2002. We use Subpart J (of 21 CFR part 1)
and BT Act interchangeably.
BT rule
Establishment and Maintenance of Records Under the Public
Health Security and Bioterrorism Preparedness and Response
Act of 2002 final rule (2004)
CDC
The Centers for Disease Control and Prevention
CORE
FDA’s Coordinated Outbreak Response and Evaluation Network
CTE
Critical tracking event
Establishment,
We use these terms interchangeably. Each firm may operate one
facility
or more establishments.
FD&C Act
Federal Food, Drug, and Cosmetic Act
FSMA
FDA Food Safety Modernization Act of 2011
FTL
Food Traceability List
FTL foods, FTL
Foods listed on the FTL and foods that contain listed foods as
products, covered
ingredients, provided that the listed food that is used as an
foods
ingredient remains in the same form (e.g., fresh) in which it
appears on the list.
FTE
Full-time-equivalent employee
KDE
Key data element
NAICS
North American Industry Classification System
O&M
Operation and Maintenance
Persons, entities
We use these terms interchangeably to refer to businesses
covered by the rule
PRIA
Preliminary Regulatory Impact Analysis of the proposed rule
RIA, FRIA
Regulatory Impact Analysis of the final rule
Small businesses,
In this RIA except in section II.E.2, but not elsewhere in the
small entities
docket for this rule, we use these terms to refer to small
businesses as defined by the Small Business Administration6
UPC
Universal Product Code
USDA
The U.S. Department of Agriculture
We, our, us, FDA,
We use these terms to refer to the U.S. Food and Drug
the Agency
Administration

https://www.sba.gov/document/support-table-size-standards

D. Comments on the Preliminary Economic Analysis of Impacts and Our Responses
On September 23, 2020, we published the proposed rule “Requirements for Additional
Traceability Records for Certain Foods” (85 FR 59984). Accompanying the proposed rule was a
preliminary regulatory impact analysis document on which we requested public comments (Ref.
[3]). We received many comments, including a large number of comments on the estimation of
costs. We organize these comments and our responses by topic in the paragraphs below. The
number assigned to each comment is purely for organizational purposes and does not signify the
comment’s value, importance, or the order in which it was received.
Comment 1 (Underestimation of Costs)
Many comments stated that costs are substantially underestimated. Some comments
elaborated on specific types of costs, including the time to learn the rule, capital investments,
training, and recordkeeping. Others noted that correcting this underestimation would cause the
costs of the rule to outweigh its benefits.
Response: After reviewing these comments, FDA determined a need to obtain additional
data for cost estimates and to revise requirements of the rule to reduce their burden. FDA
contracted with Eastern Research Group (ERG) to research additional literature and elicit
information from a panel of industry experts to further inform the costs of the rule to various
covered entities based on their baseline traceability practices (Ref. [4]). Experts based their input
on the rule as proposed (with additional brief definitions of some new CTEs in their draft-final
state at the time of the elicitation). In addition, we updated our estimates for the number of
covered entities. Revised cost estimates consistent with revisions to the rule are explained in
detail in section II.F of this analysis. We discuss changes to cost estimation from the Preliminary

Regulatory Impact Analysis (PRIA) in appendix E of this analysis. Revised coverage estimates
are explained in detail in section II.D.2 and appendix D.

Comment 2 (Costs of Reading the Rule)
Many comments stated that the time to read and understand the rule is substantially
underestimated. Due to the rule’s complexity and detailed requirements, comments stated that
reading and understanding the rule would require more than one employee per covered entity and
significantly more time than assumed in the preliminary economic analysis. Relatedly, several
commenters stated that they had already incurred labor costs to read and understand the proposed
rule (in addition to the time they will need to spend when the final rule publishes).
Response: In estimating the time to read and understand the rule, we have used methods
consistent with previous FDA analyses of the economic impacts of rulemakings. In this final
analysis, we have accounted for multiple employees reading the rule at larger companies. Our
estimate is an average over all firms, and now includes an assumption that in small firms one
employee will read the rule and in large firms three employees will read the rule.
Note also that we consider reading costs alone in the section II.F.2 “Reading and
Understanding the Rule” to be separate from the costs to identify FTL products and plan for
compliance, which we estimate below in section II.F.5.a “Traceability Plan.”

Comment 3 (Capital Investment Costs)
Many comments stated that the capital investment costs required to comply with the rule
are substantially underestimated. Several comments proposed higher estimates of capital
investment costs particularly for small businesses, for example $45,000. Comments also

challenged the PRIA’s estimates of capital investment costs on the basis that FDA did not
consult with small businesses in forming those estimates.
Response: After reviewing received comments, FDA sought additional information on
existing industry practices to improve our capital investment cost estimates. FDA contracted
with ERG to research additional literature and elicit information from a panel of industry experts
to further inform our estimates of capital investment costs faced by covered entities of various
sizes based on their baseline traceability practices (Ref. [4]). The estimates of capital investment
costs in this final Regulatory Impact Analysis (RIA) consider both one-time investments and
recurring operating and maintenance costs to affected businesses across several broad industry
categories. Revised capital investment cost estimates are explained in detail in section II.F.3 of
this analysis.

Comment 4 (Need for Costly New Systems)
Several comments stated that the rule would require businesses to adopt new practices
and systems to identify and track traceability lot codes upon receipt and shipment of FTL foods.
Comments suggested that such systems could include processes to maintain consistency of
records, methods, and storage, procedures for internal verification of records, and a measurable
and consistent recall process. Comments suggested changes might involve new technology,
operations, and management. Moreover, some commenters stated that businesses would be
forced to implement these new systems for all foods, not just foods on the FTL, because it would
not be practical to maintain two separate recordkeeping systems.
Response: Due to commenters’ concerns that the proposed rule would impose costly
drastic changes to existing practices and systems, FDA revised the requirements in this final rule,

including removing requirements for certain data elements not typically captured or
communicated between supply chain entities to better align with existing best business practices.
Such changes concerned requirements that did little to enhance traceability (especially in the
context of other requirements) but would be burdensome to industry. For example, FDA
removed requirements to record the time of receipt and the name of the transporter of received
food, and, for imports, the entry number. FDA also removed the requirement to generate, send,
and record unique location and product “identifiers.” These are not always part of existing
practices and FDA did not consider them to contribute enough to public health to warrant
requiring their introduction, collection, and sharing. Additionally, FDA removed the requirement
to generate traceability lot codes when growing foods and simplified the transmission of
traceability lot code source information when sending and receiving.
To gain more insight into industry’s possible adoption of new practices and systems in
response to the rule, FDA contracted with ERG to elicit input from an external panel of industry
experts. We have incorporated their input in section II.F.5.a “Traceability Plan,” in which we
estimate the costs of planning new procedures to comply with the final rule. Experts expressed
mixed expectations on whether and to what extent businesses would conform recordkeeping of
non-FTL foods to the requirements for FTL foods (Ref. [4]). We expect that it will be possible
for businesses to implement changes on an as-needed basis for compliance purposes, though
some might voluntarily opt to enhance traceability more broadly. In section II.F.5 on “Costs of
Recordkeeping,” we therefore estimate recordkeeping costs based on entities’ volume handled of
traceability lots specifically of foods on the FTL.

Comment 5 (Electronic Records)
Though the rule does not require use of electronic records, some comments stated that it
creates a de facto requirement of electronic recordkeeping due to the number of attributes needed
for each record, the need to send records downstream, and the 24-hour response time for
providing a sortable spreadsheet to FDA when requested. Comments stated that electronic
recordkeeping would entail significant financial cost to small businesses who currently keep
paper records, including costs for data storage and management, as well as costs to acquire
equipment necessary for generating records. A comment stated that the 24-hour response time
for providing FDA with a sortable spreadsheet would necessitate maintaining electronic records
in the course of business, and that, in switching from paper-based recordkeeping to electronic
recordkeeping, the commenter’s business incurred $10,000 in upfront costs and $2,500 in annual
costs. Relatedly, comments stated that Amish-owned businesses do not use electrical devices,
and consequently may face particular difficulty complying with the requirement to produce a
sortable spreadsheet within 24 hours.
Response: The final rule does not require electronic recordkeeping. Firstly, the final rule
simplifies the attributes needed for each record to align them more closely with data elements
already captured and communicated in standard business practices. Although FDA encourages
the use of electronic recordkeeping for traceability, persons subject to the rule may keep their
records in paper or electronic form. In response to comments on the proposed rule, the final rule
also expands the exemption from producing an electronic sortable spreadsheet to farms with less
than $250,000 in annual sales and all other businesses with less than $1 million in annual sales.
Finally, we note that the final rule, like the proposed rule, states that FDA will withdraw a
request for an electronic sortable spreadsheet to accommodate a religious belief (see §

1.1455(c)(3)(iv)), and includes provisions under which persons may request a waiver of subpart
S requirements (see §§ 1.1405-1.1450 of the rule) or an exemption from (or modification of) the
requirements (see §§ 1.1360-1.1400).

Comment 6 (Training)
Many comments said that the training required to comply with the rule is substantially
underestimated. Comments asserted that training was likely to apply to all employees, rather than
a limited number of employees as assumed in the preliminary economic analysis, and that it
would be extensive and ongoing, instead of one-time. Some comments stated that training is
likely to vary depending on job role. One comment suggested that annual training on the
requirements of the proposed rule would take five hours of each employee’s time but did not cite
any associated references or data to support this estimate. One comment, summarizing feedback
from retailers, estimated training costs would range from $15,000 to nearly $3 million, but did
not cite any associated references or data.
Response: In the PRIA, we assumed that training would be a one-time cost to train only a
limited number of current employees on the new requirements and traceability practices. We also
assumed that, for training new employees, some outdated training content will be replaced with
training related to this rule. We note that commenters did not provide additional data in support
of alternative estimates. However, after reviewing public comments on our estimates of training
costs, FDA determined a need for and sought additional data and information to improve our
estimates. FDA contracted with consultants to survey a panel of external industry experts to
further inform training costs to various covered entities based on their size and baseline industry
practices (Ref. [4]). In this final analysis, we estimated the number of trainees for entities of

different sizes across different industry sectors based on input by the expert panel. Revised
training cost estimates are explained in detail in section II. F.4. of this analysis.

Comment 7 (General Recordkeeping)
Many comments said that the time to establish and maintain general records and/or other
records to comply with the rule is substantially underestimated.
Response: FDA contracted with ERG to research additional literature and survey a panel
of external experts to better inform the costs of the rule to various covered entities based on their
baseline traceability practices (Ref. [4]). In estimating recordkeeping time for general records in
this final RIA, we used the results of this expert elicitation to update our estimates of the burden
per traceability lot for each critical tracking event (CTE) for affected businesses across several
broad industry categories. Revised cost estimates of recordkeeping are explained in detail in
section II. F.5. of this analysis.

Comment 8 (Recordkeeping per Lot)
Several commenters expressed that the PRIA underestimated the costs of recordkeeping
per lot. Comments stated that the time spent breaking a pallet or shipment down into lots for data
entry would increase the time needed to process each lot. One commenter also stated that
capturing and sending information should be treated as distinct activities. Some commenters
estimated that recordkeeping costs would be at least $1 per case.
Response: After reviewing public comments, FDA revised the requirements of this final
rule to better reflect current industry practices. We have updated our estimates of recordkeeping
burden per traceability lot, accounting both for changes between the proposed and final rules and

input from industry experts external to FDA. Additionally, unlike the PRIA, this RIA treats
capturing and sending information as distinct activities. We used the results of the expert
elicitation regarding the time needed to capture and send the relevant data for each CTE (Ref.
[4]). Revised cost estimates of recordkeeping are explained in detail in section II.F.5. of this
analysis.

Comment 9 (Number of Lots)
Comments stated that we underestimated the number of covered lots per entity. In
particular, commenters stated that warehouses and distribution centers receive more than the
primary estimate of 1,000 lots of FTL foods.
Response: We thank the commenters for raising this concern. We agree and for this Final
Regulatory Impact Analysis (FRIA) we contracted with ERG to survey a panel of external
industry experts to further inform the number of traceability lots handled by various covered
entities based on their size and role in the supply chain (Ref. [4]). Estimates of recordkeeping
costs, now accounting for revised numbers of FTL lots, are explained in detail in section II.F.5.
of this analysis.

Comment 10 (Product Identifier)
A comment stated that obtaining a product identifier, one of the proposed KDEs, imposes
minimal costs on both small and large entities. The comment notes that a single global trade item
number has a one-time cost of $30.
Response: We appreciate public input on the cost of product identifiers. The final rule,
unlike the proposed rule, no longer includes the product identifier as a KDE.

Comment 11 (Effect on Food Prices, Availability, and Product Diversity)
Many commenters stated that the rule would increase prices of FTL foods because
producers will pass their compliance costs on to consumers. Some commenters suggested that
the increase in price could reduce food availability. Relatedly, commenters expressed concern
that the rule would discourage product diversification in pursuit of lower costs. Among these
comments, some stated that the rule would impose higher costs on growers with diverse products
relative to growers with one or few products, irrespective of farm size. Also, commenters
suggested that costs to food hubs (which collect products from multiple farms and sell them to
consumers) in their capacity as first receivers were not adequately considered and would be high
due to their product diversity. Some of these comments claimed that the rule would thus reduce
or eliminate consumer access to locally grown produce, namely by rendering local food hubs
unprofitable, and that this might lower community resilience to events like COVID that interrupt
longer supply chains. Comments also suggested that third party logistics providers, importers,
and distributors may opt not to handle foods on the FTL, thereby preventing small specialty food
makers from reaching retail markets.
Response: We agree that producers might pass some of their compliance costs on to
consumers through higher prices. The RIA attempts to represent the total costs of compliance
consistent with the rule to industry and society as a whole. Section II.F of the RIA estimates
compliance costs to various covered domestic entities depending on their size and role in the
supply chain and section II.H discusses costs to foreign entities. However, we do not determine
the exact incidence of those costs, which might be passed on to other entities in the supply chain.
We do not think that the rule will cause food and ingredient prices to rise substantially, although

depending on entities’ market power some costs of the rule might be passed all the way to
consumers and retail buyers.
We agree it is possible that some producers may cease to offer some products. We expect
that this would occur when the additional traceability requirements cause that product to become
unprofitable — that is to say that the baseline costs of producing that product, together with the
incremental cost of traceability, exceed consumer willingness to pay for the product. We do not
predict individual product discontinuation, which would require detailed knowledge of markets
for an unknown number of products.
We note that the final rule shifts traceability lot code assignment from growing to initial
packing. Furthermore, the final rule states that instead of maintaining records of the growing area
coordinates for each traceability lot of food grown (as proposed), growers will only need to
maintain a farm map with field names, which are assigned once only and do not update based on
what is grown. Additionally, the final rule replaces the requirements of the first receiver critical
tracking event (CTE) with requirements for an initial packing CTE, and therefore food hubs will
not be first receivers. We updated our estimates of these burdens in this RIA accordingly.

Comment 12 (Costs to State and Local Jurisdictions)
Several comments suggested the rule would increase burdens on state and local
jurisdictions, particularly with respect to monitoring and enforcement.
Response: While we anticipate having states continue to do inspections under contract
with FDA (especially for farms), we also expect that monitoring and enforcement of the new
traceability requirements would occur during FDA inspections and following tracebacks of
foodborne illness outbreaks by FDA’s Coordinated Outbreak Response and Evaluation (CORE)

investigators. As such, CORE would discover non-compliance and follow up as needed. We do
not yet know whether or to what extent states will be expected to engage in education and
outreach.

Comment 13 (Identifying FTL Foods)
Commenters stated that the preliminary analysis did not sufficiently account for the cost
of time spent identifying which of foods that are subject to the rule. In particular, a comment
states that the rule would require foodservice distributors to determine the ingredients in each
product they receive on a shipment-by-shipment basis. The comment notes that distributors do
not currently collect such information, which may also change frequently when suppliers change
or substitute ingredients.
Response: We thank the commenters for raising this issue and agree that the analysis
should include the cost to identify FTL foods. In section II.F of this final RIA, we estimate the
cost to covered businesses to identify products containing FTL foods. To update our estimates,
we used information elicited from a panel of external industry experts (Ref. [4]). We note also
that the rule requires producers who ship FTL foods, to a distributor or other supply chain entity,
to provide traceability information to the recipient, which might help in identifying FTL foods.

Comment 14 (Additional Employees)
Some comments said that the rule represents substantial changes from current practice
and would cause businesses to hire additional full-time employees, create new job positions in
sectors where workers are not equipped for administrative duties, or assemble teams dedicated

solely to recordkeeping requirements. Among these comments, some stated that small businesses
in particular would need to hire additional staff to perform traceability-related duties.
Response: We thank the commenters for their input. We do not believe in general that the
final rule, as revised from the proposed rule, will necessitate hiring additional employees
dedicated to compliance. To further inform our estimates of impacts of the rule on various
covered entities, including small businesses, based on their baseline traceability practices, FDA
contracted with consultants to research additional literature and survey a panel of external
industry experts (Ref. [4]). This final RIA uses the elicited information on baseline prevalence of
traceability recordkeeping among businesses by size and the steps that businesses of different
sizes will take to establish traceability procedures compliant with the rule’s requirements,
including expected amounts of employee labor required. We explain cost estimates in further
detail in section II.F of this RIA.

Comment 15 (Benefits and Electronic Systems)
Comments also suggested that not requiring all firms to submit electronic records may
undermine benefits.
Response: We thank the commenters for their concerns but disagree that estimated
benefits depend on electronic recordkeeping. Although we encourage the use of electronic
records and communications for traceability, effective traceability under the final rule does not
require electronic recordkeeping or any specific technologies for records maintenance or supply
chain communications.

Comment 16 (Benefits and Epidemiology)
A comment asserted that benefits are overstated because FDA did not consider
epidemiological complexities associated with the outbreak.
Response: We thank the commenter for their concerns but disagree with their assessment
of our benefits estimates. Our estimates, which considered the outbreak epidemiological curve
and incorporated feedback from FDA epidemiologists, use assumptions that account for
complexities associated with disease outbreaks and investigations.

Comment 17 (Benefits Will be Marginal as Compared to Costs of the Rule)
A comment claimed that the added costs of this rule will offer only marginal public
health benefits since foodservice distributors already have a demonstrated record of being able to
quickly and effectively conduct recalls and tracing activities. Another comment noted that given
the complexity of the proposed rule, the benefits of the proposed rule would be very limited for
the baking industry which already has a track record of conducting timely trace-back and traceforward activities using their current recordkeeping systems.
Response: We note that FDA’s efforts to ensure food safety are largely incremental. The
new requirements complement Subpart J that put in place ‘one-up, one-back’ tracing
requirements. This rule will enable FDA and the food industry to more quickly and efficiently
trace covered foods to the sources of an outbreak. In updating our estimates, we took into
consideration existing industry practices and baseline compliance, as discussed in section II.D
and throughout this analysis.

Comment 18 (Benefits Estimates are Flawed)
One comment claimed that the benefits estimates were flawed. The commentor suggested
that FDA’s assessment of the benefits assumed that the rule would quickly reduce the impact of
illness outbreaks and prevent overly broad recalls, when in most foodborne illness outbreaks,
FDA has prevented overly broad recalls without this rule.
Response: We thank the commenter for their concerns but disagree with their assessment
of our benefits estimates. The analysis is informed by historical data and incorporates FDA’s
extensive experience of conducting outbreak investigations and issuing recalls. We discuss
updated benefits of this rule both qualitatively and quantitatively in section II.E, including public
health benefits, benefits from avoiding overly broad recalls, and efficiency savings to FDA.

Comment 19 (Feedback Supporting Benefits Estimates)
Some comments affirmed the benefits estimates in the PRIA and stated that the health
benefits resulting from the rule would outweigh the costs of implementing and enforcing the
rule.
Response: We thank the commenters for their feedback on our benefits estimates. We
estimate that implementation of this rule will have significant public health benefits and help
streamline more targeted removal of implicated covered foods from the market during foodborne
illness outbreaks.

Comment 20 (Traceback Time)
Several comments challenged our estimate of an 84 percent reduction in traceback time,
which informed the estimated benefits in the PRIA. Commenters suggested that we
overestimated this improvement in traceback time and thereby overestimated benefits.
Response: We thank the commenters for their concerns but believe our estimate of the
expected reduction in FDA’s traceback time is realistic and not overstated. We previously based
our 84 percent estimate on traceback data from FDA CORE. We have since consulted with field
experts and received additional traceback data from FDA CORE. Using the additional data, we
have updated the estimated traceback time reduction to 83 percent. Although the estimate has not
changed significantly, we nevertheless revised our benefits estimates to reflect the slightly lower
percent improvement.

Comment 21 (Cost Savings)
One comment provided a case study and supporting data demonstrating the effectiveness
of their food traceability solutions, which resulted in cost savings from reduced labor hours to
identify contaminated foods.
Response: We thank the commenter for providing this detailed example and agree that in
the long run this rule might lead to private cost savings to some businesses. We acknowledge
that not all entities will likely experience private cost savings.

Comment 22 (Baseline and COVID-19)
Some comments suggested that practices implemented in response to COVID-19 would
require changing the baseline estimates of the rule. Relatedly, comments stated that this rule will

lower resilience to COVID-19 by inhibiting growth and innovation between farmers and food
hubs. In particular, commenters expressed concern that the rule will reduce product variety and
availability already impacted by the pandemic. Comments also suggested that, in light of the low
risk of foodborne illnesses posed by small farms and retailers, this rule would unduly burden a
large number of businesses already adversely affected by the COVID-19 pandemic.
Response: We agree with the comment that our baseline estimates may not accurately
reflect a baseline during or immediately following the COVID-19 pandemic. However, as we
anticipate that pandemic restrictions and the circumstances they created will be temporary, and
since the rule will not take effect until more than three years after publication, the period before
COVID-19 informs the baseline for this analysis.
We expect the greater proportion of uncertainty in our analysis to concern knowledge of
typical traceability practices that were in place before COVID-19. In order to reduce some
uncertainty around our estimates, FDA contracted with ERG to research additional literature and
interview a panel of industry experts to further inform the costs of the rule to various covered
entities based on their ordinary baseline traceability practices (Ref. [4]). In addition, FDA has
revised the rule to provide additional exemptions to many small entities. We explain our revised
baseline in detail in section II.D of this analysis.

Comment 23 (PTI and Current Traceability Practices)
Several comments addressed the rule’s implications for existing baseline practices.
Comments stated that new traceability requirements should recognize, support, and align with
voluntary (already existing) efforts, which have also shown to be effective. One comment stated
that more than 50% of respondents of the 2020 Leafy Greens Marketing Agreement (LGMA)

survey indicated that they are utilizing the Produce Traceability Initiative (PTI) for labeling and
traceback. A commenter stated that results of a 2020 LGMA survey showed that produce
growers are capable of quickly tracking product in 2 hours or less, regardless of whether they
had a paper-based or electronic system. Some comments mentioned that while technologies exist
to meet new traceability requirements, these technologies are inoperable because the data is not
standardized, normalized, and harmonized (suggesting that many have already made capital
investment in technology). Other baseline related comments provided background information
describing current traceability practices pertinent to issues related to interoperability,
recordkeeping, capital investment and industry specific practices such as seafood producers,
farmers, and growers.
Response: To the extent that PTI practices overlap with FDA traceability requirements,
any incremental costs incurred by those who have implemented a similar traceability program
would be less than if they had no traceability program at all. We adjust our baseline (and
therefore costs) estimates to reflect the costs to the estimated proportion of entities that have
instituted similar traceability requirements to those in this rule. Our updated estimates are
addressed in sections II.D and II.F of this document.

Comment 24 (Coverage Underestimated)
Comments claimed that the numbers of farms and small retailers affected by the rule are
underestimated.
Response: We thank the commenters for raising this concern. After reviewing comments,
FDA determined a need for additional data to improve coverage estimates. FDA revised the
numbers of covered entities affected by the rule by using newer data from the 2017 SUSB and

2017 North American Product Classification System (NAPCS) data from the U.S. Census, and
the 2017 USDA National Agricultural Statistics Service (NASS) data. We explain revised
coverage estimates in detail in section I.E, section II.D.3, and appendix D of this analysis.

Comment 25 (Spillover on Non-FTL Foods)
Some comments stated that the rule will likely affect all foods and that FDA thus
underestimated the effect of the rule on covered entities. One comment said that although the
PRIA’s assessment mirrors the scope of the proposed rule, in practice the new recordkeeping
requirements will likely affect entities handling all foods. Covered entities will be required to
revise their recordkeeping systems to comply with the rule, and it would be more time- and
energy-intensive to maintain two sets of recordkeeping systems (one for covered foods and one
for non-covered foods), than to apply the recordkeeping system necessary for compliance with
the rule to all foods. The commenter argued that since covered entities will expand their
recordkeeping systems to all foods they handle, they will in turn require their suppliers to adopt
similar practices, whether those suppliers handle covered foods or not.
Response: We thank the commenters for raising these concerns but do not believe that
these issues impact our estimated costs to covered entities or that we have underestimated the
number of entities affected by the rule. Concerning firms who handle both covered and noncovered foods, we do not believe the decision on their part to keep KDEs for non-covered foods
would affect our estimates. In the first place, our accounting of new equipment, software,
services, training, and procedures—which we grant might necessarily displace existing such
systems rather than operate in parallel with them—considers these to be fixed costs with respect
to the number of foods handled. Second, we estimate the variable costs of recordkeeping as

labor, and we do not believe in general that requiring an employee to perform an action for
certain foods creates a need to perform that action for all other foods. For a firm to decide to
perform new traceability steps for all foods, doing so must cost less than distinguishing between
FTL and non-FTL foods and subsequently performing new traceability steps for just FTL foods.
Because the FTL is limited and additions take two years to become effective, and because in
practice FTL foods will also come shipped with traceability KDEs such as the Traceability Lot
Code, we do not find it plausible that this will hold generally as a consequence of the rule. We
would thus not attribute to the rule the additional labor cost of performing traceability
recordkeeping on all other foods.
Concerning the possibility that firms who do not handle covered foods might nonetheless
adopt certain traceability practices of business partners who do, we would not generally attribute
such behavior to the rule. When certain practices prove optimal on business grounds, or when
large firms—including those not subject to the rule—exert influence over supplier practices via
market power, practices might converge over time for reasons other than regulatory compliance.
Moreover, as documented in the product tracing pilots, firms with widely varying traceability
practices already conduct business with each other while serving the traceability demands of
downstream customers and industry initiatives (Ref. [5]). Since the rule does not prescribe
specific technologies for records maintenance, and since KDEs mostly consist of information
already commonly communicated between business partners, we expect supply chains to
continue to accommodate widely varying traceability practices.

Comment 26 (Effects on Global Supply of Seafood)
One comment said that because seafood is globally sourced, the rule will have a major
impact on U.S. trading partners. The commenter stated that the seafood “originator” or even the
“first receiver” often does not know the destination of the finished products and that the
regulation will therefore impose a recordkeeping burden on companies with respect to seafood
products that will never enter the United States.
Response: We disagree with the commentor that seafood “originator” or even the “first
receiver” often do not know the destination of the finished products. This rule applies equally to
both foreign and domestic firms which are expected to work with their supply chain partners to
determine whether their products will be sold in the United States as they already must be doing
to comply with several other FDA existing regulations. The rule provides exemptions for those
directly selling their products to consumers or products covered by the requirement of the
National Shellfish Sanitation Program (NSSP).

Comment 27 (Mung Bean Sprouts)
One comment asks for a more comprehensive economic analysis on the impacts of the
proposed rule on the sprout supply chain, including the sprout seed supply chain. The commenter
said that requiring sprout growers to trace mung bean seeds back to individual farms would
effectively prohibit the importation and sale of almost all internationally sourced mung bean
seeds and thereby virtually eliminate the mung bean seed/sprout market in the U.S.
Response: We thank the commenter for raising this concern. We have revised this
requirement so that the final rule only requires initial packers of sprouts to maintain records
related to the grower of sprout seeds when that information is available. As the final rule does

not require sprout growers to trace seeds back to individual farms when such information is
unavailable, we do not expect it to create a significant obstacle to the importation and sale of
internationally sourced sprout seeds. Because the rule includes sprouts as covered food (but does
not include “seed for sprouting”), sprout growers are required to comply with the subpart S
requirements.
In addition, according to Observatory of Economic Complexity (OEC) data, in 2019
mung bean sprouts constituted about $1.1 billion worth of world trade. Myanmar, China,
Uzbekistan, and Indonesia accounted for 73 percent of all mung bean exports with India,
Vietnam, China, Japan, Indonesia, and Pakistan accounting for the largest share of mung bean
imports totaling 66 percent. The U.S. accounts for roughly 3.5 percent of world mung bean
imports and only 0.7 percent of mung bean exports. Given the small trade volume of mung bean
imports in the U.S., it is highly unlikely that this rule will substantially impact the global market
or the U.S. mung bean industry. We believe that given the relatively low volume of mung bean
trade, it is unlikely for internationally sourced mung bean seeds producers to be negatively
affected because only sprout growers that supply the U.S. market will be affected by the rule.

Comment 28 (Disproportionate Effects on Consumers and Small Businesses)
Comments variously stated that the costs and benefits of this rule will not be distributed
evenly. Some commenters stated that the general public bears the brunt of the health impacts
caused by poor traceability, with poor and minority communities paying an especially heavy toll.
Comments also noted that a constriction in the U.S. mung bean seed/sprout market would
disproportionately impact the Asian American community in the U.S., for whom mung
beans/sprouts are a staple of the daily diet. Another comment mentioned that the undue burden of

this rule on industry manufacturers and distributors might raise consumer prices, making certain
food items, including many fresh fruits and vegetables, less affordable. This comment suggested
that such price increases could harm low-income consumers with a limited food budget.
Response: To further inform our understanding of the distributional impacts of the final
rule, FDA analyzed a nationwide cross-section of diet data to understand FTL consumption rates
of various demographic groups. While we find some differences in FTL foods consumption, we
have no data on substitution patterns for non-FTL foods, and thus on the effect the final rule will
have on the overall diet quality of consumers. Similarly, FDA contracted with ERG to
understand the anticipated effect of the final rule on costs and, therefore, on consumer prices
(Ref. [4]). Industry subject matter experts said they expect consumer prices to increase as a result
of the final rule. However, FDA found no evidence on the magnitude of the cost pass-through,
the incidence of cost pass-through on non-FTL items, substitution patterns of different segments
of consumers, or price elasticity estimates for FTL items for different demographics. Without
this information, we cannot assess the distributional effects of the final rule on various
consumers. We nonetheless acknowledge that the costs and benefits of the rule may accrue
unequally to various consumer segments. We addressed distributional impacts of this rule in
Section II.G.

Comment 29 (Gasoline Sales)
A comment stated that FDA’s estimated number of covered small retailers should
account for gasoline sales. The commenter stated that, by inflating the sales of gas station
convenience stores without adding significantly to profits, gasoline sales cause these stores to
exceed the sales threshold for retail exemption.

Response: The proposed rule did not contain a sales-based exemption for retail food
establishments. The final rule specifies that the exemption threshold for retail food
establishments, $250,000 (during the previous 3-year period on a rolling basis, adjusted for
inflation using 2020 as the baseline year), and similarly the threshold for exemption from the
sortable spreadsheet, $1 million, are based on the value of food sold or provided to consumers.
We have updated our estimate of the number of covered gasoline stations with convenience
stores based on approximating the share of food sales. We present our updated analysis of
impacts on small businesses in section III.

Comment 30 (Small Business Exit)
Several comments stated that restaurants and small farms in general do not generate
sufficient profit margins to absorb the costs of compliance with the proposed rule and remain in
business.
Response: We appreciate public input on the viability of small food businesses. We note
that the final rule extends full exemptions from all requirements to retailers, including
restaurants, with under $250,000 in annual food sales. We have updated our estimates to reflect
these exemptions, including the revised exemptions for restaurants. In addition, the final rule
includes an exemption for retailers, including restaurants, with less than $1 million in annual
food sales, from providing the information requested by FDA in the form of an electronic
sortable spreadsheet.
We note also that the final rule shifts traceability lot code assignment from growing to
initial packing. The final rule specifies that farms are required to assign the field names on the
farm map only once and do not need to update them based on what is grown. Additionally, the

final rule provides an exemption from producing an electronic sortable spreadsheet to farms with
less than $250,000 in annual sales. We present revised estimates of impacts on small businesses
in section III.B of this RIA. Section III.B breaks down one-time and annualized costs, in dollar
values and as a percentage of revenue, across broad industry categories.

Comment 31 (Input from Small Businesses)
Several comments claimed that the PRIA inadequately addresses small business impacts
and, in particular, that FDA did not estimate costs specifically for small entities in its economic
impact analysis. Among these, commenters stated that FDA underestimated small retailers’
compliance costs under the proposed rule, including costs for traceability systems and training.
Commenters suggested that FDA should inform its estimates with data or stakeholder input
specifically representing small businesses, including by consulting with organizations that
comprise or represent small businesses regarding compliance costs.
Response: After reviewing public comments, FDA determined a need for additional data
to improve cost estimates, including costs to small businesses. FDA contracted with ERG to
research additional literature and interview a panel of food industry experts to further inform the
costs of the rule to various covered small entities based on their baseline traceability practices
(Ref. [4]). Section III of this final RIA now incorporates the expert elicitation results specifically
addressing costs to small businesses.

Comment 32 (Proposed Option 1 - Very small, Retail)
One comment stated that if retail food regulatory jurisdictions will need to include this
rule as part of their inspection process the inclusion of very small retail operations will have an
additional financial impact on the regulatory jurisdiction.
Response: We don’t anticipate additional financial impact on food regulatory
jurisdictions from including very small retail operations to their inspection process because very
small retail operations are exempt from the requirements of this rule.7 We are also still
considering the best approach for structuring and conducting inspections for compliance with the
subpart S recordkeeping requirements, including the roles that FDA and State investigators
should play. While FDA anticipates conducting periodic, routine inspections of traceability
records outside of an outbreak traceback investigation, we will work with state and local partners
to consider mechanisms for how to conduct routine records checks of retail food establishments
and restaurants. We will consider obtaining additional funding for our regulatory partners
through various mechanisms, such as grant programs. In addition, we intend to publish guidance
for industry and provide training to regulatory partners who will be conducting inspections of
records under this rule ahead of the compliance dates.

Comment 33 (Number of Foods by FTL Commodity)
One comment stated that we have underestimated the number of covered foods within
each designation in light of having covered commodities as foods on the FTL.

1.1305(i) of the final rule provides that subpart S does not apply to RFEs and restaurants with an average annual
monetary value of food sold or provided during the previous 3-year period of no more than $250,000 (on a rolling
basis), adjusted for inflation using 2020 as the baseline year for calculating the adjustment.

Response: Both in the preliminary and final regulatory impact analyses, we base our
estimates of regulatory impacts on the number of entities that manufacture, process, pack or hold
FTL foods and not on the number of foods by FTL commodity. As discussed in the Preliminary
Regulatory Impact Analysis for the proposed rule (Ref. [3]), in estimating the impact of the
proposed rule, we accessed data from multiple sources, including the U.S. Economic Census 2012 Statistics of U.S. Businesses (SUSB), FDA’s Food Facility Registration Module, and the
USDA National Agricultural Statistics Service (NASS) 2017 Census of Agriculture. We have
since updated our estimates in the final rule RIA using additional data sources, including the
2017 SUSB and the 2017 North American Product Classification System (NAPCS) from the
U.S. Census to better inform the number of covered entities that manufacture, process, pack or
hold various FTL foods. We discuss our revised coverage estimates in greater detail in sections
I.E.2, II.D.2, and appendix D of this analysis.

Comment 34 (General)
Some comments stated that the measures needed for compliance with the proposed rule
would vary widely as some entities rely almost exclusively on paper records. Others requested
that FDA adopt a more universal method of data storage and dissemination.
Response: We agree with the comment that measures needed for compliance with this
rule will vary widely among different types and sizes of entities. This final rule does not
prescribe any specific technology for maintaining records, nor does it preclude anyone from
using paper records if that is their business practice. This final rule sets forth minimum standards
for data elements used in records, but otherwise provides flexibility to accommodate a variety of
existing practices and business needs. Although FDA strongly encourages the use of electronic

recordkeeping for traceability, persons subject to the rule may keep their records in paper or
electronic form. Firms may also contract with others to establish and maintain records required
under subpart S on their behalf as long as the firm can provide the information to FDA in
accordance with the rule. To protect certain confidential business information, the rule allows
firms to provide their customers with a reference to the information instead of directly
identifying the traceability lot code source of an FTL food they handle. For this final analysis,
FDA contracted with ERG to research additional literature and interview a panel of food industry
experts to further inform the costs of the rule to various covered entities based on their size,
broad industry category, and baseline traceability practices (Ref. [4]).

E. Summary of Changes
Compared to the preliminary economic analysis (Ref. [3]), the final regulatory impact
analysis reflects revisions to the rule and to our analytical methodology. It includes updates and
revisions to our discussion of baseline conditions, estimated health and non-health benefits, costs
to domestic entities, estimates of international impacts, distributional effects, and impacts to
small entities in the Regulatory Flexibility Analysis (RFA) section III of this analysis as
summarized below.
1. General Changes to the Rule


We have adjusted for inflation using the GDP deflator and report all benefits and costs in
2020-year dollar values.



FDA extended the effective date of compliance from 2 years to 3 years from the publication
of this final rule. Section II.J. of this RIA quantifies the impact of this change in the Analysis
of Regulatory Alternatives to the Rule (“Reduce compliance date to two years”).



Given expected long-term impacts of this rule, we have extended the time horizon for
estimating annualized costs and benefits from 10 years to 20 years.



To inform our analysis, FDA contracted with ERG to research additional literature and elicit
information from two sets of panels of industry experts (Ref. [4]). The first set of experts
further informed the benefits of avoiding overly broad recalls (hereinafter referred to as the
recall elicitation). The second set of experts further informed the costs of the rule as
proposed, with additional brief definitions of some new CTEs in their draft-final state at the
time of the elicitation (hereinafter referred to as the traceability costs elicitation).



The final rule replaces the threshold for full exemption for retail food establishments (RFEs)
from 10 full-time equivalent employees (Option 1 in the proposed rule) to monetary
threshold of no more than $250,000 in average annual monetary value of food sold or
provided during the previous 3-year period (on a rolling basis), adjusted for inflation using
2020 as the baseline year for calculating the adjustment. Appendix D discusses differences in
coverage estimates between the proposed and the final rule. In addition, section II.J of this
RIA quantifies the impacts of an alternative scope and an alternative exemption policy in the
Analysis of Regulatory Alternatives to the Rule (“Cover all firms in the same broader
industry (NAICS) category as covered firms” and “Broader exemption for retail food
establishments and restaurants”).



The final rule expands the exemption from producing an electronic sortable spreadsheet to:
a. Farms whose average annual sum of the monetary value of their sales of raw agricultural
commodities and the market value of raw agricultural commodities they manufacture,
process, pack, or hold without sale (e.g., held for a fee) during the previous 3-year period

is no more than $250,000 (on a rolling basis), adjusted for inflation using 2020 as the
baseline year for calculating the adjustment.
b. Retail food establishments or restaurants8 with an average annual monetary value of food
sold or provided during the previous 3-year period of no more than $1 million (on a
rolling basis), adjusted for inflation using 2020 as the baseline year for calculating the
adjustment.
c. Persons (other than a farm, retail food establishment, or restaurant) whose average annual
sum of the monetary value of their sales of food and the market value of food they
manufacture, process, pack, or hold without sale (e.g., held for a fee) during the previous
3-year period is no more than $1 million (on a rolling basis), adjusted for inflation using
2020 as the baseline year for calculating the adjustment.
2. Baseline Conditions: Coverage and Current Industry Practices


We updated our estimates of covered entities using additional data sources, including the
2017 Statistics of U.S. Businesses (SUSB) and the 2017 North American Product
Classification System (NAPCS) of the U.S. Census, and the 2017 USDA National
Agricultural Statistics Service (NASS) data.



We estimate the number of covered entities that produce, manufacture, process, pack, or hold
foods on the FTL by counting the number of establishments by NAICS industry sector. In
doing so, we derive a share of FTL establishments by approximating the share of the covered
entities by each NAICS industry using the NAPCS data.

8
While the PRIA included non-restaurant retailers and restaurants in the same category of retail food
establishments, the FRIA reports separate estimates of the non-restaurant retailers and restaurants provided by an
elicitation of industry experts (Ref. [4]). Both non-restaurant retailers and restaurants have the same requirements as
each other under the final rule. Experts assumed the non-restaurant retailer includes mostly grocery stores,
specialized grocery stores (fish markets, fruit markets, bakeries, etc.), supercenters, and club stores.



Some NAICS industry codes that were included in the preliminary estimates are now
excluded from our analysis due to clarifications to the FTL and revision to the FTL
definition. For example, we now exclude from our analysis some entities in each NAICS
industry if the form of the FTL food is no longer fresh and has been changed (e.g., frozen
pizza with a spinach topping).



To estimate the share of entities who currently have traceability programs in place and the
degree to which their programs meet the requirements of this rule, ERG completed the
traceability costs elicitation in December 2021 and January 2022 providing both qualitative
and quantitative input on current traceability practices (Ref. [4]). We use the information
provided by experts to better characterize current traceability practices and update our
estimates on the proportion of firms that will incur costs from this rule and per entity costs as
explained in Section II.D.



We estimate the number of covered gasoline stations with convenience stores by
approximating the share of food sales only.

3. Benefits


We use additional data on disease outbreaks associated with covered foods to extend our
outbreak data from January 2009 to December 2020 (see Appendix C).



We revise estimated benefits from avoiding overly broad recalls by using new information
from an industry expert elicitation study conducted by ERG (Ref. [4]) from December 2021
through January 2022. Experts provided information on labor and non-labor costs incurred
by firms when responding to an overly broad recall.

4. Costs


Our revised coverage estimates that we use for estimating compliance costs now explicitly
exclude fully exempt entities as well as those not handling FTL foods.



Cost estimates now reflect the revised requirements of the final rule, which redefine what
activities are CTEs and exclude or redefine several proposed KDEs. These updates are
further summarized in appendix E. Among these updates there is one, for example, that shifts
traceability lot code assignment requirements from growers to initial packers. Under the
final rule, entities that grow or raise FTL foods (other than eggs) will need to maintain a farm
map(s) with field names, rather than having to assign traceability lot codes and link those
codes to the geographic coordinates where each lot is grown. The farm map(s), which must
show the area(s) in which food is grown or raised, do not need to be updated based on what is
grown.



We incorporate new inputs throughout our analysis from the literature review and multiple
industry experts, elicited by ERG in December 2021 and January 2022, who described
anticipated cost-incurring compliance activities and expenditures, estimated variables related
to cost calculations, and further commented on factors likely to influence costs of the rule
(Ref. [4]).



We consider recurring capital costs for those cases where capital investments made towards
compliance with the rule result in higher operation and maintenance expenses than covered
entities would otherwise face (section II.F.3 “Costs of Capital Investment”).



We consider recurring training costs for those cases where new training is more time
consuming than what covered entities would otherwise have implemented as a refresher for
continuing employees and because of turnover (section II.F.4 “Costs of Training”).



We now separately estimate costs specific to small versus large entities in different
categories of industries.



We updated wage data (average wages for various occupations) using the Bureau of Labor
Statistics’ (BLS) 2020 Occupational Employment Statistics (OES).

5. International Effects


We updated the number of registered foreign food facilities using internal FDA data. We
revised it from 127,925 to 68,566 establishments due to double counting error discovered
earlier in the PRIA estimates.



We updated our estimates of costs to foreign facilities based on additional exemptions of
firms granted by FDA (Section II. H).

6. Distributional Effects


We updated our analysis of distributional impacts of the rule by considering additional data
on FTL consumption rates across consumers, geographic concentration of covered retail food
establishments, and the distribution of costs across affected entities.

7. Regulatory Flexibility Analysis


In the Initial Regulatory Flexibility Analysis, we assumed that the compliance costs faced by
small businesses would fall between the low end and middle of the range of costs estimated
for all businesses overall. We now use input from external experts (Ref. [4]) to separately
estimate costs specific to small versus large businesses in different categories of industries.



Unlike in the Initial Regulatory Flexibility Analysis, our revised coverage estimates, which
inform both the main cost analysis and Regulatory Flexibility Analysis, now explicitly
exclude fully exempt entities as well as those not handling FTL foods.



Finally, we updated small firm revenue data previously sourced from the 2012 Statistics of
U.S. Businesses (SUSB) with data from the 2017 SUSB.

II. Final Economic Analysis of Impacts
A. Background
Current recordkeeping requirements that stem from the Bioterrorism Act (BT Act) of
2002 require firms to know and record the immediate previous source of their food products and
the immediate subsequent recipient (commonly referred to as one-up, one-back recordkeeping).
Since these requirements took effect, FDA has encountered significant limitations in the
available food tracing-related information upon which government agencies and industry rely for
rapid and effective tracing of food products in the event of an outbreak investigation. These
limitations arise from gaps in recordkeeping requirements, including: a requirement to maintain a
record of the lot code or other unique identifier only if it exists, no requirement to link incoming
and outgoing product within a firm and from one point in the food supply chain to the next, and
address requirements that do not distinguish between corporate headquarters and the physical
location where the food was produced.
Inadequate traceability information and the challenge of having many point-of-service
firms (retail and foodservice) excluded from subpart J requirements has hampered recalls of
potentially contaminated foods. In 2015, for example, an outbreak of Shiga toxin-producing
Escherichia coli (E. coli) O26 (STEC O26) resulted in 55 illnesses in 11 states, leading to 21
hospitalizations (Ref. [6]). Though an investigation conducted by the CDC, FDA, and the
USDA’s Food Safety and Inspection Service linked a specific restaurant chain to the outbreak as
early as October of 2015, investigators could not identify a particular ingredient or food item as
47

the likely source of contamination. The lack of information in the records maintained by the
restaurant caused an inability for regulatory officials to use traceability to narrow the ingredients
to further investigate which ingredients came from common sources.
Inadequate traceability and the exclusion of farms from Subpart J requirements also
triggered the need for broad recalls that inadvertently affected non-contaminated product. In
2015, for example, FDA identified 36 farms as potentially having produced leafy greens for a
leafy greens mix linked to an E. coli outbreak. Without being able to identify specific lots and
growers of contaminated product, it was not possible to narrow investigative efforts to the source
of the outbreak which would have allowed the Agency to narrow the scope of the recall (Ref.
[7]).
On January 4, 2011, FSMA (Public Law 111-353) was signed into law. Section
204(d)(1) of FSMA requires FDA to establish recordkeeping requirements for facilities that
manufacture, process, pack, or hold foods that we designate as high-risk foods. These
recordkeeping requirements will be additional to the traceability recordkeeping requirements in
21 CFR Part 1, Subpart J (the Subpart J requirements), which were promulgated in accordance
with the BT Act of 2002. Section 204(d)(2) of FSMA requires the Agency to designate the foods
for which these additional recordkeeping requirements are appropriate and necessary to protect
the public health, and to publish the list of such foods (the FTL) on our web site.
On September 23, 2020, FDA published the preliminary regulatory impact analysis
(PRIA) of the proposed rule “Requirements for Additional Traceability Records for Certain
Foods” (85 FR 59984). The preliminary regulatory impact analysis included two proposed
options for how this rule would apply to retail food establishments. The first option proposed a
full exemption from the requirements of the rule for retail food establishments (RFEs) that

employ 10 or fewer full-time equivalent employees (FTEs). The second option proposed only
partial exemption to RFEs that employ 10 or fewer FTEs, stating that they would be exempt
from the requirement to provide FDA, under specified circumstances, with an electronic sortable
spreadsheet containing certain traceability information but that they would be required to comply
with all other aspects of the rule.
This FRIA analyzes the economic impact of the traceability requirements set forth in this
final rule by making revisions and updating the 2020 PRIA, based on information from
comments, changes to the rule, and new information not formerly considered in the PRIA.

B. Potential Need for Federal Regulatory Action
A traceability system or program is a method that allows for information about product
attributes to flow among entities in a supply chain.9 Industry reports offer some evidence that
implementing intra-firm traceability is beneficial for the firm because it helps in improving
supply chain efficiencies and minimizing the impact of food safety hazards (Ref. [8]). However,
private incentives to implement traceability systems vary among different industries and industry
sectors and depend on many factors. Aside from return-on-investment decisions by individual
firms, reasons for adopting a traceability system or program may include reasons other than food
safety. The motivations also depend largely on private decisions of the firm which may also
depend on specific attributes of products, whether they are included in the FTL or not. However,
depending on the level of sophistication, traceability can be costly and considering the
differences in the characteristics of inter-firm or inter-industry traceability systems, benefits may

For purposes of this analysis, we use the terms traceability system, program, or method interchangeably.

not be evenly distributed across firms in the supply chain, thus lowering incentives for some
firms to adopt a socially optimal level of traceability.10
The effectiveness of a traceability system depends on the accuracy, quality, uniformity,
and extent of collected information. Firms generally have private incentives to avoid the
deliberate or accidental contamination of food linked to their products or facilities. Nevertheless,
those incentives may not be enough for all firms to provide others with the socially optimal
amount of information about their entire production and distribution network. Because firms’
revenues may not capture all of the benefits that accrue to the public from improved food
traceability, firms may collect and supply less information than would be socially optimal for
adequate protection of public health.
Several types of market failure may impact current traceability efforts. First, private
incentives for investing in socially optimal level traceability may be low. When traceability is
limited for a product, producers and other supply chain participants can remain anonymous and
therefore not as accountable for the quality or safety of their product. Anonymity can also allow
someone to freeride on the reputation of other competing producers with better traceability and
maybe even safer food, thus creating a market failure. Since unsafe food can originate at
different levels of the supply chain, the probability of a prolonged foodborne outbreak due to
inefficient tracing is more likely with imperfect inter-firm participation (Ref. [9]).
Second, firms currently utilize various traceability methods at times with competing data
standards, creating interoperability challenges and making costs of coordination prohibitively
high, reducing the incentives to adopt uniform standards. Without broad adoption of

The socially optimal level of traceability considers all private costs and benefits (those faced by firms) as well as
public costs and benefits (those faced by everyone other than firms). In other words, the socially optimal level of
traceability maximizes the aggregate welfare of society, which includes firms and non-firms (e.g., consumers).

interoperable traceability methods and standards, the system will not work as effectively to allow
fast and efficient traceback when an outbreak occurs. When the value of adopting a good or
service (such as a new technology or traceability system) depends on the number of users
adopting the good or service, it is sometimes referred to as a network effect (Ref. [10]). In the
same manner, the value of adopting traceability would depend on the number of users adopting
traceability. Interoperability challenges caused by the absence of uniform standards are
considered a network externality that affects the entire industry and reduces the value of adopting
traceability among individual potential users, therefore reducing their incentives to adopt.
Third, the return on each firm’s additional investment in traceability depends on the level
and type of investment made by other firms, potentially causing a disincentive for some firms to
invest. According to a study by Mai et al. (Ref. [11]), costs and benefits of current traceability
investments are not evenly shared among different entities in the supply chain. For example,
while processors and producers in the seafood industry may incur a greater share of the costs of
implementing traceability than retailers and distributors, retailers and distributors may capture a
larger portion of the benefits in the form of market growth. This misalignment of benefits and
costs might explain the lack of incentives for some entities in the supply chain to adopt
traceability. As a result, the risk of foodborne illnesses that can sometimes be attributed to FTL
foods is likely not fully priced into FTL products (Ref. [11]).
While the last decade has experienced growth of technology enabled traceability
methods, current traceability systems may in part originate from requirements of the BT Act.
Economic incentives, such as improved supply-side management and safety and quality control,
may have motivated some producers to develop traceability systems of varying sophistication
and comprehensiveness. Food producers in the U.S. use a variety of systems to trace the

movement of food in the supply system. Tracing systems vary by the type and amount of
information they collect and record, the record medium (e.g., paper vs. electronic), and the extent
of the supply chain is covered (e.g., the immediate previous and next steps vs. the entire chain
from farm to retailer). In some instances, owners of large supply chains (e.g., major retailers,
major restaurant operators, brokers of different size that represent farms, food processors with
many ingredient suppliers, importers of seafood from many vessels) compete on supply chain
efficiency and consumer transparency, which requires traceability as a component of that
strategy. However, to maintain competitive advantage, most supply chain owners require their
suppliers to share traceability data through private portals. This leads to a proliferation of
different portals and data standards, which reduces the potential for interoperability. A universal
standard for traceability would enable suppliers to insist their customers (and portals) accept, at
minimum, a standard list of certain CTEs and KDEs, which would lower the cost for suppliers
across and within industries among other benefits.
So far FDA has often experienced the significant limitations in the available tracingrelated information on which government agencies and industry currently rely to conduct tracing
operations. Industry often does not fully understand what data the FDA needs to effectively
investigate foodborne illness outbreaks. Further, while standard production and distribution
records carry a lot of useful information, they do not necessarily capture the complete set of
information, in any standard format, that FDA would need to efficiently investigate a
contamination of unknown origin. The result is that many of the systems and approaches that
firms currently use for voluntary traceability are not interoperable, which results in potentially
avoidable costs for all entities in the food supply chain. This failure of interoperability also slows
outbreak investigations, sickens more consumers, and reduces trust in the U.S. food supply.

Although some supply chain owners have rapidly adopted traceability technology,
recordkeeping practices lack uniformity across supply chains. Different supply chain entities
such as growers, harvesters, shippers, distributors, retailers, and restaurants lack incentives to
standardize recordkeeping in the form of common key data elements. This rule would ensure that
producers, distributors, retailers, and other covered entities know what information they need to
keep and provide. Without uniform recordkeeping standards, competing traceability solution
providers promote mutually exclusive, proprietary frameworks, whose incompatibility increases
traceability costs. The current lack of system interoperability impedes collaboration in
identifying sources and/or recipients of potentially contaminated covered food. While the
effectiveness of each traceability system increases with the number of participants throughout the
supply chain, the lack of standardization in recordkeeping and data sharing among incompatible
systems causes duplication of efforts. In addition, high transaction and coordination costs of
setting up a complete farm-to-retail national traceability system may even disincentivize some
firms from investing in traceability systems, particularly those firms that are not vertically
integrated. This final rule will standardize the key data elements and critical tracking events,
significantly reduce the private coordination and transaction costs of setting up a complete
tracing system and enable FDA and other entities involved in a tracing investigation to accelerate
and enhance the acquisition of robust product tracing information.
Underscoring the need for standardized data elements, food trade associations,
technology providers, consumer advocacy groups, standards bodies, multi-unit restaurant
operators, retailers, distributors and food producers have asked FDA to describe the types of data
we need, and the format in which we prefer to receive such data during an outbreak

investigation.11 This information would enable companies and solution providers to develop
systems and procedures to efficiently collect that data, so it can be shared with the FDA when
needed. Ultimately this might lower traceability costs for most members of the food supply
chain because it would encourage the development of interoperable traceability systems.
From public comments12 received as part of FDA’s New Era for Smarter Food Safety
Public Meeting held on October 21, 2019, one large food industry trade association representing
food companies from around the world commented that one of the most foundational and
significant actions FDA could take is identifying the key data elements that should be
communicated throughout the global supply chain. Similar comments echoed the need to
establish a common set of key data elements and to have clarification from FDA on the key data
elements needed to provide effective and rapid tracing.
In addition to standardized key data elements and critical tracking events, the
effectiveness of a tracing system depends on the extent to which firms throughout the supply
chain participate. Unfortunately, even a small number of gaps in tracing information through the
supply chain can prevent the FDA and others from being able to trace or to effectively trace
contaminated products to their source. Full supply chain traceability requires policy intervention
as some firms do not have an immediate financial incentive to institute traceability systems (Ref.
[12]). For example, the Institute of Food Technologists (IFT) noted in the Product Tracing Pilots
report (Ref. [5], see page 217, subtitled “Lack of Standards Results in Fragmented
Requirements”) that traceability is likely to stay in a state of perpetual flux until FDA clearly

https://www.regulations.gov/searchResults?rpp=25&po=0&s=FDA-2019-N-4187&fp=true&ns=true
This section references public comments from the New Era for Smarter Food Safety Public Meeting - Docket ID:
FDA-2019-N-4187(https://www.regulations.gov/searchResults?rpp=25&po=0&s=FDA-2019-N4187&fp=true&ns=true), including comments by the Grocery Manufacturers Association, the US Apple
Association, National Fisheries Institute, United Fresh, Produce Manufacturers Association, among others.
12

defines the data requirements and establishes a framework for full supply chain traceability. IFT
found that producers were reluctant to invest in tracing systems if their fellow producers were
not similarly investing, since tracing is not an isolated exercise.
Comments received as part of FDA’s New Era for Smarter Food Safety Public Meeting
indicated that there are inconsistencies among suppliers and buyers in terms of the level of
capability for traceability and that food supply chain companies cannot control the recordkeeping
by entities that repackage product further up the supply chain. One comment from a large trade
association indicated that “off the record” conversations with their broad membership indicated
consensus that the time of hoping for voluntary adoption of effective traceability systems has
passed. As discussed in section I.D and the preamble of the final rule we also received
comments in response to this proposed rule in support of the need for the rule.
In sum, market prices convey most of the necessary information for the ordinary
production and distribution of foods, including the foods on the FTL.13 However, an actual or
suspected contamination of unknown origin requires more complete and standardized
information as well as the ability to rapidly access and consolidate that information. In order to
protect consumers from further exposure and to quickly and efficiently find the source and cause
of contamination, FDA must be able to trace covered food backward and forward through the
entire supply chain. Although the nation’s food manufacturers, processors, distributors, retail
food establishments, and others may benefit from such a system, the private costs of creating it
would be prohibitively expensive for any single firm or third-party organization. As discussed in
the following section II.C, this rule addresses many of the known limitations of current

Prices provide most information about goods and services without the need for buyers and sellers to know much
about each other. However, prices do not always communicate the difference between contaminated versus not
contaminated product in the market, which explains the potential need for government intervention.

traceability systems by requiring a rigorous and consistent approach to food tracing across
different industry sectors for more efficient traceability of foods on the FTL. The final rule will
also enable FDA, its regulatory partners, and industry to better identify and remove contaminated
FTL foods from the marketplace in the case of an outbreak, as well as to develop mitigation
strategies to prevent future contamination.
We have not received any comments that would refute the market failure claims above
nor have we received any information in comments that may be used to quantify the scope of
incomplete internalization of relevant baseline costs.14

C. Purpose of the Rule
The purpose of this rule is to ensure that contaminated FTL foods covered by this rule
can be swiftly identified and removed from the market to prevent or mitigate a foodborne illness
outbreak. In order to improve FDA’s ability to follow the movement of FTL foods through the
supply chain, the rule will establish traceability recordkeeping requirements for persons who
manufacture, process, pack, or hold FTL foods. Namely, the rule specifies the data elements and
information firms must keep, along with information they must provide to the next entity in the
supply chain. The core requirements are to establish and maintain a traceability plan and keep
and provide records of KDEs associated with different CTEs in a covered food’s supply chain,
including the harvesting, cooling, receiving, initial packing, transforming, and shipping of the
FTL food. Required traceability plans include: a description of the procedures used to keep
required subpart S records; a description of the procedures used to identify foods on the FTL

Industry internalizes costs when they incorporate society-wide costs as part of a pricing structure (to be specific,
social costs that are born from their economic activities).

they manufacture, process, pack, or hold; a description of how they assign traceability lot codes
to foods on the FTL; a statement identifying a point of contact for questions about the
traceability plan and records; and, for persons who grow or raise an FTL food (other than eggs),
a farm map showing the area in which the food is grown or raised. To protect certain confidential
business information that some firms may be reluctant to share, the rule allows the flexibility for
firms to provide a reference to information on the traceability lot code source instead of directly
identifying this information for an FTL food.
The rule also provides consistent food tracing terminology, encourages a transition from
paper-based recordkeeping to electronic records (although persons subject to the rule may keep
their records in paper or electronic form) and promotes a broader understanding of the data
elements needed for efficient traceability and product recalls. Further, this rule enables FDA and
industry to faster and more effectively identify the source of an outbreak or other contamination
event, expedite removal of contaminated food from the marketplace, and prevent additional
consumer exposures, as well as develop mitigation strategies to prevent future contamination.
This rule further helps the Agency deter and limit the effects of foodborne outbreaks from FTL
foods and thereby improve the safety of the food supply in the United States.

D. Baseline Conditions
We consider the current state of the world, including current trends towards greater
traceability capabilities, as a reasonable approximation of the baseline (the projected future
without the rule) against which to measure the costs and benefits of the rule and the regulatory
alternatives discussed in section II.J.15 While we are not able to explicitly estimate trends in

We acknowledge that health benefits of this rule are linked, at least in part, to other rules issued by FDA. Some of

industry investment in traceability capabilities,16 which might result in some overestimation of
costs as discussed in sections II.F.3 “Costs of Capital Investment” and II.I “Uncertainty and
Sensitivity Analysis,” many food businesses have been increasingly adopting traceability
systems or practices, including technologies, for reasons other than regulatory mandate. Such
reasons include operational efficiency, transparency, and customer demand. Existing supply
chain practices of many affected businesses already partially satisfy this rule’s requirements for
traceability records of FTL foods. In estimating the baseline, we use the information provided by
the traceability costs elicitation (Ref. [4]) in assuming that a portion of covered entities is already
moving toward implementing traceability with or without the rule and that another portion of
covered entities would not implement traceability without the rule.
1. Bioterrorism Act of 2002 and the 2004 BT Final Rule Recordkeeping Requirements
Before the enactment of FSMA, FDA implemented recordkeeping requirements (Subpart
J) related to product tracing under authority of the BT Act of 2002. Thus, the current estimated
baseline includes the costs and benefits of the pre-FSMA Establishment and Maintenance of
Records Under the Public Health Security and Bioterrorism Preparedness and Response Act of

these rules addressing foodborne illnesses have not taken effect or are not captured in the data used to characterize
the baseline scenario of this analysis. In particular, FDA’s Standards for the Growing, Harvesting, Packing, and
Holding of Produce for Human Consumption Relating to Agricultural Water proposed rule
(https://www.federalregister.gov/documents/2021/12/06/2021-26127/standards-for-the-growing-harvesting-packingand-holding-of-produce-for-human-consumption-relating), if finalized, might prevent some of the illnesses that this
RIA estimates will occur at baseline absent food traceability. We cannot confidently predict the impact of the
Agricultural Water rule on the baseline for traceability, partly because the Agricultural Water rule has not been
finalized at the time of this writing. We thus possibly overestimate benefits of the traceability rule in section II.E.1
of this document.
16
We acknowledge that data available prior to finalization of this rule may show substantial changes relative to the
present — due, for example, to other FSMA regulations increasingly taking effect and to societal changes associated
with the COVID-19 pandemic. We haven’t received comments on estimating the baseline trajectory, given the
dynamic nature of the regulatory environment. However, to address uncertainty in our estimates we use the results
from an expert elicitation in which ERG asked experts not to consider as impacts of the rule current and probable
future traceability practices that would occur in any case (Ref. [4]).

2002 final rule issued in 2004, as estimated in the economic impact analysis for that rule and
further modified by updated assumptions discussed below.17
The 2004 economic impact analysis of the BT final rule estimated annual and first-year
costs of requiring establishment and maintenance of records to trace the transportation of all food
to both foreign and domestic entities, as well as costs for future entities entering the market each
year.18 Benefits of the 2004 BT final rule were estimated as the number of averted illnesses due
to improved recordkeeping practices. Nevertheless, in almost twenty years since implementation
of these recordkeeping requirements, FDA has learned that there are critical gaps in the
requirements that limit the ability of regulatory agencies to conduct prompt, effective product
tracing, especially in response to foodborne illness outbreaks. These critical gaps, which are
discussed in sections III.A and III.B of the preamble of the final rule, suggest that the benefits of
the 2004 BT rule may have not been realized and were consequently overestimated. However,
as described elsewhere in this document, advances in information technology in the last decade
are such that private incentives have led some entities to implement food traceability beyond the
2004 requirements. This suggests that the annualized costs of the BT rule estimated in 2004 may
not have fully accounted for baseline trends towards increased traceability and thus were also
overestimated.

2. Coverage of the Rule
Covered entities (firms or establishments19) will incur costs from the final rule to the
extent that compliance requires them to change their current practices. Covered entities are those

Federal Register /Vol. 69, No. 236 / Thursday, December 9, 2004 / Rules and Regulations, page 71611.
Federal Register / Vol. 69, No. 236 / Thursday, December 9, 2004 / Rules and Regulations, page 71640.
19
Each firm may operate one or several establishments. Some costs are estimated on per-firm level and some are on
per-establishment level.
18

that manufacture, process, pack, or hold foods that FDA has designated as requiring additional
recordkeeping and placed on the FTL.20 The traceability recordkeeping requirements will
generally not apply to:


Farms or the farm activities of farm mixed-type facilities with respect to the produce
they grow that are not covered under the FSMA Produce Safety Rule, 21
CFR 112.4(a).



Produce farms and producers of raw agricultural commodities other than produce or
shell eggs (e.g., aquaculture operations) when the average annual sum of the
monetary value of their sales of food and the market value they manufacture, process,
pack, or hold without sale (e.g., held for a fee) during the previous 3-year period is no
more than $25,000 (on a rolling basis), adjusted for inflation, using 2020 as the
baseline year for calculating the adjustment.



Shell egg producers with fewer than 3,000 laying hens at a particular farm, with
respect to the shell eggs they produce at that farm.



Covered foods produced on a farm (including food that is also packaged on the farm)
that is sold or donated directly to a consumer by the owner, operator, or agent in
charge of the farm.



Covered foods produced and packaged on a farm if the packaging of the food remains
in place until the food reaches the consumer and maintains the integrity of the
product, prevents subsequent contamination or alteration of the product, and has
labeling that includes the name, address, and business phone number of the farm.

The list of applicable foods can be updated by publishing a notice in the Federal Register, using the process
described in final § 1.1465. See Appendix A for the list as of this writing.



Covered produce that receives commercial processing to adequately reduce the
presence of microorganisms of public health significance if the conditions set forth in
21 CFR 112.2(b) are met (regarding the commercial processing exemption to the
FSMA Produce Safety rule).



Shell eggs when all eggs produced at the particular farm receive a treatment (as
defined in 21 CFR 118.3) in accordance with 21 CFR 118.1(a)(2).



Foods on the FTL subject to a kill step, provided that records document the receipt of
the food (as specified in § 1.1345) and application of the kill step.



Covered foods that are changed such that the foods are no longer on the FTL,
provided that records document the receipt (as specified in § 1.1345) of the food to be
changed.



Food that has previously been subjected to a kill step or that has previously been
changed such that the food is no longer on the FTL.



Food that will be subjected to a kill step or will be changed by an entity other than a
retail food establishment, restaurant, or consumer such that the food will no longer be
on the FTL, provided that there is a written agreement between the shipper of the
food and the receiver as specified in § 1.1305.



Produce that is listed as rarely consumed raw in 21 CFR 112.2(a)(1).



Raw bivalve molluscan shellfish that are covered by the requirements of the National
Shellfish Sanitation Program (NSSP), subject to the requirements of part 123, subpart
C, and 21 CFR 1240.60, or covered by a final equivalence determination by FDA.



Persons who manufacture, process, pack, or hold covered foods during or after the
time when the food is within the exclusive jurisdiction of the U.S. Department of
61

Agriculture (USDA) under the Federal Meat Inspection Act (21 U.S.C. 601 et seq.),
the Poultry Products Inspection Act (21 U.S.C. 451 et seq.), or the Egg Products
Inspection Act (21 U.S.C. 1031 et seq.).


Commingled raw agricultural commodities and raw agricultural commodities that
will become commingled, provided that there is written agreement between the
shipper of the food and the receiver as specified in § 1.1305.21



Retail food establishments and restaurants whose average annual monetary value of
food sold or provided during the previous 3-year period is no more than $250,000 (on
a rolling basis), adjusted for inflation using 2020 as the baseline year for calculating
the adjustment.



Retail food establishments or restaurants with respect to covered foods that are
produced on a farm, and both sold and shipped directly to the retail food
establishments or restaurants by the owner, operator, or agent in charge of that farm.22



Either entity when a retail food establishment or restaurant purchases a covered food
from another retail food establishment or restaurant, when such purchases occur on an
ad hoc basis outside of the buyer’s usual purchasing practice (e.g., not pursuant to a
contractual agreement to purchase food from the seller).23

A written agreement must include the effective date, printed names and signatures of the persons entering into the
agreement and the substance of the agreement, and the agreement must be maintained by both parties as long as it is
in effect and must be renewed at least once every 3 years. If registered under section 415 of the Federal Food, Drug,
and Cosmetic Act with respect to the manufacturing, processing, packing, or holding of the applicable foods, such
person must maintain records identifying the immediate previous source of such raw agricultural commodity and the
immediate subsequent recipient of such food. Such records must be maintained for 2 years.
22
The only records retail food establishments or restaurants must maintain in such cases are the name and address of
the source farm. They must maintain such records for 180 days.
23
The buyer must keep a record (e.g., a sales receipt) containing the name of the product purchased, the date of
purchase, and the name and address of the place of purchase.



Farm to school or farm to institution programs, with respect to a food that is produced
on a farm and sold or donated to the school or institution.24



The owner, operator, or agent in charge of a fishing vessel with respect to foods
obtained from the fishing vessel; and any entities that manufacture, process, pack, or
hold the food until such time as it is sold by the owner, operator, or agent in charge of
the fishing vessel.25



Transporters of covered foods.



Nonprofit food establishments.



Persons who manufacture, process, pack, or hold covered foods for personal
consumption.



Persons who hold covered foods on behalf of specific individual consumers, provided
that these persons are not parties to the transaction involving the food they hold and
are not in the business of distributing food.



Food for research or evaluation use, provided that such food is not intended for retail
sale and is not sold or distributed to the public and is accompanied by the statement
“Food for research or evaluation use.”

To estimate the number of domestic covered entities, we use several data sources. These
sources include the 2017 Statistics of U.S. Businesses (SUSB) from the U.S. Census Bureau, the
2017 North American Product Classification System (NAPCS) from the U.S. Census Bureau,
24

The school food authority or relevant food procurement entity must maintain records documenting the name and
address of the farm that was the source of the food. We believe that this is the same location description data
element that is typically stored in distribution and shipping recordkeeping systems. This record must be maintained
for 180 days, which is the same retention period for retail food establishments purchasing foods on the FTL directly
from farms.
25
If registered under section 415 of the Federal Food, Drug, and Cosmetic Act with respect to the manufacturing,
processing, packing, or holding of the applicable food, such person must maintain records identifying the immediate
previous source of such food and the immediate subsequent recipient of such food. The records must be maintained
for 2 years.

and the summary reports of the 2017 Census of Agriculture and the 2018 Census of Aquaculture
from the USDA NASS (Ref. [13] [14]).26 Appendix D of this document discusses our estimates
in detail. We use number of registered facilities from FDA’s Food Facility Registration Module
(FFRM) to estimate the number of foreign facilities affected by this rule and use those estimates
only in the international effects section II.H.2.
The U.S. Census Bureau’s SUSB publishes the number of firms, establishments,
employment by firm size and industry on an annual basis, and annual payroll for most U.S.
business establishments. The most recent data available are from 2017. Many SBA small
business size standards covered in section III are based on the number of employees, so the 2017
SUSB employment size categories are additionally useful for identifying the number of small
entities in each affected industry. The data are tabulated by geographic area, industry, and
employment size of the enterprise. The industry classification is based on 2017 North American
Industry Classification System (NAICS) codes.
We estimate that the final rule will cover approximately 323,872 domestic firms
operating 484,124 establishments. These entities include 11,760 farms, 7,991 manufacturers,
12,007 wholesalers, 2,504 warehouses, 102,424 non-restaurant retailers and 187,185 restaurants.
We estimate that the total number of domestic farms that produce foods on the FTL (including
produce, eggs, and seafood) and will thus be affected by the final rule is 11,760.27 This includes
7,089 produce farms, 95 sprout growers28, 2,521 shell egg farms, and 2,055 aquaculture farms.
These numbers include only domestic entities (firms or establishments) that manufacture,
26

All datasets used in this analysis were the latest available to us as of January 2022.
The first domestic entity that takes physical possession of an imported product will receive the required
information from foreign farms – these firms are affected entities and included within other sector categories
(manufacturers/processors/packers/holders, wholesalers/distributors, or warehouse and storage).
28
In this analysis, we use the inventory of sprout farms and operations used by the FDA’s Office of Regulatory
Affairs. Excluding very small sprout growers, this internal inventory counts 95 sprout growers. The number of
sprout growers is included in the total number of farms in Table 3.
27

process, pack, or hold FTL foods destined for consumption or use in the United States. Table 3
contains a summary and breakdown of by NAICS codes.29
Table 3. Number of Affected Entities by Industry Sector
Number of
Number of
Type
NAICS Codes
Firms
Establishments
Farms /Aquaculture
111219, 111339, 111419, 112310,
11,760
11,796
/ Growers
112511, 112512, 114111, 114112
311340, 311351, 311352, 311411,
Manufacturers /
7,991
8,650
311513, 311710, 311811-311813,
Processors / Packers
311821, 311911, 311941, 311991
Wholesalers /
424410, 424420, 424430, 424450,
12,007
15,101
Distributors
424460, 424480, 424490
Warehouse and
2,504
5,176
493110, 493120, 493130
Storage
445110, 445120, 445220, 445230,
Retail Food
445291, 445292, 445299, 447110,
Establishments
102,424
171,380
452311, 454110, 454210, 722310,
(Non-restaurants)
722320, 722330, 722410, 722514722515
Restaurants

187,185

272,021

Total

323,872

484,124

722511, 722513

As discussed in section II.H, we use FDA’s FFRM biennial registration data for our
estimates of foreign entities affected by this final rule. While the data do not include farms and
retail establishments, we believe the number of foreign retail food establishments affected by this
rule to be small. Thus, we estimate that in addition to entities summarized in table 3, the rule will
cover approximately 61,741 firms and 68,784 foreign establishments. We do not have a detailed
breakdown of foreign firms and establishments by industry, and instead assume the same
proportional breakdown as in the main analysis.

https://www.census.gov/naics/
In each NAICS code category, only entities that manufacture, process, pack, or hold foods on the FTL will be
affected by this rule. Some NAICS codes that were included in the PRIA estimates are now excluded from our
analysis due to updates to the text of the rule between the proposed and final (NAICS 311412, 311421-311423,
311520, 311824, and 311942).

3. Current Industry Practices
For purposes of this analysis, we assume that firms in the food supply system already
adhere to the Subpart J traceability recordkeeping requirements stemming from the BT Act.30
Subpart J requires that non-transporters of food (persons who hold, manufacture, process, pack,
import, receive, or distribute food for purposes other than transportation) maintain records
regarding their receipt and release of food. More limited requirements apply to transporters of
food. In accordance with section 414(b) of the FD&C Act, Subpart J does not apply to many of
the entities covered by this rule, such as farms and restaurants.
Under existing Subpart J requirements (whose coverage, unlike in this rule, applies to all
foods, not just those on the FTL), firms must know and record information regarding the
immediate previous sources of foods and the immediate subsequent recipients of the products
they make or distribute or both (commonly referred to as one-up and one-back recordkeeping).
This entails recording the name, address, and telephone number of source and receiver firms and
of transporters; a description of the type of food; the date it was received or released; the quantity
of the food; and how it is packaged. Firms covered by Subpart J that manufacture, process, or
pack food also must record a lot or code number or any other identifier when available, though
there is no standardized format for either records or identifiers (Ref. [5]).
Additionally, some firms already voluntarily conform to traceability standards and best
practices developed by outside groups, for example, the consensus standards developed by GS1,
an international non-profit organization that develops and maintains standards for barcodes31.

We use the results of the expert elicitation to estimate the portion of entities that would implement traceability
without the rule. We have not received comments on whether compliance with the earlier regulations may increase
as a result of this rule nor on how to quantify the impact.
31
https://www.gs1.org/

Other examples of business communication standards include QR codes32, data matrices, and
radio frequency identification (RFID) codes. At present, the Foodservice GS1 U.S. Standards
Initiative, which promotes traceability standards, has 130 food service companies among its
membership (Ref. [15]). GS1 has developed standards and best practices for various entities in
the food supply system. For example, GS1 standards for farmers include encoding and
communicating trace lot codes, location identification, and other harvest information in order to
link individual cases of product to harvest sites (Ref. [16]). GS1 standards for subsequent supply
chain entities enable enhanced forward and backward traceability between farms and retailers.
When describing traceability practices that were prevalent in the U.S. in the early years of
the first decade of the 2000s, a 2004 report from the USDA Economic Research Service (ERS)
found that private sector food firms had already developed substantial capacity to trace food
products by the time the BT final rule was published (Ref. [17]). According to the report, food
producers, manufacturers, and retailers were typically keeping traceability records for a wide
range of foods and food attributes including elements concerning food safety. Recordkeeping
systems needed for the Subpart J traceability requirements resembled the systems that already
existed for recording receipts and bills. For many of those firms, Subpart J one-up-and-one-back
traceability for a standard set of data elements required little change to their existing systems.
While some entities covered by this rule already maintain, to varying degrees, KDEs
required by this rule and will likely incur little cost to comply, other covered entities might face
more substantial changes to their existing recordkeeping systems. For example, some firms
might need to establish a traceability plan and assign traceability lot codes to FTL food where
these codes do not already exist. Additionally, growers of sprouts, whom we assume to perform

A Quick Response code – so-called QR code – is a machine-readable code consisting of an array of black and
white squares normally for storing smartphone readable URLS or other information.

initial packing of the sprouts, might need to collect additional documentation from seed vendors.
Firms that transform covered products but do not currently link supplier lots (of ingredients) to
manufactured lots (of output) would need to add this step to their recordkeeping process.
In section II.F “Costs of the Rule,” we aim to estimate only the costs that businesses will
incur specifically for compliance with the rule. We therefore estimate the extent to which current
industry practices align with the requirements of the rule using information from the ERG expert
elicitation study (Ref. [4]). ERG asked food industry experts for estimates of the proportion of
industry, by business size, that would have required new traceability-related capital (e.g.,
equipment, software, etc.), training, and recordkeeping specifically for compliance with the
proposed version of the rule, with additional brief definitions of some new CTEs in their draftfinal state at the time of the traceability costs elicitation. We apply these estimates based on the
revised requirements of the final rule.
In consideration of existing baseline trends towards greater traceability for business
purposes, ERG asked experts to consider costs specifically to comply with the proposed rule
(with additional brief definitions of certain CTEs in their draft-final state at the time of the
traceability costs elicitation) in excess of what they spend in the absence of the rule. Reasons for
adopting a traceability system or program vary and may reflect concerns other than food safety.
These decisions of the firm might also depend on private incentives relating to specific attributes
of products, whether they are included in the FTL or not. Therefore, we consider firms affected
by the rule to be a subset of firms who are covered by this rule. We present the estimated number
of entities affected by each provision of the rule in Table 4 below.
We acknowledge the large variability in the characteristics of traceability systems within
and across industries. Due to the large number of NAICS industries relevant to the rule, we

group these industries by similar business activities with respect to the CTEs we have deemed
them likely to perform. ERG elicited incremental estimates of necessary traceability-related
capital, training, and recordkeeping for small and large businesses in each of these broader
groupings of NAICS industries while attempting to account for baseline. Though our resulting
estimates of costs per entity represent averages spanning broad industry and size groups, we
acknowledge that baseline traceability capabilities of individual covered entities, and hence the
degree of changes needed for compliance, vary widely.

Table 4. Entities Incurring Costs Due to Provisions of the Rule
Provision

Entities
incurring onetime costs

Entities
incurring
recurring costs

Firms or
establishments

Firms
323,872
Reading and Understanding the Rule
1
Establishments
17,615
15,854
Capital Investment
Establishments
34,737
26,053
Training
Firms
212,368
§ 1.1315 Traceability Plan
Establishments
95
Seed lot records (Growers of sprouts)2
Establishments
6,058
§ 1.1325 Records of Harvesting
Establishments
3,511
§ 1.1325 Records of Cooling
Establishments
4,218
§ 1.1330 Records of Initial Packing
§ 1.1335 Records of First Land-Based
Establishments
367
Receiving
Establishments
31,434
§ 1.1340 Records of Shipping
Establishments
470,580
§ 1.1345 Records of Receiving
Establishments
8,574
§ 1.1350 Records of Transformation
§ 1.1455(c)(3)(ii) Electronic Sortable
Establishments
75
Spreadsheet Upon Request
1
With respect to Capital Investment and Training, we assume the entities facing recurring costs
to be a subset of the entities facing one-time costs.
2
Although seed lot records fall under § 1.1330 Records of Initial Packing, we assume the
incidence of these costs will fall on growers of sprouts.

E. Benefits of the Rule
We expect the following benefits from this rule:
1) Public health benefits from averted foodborne illnesses and deaths caused by foods
covered by the rule.
2) Benefits from avoiding overly broad recalls following FDA issued public health
advisories.
3) Other benefits discussed qualitatively.
To account for public health benefits of this rule, we adopt the framework developed by
the Institute of Food Technologists (IFT) in their 2012 commissioned report to FDA (Ref. [5]).
We estimate and quantify two types of benefits. These include benefits from averted illnesses
and benefits from preventing overly broad recalls following FDA issued public health
advisories.33 We discuss other benefits qualitatively.
This rule will improve FDA’s ability to: (1) quickly and efficiently trace the movement of
covered foods through the supply chain and (2) identify and remove contaminated food from the
marketplace during an outbreak. In the event of a foodborne outbreak, the ability to trace a food
back through the supply chain from the point of sale or service to a common source is important
for identifying contaminated foods or ingredients and removing such products from the
marketplace to prevent additional illnesses. The ability to trace covered foods forward can help
FDA ensure timely removal of all affected products from the marketplace. It can also help FDA
understand how the distribution of a covered food product relates to illnesses or illness clusters,

A public health advisory is issued for an outbreak investigation that has resulted in specific, actionable steps for
consumers to take to protect themselves.

especially for outbreaks that are challenging to resolve, such as those involving multiple foods
and foods with multiple ingredients.
The mechanisms through which health benefits are realized are through disease outbreaks
averted or reduced-duration foodborne illness outbreaks from covered foods. Benefits from
avoiding overly broad recalls may be realized only when recalls are initiated in response to an
FDA public health advisory. Meanwhile, other benefits may be realized regardless of whether an
outbreak had or had not occurred. 34
In the absence of standardized records, the time needed to identify implicated foods by
linking shipments through the supply chain and back to their sources can be unacceptably long,
leading to a larger and more costly disease outbreak. Yet, for public health benefits to be
realized, the Agency and industry must take timely preventive actions. Executing effective and
timely recall of contaminated foods is important but difficult to achieve (Ref. [18]). For example,
the 2010 shell eggs Salmonella contamination illustrates this point as it shows how conducting a
food recall can be a complex process. According to CDC, the shell eggs outbreak was first
reported in May 2010 and the recall was issued in August 2010. However, the outbreak
continued until October 2010 when all contaminated food vehicles were identified and recalled
(Ref. [19]). Because of the time it took to identify the food vehicle, this outbreak became the
largest reported foodborne disease outbreak since the early 1970s, when outbreak surveillance
was first established (Ref. [20] [21]). With better tracing tools and standardized records, the
illnesses and costs associated with this outbreak could have been mitigated or avoided.35

Costs of the rule will be incurred by all covered entities regardless of whether there is an outbreak investigation or
recall underway, and regardless of whether they are implicated in the outbreak.
35
More details about this outbreak can be found in Appendix F.

1. Public Health Benefits from Averted Illnesses
Public health benefits from this rule are possible only if the following two conditions
hold: (1) a foodborne outbreak occurs and (2) the traceability records required by this rule help
FDA to locate a commercially distributed violative product quickly and accurately to ensure it is
removed from the market. Therefore, primary public health benefits from this rule arise when
foodborne illnesses from covered foods are averted. To assess these benefits, we must first place
a value on risk reduction and health-related costs for illnesses that may be averted. To quantify
health benefits of this rule we estimate the following:
i. Baseline risks of foodborne illness attributable to foods covered by the rule
ii. Economic burden of foodborne illness associated with covered foods
iii. Value of foodborne illnesses reduced from improved traceability
The baseline risk of foodborne illness is a component in estimating the economic burden
of foodborne illness. The value on reducing the risk of foodborne illness through this rule is
expressed as a reduction of the economic burden of foodborne illness associated with covered
foods.

i. Baseline Risk of Foodborne Illnesses Attributable to FTL Foods
As explained in more detail in Appendix B, we estimate that on average 153,807 cases of
foodborne illnesses per year in the U.S. are caused by known pathogens associated with foods
that are covered by this rule. Though we use data from multiple past years to obtain this average,
we do not estimate trends in illnesses, hospitalizations, or deaths. Each year’s numbers reflect
changes in measurement, such as increased testing, improved technologies, and changes to
outbreak surveillance, which would confound estimation of trends in the number and size of

outbreaks. We are also not able to account for likely drivers of coming increases in foodborne
illness, such as new pathogens, new food vehicles, climate change, and density of served
populations. Furthermore, increasing globalization of the food supply and production
concentration continue to create new challenges for detection, investigation, control, and
prevention of foodborne illnesses, making quantification of future foodborne illnesses even more
challenging (Ref. [22]). As such, our constant-trend assumption reflects an analytic limitation
rather than confidence in the constant-trend projection.
From FDA CORE foodborne illness outbreak data (Ref. [23]), we identified a total of
four microbial pathogens responsible for most (98%) of illnesses attributable to foods covered by
this rule. Of these pathogens, E. coli accounts for approximately 68 percent of the foodborne
illnesses, Salmonella accounts for approximately 13 percent, Cyclospora cayetanensis accounts
for approximately 16 percent and Listeria monocytogenes accounts for less than one percent. In
terms of annual hospitalizations, the leading pathogen is Salmonella, which accounts for 57
percent, followed by E. coli, which accounts for 21 percent, Listeria which accounts for 14
percent and Cyclospora which accounts for 6 percent of total annual hospitalizations (see
Table 5). Other pathogens within FDA data that cause foodborne illnesses from FTL foods
include yersinia entorocolitica, hepatitis A, norovirus, campylobacter, and vibrio spp. In
addition to pathogens, naturally occurring chemical contaminants such as ciguatoxin and
scombrotoxin can also cause foodborne illnesses from FTL foods.
The foods covered by this rule have a higher public health risk because of their associated
frequency and severity of illness outbreaks as well as their frequency of consumption.
Table 5 summarizes the estimated number of illnesses, hospitalizations, and deaths
attributable to covered foods. As described in Appendix B, to obtain these numbers, we reviewed
73

FDA’s CORE outbreak data for FTL foods covering the period from January 2009 to December
2020. We used several multipliers from peer-reviewed literature to account for underreporting
and underdiagnoses of foodborne illnesses (Ref. [24] [25]). We use the methodology outlined in
the Scallan et al. follow-up article to account for unspecified and unknown agents36 (Ref. [26]).
Based on this article, the estimated number of annual illnesses in
Table 5 (153,807) before adjusting it for the number of illnesses from unspecified and
unknown agents constitute only 20 percent of all illnesses. Therefore, the total number of
illnesses from known toxins and pathogens is 153,807 and from unspecified and unknown agents
is 615,226 (= 153,807 x (1-0.2)/0.2), which yields a total of 769,033 illnesses (= 153,807 +
615,226).37
Table 5. Estimated Baseline Illnesses, Hospitalizations, and Deaths Attributable to Foods
Covered by the Final Rule38
Annual
Percent of
Type of Pathogen
Hospitalization Deaths
Estimates
total cases
Campylobacter
63
0
0
0.04
Ciguatoxin
31
0
0.0
0.02
Cyclospora cayetanensis
25,332
20
0.0
16.47
E. coli (STEC) O157
55,074
66
1.7
35.81
E-coli (STEC) non-O157
50,156
10
0.0
32.61
Hepatitis A Virus
13
1
0.3
0.01
Listeria Monocytogenes
62
52
9.7
0.04
Norovirus
803
1
0.0
0.52
Salmonella typhoidal
200
3
0.0
0.13
Salmonella non-typhoidal
20,005
207
2.3
13.01
Scombrotoxin
107
0
0.0
0.07
Vibrio-parahaemolyticus
1,279
2
0.0
0.83
36

As explained in Appendix B, according to Scallan et al., (2011b), apart from foodborne illnesses caused by major
known pathogens, nearly 80% of additional episodes of foodborne illness are caused by unspecified agents,
including known agents about which we lack sufficient data to estimate agent-specific illness. There are also
illnesses caused by known agents that are not yet recognized as causing foodborne illness as well as substances
known to be in food but of unproven pathogenicity, and unknown agents.
37
Figures are rounded to nearest whole number
38
This table is compiled using FDA’s foodborne illnesses outbreak data that was also used in the risk-ranking
model, which was the basis for designating the FTL.

Vibrio-Cholerae
Yersinia enterocolitica
Estimated total from
known pathogens39

36
63

0
0

0.0
0

0.02
0.42

153,807

364

100

ii. Economic Burden of Foodborne Illnesses Associated with FTL Foods
We estimate the total burden of foodborne illnesses attributed to FTL foods by
multiplying the estimated annual number of illnesses per pathogen from
Table 5 by the updated burden of illness estimates. We update burden of illness estimates first
published in 2015 (Ref. [27]) to 2020-dollar values40 (Ref. [28]). This burden includes both direct
costs and indirect costs, and accounts for variations in the level of severity of foodborne
illnesses. The direct costs are associated with doctor visits and hospitalization. Indirect costs are
from the loss in quality of life (of which loss in productivity is a subset) because of the
symptoms and severity of the foodborne illness. The burden is monetized using the value of a
statistical life (VSL) as provided in HHS guidelines (Ref. [29]). The total economic burden of
illness is therefore estimated by computing and combining for each illness the average monetized
acute health loss, the average monetized secondary health loss (from long-term health effects),
the average monetized loss of life years, and the acute and secondary medical costs.41 We rely on

Our estimates of hospitalization and death counts include specific pathogen multipliers according to Scallan’s
methodology. Except for Norovirus, the Scallan methodology doubles both death and hospitalization numbers,
which we accounted for in this table. Chemicals or substances like ciguatoxins, scombrotoxins and tetrodotoxins
were not available in the original Scallan (2011a) article but have since been established as causes of foodborne
illnesses. We obtain underdiagnosis multipliers for ciguatoxins and scombrotoxins from Pennotti et al., (2013). We
assume their hospitalization and death rates were constant.
40
The model updates included revised annual dollar estimates of foodborne illness costs from the 2015 estimates to
2020 values using BLS deflators. Additionally, new foodborne illness causing agents, which include scombrotoxin
fish poisoning and ciguatera fish poisoning, were added into the model. For these multipliers, we cite Pennotti et al.
(2013) [25] because they were not included in Scallan et al. (2011a) [24].
41
Minor et al., (2015) [27] present their estimates of cost per illness using illness data from Scallan et al., (2011a)
[24] which uses data from 2000 to 2008 from several sources. By contrast, the estimates in this RIA are based on
FDA CORE data from 2009 to 2020. While the Scallan et al., (2011a), primary estimate of 11,407 cases annually
across all food sources is far lower than what is attributed here just to FTL products, Tack et al. (2020) [52] indicate

these estimates for our analysis and refer the reader to Minor, et al., 2015 for more detailed
discussion of these computations.
Table 6 shows the estimated burden of illnesses associated with outbreaks attributable to
foods covered by this rule. We list the common microbial pathogens associated with covered
foods and the estimated average annual number of illnesses associated with these pathogens. The
number of illnesses per pathogen is then multiplied by its expected burden of illness to produce
the economic burden of foodborne illnesses caused by each pathogen. We estimate the economic
burden of foodborne illnesses associated with foods covered by this final rule as approximately
$5.8 billion dollars per year (Table 6). More detailed calculations of the estimated range of the
economic burden and cost per illness can be found in table B2 of Appendix B.
Table 6. Estimated Economic Burden of Foodborne Illnesses Associated with Foods
Covered by this Final Rule (2020$)
Estimated Annual
Illnesses

Monetized Burden
per Illness

Campylobacter

$4,748

$300

Ciguatoxin

$31,402

$985

Cyclospora cayetanensis

25,332

$4,451

$112,751

E. coli (STEC) O157

55,074

$13,757

$757,657

E-Coli (STEC) non-O157

50,156

$2,506

$125,691

Hepatitis A Virus

$58,440

$780

Listeria Monocytogenes

$1,987,005

$123,774

Norovirus

803

$487

$391

Salmonella typhoidal

200

$7,116

$1,420

20,005

$7,248

$144,993

107

$548

$59

$1,675

$61

Pathogen

Salmonella non-typhoidal
Scombrotoxin
Vibrio-Cholerae

Total:
Primary ($1,000)

that compared with 2016-2018, the incidence of Cyclospora increased significantly (1,209%). This increase has
been seen in previous years as well – CDC notes that the incidence of Cyclospora infections increased markedly in
2018, in part because of large outbreaks associated with produce
(https://www.cdc.gov/mmwr/volumes/68/wr/mm6816a2.htm?s_cid=mm6816a2_w). The increase in Cyclospora
infections might be partly due to increased detection (by more labs using new tests) but also partly due to increased
exposure to this pathogen, particularly contaminated FTL products.

Yersinia enterocolitica
Campylobacter
Subtotal Illnesses from Known
Pathogens
Adjusting Subtotal for
Unidentified/Unspecified Pathogens
Total Illnesses

645

$6,255

$4,033

$4,748

$300

153,807
615,226
769,033

$1,276,266
$5,105,063
$6,381,329

iii. Value of foodborne illnesses reduced from improved traceability.
We estimate the value of foodborne illnesses reduced from improved traceability (i.e.,
public health benefits) using the model provided in the IFT report (Ref. [5]). In describing public
health benefits related to tracing, the IFT report presents an analysis based on eight outbreaks.
Seven of the outbreaks were from Salmonella infections and one was from Listeria
monocytogenes. Each outbreak provided information on 1) the pathogen associated with the
outbreak; 2) the investigation description; 3) the potential improvement from the estimated date
of the initiation of traceback to the estimated date of recall or other intervention; and 4) total
illnesses and deaths for the duration of the outbreak.
Table 7 below shows the estimated percentage of illnesses prevented assuming 100%
product tracing improvement (a hypothetical maximum of instantaneous traceability) during the
investigation of these foodborne outbreaks. Table 7 includes results from four of the eight case
studies from the IFT report (excluding two involving foods not on the FTL). It also includes 17
case studies using epidemic curve data from CDC and investigation and intervention data from
FDA as explained in Appendix C of this analysis. All cases used in this analysis cover outbreaks

associated with four pathogens: Cyclospora, E. coli (STEC), Listeria monocytogenes and nontyphoidal Salmonella.42
Appendix S of the IFT report describes in detail the analytical process and the
applicability of the analysis43 that we use to estimate the percentage of illnesses that are
potentially preventable with the tracing requirements of this rule. We use the same process in
estimating the percentage of illnesses potentially prevented assuming FDA had 100% tracing
improvement (i.e., instantaneous product tracing) resulting from the traceability recordkeeping
requirements from this rule. However, FDA outbreak investigation processes and outbreak data
collection have changed since the IFT report. In 2011 the CORE Network was created with the
purpose of providing a structured process for responding to an outbreak which includes an
outbreak response phase that centers on traceback of product, removal of product from the
marketplace, and investigation of how the outbreak may have occurred. Due to this established
framework, the best consistent date range for post-2011 traceback investigations is the initiation
and completion dates of CORE traceability activities, as described in Appendix C.44 We
therefore use “initiation” and “completion” dates provided by CORE in estimating the
percentage of illnesses potentially prevented for post-2011 outbreaks. We extend the IFT
analysis by including additional pathogens and using multiple case studies per pathogen to

FTL associated outbreaks caused by these four pathogens represent about 98% of all FTL associated illnesses.
In Appendix S of the IFT report, the estimated number of reduced illnesses potentially prevented is calculated by
using the epidemic curve data for each associated outbreak. Over the outbreak timeline, the IFT report estimates the
number of days (and illnesses) between the initiation of the traceback and the initial or final intervention date
(depending on the outbreak). The number of illnesses over the time period is divided by the total number of illnesses
during the outbreak to obtain the ratio of illnesses potentially prevented assuming 100% tracing improvement (i.e.,
instantaneous tracing).
44
Since each outbreak presents unique circumstances, such as availability of product on the market to recall and the
potential for multiple sequential recalls during one outbreak, using the initial date of recall may not represent the
best end date to represent the end of traceability activities. The CORE traceback initiation date represents a point in
time when traceability activities began, and the CORE traceback completion date represents a point in time in which
the traceback activities and interventions such as a recall have ended.
43

estimate the average percentage of preventable illnesses by pathogen. Minimum and maximum
preventable illnesses in Table 7 represent variable potential impact of traceability among case
studies involving the same pathogen.

Table 7. Estimated Percentage of FTL Associated Illnesses Preventable with Product
Tracing Improvement.
Year

Commodity

2008

Hot Peppers

2008

Cantaloupe

2009

Alfalfa
Sprouts

2010

Shell eggs

20082009

Peanut Butter
and peanut
butter products

2018

Shell eggs

2018

Tahini

2019

Ground Tuna

2019

Tahini

2019

Cantaloupe

Percentage range of
cases prevented for
Salmonella

2012
2016

Spinach

Alfalfa sprouts
Romaine
2018
Lettuce
Romaine
2019
Lettuce
Percentage range of
cases prevented for
E.coli O157:H7

Total
Illnesses per
Epidemic
Curve

Preventable
illnesses

Percentage
of Illnesses
That Are
Preventable

1,442

790

55%

235

31%

3,578

120

636

188

30%

CDC (1)

24%

Appendix C

25%

Appendix C

43%

Appendix C

17%

Appendix C

163

15%

Appendix C

618
6
3,578

122
1
790

24%
2%
55%

E. coli O157: H7

14%

Appendix C

E. coli O157: H7

Appendix C

E. coli O157: H7

Appendix C

E. coli O157: H7

167

17%

Appendix C

Average
Minimum
Maximum

68
11
167

9
1
28

11%
3%
17%

Pathogen
Salmonella
Saintpaul
Salmonella
Litchfield
Salmonella
Saintpaul
Salmonella
Enteritidis
Salmonella
Typhimurium
Salmonella
Braenderup
Salmonella
Concord
Salmonella
Newport
Salmonella
Concord
Salmonella
Javiana
Average
Minimum
Maximum

Source

IFT report,
Table 48 and
Appendix C

Year

Commodity

Pathogen

Total
Illnesses per
Epidemic
Curve

2012

Clover sprouts
Romaine
Lettuce

E. coli O26

34%

Appendix C

E. coli O145

Appendix C

12%

29
26
33

5
0
10

16%
0%
34%

139

50%

Appendix C

63%

Appendix C

74
8
139

37
5
69

56%
50%
63%

241

Appendix C

631

146

23%

Appendix C

436
241
631

78
9
146

13%
4%
23%

2010
2019

Ground Bison

Percentage range of
cases prevented for
other E. coli
2011

Cantaloupe

Hard boiled
eggs
Percentage range of
cases prevented for
Listeria
monocytogenes
2019

2019

Basil

2013

Leafy Greens,
Cilantro

Percentage range of
cases prevented for
Cyclospora

E. coli O121:H19;
O103:H2
Average
Minimum
Maximum
Listeria
monocytogenes
Listeria
monocytogenes
Average
Minimum
Maximum
Cyclospora
cayetanensis
Cyclospora
cayetanensis
Average
Minimum
Maximum

Preventable
illnesses

Percentage
of Illnesses
That Are
Preventable

Source

Appendix C

(1) https://www.cdc.gov/salmonella/2009/peanut-butter-2008-2009.html
After estimating the number of illnesses that may be prevented with better tracing, we
then multiply the percentage range of preventable illnesses from Table 7 by the estimated
number of annual illnesses for each pathogen in Table 8. The numbers of annual illnesses below
are from the baseline number of illnesses as described in Appendix B, adjusted for unspecified,
underreported, and undiagnosed illnesses (Ref. [24] [26]). We believe that accounting for such
cases is critical because not all illnesses caused by outbreaks are ultimately documented and
attributed to those outbreaks. This approach is consistent with FDA’s past regulatory impact
analyses.

We make one more adjustment to our estimates to account for certain establishments that
are now fully exempt from this rule. As mentioned in Section 2.D.2 Coverage of the Rule, retail
food establishments and restaurants with under $250,000 in annual sales and farms with less than
$25,000 in annual sales are fully exempt from this rule. We adjust our benefits estimates to
account for this exemption assuming that the reduction in illnesses averted, and therefore the
reduction in benefits is commensurate in proportion with food revenues by these exempt entities.
We use 2017 Statistics of U.S. Businesses (SUSB) data from the U.S. Census and estimate that
about 2% of revenues would correspond to 2% reduction in cases averted. This means that by
exempting these entities, health benefits from this rule are slightly less than benefits if the rule
had no exemptions. To account for this reduction, the number of annual illnesses in Table 8
represent 98% (= 100% - 2%) of all annual illnesses.
Table 8. Estimated Annual Cases of Foodborne Illness That Are Preventable with Product
Tracing Improvement
Estimated Annual Preventable Illnesses
Annual
Pathogen
Illnesses
Primary
Minimum
Maximum
24,838
3,337
928
5,747
Cyclospora cayetanensis
49,179
7,740
0
16,958
E. coli (STEC) non-O157
54,001
5,782
1,714
9,054
E. coli (STEC) O157
61
34
30
38
Listeria monocytogenes
19,615
4,805
370
10,746
Salmonella (non-typhoidal)
147,694
21,598
3,042
42,543
Subtotal
36
590,775
86,390
12,169
170,174
Unspecified unknowns
738,469
107,988
15,211
212,717
Total
In Table 9, we estimate the burden of foodborne illnesses attributed to FTL foods by
multiplying the estimated total annual number of illnesses from Table 8 by the weighted average
burden per illness (based on the prevalence of preventable illnesses related to each pathogen).
We use the same burden of illness estimates for the selected pathogens as in section II.E.1.vi.

Hence, the columns in Table 9 rely on the primary, minimum, and maximum possible burden of
illness for the selected pathogens from our burden-of-illness model. 45
Table 9. Annual Benefits Based on 83% Improved Product Tracing Time
Pathogen

Annual Undiscounted Benefits from Tracing Time Reduction
by 83% to 6 Days (Based on Weighted Burden per Illness) (1)(2)
Primary
Minimum
Maximum

Weighted burden per illness for
four pathogens

$10,020

$5,377

$14,586

Total annual benefit from faster
tracing

$896,548,792

$67,775,664

$2,570,829,560

Annualized Present Value
Benefit (7%, 20 years)

$780,498,704

$59,002,721

$2,238,059,052

Annualized Present Value
Benefit (3%, 20 years)

$809,640,406

$61,205,722

$2,321,622,098

Present Value Benefit (7%, 20
years)

$8,268,614,390

$625,075,667

$23,710,029,480

Present Value Benefit (3%, 20
years)

$12,045,404,78
3

$910,586,595

$34,539,874,403

(1) The estimated range of values (primary, minimum, and maximum) represent the variability in the valuation
of illness per pathogen.

(2) Foodborne Illnesses caused by Cyclospora Cayetanensis, E. coli (STEC), Listeria monocytogenes, and
Salmonella (non-typhoidal).

We estimate the percent improved traceback time resulting from better tracing
requirements using information provided by FDA’s CORE that includes a case study from the
2019 E. coli Romaine lettuce outbreak (Ref. [30]). The outbreak illustrates the difference in time
to identify implicated farms from points of sale (POS) where lot codes were available versus
POS where lot codes were not available. For POS where lot codes were available because
product packaging from a sample that tested positive and matched the outbreak strain was

For estimating discounted costs and benefits, with a 3-year effective compliance date, we assume the publication
date of the rule is year 0 and that one half of first year costs will be incurred in year 1 (which is two years before the
3-year effective date) and that the other half of first year’s costs will be incurred in year 2. Therefore, we assume
half of benefits will begin to accrue on year 2 (one year after half the requirements were implemented in year 1) and
full annual benefits will begin to accrue on year 3 (two years after requirements would have been implemented).

available from an ill consumer, farms were identified within 24 hours compared to 29 days for
those where no lot code information was available (1 day over 29 days would represent a 96
percent improvement). Although every case is unique, this provides an example of how the
availability of lot code information at the POS could significantly shorten the time in
determining the source of the contaminated product. Given FDA CORE’s combined years of
experience in conducting traceback investigations associated with foodborne outbreaks at a
national level, it is FDA CORE’s expert judgment that access to lot codes, traceability lot code
sources, and other key data elements throughout the supply chain would likely enable FDA to
identify common product sources in about five to seven days, for an average of six days (Ref.
[30]).46 Given that product packaging is often discarded by consumers and not available to
outbreak investigators, the five to seven days estimate assumes that the product package would
not be available. We use this information to estimate the resulting percent improvement
(reduction) based on the median number of days to reach maximum improvement from our
sample outbreaks.
The average number of days used for identifying a product source without lot codes is
about 35 (ranging from about 0 to 84),47 whereas the average number of days needed to identify
a product source from a product with lot codes is six days (ranging between five and seven days).
We estimate that the percent improvement that would result from identifying common product

This time period does not account for the time needed for the epidemiologic information (i.e., food exposures)
provided by public health officials to identify POS clusters and be provided to FDA for tracing. Additionally, this
timeframe may vary depending on the complexity of a food’s supply chain. For example, if a food is transformed
multiple times before it reaches an RFE, more time may be needed to identify source information. However, we
account for this in our analysis of outbreak examples as we used the CORE traceback initiation date which
represents a clear point in time when traceability activities began.
47
We estimate the average number of days used for identifying a product source without lot codes by taking the
average difference between FDA traceback completion date and FDA traceback initiation date of the 23 outbreaks
identified in Table C.1 of Appendix C.

sources in five to seven days would range between 80% to 86%. We use the middle estimate of
six days which is equivalent to an 83% improvement (Table 9).
We estimate that corresponding (undiscounted) public health benefits would range
between $68 million and $2.6 billion per year with a primary estimate of $897 million per year
(Table 9). The present value of health benefits with seven percent discounting over 20 years
(Table 9) ranges from about $625 million to $23.7 billion, with a primary estimate of $8.3
billion. The present value of health benefits with three percent discounting over 20 years (Table
9) ranges from about $911 million to $34.5 billion, with a primary estimate of $12.0 billion. At a
seven percent discount rate over twenty years, the annualized monetized health benefits of the
rule range from $59 million to $2.2 billion with a primary estimate of $780 million. At a three
percent discount rate over twenty years, the annualized monetized health benefits range from $61
million to $2.3 billion with a primary estimate of $810 million.
These benefits are slightly underestimated as the annual dollar value of the burden
associated with outbreaks caused by these four pathogens represents about 99% of the total
annual burden of all FTL associated illnesses. These benefits may also be overestimated due to
uncertainty in adjustments accounting for under-reported, undiagnosed, and unspecified illnesses
from all FTL-associated illnesses.

2. Benefits from Avoiding Overly Broad Recalls Following FDA Issued Public Health Advisories
In addition to the public health benefits discussed above, implementation of more precise
food recalls due to improved tracing may result in social benefits realized by avoiding overly
broad recalls when contaminated FTL foods covered by the rule are identified. Overly broad
recalls may occur when the source of an outbreak cannot be promptly identified or is originally

misidentified, so that the recall extends to product lots and products beyond the implicated
product.
Market withdrawals and recalls are expensive and commercial distribution of
contaminated food can result in economic harm to consumers. In the event of a food recall or
market withdrawal, the records required by the regulation may help us to more quickly and
accurately locate a violative product that was commercially distributed, which could reduce the
likelihood of conducting an overly broad recall. Costs of conducting a recall or market
withdrawal include lost sales (lost retail value of product), expenses associated with notifying
retailers and consumers, collection and shipping costs, disposal costs, and legal costs, among
others.48 In addition to costs of conducting a recall, aggregate costs include spillover costs to
shareholders, competitors, wholesalers, retailers, and customers. While well-established, profitmaximizing food manufacturers and distributors and retailers may be able to consider in their
decisions the costs associated with recalling a product beyond the value of recalled units to
include expenses associated with notifying retailers and consumers, collection, shipping, disposal
and legal costs, there are spillover or negative externalities associated with a recalled product
that may be larger in the aggregate than the losses of the recalled product to the producer.49
Although recall of rightly implicated foods is necessary and costly, overly broad recalls
that involve entire industries or loosely or unrelated products can be extremely expensive.
According to a survey of companies conducted by the Grocery Manufacturers Association

One of the steps of a recall process involves disposing or destroying the recalled food product. While it may be
possible for some companies to recover costs by repurposing their recalled products into pet or animal feed or even
fertilizer, this practice is more common with meat producers. In the wake of foodborne illness outbreak-related
recalls, repurposing or diverting recalled foods to recover losses is not a conventional response within our review of
case studies and pilot projects. To the extent any repurposing does occur, the overall costs from lost retail sales
would be defrayed by the value of the repurposed products.
49
Jarrel, Gregg; Peltzman, Sam; The Impact of Product Recalls on the Wealth of Sellers. Journal of Political
Economy, Vol. 93, No. 3 (1985), pp 512-536.

(GMA), 77 percent of respondents that faced a recall in the past five years estimated the financial
impact of the recall to their company to be up to $30 million, with 23 percent reporting even
higher costs (Ref. [21]). The GMA study suggests that the average cost of recalls to their
members is about $2 billion over a five-year period or an annual loss of about $400 million.
These costs represent an estimated fraction of 0.3 percent of the GMA sample’s annual revenues
and result from business interruption, product disposal costs, customer reimbursement,
transportation, investigation, external professional fees, sanitizing production facilities,
warehouse costs, decreased sales of the brand name product identified, internal time, and other
expenses. This final rule may help ensure recalls are conducted in a more precise manner and
unnecessary costs to both industry and consumers are mitigated.
In cases when the firm cannot be identified in a timely manner, FDA is more likely to
issue an advisory, which has much greater potential spillover effects than a manufacturerinitiated recall. Rolling recalls can also occur, such as when a recall continues to grow because
implicated ingredients are found in other products, as with the peanut butter recall in 2009.
To inform our analysis of the benefits of avoiding overly broad recalls, ERG completed a
literature review and a recall elicitation of industry experts in December 2021 and January 2022
(Ref. [4]). One purpose of the literature review was to find data on the three overly broad recalls
presented in Appendix F and identify other possible case studies to help structure the expert
elicitation instrument for developing data on the spillover effects from an overly broad recall.50
We present the case study data summarized by ERG in

Case studies of overly broad recalls were used to help experts consider how more-targeted recalls might affect
some of the identified case study costs.

Table 10 below.

Table 10. Findings of Recalls Case Studies
Recall

Data51

Source

2018 Romaine
Recall



Romaine growers who sell through spot sales lost $750,000
due to the advisory



Romaine growers who sell through contracts gained
$690,000 due to the advisory

Kiesel et
al. (2021)
(Ref. [31])



Processors and shippers lost $23.8 million in pulling
harvested lettuce from the supply chain, and $43.1 million
from lettuce under their control that could not be sold,
totaling $60.9 million



Grocery retailers lost $21.2 million from pulling product and
$8.4 million from price changes, totaling $29.7 million



Food service operators lost $4.2 million from an inability to
sell but gained $2.8 million from reduced acquisition costs.
Food service operators lost a total of $1.4 million



Total societal loss (including loss on the parts of consumers,
damages to suppliers of labor and materials, employment
reduction in related industries, etc.) was $320-$400 million



Large, publicly owned distributors, retail outlets, and
restaurant chains saw no drop in stock value during the
advisory



National Restaurant Association claimed members lost $130
million



Florida tomato industry estimated loss of $660 million
(including $50 million of tomatoes from Florida removed
from the distribution system)



Over 33,000 farm workers hired daily to pick tomatoes in
Florida lost work



One Texas distributor estimated $660,000 in their own
company’s losses



Georgia claimed losses of $11 million



California lost $400,000 in the destruction of good product;



$1.3 million in product sales following the announcement of
the advisory; and

2008 Tomato
Recall

Meyerson
(2009)
(Ref. [32])

Kawamura
(2008)
(Ref. [33])

All estimates have been adjusted to December 2021 dollars (BLS, 2022).



$26-$32 million in indirect losses due to low demand and
poor prices over time



U.S. tomato farms may have lost $33 million



Total societal loss is estimated at $330 million

Palma et
al. (2010)
(Ref. [34])
Hussain
and
Dawson
(2013)
(Ref. [35])

In the literature review, ERG found that while much has been written qualitatively to
address the benefits of traceability in food production and recall processes, little has been done to
quantify those effects and that even less has been done to assess the spillover costs of broad
recalls. Existing research, as presented in

Table 10, confirms that broad recalls can cost society hundreds of millions of dollars,
with economic harm felt along entire supply chains. The recall elicitation (designed to develop
estimates of the reduction in spillover costs) provided quantitative information on the potential
effects of traceability on food recalls due to FDA advisories and also insights on how firms react
under different recall scenarios.
Experts made a noteworthy distinction between two scenarios in which an overly broad
recall can occur:


The first scenario is when a recalling firm or its downstream customers cannot
definitively track the movement of outbreak causing recalled product and
therefore recall additional products as a “buffer”. In this situation, when the
recalling firm is known, some of these costs through the supply chain may be
partially or wholly reimbursed by the firm (depending on existing contracts).



Under the second scenario firms recall products following FDA-issued public
health advisories about products when no one can identify the specific firm,
brand, or production dates of the product that is causing a foodborne illness
outbreak. A public health advisory is issued for an outbreak investigation that has
resulted in specific, actionable steps for consumers to take to protect themselves.

Under both scenarios the source of an outbreak cannot be promptly identified or is
originally misidentified, so that the recall extends to product lots and products beyond the
implicated product. The main difference is that under the first scenario some of these “spillover”
costs are internalized by the recalling firm, and under the second scenario these “spillover” costs
are borne by others in society (including processors, distributors, retailers, consumers, and
producers other than the implicated producers). In summary much greater spillover impacts

occur when an FDA advisory is issued and an implicated firm is not identified. Under the second
scenario, firms may not be able to recover their losses. Generally, the faster a recall is
completed, the less it costs, and fewer firms will be affected.
Another insight by experts is that by requiring firms to implement traceability for FTL
foods, this rule will reduce the scope of advisories and possibly even eliminate the need for some
advisories. We therefore identify as a benefit of this rule the reduction in spillover costs incurred
by manufacturers, producers, distributors and retailers from inadvertently being part of a broad
scope or undifferentiated product recall following an FDA advisory. The benefits of avoiding
overly broad recalls following an FDA advisory is the reduction in spillover costs.
We estimate the reduction in spillover costs as the difference between total costs incurred
by firms affected by an FDA advisory and total costs to firms affected under a more targeted
recall.

Costs per Firm
During the recall elicitation, ERG asked experts to consider the overly broad recalls that

occur when an FDA advisory has been issued but no firm has been identified yet as the source of
the problem. As presented below, they were asked to estimate the per firm costs incurred by type
of entity or industry category (e.g., processor, distributor, retailer, etc.), to respond to an FDA
advisory that implicates a food from FDA’s food traceability list.
Keeping in mind that answers to questions about labor and non-labor costs often include
some variability due to different factors (firm’s response to FDA advisories, size of operations,
type of product, variation in FDA advisories, etc.), ERG therefore asked for ranges of estimates

(low to high) and average that reflect that variability for each industry category. ERG also asked
that they explain how they arrived at their answers (for example, to describe the scenario that
they were thinking about in generating their answers, how they estimated the low, most likely,
and high estimates, etc.)52. They were also asked to note any costs that were not specifically
captured in the recall elicitation. All estimated per firm costs from the expert elicitation include
labor and non-labor costs by industry category affected by an FDA advisory.
Assumptions
Industry categories used by experts included 1) producer/processor, 2) shipper distributor
3) non-restaurant retailers and 4) restaurants. For context, the term producer/processor also
applied to manufacturers, farms, packers and growers. The responses under the category for
shipper/distributor also applies to wholesalers and warehouses. For the term “non restaurant
retailer” experts assume the term includes mostly grocery stores, specialized grocery stores (fish
markets, fruit market, bakeries, etc.) supercenters and club stores.
In describing low and high costs estimates, experts treated the range for low and highcost estimates as costs incurred by small and large firms respectively. Experts participating in
the recall elicitation did not necessarily use SBA definitions when referring to certain entities as
small. The assumptions from experts on the criteria of low, most likely, and high estimates
varied among experts and also varied considerably by industry. For small firms across different
industry categories, experts were generally thinking about single establishment firms or
independent operators (not part of a chain). For large firms some experts considered firms with
50 or more establishments. Other experts provided specific size breakdowns by industry
category. For producers/processors some considered their low to high estimates applied to one
52

Although experts provided low, most likely and high estimates, they didn’t include their rationale for the most
likely estimate.

establishment firm to firms with two or more manufacturing facilities or with annual revenues in
excess of $250 million. For restaurant and retail firms, some experts considered them small if
they had at most two to three establishments with two or three employees per establishment.
Other experts considered small regional store chains with two to three employees per
establishment as small. For restaurants, some experts considered small firms could range from
being a single establishment to a chain with 2 to 3 sites while others consider a small firm from
being a single establishment to a chain with up to 150 locations. For non-restaurant firms experts
considered the range of small to large firms from respectively having 50 or fewer locations to
more than 1,000 locations.
For the most likely estimate some experts considered costs to be also small because the
average number of locations included in a firm is skewed toward small firms. Others considered
that under this cost category, the firms would be part of a regional chain.
We describe estimates for labor, non-labor and total costs per firm by industry category
below.
Labor hours
From the recall expert elicitation by ERG experts provided estimated ranges and average
labor hours expended per firm by type of labor category and type of firm when responding to an
FDA advisory in which:
•

No firm(s) has/have been identified as the source of the problem.

•

The implicated product is on the FTL food list (e.g., romaine lettuce or peanut butter).

The type of labor hours is broken down by industry category: producers/processors,
shippers/distributors, restaurants and non-restaurant retailers. Depending on the industry
category, the type of labor hours may include costs to identify, remove, quarantine, store and
93

dispose product, and resupply customers. Labor hours are also broken down by labor type
including executive management, mid-level management, legal, administrative, expert
consultant, and production hours. Table 11 provides a general description of the types of costs
and labor incurred by industry category.
Table 11. Type of Labor and Wages by Industry and Occupation Expended During an
FDA Advisory
Industry
Type of cost
Type of labor (2)
category
Incur costs to:
- identify product to be removed
- if large, contact each processing
facility
- public relations/monitor public
Executive management $110.14
response
Midlevel management: $59.93
-meet and determine public
Legal $80.39
Producers/
response
Production labor hours: $17.32
processors
- contact distributors/retailers
Administrative labor hours $20.16
- hire legal experts and other expert
Expert consultant: $75.50
consultants
Other: $32.47 (1)
- quarantine and store or dispose of
product
- resupply customers
- effectiveness checks
Executive management $105.64
Incur costs to:
Midlevel management: $56.56
- identify product to be removed
Legal $74.42
- clean and sanitize if bulk product
Shippers/
is affected
Production labor hours: $18.41
distributors
- quarantine and store or dispose of
Administrative labor hours $22.90
product
Expert consultant: $75.50
- resupply customers
Other: $32.62 (1)
Incur costs to:
Executive management $74.58
-identify product to be removed
Midlevel management: $40.75
-if large, contact each restaurant
Legal $74.42
- public relations/monitor public
Restaurants
Production labor hours: $14.30
response
Administrative labor hours $15.44
-remove product from inventory
Expert consultant: $75.50
-quarantine and store or dispose of
Other: $23.50 (1)
product

Nonrestaurant
retailers

Incur costs to:
- identify product to be removed
- if large, contact each retailer
- remove product from shelves and
storage
-restock product
- quarantine and store or dispose of
product
- public relations/monitor public
response

Executive management $85.99
Midlevel management: $40.75
Legal $74.42
Production labor hours: $16.60
Administrative labor hours $16.49
Expert consultant: $75.50
Other: $24.61(1)

(1) Assume "other" labor hours are a blend of administrative, production, and midlevel
management labor hours
(2) Source for wages: https://www.bls.gov/oes/current/oessrci.htm
From the range of labor cost responses from all nine experts, presented through the
columns in Table 12, firms in the producer/processor category lead in labor cost shares at every
estimate level and restaurants appear to have the smallest share of labor costs. According to
some experts, managerial and administrative costs are largely fixed across scale of operations,
but production hours may vary the most according to the size of the operation. Spillover costs
can include labor costs from meetings to determine response, costs to notify downstream entities
and the public, costs to hire legal and food safety experts, identification and disposition of
product, and resupplying customers.
Table 12. Labor Costs of Advisory Event per Firm ($1,000).(1)
Cost Range by Industry
Producer/ Processor
Low estimate
Most likely estimate
High estimate
Shipper/ Distributor
Low estimate
Most likely estimate
High estimate
Restaurant
Low estimate

Min

Median

Average

Max

$3
$9
$15

$11
$37
$59

$30
$76
$1,486

$90
$242
$11,585

$1
$1
$2

$8
$17
$25

$9
$45
$511

$23
$230
$2,230

$0.1

$35
95

Most likely estimate
High estimate
Non-Restaurant Retailer
Low estimate
Most likely estimate
High estimate

$0.1
$1

$19
$29

$36
$232

$114
$1,259

$2
$5
$8

$7
$22
$33

$11
$34
$558

$30
$82
$4,086

Total labor cost per firm per advisory (1)
Low estimate
$9
$34
$57
$166
Most likely estimate
$24
$94
$183
$474
High estimate
$38
$148
$2,700
$13,815
(1)
The total labor cost per firm per advisory is the range of the sum of individual expert estimates
by industry category and is not additive.

Non-labor hours
Experts also provided estimates of non-labor costs expended per firm by type of nonlabor costs and type of firm when responding to an FDA advisory in which:


No firm(s) has/have been identified as the source of the problem.



The implicated product is on the FTL food list (e.g., romaine lettuce or peanut butter).

Similar to labor costs, they are broken down by industry (Table 13).
Table 13. Type of Non-Labor Costs by Industry Type Expended During an FDA Advisory
Industry type

Producers/ processors

Shippers/ distributors

Type of cost
Incur costs:
- for lost sales
- product destruction
- product restocking
- product storage
- public relations and advertising costs
Incur costs:
- for lost sales
- product destruction
- product restocking
- product storage

Restaurants

Non-restaurant retailers

Incur costs:
- for lost sales
- product destruction
- product restocking
- product storage
Incur costs:
- for lost sales
- product destruction
- product restocking
- public relations and advertising costs

The estimated range of non-labor costs varies even more widely than labor costs. The
share of costs appears to be at the producer/processor level and disproportionately much larger
than other industry categories. For example, reduced demand for products (due to the recall)
may result in growers having to plow product under. One possible explanation is that many nonlabor costs in the producer/processor category are estimated by a discrete unit such as a batch or
field (Table 14).
Table 14. Non-Labor Costs of Advisory Event per Firm ($1,000)(1)(2)
Cost Range by Industry
Producer
Low estimate
Most likely estimate
High estimate
Shipper Distributor
Low estimate
Most likely estimate
High estimate
Restaurant
Low estimate
Most likely estimate
High estimate
Non-Restaurant Retailer
Low estimate
Most likely estimate
High estimate
Total non-labor cost per advisory

Min

Median

Average

Max

$0
$2,400
$5,800

$3,020
$7,925
$66,600

$8,885
$21,965
$59,612

$48,000
$82,000
$127,000

$0
$0
$0

$75
$200
$2,275

$1,214
$3,186
$7,041

$10,000
$16,000
$22,000

$0
$4
$15

$29
$117
$925

$63
$695
$2,825

$276
$2,500
$14,200

$0
$45
$165

$113
$1,208
$2,975

$343
$2,139
$8,576

$1,001
$10,500
$33,000

Low estimate
$0
$3,148
$10,461
$58,750
Most likely estimate
$3,175
$7,925
$27,669 $111,000
High estimate
$9,315
$66,600
$76,788 $186,000
(1)
The total non-labor cost per firm per advisory is the range of the sum of individual expert
estimates by industry category and is not additive.
(2)
Estimates include the addition of short- and long-term sales losses reported by all experts
for consistency.
There are multiple reasons that could explain the wide variation among expert response,
the main one being that outbreaks and therefore FDA advisories vary in characteristics. Costs
can also vary widely depending on the size and magnitude of the recall but also by industry. For
example, according to experts, large firms are more likely to react to an advisory as they have
more resources to address publicity that may be related to the advisory. Small firms, on the other
hand, are not as likely to be prepared and/or have similar resources that support them to take an
action.
Total Costs per Firm
Costs incurred by firms is essentially driven by how they think they should respond to the
FDA advisory. Total costs of an advisory event per firm are the sum of labor costs and non-labor
costs) are summarized in Table 15.
Table 15. Total Cost of Advisory Event per Firm ($1,000)
Cost Range by Industry
Producer
Low estimate
Most likely estimate
High estimate
Shipper Distributor
Low estimate
Most likely estimate
High estimate
Restaurant
Low estimate
Most likely estimate

Min

Median

Average

Max

$7
$2,471
$5,859

$3,031
$8,167
$72,658

$8,915
$22,041
$61,098

$48,006
$82,011
$127,015

$1
$2
$7

$98
$300
$2,311

$1,223
$3,231
$7,552

$10,006
$16,011
$22,016

$2
$7

$36
$163

$73
$731

$311
$2,518
98

High estimate
$20
$1,118
$3,057
$14,368
Non-Restaurant Retailer
Low estimate
$3
$129
$354
$1,031
Most likely estimate
$52
$1,220
$2,172
$10,524
High estimate
$179
$3,068
$9,134
$33,034
Total cost per advisory
Low estimate
$17
$3,182
$10,518
$58,784
Most likely estimate
$3,649
$8,212
$27,852 $111,063
High estimate
$9,463
$80,028
$79,487 $186,091
(1) The total cost per firm per advisory is the range of the sum of individual expert estimates by
industry category and is not additive.
While non-labor costs vary widely by industry and size of firms, they also make over
99% on average of total costs across all industries and about half of non-labor costs are incurred
by producers/processors in lost sales.

ii.

Affected Firms
Experts did not have information to help estimate the reduction in scope (number of firms

affected) of an overly broad recall, as they did not know the number of firms affected overall.
The following steps we use in estimating the number of affected firms and then characterizing
the firms in a way that is closest to the criteria laid out by experts for firms with a low, most
likely, and high estimate require a series of calculations described below. Once we have laid out
the series of calculations, our final step involves using Monte Carlo simulations. A Monte Carlo
simulation requires assigning parameters to corresponding probability density functions to 1)
characterize the variability inherent in the cost estimates, and 2) to characterize the inherent
uncertainty in the estimated number of firms by their respective cost category. Parameter
estimates and results of the simulation are presented in Appendix G.

To estimate the scope of a broad recall, we assume that the annual number of firms
affected by an advisory is dependent on the probability that:
1) at least one firm in any given industry is affected by an outbreak and
2) a number of firms in any given state in the country will experience an outbreak.
To estimate these probabilities, we selected nine outbreaks of the 23 Outbreaks in
Appendix C. We only use those with 100 or more illnesses and for which there was an FDA
advisory (such as in the case studies used by experts) and present them in Table 16.53
Table 16. Outbreak Used for Estimating the Probability of a Large Outbreak.
Year
20082009
2008

Commodity

Pathogen

Peanut Butter and peanut butter products Salmonella Typhimirium

Total
Illnesses
per
Epidemic
Curve
636

Peppers and tomatoes

Salmonella St Paul

1,442

2009

Sprouts

Salmonella St Paul

235

2010

Shell eggs

Salmonella Enteritidis

2011

Cantaloupe

Listeria Monocytogenes

139

2013

Salad mix

Cyclospora cayetanensis

631

2019

Cantaloupe

Salmonella Javiana

163

2019

Romaine Lettuce

E. coli O157: H7

167

Cyclospora cayetanensis

241

2019
Fresh Basil
Source: www.cdc.gov

3,578

We created two matrices: one outlines the number of states affected by each advisory while
the other outlines the number of FTL related industries associated with each outbreak.
1) To estimate the probability that a producer/processor firm (associated with production or
farming) is affected by an outbreak advisory we use the matrix of outbreaks affecting

Links with information on each outbreak can be found in Appendix C.

100

industry categories by matching their corresponding NAICS code. We use the ratio of the
sum of outbreaks per NAICS to estimate the probability range (min, average, max) of an
outbreak by multiplying the ratio by the number of firms in each firm category. The
probability that any producing firm is associated with an outbreak in a 12-year period
ranges from 0 to 25 percent with an average of 3 percent. In a similar manner we estimate
the probability that a farm is associated with an outbreak in a 12-year period ranges from
0 to 29 percent with an average of four percent (Table 17).54
Table 17. Probability of Outbreak by Industry Category (NAICS) Over 12-year Period
Year / Commodity

Pathogen

Salmonella St
Paul
2008-2009 / Peanut Butter
Salmonella
and peanut butter products
Typhimirium
Salmonella St
2009 / Sprouts
Paul
Salmonella
2010 / Shell eggs
Enteritidis
Listeria
2011 / Cantaloupe
Monocytogenes
Cyclospora
2013 / Salad mix
cayetanensis
Salmonella
2019 / Cantaloupe
Javiana
2019 / Romaine Lettuce
E. coli O157: H7
Cyclospora
2019 / Fresh Basil
cayetanensis
Ratio of selected outbreaks Min
by industry category (sum of Average
large outbreaks/ total
number of large outbreaks) Max
2008 / Peppers and tomatoes

Number of Large Outbreaks by Pathogen
Farms

Producers

0%
4.2%

0%
2.9%

29%

25%

Source: Statistics of U.S. Businesses SUSB is an annual series that provides national and subnational data on the
distribution of economic data by enterprise size and industry. https://www.census.gov/programs-surveys/susb.html

101

NAICS

111219, 111339,
111419, 112310,
112511, 112512,
114111, 114112,
115114, 115115

311340, 311351,
311352, 311411,
311412, 311423,
311513, 311520,
311710, 311811311813, 311821,
311824, 311911,
311941, 311942,
311991

2) To estimate the probability that firms in the distributor, wholesaler, retailer and restaurant
categories are affected by an outbreak advisory, we use the matrix of the same selected
outbreaks as in Table 17 by states (depicted best using a map, see Figure 1). We use the
ratio of the sum of outbreaks per state to estimate the probability range (min, average,
max) of an outbreak by multiplying the ratio by the number of firms in each state. The
probability that any firm in the retail, restaurant, or distributor category in any given state
is associated with an outbreak in a 12-year period ranges from 0.4 percent to 3.2 percent
with an average of 2 percent (Figure 1).

102

Figure 1. Probability of a Large Outbreak from Covered Foods by State Over a 12-Year
Period.
The estimated annual range of firms affected by an FDA advisory is shown in Column D
of Table 18 and is expressed as the product of columns A x B x C divided by 12 (for 12 years of
outbreak data) where:


A is the number of firms by NAICS category,



B is the percentage of firms associated with covered products; and



C is the probability of an outbreak by firm category.

To arrive at the estimated range on column D, we first use the estimated number of
covered firms by their corresponding NAICS using SUSB data from the U.S. Census in Column
A. Column B in Table 18 represents the share of firms associated with covered products by the

103

ratio of sum of covered NAPCS establishments over total NAICS establishments, derived from
the Coverage section estimates on Appendix D and in section II.D.2 Coverage of the Rule.
The product of Columns A, B and C divided by 12 in Table 18 gives us Column D: The
annual number of firms affected by advisory. The total annual number of firms is, in effect, the
weighted sum of the number of firms affected by an advisory.55 The estimated annual number of
firms potentially affected by an advisory range from 141 to 1,278 with an average of 710.

Table 18. Annual Number of Firms Affected by Advisory (1).

Category of Firms

(A)
NAICS
Firms

(B)
Perce
nt
FTL
Firms

(D) = (A x B x C)/12
Annual number of Firms
Affected by Advisory
min
(all)

average
(all)

max
(all)

Farms
Producers/
Manufacturers
Wholesaler/
Shipper/Distributors

50,756

67%

0.0%

0.6%

4.6%

130

18,907

55%

0.0%

0.5%

4.0%

32,160

43%

0.4%

2.0%

3.2%

Retailers

485,893

66%

0.4%

2.0%

3.2%

106

524

845

Restaurants

180,577

49%

0.4%

2.0%

3.2%

145

233

Weighted Sum

768,293

141

710

1,278

61%

(1) Sums and sums of products may not be exact due to rounding.

Using ratio of outbreaks per NAICS for farms and producer/processor firm (0%,0.6% and 4.6%) divided by 12
and using ratio of outbreaks per state for firms in the categories for distributor/shipper, restaurants, and nonrestaurant retail (0.4%,2% and 3.2%) divided by 12.

104

iii.

Firms Affected by Low, Middle (Most Likely) and High Estimate.
The next step of this calculation involves finding the corresponding proportion of firms in

Column D that would most likely be affected by either the low, most likely or high-cost estimate
provided by the recall expert elicitation in Table 12, Table 14 and Table 1556.
To match the four product categories in the cost estimates from the recall expert elicitation, and
as discussed in the Averted costs per firm section, we assume farms and producers will fit in the
same cost category.

Table 19. Proportion of Firms Associated with Low, Middle and High Estimate Categories
(E)
(F)
(G)
Category
% Firms for low % Firms for middle % Firms for high
estimate
estimate
estimate
Farm

45.43%

51.30%

3.26%

Producer/processor

45.43%

51.30%

3.26%

Shipper/ Distributor

49.41%

45.41%

5.18%

Restaurant

36.76%

62.77%

0.63%

Non-Restaurant Retailer

61.20%

38.17%

0.48%

Assigning the percentages in Table 19 to low, most likely, and high-cost ranges used in the
recall elicitation, requires that we make the following assumptions:
1. The proportion of firms affected by the low estimate is equivalent to the number of firms
from SUSB Census with five or less employees.
2. The proportion of firms affected by most likely estimate is equivalent to the number of
firms from the SUSB with 6 to 499 employees.

Some experts may have interpreted the middle estimate as the most likely, while others may have interpreted it to
be the same as the median.

105

3. The proportion of firms affected by the high estimate is equivalent to the number of firms
from the SUSB with 500+ employees.
We multiply the range in column D of Table 18 by Columns E, F and G in Table 19 to obtain the
annual number of firms affected by advisory according to cost category characterized by the
range of H, I and J in Table 20.

Table 20. Annual Number of Firms Affected by Advisory by Cost Category.

Category of Firms
Producers/
Processor
Shipper/Distributors
Restaurants
Non restaurant
retailers
Weighted Sum

iv.

H
I
J
Firms affected by low
Firms affected by middle
Firms affected by high
estimate
most likely estimate
estimate
min
average max
min
average max
min average max
(small) (small) (small) (small) (small) (small) (large) (large) (large)
2
2
39

8
11
193

73
18
311

2
2
66

9
10
329

83
17
530

2
0
1

2
1
3

4
2
4

18
61

88
300

143
544

11
82

55
403

89
719

0
3

1
7

1
12

Estimated Benefits from Reduction in Overly Broad Recalls

Table 21 shows a central point estimate for the average (middle) number of affected firms from Table 20
Columns H, I & J by industry and cost category per firm from Table 15. The estimated annual
benefit of $862 million in

Table 21 is illustrative as it does not account for variability and uncertainty.

106

Table 21.- Average Estimated Benefits by Industry ($1,000)
Industry Category

Average Cost per
Firm

Firms
(Average
Estimate)

Product (Cost per
Firm x Number of
Firms)

Producer
$8,915
Low estimate
8
$22,041
Most likely estimate
9
$61,098
High estimate
2
Shipper Distributor
Low estimate
$1,223
11
Most likely estimate
$3,231
10
High estimate
$7,552
1
Restaurant
Low estimate
$73
193
Most likely estimate
$731
329
High estimate
$3,057
3
Non-Restaurant Retailer
Low estimate
$354
88
Most likely estimate
$2,172
55
High estimate
$9,134
1
Total
Low estimate
$10,565
300
Most likely estimate
$28,175
403
High estimate
$80,842
7
Sum
710
Annual Benefit from avoiding overly broad recalls following
FDA advisories

$68,760
$193,090
$122,197
$13,732
$33,320
$8,886
$13,969
$240,326
$7,650
$31,359
$119,897
$8,323
$127,820
$586,633
$147,055
$861,508
$861,508

The final step in this series of calculations involves using Monte Carlo simulations to
estimate the range of possible benefit estimates using the value ranges discussed so far.57
Parameter estimates and results of the simulation are presented in Appendix G.

A Monte Carlo simulation is a probabilistic technique used to model the probability of different outcomes in a
process that cannot easily be predicted.

107

Table 22 presents simulated results of primary, low, and high (undiscounted) cost
estimates expected to be averted by improved traceability. We use the median as our primary
(central estimate), bounded by the 5th percentile and the 95th percentile for the low and high
estimates. The estimated annual benefit from avoiding overly broad recalls after an FDA
advisory ranges from $270 million to $2 billion dollars with a primary estimate of $661 million.

Table 22.Total Averted Costs from Overly Broad Recalls Resulting from Improved
Traceability (Undiscounted, $1,000)
Primary Estimate
Low
High
Total Averted Cost, including
‐ Cost from Labor
‐ Cost from Non-labor

$660,509
$18,521
$638,059

$268,103
$7,433
$248,279

$2,032,980
$50,094
$2,005,029

The present value of benefits from avoiding overly broad recalls with seven percent
discounting over 20 years (not shown in Table 22) ranges from about $2.5 billion to $18.8
billion, with a primary estimate of $6.1 billion. The present value of these benefits with three
percent discounting over 20 years (not shown in Table 22) ranges from about $3.6 billion to
$27.3 billion, with a primary estimate of $8.9 billion. At a seven percent discount rate over
twenty years, these annualized monetized benefits from avoiding overly broad recalls range from
$233 million to $1.77 billion with a primary estimate of $575 million. At a three percent discount
rate over twenty years, these annualized monetized benefits range from $242 million to $1.84
billion with a primary estimate of $596 million.
The estimated benefit from improved traceability, especially in averted lost sales due to
reduced consumer confidence, may not fully capture costs of unnecessary preventive actions by
consumers. Such preventive measures by consumers may involve throwing away food, stopping

108

their consumption of the suspect food item resulting in reduced consumer welfare, or visiting
physicians or emergency rooms to determine if they have been exposed to a pathogen.
3. Other Benefits
There are additional benefits that may arise from this rule. We briefly discuss these
benefits qualitatively hereunder. These may include:


Avoidance of costs due to unnecessary preventive actions by consumers.



Reduction in food waste by both consumers and producers.



More expedient initiation and completion of recalls of covered foods.



Improvements in supply chain management and inventory control.



Other supply system efficiencies due to a standardized approach to traceability,
including greater transparency, an increase in trust among food supply system
participants, and potential deterrence of fraud.

Broad recalls that take long to implement may have devastating consequences to both
businesses and consumers, especially in situations when a recalling firm or its downstream
customers cannot definitively track the movement of outbreak causing recalled product and
therefore recall additional products as a “buffer”. In this situation, when the recalling firm is
known, some but not all of these costs through the supply chain may be partially or wholly
reimbursed by the firm (depending on existing contracts). Market withdrawals and advisories
result in additional costs to FDA and industry and may force consumers to take unnecessary
costly preventive actions (Ref. [36]). For example, the disposal of unimplicated food increases
the amount of food waste generated which may be associated with a negative externality. The
social impact of higher food waste may lead to an increase in the price level of food in the longterm (Ref. [37]).
109

Without having prompt access to records that provide timely information, FDA, state,
and local authorities might spend additional time and resources on tracing the source of an
outbreak, initiating broad recalls, and communicating to industry and consumers. The traceability
records required by this rule could help firms in initiating recalls earlier and shortening recall
periods, which can significantly reduce the costs associated with management and support of
recall activities. Shorter recall periods may also mitigate the loss of future product sales by
shortening the length of negative publicity these products received because of longer recalls.
Another benefit relates to improved rate of recovery for contaminated products. When
recall orders are issued and include a precise list of products suspected to be contaminated, the
rate of recovery of this product may be substantially improved under traceability rule. The ability
to precisely identify the lot numbers, production dates and other related information may help
those collecting the product achieve a higher recovery rate than would be otherwise achievable in
situations where robust traceability is unavailable. When this is done timely, issuance of warning
notices to the public to avoid certain contaminated foods may be unnecessary and the demand for
those goods may be unaffected. The 2015 Blue Bell ice cream recall is a good example. Since it
took so long to be able to identify the contaminated products and the recalls were issued in
piecemeal, the rate of recovery was only about 8 percent. Of the 493 million Blue Bell products
recalled because of Listeria contamination, only 39 million (or about 8 percent) were recovered
and successfully destroyed (Ref. [38]).
Finally, other benefits may include supply chain management improvements, increase in
transparency and food system trust, and potential deterrence of fraud (Ref. [2, 1]). While this
rule’s primary focus is not to deter fraud, it might result is some potential ancillary benefits that
may be realized through a reduction in fraudulent activity or an exit from the market by some

110

firms because of increased accountability due to more standardized and improved traceability.
Since the rule requires improved recordkeeping that includes product name, lot codes, and who
put the lot code on the product, such requirements would potentially limit food fraudsters and
consequently result to spillover benefit effects for the entire food industry.
i. Costs Savings to FDA
In addition to outlined benefits above, this final rule will result in substantial cost savings
to FDA due to outbreak investigation efficiencies that would be achieved. Currently, nearly 18
percent of FDA’s budget is spent on the agency’s foods program, including food safety
surveillance, risk assessment, research, inspection, and food safety related education (Ref. [39]).
The costs of conducting outbreak investigations can be substantial especially when mechanisms
to trace contaminated foods are inefficient. The costs of outbreak investigation can be
multifaceted to include epidemiological, laboratory confirmation, records collection activities,
traceback investigation, coordinating the environmental related investigations as well as issuing
recalls once the implicated food is established. Depending on the nature of the investigation and
the time it takes to investigate, identify and recall an implicated food product, the costs can be
substantial. FDA outbreak investigation experts believe that some of these costs can be reduced
significantly resulting in substantial cost savings to FDA. Investigations will require less
personnel hours to complete, resulting in significant savings to FDA. Some of these savings can
be used to address other FDA public health priority needs.

F. Costs of the Rule
To better inform our analysis of the costs of this final rule, ERG completed a literature
review and an elicitation of industry experts, the traceability costs elicitation, on current industry
practices and traceability costs in December 2021 and January 2022 (Ref. [4]). In the elicitation,
111

experts provided both qualitative and quantitative input based on the proposed version of the
rule, with additional brief definitions of some new CTEs in their draft-final state at the time of
the elicitation. Input included describing anticipated cost-incurring compliance activities and
expenditures, estimating variables related to cost calculations, and further commenting on factors
likely to influence costs of the rule. We have incorporated their input, as well as input from the
ERG literature review, while updating our analysis to reflect changes in requirements under the
final versus the proposed rule.
Throughout this analysis, we refer to small firms as defined by the Small Business
Administration.58 Where we report estimates for small establishments, we refer to those
establishments owned by small firms.
We consider baseline food traceability practices of all covered entities (firms or
establishments) and estimate the incremental costs related to changes in these practices that these
entities choose to implement to meet the requirements of the final rule. We estimate the
following costs of the rule to industry:
•

one-time labor costs of reading and understanding the rule;

•

one-time labor costs of searching inventory for covered products and planning new
traceability practices as necessary;

•

one-time capital investment costs to establish and maintain a traceability plan and
maintain and make available required records;

•

recurring costs of operating and maintaining capital;

•

one-time costs of training personnel to implement new recordkeeping systems;

•

recurring costs of refresher training, and;

https://www.sba.gov/document/support-table-size-standards

112

•

recurring costs of maintaining and making available records required by this rule
during critical tracking events.

In tables throughout this and other sections, figures might not sum to the displayed totals
due to rounding.
1. Main Assumptions of the Cost Analysis
For discounting purposes, we assume the publication date of the rule is year zero and that
costs will gradually begin to be incurred in year one (which is two years before the three-year
compliance date).59 We expect that one-time costs of the final rule will occur evenly over the
first two years after the rule becomes effective. We expect that recurring costs will begin in the
second year, though at only half the estimated amount, lagging by one year behind the half of
one-time costs occurring in year one.
We estimate that the costs of this final rule will mainly arise from recordkeeping
requirements, which represent new and substantive responsibilities for covered entities. Using
the current version of the FTL, we identify the entities and food supply system sectors affiliated
with covered foods. Because each provision of the final rule requires different entities within the
food supply system to maintain different sets of records and key data elements, we list estimated
costs by provision. First, all covered entities must establish and maintain a traceability plan as
outlined in §1.1315. The following provisions, described in more detail in the sub-sections
below, apply differently to entities depending on whether they grow (or raise), harvest (or cool),
initially pack, first receive on land from fishing vessels, ship, receive, or transform foods on the

The year of publication is year zero and the effective date is year three. For covered entities to comply with the
requirements of this rule by the effective date (year three), we assume they will begin to incur compliance costs in
year one.

113

FTL. If an entity performs more than one of these activities, it must keep the records required by
all relevant provisions.
The total cost of each statutory requirement depends on the number of entities affected
and the additional burden of each requirement relative to baseline practices. In this final rule,
FDA specifies the records and information a covered entity must maintain and provide but does
not specify the form or system in which those records should be maintained. We expect that, to
the extent possible, firms would satisfy this rule’s recordkeeping requirements using existing
systems. Furthermore, we assume that firms will comply with new recordkeeping requirements
by modifying existing shipping or purchase records such as Advance Shipping Notices, Bills of
Lading, Invoices, or Purchase Orders.
Some entities already follow certain traceability-related recordkeeping practices as part of
their baseline business practices. These entities include those covered by the Subpart J provisions
as well as those following certain industry conventions and best practices. Additionally, food
businesses have increasingly adopted traceability-related technologies and practices as part of
their internal business decisions, regardless of the rule. The likelihood of an entity already
adhering to certain recordkeeping practices or planning to adopt them depends on its size,
industry, and position in the food supply system. We used input by food industry experts in
estimating the expected baseline prevalence of the recordkeeping practices required by the rule
(Ref. [4]).
Finally, as discussed in the baseline section of this document and in further detail below,
certain entities covered by the final rule are exempt or partially exempt from new recordkeeping
requirements. In all our cost estimates below, we exclude produce farms, shell egg producers,
and certain other producers of raw agricultural commodities that would be fully exempt from the

114

final rule because of their very small size. Additionally, this final rule also exempts retail food
establishments and restaurants with average annual food sales below $250,000, which we
therefore exclude from our analysis.60 Some entities or types of products are partially exempt,
which we describe in further detail below. Of these partially exempt entities, we exclude farm-toschool and farm-to-institution entities from our cost estimates because we do not expect these
entities to change their current practices due to the final rule.
We conduct our analysis by provision to parallel the structure of the rule. We cover a
twenty-year horizon following the final rule’s compliance date and assume one-time costs to
occur evenly over years one and two. All wage rates come from the Bureau of Labor Statistics,
Occupational Employment Statistics (OES), May 2020, National Industry-Specific Occupational
Employment and Wage Estimates. We estimate hourly labor costs by doubling the OES reported
mean wage to account for benefits and overhead.
2. Costs of Reading and Understanding the Rule
We estimate that all covered firms will read and understand the rule. Note that we
consider reading costs in this section to be separate from the costs to identify FTL products and
plan for compliance, which we estimate below in section II.F.5.a “Traceability Plan.”
To estimate the number of firms that will need to read the rule, we use the 2017 SUSB
from Census Bureau and the 2017 Census of Agriculture from the National Agricultural
Statistics Service. We first identify the NAICS categories of firms that likely manufacture,
process, pack or hold foods on the FTL. Because not all firms in each NAICS industry would be

An average annual monetary value of food sold or provided during the previous 3-year period of no more than
$250,000 (on a rolling basis), adjusted for inflation using 2020 as the baseline year for calculating the adjustment.

115

affected by this rule, we remove exempt firms and those not likely to handle foods on the FTL.
We estimate that about 323,872 firms will incur a one-time cost to read and understand the rule.
Table 23 summarizes the estimated costs of reading and understanding the rule. The
preamble and regulatory text of the rule contains approximately 219,000 words. Per HHS
guidelines on reading speed (Ref. [29]), we estimate that an adult reads 200 to 250 words per
minute, with an average speed of 225 words per minute. We divide the number of words in the
preamble and codified by reading speed, producing an estimate of 16.2 hours (= (219,000 / 225) /
60) to read the rule, with a lower bound of 14.6 hours (= (219,000 / 250) / 60) and an upper
bound of 18.3 hours (= (219,000 / 200) / 60).
We assume that one employee will read the rule at small firms and that three employees
will read the rule at large firms. We expect that the type of employees reading the rule at small
firms will be roughly equivalent to supervisors of food preparation (occupation code 35-1010 in
NAICS 445), and that those at large firms will be compliance officers (occupation code 13-1041
in NAICS 311). The mean wage of a supervisor of food preparation is $20.12 per hour, which we
double to $40.24 to account for benefits and overhead. The mean wage of a compliance officer is
$32.71 per hour, which we double to $65.42 to account for benefits and overhead. We obtained
this wage information from the 2020 BLS Occupational Employment and Wage Statistics. We
therefore estimate that the one-time cost of reading and understanding the rule is approximately
$507 per covered small firm (= 12.6 hours x 1 employee x $40.24) and $2,471 per covered large
firm (= 12.59 hours x 3 employees x $65.42). We estimate the total cost of reading and
understanding the rule (sum of costs to small and large firms in the table below) to range from
about $203 million to $254 million, with a primary estimate of $226 million.
Table 23. One-time Costs of Reading and Understanding the Final Rule (2020$)
Small firms
Large firms
116

(a) Average reading
speed (words/min)
(b) Words to read
(c) Hours = (b/a)/60
(d) Employees
(e) Hourly labor cost
(f) Per firm cost = c x d
xe
(g) Number of firms
Total = f x g

Primary

Low

High

Primary

Low

High

225

250

200

225

250

200

219,000
16.2

219,000
14.6

219,000
18.3

219,000
16.2

219,000
14.6

219,000
18.3

1
$40.24

3
$65.42

$652.78

$587.50

$734.38

$3,183.7
7

$2,865.4
0

$3,581.75

318,252
$207,749,
096

318,252
$186,974,
187

318,252
$233,717,
733

5,620
$17,892,6
13

5,620
$16,103,3
52

5,620
$20,129,190

3. Costs of Capital Investment
Capital investment refers to equipment and software and may include food traceability
software, scanners or barcode readers, barcode printers, and increased data storage (hard disk or
cloud storage) to handle the increased recordkeeping requirements of the final rule.
As discussed above, some entities will be able to comply without additional capital
investments, while others will need to invest in traceability-related capital. Case studies of prior
traceback efforts in 2012 and earlier show a wide range of existing tracing capabilities across
sectors and firm size (Ref. [5]). For example, according to the 2012 IFT study of pilot projects
for improving product tracing, which surveyed 22 entities, selected large growers, distributors,
and processors already had the capacity to scan KDEs, while only 30-50 percent of selected
small distributors, small processors, and large retailers had this capability at that time. In general,
traceability technologies, adoption, and implementation have continued to expand since 2012,
supported by efforts in multiple industries to integrate standards like those of GS1 into common
practice.61 More recently, ERG’s literature review revealed that traceability accounts for 20 to 45

See, for example, the Produce Traceability Initiative (https://www.producetraceability.org/), the International
Dairy-Deli-Bakery Association (https://www.iddba.org/initiatives/industry-initiatives/food-traceability), and the
National Fisheries Institute (https://www.aboutseafood.com/traceability-and-sustainability-in-the-supply-chain-4/).

117

percent of the roughly $80 to $91 million that California Leafy Green Marketing Agreement
(LGMA) members spend on meeting LGMA safety standards each year (Ref. [4] ).
The following Table 24 summarizes baseline costs of operating product traceability
systems and the additional investments related to specific system improvements, as reported by
the 2012 pilot project participants (Ref. [5]). Participants reported that total costs for traceability
systems they used at the time of survey to capture KDEs ranged from tens of thousands to
millions of dollars. They reported that additional improvements in traceability would range from
minimal cost to hundreds of thousands of dollars.
Table 24. 2012 Costs of Traceability-Related Investments, as Reported by Selected Entities
Large
Large
Small
Large
Small
Large
Type of firm
Grower
Processor
Processor Distributor Distributor
Retailer
Number of
3
4
3
4
4
4
firms surveyed
Base cost to
capture KDEs
$350,000$500,000$250,000$50,000$40,000-$1.5M Unknown
(manual or
$4.5M
$1.2M
$800,000
$1M
electronic)
Additional costs by activity
Incoming
KDEs by
electronic data
messages
Supply chain
link
Standardized
naming
Outgoing
KDEs
electronic to
customers
Provide data
summary

Unknown

$0-$15,000

Unknown

$0-$65,000

$0$60,000/year

Unknown

$0-$150,000

Unknown

$0-$500,000

Unknown

$0-$80,000

$5,000$150,000

Unknown

$2,000$5,000

Unknown

N/A

Unknown

$0-$2,000

$0-$10,000

N/A

118

Incoming lot
number
information

N/A

$0$60,000/year

$0-$150,000

Unknown

These estimates reported above provide some insight into the types of capital investment
costs surveyed companies needed to improve traceability systems back in 2012, such as
conversions to standardized naming and provision of data to supply chain partners. Importantly,
not all entities expected to make the same types of investments, and some entities, like large
distributors, already had the capability to integrate additional traceability practices into existing
systems. For example, in ERG’s literature review, traceability solution providers report several
recurring subscription costs that vary by volume of product or number of users, such as hardware
rental, software licenses, and user accounts (Ref. [4]).
There are several important considerations for interpreting the costs from the IFT report
and applying them to the final rule (Ref. [5]). First, the report surveyed selected companies about
costs to update system-wide traceability operations. In other words, companies may have also
accounted for training and system integration costs, in addition to capital, in their estimates. The
incremental costs of complying with the final rule may be significantly less than the cost of a
total system upgrade, particularly because we believe that over a third of covered companies
already perform some form of traceability recordkeeping (Ref. [4]).
Second, entities surveyed in the report are not fully representative of sectors and sizes
affected by the final rule. Most entities reported that they invested in and improved their tracing
systems in the five years prior to the 2012 report (Ref. [5]). The reported costs came from large
companies, which likely traditionally rely on relatively sophisticated systems. For example, the
large growers surveyed were in the top two percent of the industry in terms of revenue, and no
costs were provided for small and medium growers. The study contacted fifteen small businesses
119

from unreported sectors, most of which reported that their existing tracing capabilities sufficed
and that additional investments would be costly. Others had already invested in tracing networks
that rely on standardized nomenclature.
Third, competition among providers of traceability products and services is likely to
produce lower costs over time (Ref. [40]). Indeed, since the time of the 2012 pilot project report,
technology and vendors have evolved to provide new traceability solutions, software, and offthe-shelf packages that were not available to companies until recent years. Today companies
have access to new tools like web-based systems and blockchain subscriptions.62 For example,
web and mobile platforms are now able to streamline links between retailers and their wholesale
suppliers.63 Equipment has also become more accessible and lower-cost. As an indication of this
trend, compared to the reference year of 2007, the consumer price index for computers,
peripherals, and smart home assistants decreased from 74 percent in December 2010 to 39
percent in December 2021.64
A literature review by ERG finds several other case studies on traceback efforts and
traceability systems (Ref. [4]). Some cost categories such as traceability software and hardware
such as scanners and printers are commonly mentioned in the literature. However, capital
investments needed to comply with the rule will depend on a firm’s size, role in the supply chain,
products, and existing traceability systems, as well as whether the firm decides to adopt an
electronic recordkeeping system as a result of this rule (although the rule does not require
electronic maintenance of records). To better understand incremental capital investment costs

See, for example, IBM’s Food Trust: https://www.ibm.com/blockchain/solutions/food-trust.
See, for example, BlueCart: https://welcome.bluecart.com/.
64
U.S. Bureau of Labor Statistics, Consumer Price Index for All Urban Consumers: Computers, Peripherals, and
Smart Home Assistants in U.S. City Average [CUSR0000SEEE01], retrieved from FRED, Federal Reserve Bank of
St. Louis; https://fred.stlouisfed.org/series/CUSR0000SEEE01, February 1, 2022.
63

120

related to the rule for smaller entities, as opposed to complete system upgrades for large firms
described above, ERG elicited one-time and recurring cost estimates from a panel of external
industry experts based on the proposed version of the rule, with additional brief definitions of
some new CTEs in their draft-final state at the time of the traceability costs elicitation (Ref. [4]).
In addition to cost estimates, experts identified and described, in general terms, the primary
expected capital expenditures, such as scanners, label printers, and mobile computers, as well as
the share of covered entities that would choose to invest in capital to comply with the rule as
proposed.
Although the traceability costs elicitation asked for estimates of expected costs due
specifically to compliance with the proposed traceability requirements, some experts suggested
that their estimates included capital investments likely to occur for reasons other than regulatory
compliance, such as obsolescence (Ref. [4]). Since experts did not separately quantify the extent
to which regulatory compliance drives their estimates, we nonetheless attribute all estimated
capital investment to the proposed traceability requirements when using the elicitation to inform
our analysis of the final rule. This likely results in us overestimating the capital costs stemming
specifically from the rule.
Capital requirements for traceability will likely be more complex for some businesses
than others. For instance, only entities downstream from harvesting and cooling raw agricultural
commodities will be responsible for assigning traceability lot codes. Hence, we do not generally
expect covered produce farms, shell egg producers, and other producers of raw agricultural
commodities who do not also initially pack to incur capital costs due to the final rule. First, under
the final rule, these entities do not need to assign lot codes. We also note that with respect to the
farm map record that is a part of the traceability plan requirement, FDA has previously received

121

farm maps with field names and coordinates during outbreak investigations. Given the
widespread use of free mapping and direction websites and web applications with GPS
coordinate plotting functionality, we expect most affected entities either already keep the
required map (in paper or electronic format) or will be able to produce it in minutes without any
specialized equipment. Additionally, shipments of food that occur prior to initial packing or for
food obtained from a fishing vessel prior to first land-based receiving are not subject to the
shipping or receiving requirements of the final rule. However, for industries in which we were
unable to distinguish between initial packers and other growers or producers of raw agricultural
commodities, we use expert estimates (Ref. [4]) of the proportion of establishments that will
incur capital costs.
We also note that under the final rule’s partial exemption of sales between retailers, the
purchasing retailer needs only to keep information typically communicated via sales receipts and
there are no traceability requirements on the selling retailer. Thus, restaurants and other retailers
under the final rule are generally only receivers, but not shippers, of covered foods. Because the
final rule does not require retailers to regularly use the received information for compliance
purposes, and because compliant storage options include taking pictures of records or not
digitizing at all, we do not expect retailers to purchase equipment or services towards
compliance. While retailers need to keep the required records provided to them by suppliers,
which we expect to take the form of receipts and purchase and delivery documents, they may
keep these records as either paper or electronic originals or “true copies” including photocopies,
pictures, scans, or other accurate reproductions of the original record. As such, we do not expect
these entities to require additional capital for compliance purposes.

122

Table 25 summarizes our estimates of the incremental one-time compliance related costs
of capital for affected establishments owned by small and large firms, based on reconciling
experts’ estimates and assumptions with differences between the final and proposed rule (with
additional brief definitions of some new CTEs in their draft-final state at the time of the
traceability costs elicitation) (Ref. [4]). Experts in ERG’s panel estimated overall one-time
capital costs per small and large entity from different industry sectors and identified, but did not
individually price, component elements of capital cost. We apply experts’ input on the
proportion of establishments that will incur capital costs to the industry categories that we expect
to make capital investments faced with the revised requirements of the final rule. We estimate
that about 34,737 establishments will incur these costs. Nearly all experts mentioned software,
while a majority also mentioned label printers and scanners. Some experts also mentioned
traceability consulting and training, which we do not consider as capital costs. To the extent that
those experts did not exclude these items from their capital cost estimates, it is possible that we
count such costs both in this section and in later sections. For example, experts estimated that the
one-time, per-firm capital cost for a small manufacturing/processing facility is $21,875. We use
the experts’ estimates and assume other costs such as consulting or training are not included in
the total.
We estimate the total one-time capital investment costs (sum of costs to small and large
firms in the table below) of the final rule to range from about $278 million to $4,867 million,
with a primary estimate of about $1,139 million.
Table 25. One-time Capital Investment Costs of the Final Rule (2020$)
Small
Large
Primary
Low
High
Primary
Low
Farms/ growers (produce)

High

123

(a) One-time
cost per
establishment
(b) Percent of
establishments
needing capital
(c) Number of
covered
establishments
Subtotal =
a*b*c
(a) One-time
cost per
establishment
(b) Percent of
establishments
needing capital
(c) Number of
covered
establishments
Subtotal =
a*b*c
(a) One-time
cost per
establishment
(b) Percent of
establishments
needing capital
(c) Number of
covered
establishments
Subtotal =
a*b*c
(a) One-time
cost per
establishment
(b) Percent of
establishments
needing capital
(c) Number of
covered
establishments

$11,250

$5,000

$50,000

$300,000

$100,000

$1,000,00
0

50%

70%

50%

40%

50%

1,065

134

$5,347,4
13

$66,842,6
65

$5,988,64
5

$2,661,62 $37,262,68 $20,052,80
0
0
0
Shell Eggs Producers

$13,000

$7,500

$50,000

$300,000

$100,000

$1,000,00
0

60%

50%

2,500

$1,055,1
64

$10,551,6
44

$19,501,7
94

$11,251,0 $75,006,90
$3,165,493
35
1
Fishing/ aquaculture

$20,000

$6,250

$52,500

$200,000

$75,200

$1,100,00
0

60%

50%

20%

80%

2,021

$1,039,4
92

$60,821,3
56

$24,252,6
98

$7,578,96 $63,663,33
$6,911,518
8
2
Manufacturing/ processing

$21,875

$10,000

$50,000

$226,000

$100,000

$1,000,00
0

48%

15%

65%

50%

30%

50%

8,203

447

447
124

Subtotal =
a*b*c
(a) One-time
cost per
establishment
(b) Percent of
establishments
needing capital
(c) Number of
covered
establishments
Subtotal =
a*b*c
Total

$85,237,9 $12,305,0 $266,608,8 $50,471,67
13
22
10
6
Wholesale/ Distributors/ Warehouse/ Storage

$13,399,
560

$223,326,
000

$42,500

$10,000

$65,000

$200,800

$100,000

$1,000,00
0

50%

15%

70%

50%

33%

55%

14,053

6,224

$298,620,
957
$433,602,
007

$21,079,1
26
$54,875,7
72

$639,400,1
66
$1,081,941
,890

$624,914,6
09
$705,516,0
95

$202,288
,096
$223,129
,725

$3,423,33
7,001
$3,784,87
8,666

As noted above, not all entities will need to make the same types of investments,
depending on their sector and size. We also do not expect that all covered entities will need to
make capital investments due to the final rule. In particular, if the rule would have been finalized
as proposed, covered retailers and restaurants would have faced significant potential burden
when receiving covered foods. To address this, FDA removed requirements for data elements
that would require record creation by the retailer, such as the time of receipt. We therefore expect
retailers and restaurants to satisfy the requirements of the final rule by keeping receipts and
purchase documents provided by suppliers, as many likely already do for reasons unrelated to
regulatory compliance.
For some of the other supply chain entities, who must further use the traceability
information they receive for compliance activities, capital investment might include software
subscriptions and other systems with ongoing operation and maintenance (O&M) costs. Thus, we
also estimate recurring capital costs, which we treat as a percent of one-time costs. Through
ERG’s traceability costs elicitation, external industry experts confirmed that not all firms
incurring capital investment costs will face incremental operation and maintenance costs
125

associated with new capital (Ref. [4]). Namely, firms that replace existing capital might face
operation and maintenance costs that are higher, lower, or the same as those for the capital being
replaced.
Table 26 summarizes our estimates of recurring capital costs for those firms whose
capital investments due to the final rule will result in an incremental increase in operation and
maintenance costs. Experts estimated recurring capital costs per establishment as a percentage of
one-time costs. Similarly, experts estimated the share of establishments whose capital
investments result in higher operating and maintenance costs as a percentage of all those making
capital investments. We thus obtain an estimate of about 15,854 establishments facing recurring
capital costs as a result of the rule. We estimate the total recurring capital costs (sum of costs to
small and large firms in the table below) of the rule to range from about $15 million to $980
million, with a primary estimate of about $185 million.
Table 26. Recurring Capital Investment Costs of the Final Rule (2020$)
Small
Large
Primary
Low
High
Primary
Low
Farms/ growers (produce)
(a) Recurring
$2,250
$425
$10,000
$60,000
$8,500
cost per
establishment
(b)
Establishments
479
426
745
60
43
with higher
O&M
Subtotal = a*b

$1,077,956

$180,990

$7,452,536

High

$200,000

$3,609,504

$363,624

$13,368,53
3

Shell Egg Producers

(a) Recurring
cost per
establishment
(b)
Establishments
with higher
O&M

$2,600

$300

$12,500

$60,000

$4,000

$225,000

1,350

1,200

1,500

126

Subtotal = a*b

$360,033

$4,000

$531

$13,125

1,091

970

1,213

$4,365,486

$515,370

$4,102

$700

$15,000

3,507

615

4,799

(a) Recurring
cost per
establishment
(b)
Establishments
with higher
O&M
Subtotal = a*b
(a) Recurring
cost per
establishment
(b)
Establishments
with higher
O&M
Subtotal = a*b

$18,751,72
$569,789
5
Fishing/ aquaculture

$3,510,323

$33,765

$2,374,120

$35,000

$4,512

$247,500

$49,896

$13,684,805

$39,550

$7,000

$225,000

201

$468,985

$45,223,515

$15,915,83
$1,088,564
3
Manufacturing/ processing

$71,984,37
$7,949,289
9
Wholesale/ Distribution/ Warehouse/ Storage

$14,383,898

$430,676

$7,969

$700

$19,500

$35,140

$7,000

$175,000

6,324

1,686

9,837

2,801

1,618

3,423

Subtotal = a*b

$50,392,286

$1,180,431

Total

$73,729,949

$2,667,500

$11,328,1
33
$12,244,4
03

$599,083,97
5
$673,734,94
8

(a) Recurring
cost per
establishment
(b)
Establishments
with higher
O&M

$191,820,0
50
$305,924,5
23

$98,424,051
$111,641,19
7

4. Costs of Training
Covered establishments will also incur costs to train employees to comply with the final
rule. We expect operational changes on a day-to-day basis to be disseminated through routine
meetings and trainings, such that establishments will not face additional costs of training all
employees specifically due to the final rule. However, we assume that establishments will need
127

to conduct new education of key personnel in traceability practices, particularly relating to the
use of new investments in capital, such as equipment, systems, and software. We estimate that
about 34,737 establishments will incur these costs.
The labor cost of training depends on its duration and number of participants. Previous
case studies from 2012 contain little information on the labor cost of training employees on new
traceability systems and practices (Ref. [5]). One major food service chain with 20,000
restaurants reported a training cost of $100 per restaurant but did not specify what this training
involved. One software provider reported a training cost of $1,000 per day, stating that the
number of days could vary based on the size of the operation and nature of changes to processes.
Other entities did not include separable training costs when reporting costs of capital investment,
suggesting that the labor cost of training may be unknown, included in capital investment (such
as a software package that includes training), or that new traceability training may replace
current training that becomes obsolete.
For this final analysis, we base training costs on estimates that ERG elicited from
external industry experts on the costs of traceability training related to the rule as proposed (with
additional brief definitions of some new CTEs in their draft-final state at the time of the
traceability costs elicitation) (Ref. [4]). As mentioned in the Baseline section II.D, firms vary in
the degree to which their baseline training already aligns with the rule’s requirements. Experts
estimated the training costs imposed by the rule on small versus large establishments accounting
broadly for industry and position in the supply chain.
Traceability activities will likely be more complex for some businesses than others.
While restaurants and other retailers need to keep the required records provided to them by
suppliers, which we expect to take the form of receipts and purchase and delivery documents, the

128

rule allows keeping these records as either paper or electronic originals or “true copies” such as
photocopies, pictures, scans, or other accurate reproductions of the original record. Because the
final rule does not require retailers to regularly use the received information for compliance
purposes, we do not expect retailers to purchase equipment or services towards compliance, nor
do we expect them to develop extensive or highly formal plans for maintaining traceability
records beyond determining physical or digital storage locations. Therefore, we do not expect the
requirements of the rule to generate a new training burden for restaurants and other retailers.
Table 27 summarizes our estimates of the incremental one-time training costs for affected
establishments owned by small and large firms, based on reconciling experts’ estimates with
differences between the final and proposed rule (with additional brief definitions of some new
CTEs in their draft-final state at the time of the traceability costs elicitation). Experts in ERG’s
panel estimated the cost of developing a training program, the number of trainees needed, and
the hours for training (Ref. [4]). In addition to the cost of developing or purchasing training
courses, training costs also include the labor cost of the employees in training. While training
may involve many kinds of employees, we approximate the average wage of First-Line
Supervisors of Production and Operating Workers (code 51-1011) in Food Manufacturing from
the 2020 BLS Occupational Employment and Wage Statistics, which is $29.10, for
manufacturing and distribution establishments. We double this wage to account for benefits and
overhead, obtaining an hourly labor cost of $58.20. For establishments in initial packing of raw
agricultural commodities, we use the average wage of Agricultural Workers (code 45-2000) in
Agriculture, Forestry, Fishing and Hunting, which is $14.53. We double this wage to account for
benefits and overhead, obtaining an hourly labor cost of $29.06. We estimate the total one-time

129

training costs (sum of costs to small and large firms in the table below) of the rule to range from
about $13 million to $409 million, with a primary estimate of about $241 million.
Table 27. One-time Training Costs of the Final Rule (2020$)
Small
Primary
Low
High
Primary
Farms/ growers (produce)
(a) Cost of
$3,000
$500
$10,000
$14,000
training course
(b) Trainees per
8.25
1.00
10.00
19.25
establishment
(c) Training hours
4.00
1.50
12.00
4.50
(d) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost
(e) Number of
1,065
1,065
1,065
134
establishments w/
training costs
Subtotal = (a +
$10,051,3 $2,208,1
$4,214,920 $289,366
(b*c*d))*e
84
24
Shell Eggs Producers
(a) Cost of
$2,875
$500
$7,500
$13,250
training course
(b) Trainees per
8.25
1.00
10.00
19.25
establishment
(c) Training hours
4.00
1.00
8.00
4.25
(d) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost
(e) Number of
establishments w/
2,500
2,500
2,500
21
training costs
Subtotal = (a +
$14,738,5
$9,585,832 $793,663
$329,791
(b*c*d))*e
56
Fishing/ aquaculture
(a) Cost of
$3,000
$950
$7,500
$12,292
training course
(b) Trainees per
6.50
1.00
10.00
11.00
establishment
(c) Training hours
4.00
1.00
8.00
4.25
(d) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost

Large
Low

High

$3,024

$25,000

4.50

35.00

4.00

18.00

$29.06

134

$189,677

$2,894,80
9

$2,500

$24,596

4.00

35.00

2.00

16.00

$29.06

$28,832

$431,241

$3,012

$21,572

4.00

35.00

2.00

16.00

$29.06

130

(e) Number of
establishments w/
training costs
Subtotal = (a +
(b*c*d))*e
(a) Cost of
training course
(b) Trainees per
establishment
(c) Training hours
(d) Hourly labor
cost
(e) Number of
establishments w/
training costs
Subtotal = (a +
(b*c*d))*e
(a) Cost of
training course
(b) Trainees per
establishment
(c) Training hours
(d) Hourly labor
cost
(e) Number of
establishments w/
training costs
Subtotal = (a +
(b*c*d))*e
Total

2,021
$7,590,205

2,021

$1,187,2 $11,913,8
$943,461
42
95
Manufacturing/processing

$44,849

$2,092,56
4

$3,000

$750

$7,500

$10,750

$3,012

$17,500

8.50

1.50

10.00

19.25

5.50

35.00

4.25

1.00

8.00

5.25

2.00

16.00

$58.20

8,203

447

$489,379

$11,186,8
46

$41,857,37 $1,030,2 $64,817,9 $7,428,6
8
99
34
43
Wholesale/ Dist./ Warehouse/ Storage
$3,000

$750

$7,500

$10,750

$3,012

$17,500

8.50

1.50

10.00

19.25

4.50

35.00

4.00

1.00

8.00

4.25

2.00

16.00

$58.20

14,053

6,224

$69,965,83
6
$133,214,1
72

$1,764,9
55
$5,065,5
26

$119,577,
668
$221,099,
438

$96,547,
362
$107,457
,381

$7,152,5
02
$7,905,2
39

$171,481,
797
$188,087,
258

In addition to one-time costs to develop training, we consider establishments to incur
recurring costs if refresher training necessary for compliance with the rule is more burdensome
than what an establishment would have otherwise implemented. Table 28 summarizes our
estimates of recurring training costs for those establishments who, due to the final rule, incur an
incremental increase over what they would otherwise have spent on recurring training. We base
the numbers of establishments with recurring training costs on estimates from ERG’s expert
131

panel of the share of such establishments. In estimating the recurring costs of training, we use the
same wages and numbers of trainees as in Table 27. Experts in ERG’s panel provided separate
estimates of the hours of recurring training. We assume that establishments would reuse the
existing training course in later years and therefore count only the labor cost of the employees in
training. We estimate total recurring training costs (sum of costs to small and large firms in the
table below) of the rule to range from about $1 million to $202 million, with a primary estimate
of about $40 million.
Table 28. Recurring Training Costs of the Final Rule (2020$)
Small
Primary
Low
High
Primary
Farms/ growers (produce)
(a) Trainees per
8.25
1.00
10.00
19.25
establishment
(b) Training
3.00
2.50
10.00
3.00
hours
(c) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost
(d) Number of
798
586
798
100
establishments w/
recurring training
Subtotal =
$574,299
$42,541 $2,320,400 $168,265
a*b*c*d
Shell Eggs Producers
(a) Trainees per
8.25
1.00
10.00
19.25
establishment
(b) Training
2.00
2.00
8.00
2.00
hours
(c) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost
(d) Number of
1,875
250
1,875
16
establishments w/
recurring training
Subtotal =
$899,126
$14,531 $4,359,401 $17,708
a*b*c*d
Fishing/ aquaculture
(a) Trainees per
6.50
1.00
10.00
11.00
establishment

Large
Low

High

4.50

35.00

2.50

10.00

$29.06

100

$24,038

$1,019,785

4.00

35.00

2.00

8.00

$29.06

$491

$128,785

4.00

35.00

132

(b) Training
hours
(c) Hourly labor
cost
(d) Number of
establishments w/
recurring training
Subtotal =
a*b*c*d

2.00

8.00

2.00

8.00

$29.06

1,516

202

1,516

$572,637

$11,746

$3,523,917

$33,140

$1,607

$421,782

Manufacturing/processing

(a) Trainees per
establishment
(b) Training
hours
(c) Hourly labor
cost
(d) Number of
establishments w/
recurring training
Subtotal =
a*b*c*d
(a) Trainees per
establishment
(b) Training
hours
(c) Hourly labor
cost
(d) Number of
establishments w/
recurring training
Subtotal =
a*b*c*d
Total

8.50

1.50

10.00

19.25

5.50

35.00

2.00

8.00

2.00

8.00

$58.20

6,153

820

6,153

335

$28,595

$5,458,981

$28,646,09
$750,610
1
Wholesale/ Dist./ Warehouse/ Storage

$6,087,294

$143,230

8.50

1.50

10.00

19.25

4.50

35.00

3.00

2.50

10.00

3.00

2.50

10.00

$58.20

10,540

1,405

10,540

4,668

622

4,668

$61,340,25
8
$100,190,0
67

$15,690,0
09
$16,659,7
32

$15,641,76
6
$23,775,12
2

$306,701
$518,750

$407,533
$462,263

$95,090,96
6
$102,120,2
99

5. Costs of Recordkeeping
The final rule will require certain persons to keep records related to the CTEs identified
in the rule, such as initial packing, shipping, receiving, and transforming of covered foods. We
estimate the new costs of each recordkeeping requirement, considering such persons’ current
133

recordkeeping practices. Relevant records contain KDEs associated with different CTEs in a
food supply system. These required records and data elements, which vary across the types of
entities in the food supply system, are described in detail below.
To estimate the recordkeeping costs of the rule, including frequency of recordkeeping
and the average time spent keeping records for covered foods by record type, we consulted
estimates that ERG elicited from external food industry experts (Ref. [4]). Experts expressed a
high degree of uncertainty regarding the time burden per record across activities. In general,
experts provided estimates of manual entry times in minutes while conveying in supplemental
comments that scanning using an electronic system would take seconds. As experts also
estimated the proportion of industry that currently keeps records mostly manually, we scaled
estimated times they provided by the proportion of industry with electronic recordkeeping
capabilities in order to account for baseline practices in estimating the incremental burden of the
rule. Our estimates of time burden per record therefore represent averages between manual and
electronic recordkeeping weighted by the baseline prevalence of these practices. Finally, we
reconciled estimates of the recordkeeping burden under the proposed version of the rule (with
additional brief definitions of some new CTEs in their draft-final state at the time of the
traceability costs elicitation) to the revised requirements in the final rule. As in previous sections,
all wage rates come from the Bureau of Labor Statistics, Occupational Employment Statistics
(OES) from May 2020.
As most experts included scanning and label printing equipment and systems when asked
to describe anticipated capital investments (Ref. [4]), we account for expected adoption of these
capabilities among the proportion of establishments estimated by experts to invest in capital
towards compliance with the rule. In particular, we expect that entities making capital

134

investments will be motivated by interoperability to prioritize streamlining shipping and
receiving recordkeeping, since these are the CTEs performed by the most entities across the
supply chain. In our estimates of the time burden per record for shipping and receiving, we
therefore additionally consider the proportion of entities making capital investments to be
capable of keeping and sending records via scanning and the use of barcodes (or related
technology, e.g., RFID).
Some entities perform multiple CTEs and will be subject to more than one recordkeeping
provision. Each provision outlines the KDEs necessary to effectively trace a product based on
the CTE an entity performs (e.g., receiving, transformation, shipping). Each CTE involves a
different set of KDEs, some of which they share in common; however, entities that perform
multiple CTEs will need to maintain all KDEs that pertain to the CTEs they perform. For
example, an entity that receives a covered food and then transforms and ships it will need to
record the quantity of food received, transformed, and shipped. We estimate recordkeeping
burdens by CTE because each entity must comply with all requirements in its relevant CTE.
Some KDEs (e.g., traceability lot code) are required for multiple CTEs; however, no two CTEs
contain exactly the same KDE requirements.
a. Traceability Plan (§ 1.1315)
The final rule will require entities that manufacture, process, pack, or hold covered foods
to establish and maintain a traceability plan containing:


a description of procedures for maintaining required records,



a description of procedures for identifying foods on the FTL,



a description of traceability lot code assignment, if applicable,



a statement identifying a point of contact for questions about traceability records, and

135



for entities that grow or raise FTL foods (except shell eggs), a map showing the areas in
which FTL food is grown or raised.
Entities affected by this provision will therefore incur one-time costs at the firm level to

create a traceability plan. While the final rule requires firms to update their traceability plans “as
needed,” possible future updates to the FTL, which might require some firms to identify
additional products, will only take effect two years after publication in the Federal Register. We
expect that this delay will allow firms to make necessary updates within the scope of routine
updates to standard operating procedures in the normal course of business. We thus expect that
entities affected by this provision will incur one-time costs at the firm level to create a
traceability plan.
Some firms already follow practices that satisfy some or most of the requirements in this
provision. Experts in ERG’s panel estimated the average proportion of small and large
establishments across business types (e.g., distributors, retail, etc.) that already keep traceability
records as part of a traceability system (Ref. [4]). For these entities with existing traceability
systems and practices, we expect that the three-year compliance period prior to the effective date
of the rule will allow necessary changes to take place within the scope of routine updates to
standard operating procedures in the normal course of business. We estimate that about 212,368
firms will incur incremental costs of the rule due to this provision.
Additionally, planning for traceability will likely be more complex for some businesses
than others. For instance, only certain entities will be responsible for assigning traceability lot
codes. Additionally, we do not generally expect growers and other producers of raw agricultural
commodities to incur substantive costs in identifying covered foods, since their products would
not contain multiple ingredients and thus identifying covered foods would be straightforward.

136

Growers who do not pack also will not need to assign lot codes. Given that FDA has previously
received farm maps with field names and coordinates during outbreak investigations, and given
the widespread use of free, mapping and direction websites and web applications with GPS
coordinate plotting functionality, we expect most affected entities either already keep the
required map or will be able to produce it in minutes. We intend our estimate in Table 29 of the
time for growers to produce a traceability plan to represent an average. Individual growers who
initially pack might spend more time, while those who do not initially pack will likely spend less.
We also note that under the final rule’s partial exemption of ad hoc sales between
retailers, the purchasing retailer needs only to keep information typically communicated on
ordinary sales receipts and there are no additional traceability requirements on the selling
retailer. Thus, restaurants and other retailers under the final rule are generally only receivers, but
not shippers, of covered foods. We expect these entities will identify covered foods based on the
required records provided to them by suppliers. Alternatively, retailers might opt to keep all
receipts and delivery documents for two years or might already do so for tax purposes. Because
the final rule does not require retailers to regularly use the received information for compliance
purposes, and because compliant storage options include taking pictures or not digitizing at all,
we do not expect restaurants and other retailers to develop extensive or highly formal plans for
maintaining traceability records beyond determining physical or digital storage locations.
Table 29 summarizes our estimates of the one-time costs of creating a traceability plan,
including identifying covered foods. Experts in ERG’s panel provided descriptive comment on
the types and numbers of employees who might work on procedures for traceability (Ref. [4]).
We interpreted their comments in estimating the numbers of employees below. After reviewing
descriptions of tasks involved, we assumed the numbers of hours per employee below. While

137

different types of employees might work on traceability plans, we use the wage of a Business
Operations Specialist (occupation code 13-1000) in the 2020 BLS Occupational Employment
and Wage Statistics, taking the average from different industries for different types of firms. For
growers and other producers of raw agricultural commodities, we use the average wage from
Agriculture, Forestry, Fishing, and Hunting, $29.51. For manufacturers and distributors, we use
the average wage from Food Manufacturing, $32.19. For retailers we use the average wage from
Food & Beverage Stores, $23.48. For restaurants, we use the average for Food Service and
Drinking Places, $23.48. We double all wages to account for benefits and overhead. We estimate
the total cost (sum of costs to small and large firms in the table below) of this provision of the
rule to range from about $15 million to $345 million, with a primary estimate of about $79
million.
Table 29. One-Time Traceability Plan Costs of the Final Rule (2020$)
Small
Primary
Low
High
Primary
Farms/ growers (produce)
(a) Number of
3
1
10
10
employees needed
(b) Hours
10
10
10
10
(c) Hourly labor cost
$59.02
$59.02
$59.02
$59.02
(d) Cost per firm =
$1,771
$590
$5,902
$5,902
a*b*c
(e) Number of firms
needing new
3,303
1,651
4,624
202
traceability plan
$5,848,2 $974,71 $27,291,9 $1,194,0
Subtotal = d*e
66
1
09
34
Shell Egg Producers
(a) Number of
3
1
10
10
employees needed
(b) Hours
10
10
10
10
(c) Hourly labor cost
$59.02
$59.02
$59.02
$59.02
(d) Cost per firm =
$1,771
$590
$5,902
$5,902
a*b*c

Large
Low

High

10
$59.02

$1,180

$17,706

231

$4,093,8
32

10
$59.02

$1,180

$17,706
138

(e) Number of firms
needing new
traceability plan
Subtotal = d*e
(a) Number of
employees needed
(b) Hours
(c) Hourly labor cost
(d) Cost per firm =
a*b*c
(e) Number of firms
needing new
traceability plan
Subtotal = d*e
(a) Number of
employees needed
(b) Hours
(c) Hourly labor cost
(d) Cost per firm =
a*b*c
(e) Number of firms
needing new
traceability plan
Subtotal = d*e
(a) Number of
employees needed
(b) Hours
(c) Hourly labor cost
(d) Cost per firm =
a*b*c
(e) Number of firms
needing new
traceability plan
Subtotal = d*e

1,250
$2,213,4
54

1,000

2,000

$590,25 $11,805,0
4
86
Fishing/ aquaculture

$56,048

$0-

$261,558

10
$59.02

$1,771

$590

$5,902

$1,180

$17,706

998

599

1,297

$8,725

$628,228

$1,766,5
38

$353,30 $7,654,99
$139,606
8
7
Manufacturing/ processing

20
$64.38

$3,863

$1,288

$12,876

$2,575

$38,628

3,924

2,354

4,708

$1,665,7
71

$15,156, $3,031,3 $60,626,3
$462,714
579
16
16
Wholesale/ Distribution/ Warehouse/ Storage
3

20
$64.38

$3,863

$1,288

$12,876

$2,575

$38,628

8,002

5,334

10,669

529

235

823

$6,811,9
05

$605,50
3

$31,788,
889

$30,908,
472

$6,868,5 $137,370,
49
985
Non-restaurant retail

139

(a) Number of
employees needed
(b) Hours
(c) Hourly labor cost
(d) Cost per firm =
a*b*c
(e) Number of firms
needing new
traceability plan
Subtotal = d*e
(a) Number of
employees needed
(b) Hours
(c) Hourly labor cost
(d) Cost per firm =
a*b*c
(e) Number of firms
needing new
traceability plan
Subtotal = d*e
Total

0.5
$46.96

$70

$23

$235

$47

$704

58,168

30,349

91,046

631

252

883

$148,115

$11,849

$622,083

$4,097,3
68

$712,58 $21,377,5
6
71
Restaurants

0.5
$46.96

$70

$23

$235

$47

$704

133,983

83,162

166,324

1,309

476

1,666

$9,437,7
89
$69,428,
465

$1,952,6
46
$14,483,
370

$39,052,9
21
$305,179,
786

$307,438

$22,359

$9,119,8
61

$648,43
6

$1,173,8
55
$40,234,
217

b. Seed Lot Records for Sprouts
Under the final rule initial packers of sprouts will need to maintain records regarding the
seeds they use for sprouting, and we expect that sprout growers will incur costs to establish and
maintain these records. This required information includes:


the location description of the grower of seeds for sprouting and the date of seed
harvesting, if either is available,



the location description of the seed conditioner or processor and the associated seed lot
code,



the date of conditioning or processing,
140



the location description of the seed packinghouse, including any repackers (if applicable),



the associated seed lot code assigned by the seed packinghouse (if applicable),



the date of packing (and repacking, if applicable),



the location description of the seed supplier,



any seed lot code assigned by the seed supplier, including the master lot and sub-lot
codes,



any new seed lot code assigned by the sprouter,



a description of the seeds, including the seed type or taxonomic name, growing
specifications, type of packaging, and (if applicable) antimicrobial treatment,



the date of receipt of the seeds by the sprouter, and



the reference document type and reference document number.
We estimate that sprout growers not already performing certain recordkeeping activities

would incur new recurring recordkeeping costs for the records outlined above. Some sprout
growers might already keep some of the required records as described in the 2017 FDA draft
guidance for the sprout operations industry (Ref. [41]) or as recommended by good agricultural
practices. In this analysis, we use the inventory of sprout farms and operations used by the
FDA’s Office of Regulatory Affairs. Excluding very small sprout growers, this internal inventory
counts 95 sprout growers that we believe will incur recurring costs due to this provision. We
assume the same proportion of these growers are small as the proportion among other produce
growers.
Table 30 summarizes our estimates of the annual recordkeeping costs of the final rule on
growers of sprouts. We estimate the annual number of FTL lots based on input elicited by ERG
from the expert panel (Ref. [4]), assuming that growers of sprouts grow as many lots as the
141

number of lots that other produce growers handle. We estimate the recordkeeping times below
using experts’ input, adjusting for our expected degree of electronic recordkeeping and
differences between the requirements of the final and proposed rule (with additional brief
definitions of some new CTEs in their draft-final state at the time of the traceability costs
elicitation).65 Because we are unable to separate entities who do not initially pack from those
who do, we assume, for the purpose of estimating initial packing costs in section II.F.5.d, that all
sprout growers initially pack. However, for the purpose of estimating costs to sprout growers, we
assume that sprout growers will incur costs to provide seed lot records to initial packers. To the
extent that sprout growers initially pack their own sprouts, estimating costs to provide seed lot
records to initial packers results in an overestimate of costs. To estimate hourly labor cost, we
use the average wage of Agricultural Workers (code 45-2000) in Agriculture, Forestry, Fishing
and Hunting, which is $14.53. We double this to account for benefits and overhead, obtaining an
hourly labor cost of $29.06. We estimate the total cost (sum of costs to small and large firms in
the table below) of this provision to growers of sprouts to range from about $4,000 to about
$836,000, with a primary estimate of about $97,000.
Table 30. Annual Recordkeeping Costs of Growing Sprouts (2020$)
Small
Large
Primary
Low
High
Primary
Low
Sprout growers
(a) Number of
seed lots per
832
364
2,600
1,456
1,456
establishment
(b) Hours per lot
0.02
0.002
0.07
0.02
0.0007
to capture record

High

9,100
0.03

As explained in the beginning of section II.F.5, we scaled experts’ estimates of manual entry times by the
proportion of industry they estimated to have electronic recordkeeping capabilities in order to account for baseline
practices in estimating the incremental burden of the rule. Our estimates of time burden per record therefore
represent averages between manual and electronic recordkeeping weighted by the baseline prevalence of these
practices. For our primary estimates, we thus estimate that about 60 percent of small and large businesses will keep
records manually at about two minutes per record, while the remainder will scan records at about 2.5 seconds per
record.

142

(c) Hours per lot to
provide record
(d) Labor cost of
hourly employee
(e) Number of
covered
establishments
Total =
a*(b+c)*d*e

0.02

0.002

0.04

0.02

0.0005

0.03

$29.06

$85,700

$3,686

$734,879

$11,549

$380

$101,100

c. Records of Harvesting or Cooling a Food on the Food Traceability List (§ 1.1325)
The final rule will require entities that harvest or cool raw agricultural commodities on
the FTL other than those obtained from a fishing vessel to maintain traceability records and to
make these records available to the initial packer of the foods they harvest or cool.
Specifically, entities that harvest FTL foods will need to keep records describing the
food, including the commodity and (if applicable) variety of the food, the quantity harvested and
the unit of measure of the food, the harvest location (including the name of the field or
aquaculture container in which food was grown or raised), the location of the immediate
subsequent recipient (other than a transporter), harvest date, and the reference document type and
reference document number. Harvesters must provide this information (except the reference
document type and number) to the initial packer of the food, along with their own business name
and phone number. Entities that cool FTL foods before initial packing will need to maintain
records describing the food, including the location of the immediate subsequent recipient (other
than a transporter), the commodity and (if applicable) variety of the food, the quantity cooled and
unit of measure of the food, the cooling location, the cooling date, the harvest location, and the
reference document type and reference document number. Coolers must provide this information
(except the reference document type and number) to the initial packer of the food.

143

Some harvesters and coolers might already follow practices meeting the requirements in
this provision. Experts in ERG’s panel estimated the average proportion of small and large
harvesters and coolers already keeping records (Ref. [4]). In these cases, we expect that existing
recordkeeping practices already include date and location of harvest or cooling, and the food,
quantity, and subsequent recipient. We expect that amending existing location information to
include the name of the field or aquaculture container in which food was grown or raised will
occur via a letter or number designation (e.g., Tank A, B, etc.) and not appreciably increase
current recordkeeping time. However, entities who do not currently keep records for these
activities will incur a new recurring recordkeeping burden.
We estimate the total number of harvesters and coolers affected by identifying NAICS
categories likely to harvest or cool foods on the FTL and removing exempt and non-covered
entities. We assume that all growers and other producers of raw agricultural commodities (other
than those obtained from a fishing vessel) harvest food, but that only those growers who cool
also perform initial packing. We estimate that about 6,058 establishments that harvest and about
3,511 that cool FTL foods will incur recurring costs due to this provision of the rule.
Table 31 summarizes our estimates of the annual costs to harvesting and cooling
establishments belonging to small and large firms. Under the final rule, traceability lot codes are
not assigned prior to initial packing. Hence, we expect harvesters and coolers to be able to satisfy
the requirements of the final rule via relatively few instances of recordkeeping compared with
transformers (e.g., once per commodity per field per harvest date per immediate subsequent
recipient). While affected entities may keep and provide records with varying frequency, we
assume that they will keep one record and provide one record with each truckload delivered to a
subsequent recipient. We estimated numbers of truckloads after reviewing comments by experts

144

in ERG’s panel. We estimate the recordkeeping times below using experts’ input, adjusting for
the elicited degree of electronic recordkeeping and differences between the requirements of the
final and proposed rule (with additional brief definitions of some new CTEs in their draft-final
state at the time of the traceability costs elicitation).66 To estimate labor cost, we use the average
wage of an Agricultural Worker (occupation code 45-2000) in Agriculture, Forestry, Fishing,
and Hunting from the 2020 BLS Occupational Employment and Wage Statistics, which is
$14.53. We double the wage to $29.06 to account for benefits and overhead. We estimate the
total recurring costs (sum of costs to small and large firms in the table below) of recordkeeping
related to harvesting and cooling to range from about $0.2 million to about $36.2 million, with a
primary estimate of about $4.9 million.
Table 31. Annual Recordkeeping Costs of Harvesting and Cooling (2020$)
Small
Large
Primary
Low
High
Primary
Low
Harvesters
(a) Truckloads per
549
366
732
1,098
915
establishment
(b) Hours to capture
0.01
0.001
0.04
0.01
0.001
per truckload
(c) Hours to provide
0.02
0.001
0.13
0.02
0.001
per truckload
(d) Hourly labor cost $29.06
$29.06
$29.06
$29.06
$29.06
(e) Establishments
not already
4,782
1,435
7,173
128
32
capturing
(f) Establishments
5,738
5,738
6,934
320
256
not already
providing

High

1,281
0.03
0.11
$29.06
192
352

145

Subtotal = a*b*d*e
+ a*c*d*f

$2,862,9
12

(a) Truckloads per
549
establishment
(b) Hours to capture
0.02
per truckload
(c) Hours to provide
0.02
per truckload
(d) Hourly labor cost $29.06
(e) Establishments
not already
2,171
capturing
(f) Establishments
3,365
not already
providing
Subtotal = a*b*d*e
$1,646,4
+ a*c*d*f
70
$4,509,3
Total
81

$113,57
1

$25,255,5
44
Coolers

$251,716

$5,760

$1,747,598

366

732

1,098

915

1,281

0.00

0.05

0.02

0.00

0.02

0.00

0.07

0.02

0.00

0.11

$29.06

1,737

2,605

3,365

4,124

146

207

$115,356

$3,170

$932,450

$367,072

$8,929

$2,680,048

$93,199
$206,77
0

$8,280,65
0
$33,536,1
94

d. Records of Initial Packing of Raw Agricultural Commodities on the FTL (§ 1.1330)
The final rule will require entities that initially pack raw agricultural commodities on the
FTL other than those obtained from a fishing vessel to maintain traceability records. Initial
packing records must include:


the commodity and (if applicable) variety of the food received, and the product
description of the packed food,



the date the food was received,



the traceability lot code assigned by the packer,



the quantity and the unit of measure of the food received, and of the packed food,



the packing location,



the packing date,

146



the harvest location (including the name of the field or aquaculture container in which
food was grown or raised),



harvest dates,



business name and phone number for the harvester,



cooling location and dates (if applicable), and



the reference document type and reference document number.
To the extent that initial packers do not already maintain these records, these entities will

face a recurring recordkeeping cost at the establishment level to comply with the final rule. We
assume that some of these requirements together or in part will impose a new recordkeeping
burden on all covered establishments that initially pack and estimate the average burdens for
establishments owned by small and large firms. We estimate that about 4,218 total initial packing
establishments will incur recurring costs due to this provision.
For initial packing of sprouts, records must also include the data elements described
above in section II.F.5.b “Records of Growing a Food on the Food Traceability List.” We expect
the incidence of these costs will fall on growers of sprouts and therefore estimate them in section
II.F.5.b. In this section, we estimate costs to initial packers of sprouts similarly to the costs that
we expect other initial packers to face.
Not all growers and producers of raw agricultural commodities are initial packers. For
growers of produce other than sprouts, we estimate the number of initial packing establishments
using the USDA NASS 2017 Census of Agriculture. However, for sprout growers, shell egg
producers, and aquaculture operations, we are unable to separate entities who do not initially
pack from those who do and therefore assume that all are initial packers.

147

Table 32 summarizes our estimates of the annual recordkeeping costs of the final rule to
initial packing establishments owned by small and large firms. We estimate the annual number
of FTL lots initially packed based on input elicited by ERG from the expert panel (Ref. [4]). We
estimate the recordkeeping times below using experts’ input, adjusting for the elicited degree of
electronic recordkeeping and differences between the requirements of the final and proposed rule
(with additional brief definitions of some new CTEs in their draft-final state at the time of the
traceability costs elicitation).65 To estimate labor cost, we use the average wage of an
Agricultural Worker (occupation code 45-2000) in Agriculture, Forestry, Fishing, and Hunting
from the 2020 BLS Occupational Employment and Wage Statistics, which is $14.53. We double
the wage to $29.06 to account for benefits and overhead. We estimate the total recurring
recordkeeping costs (sum of costs to small and large firms in the table below) related to initial
packing to range from about $0.1 million to $22.8 million, with a primary estimate of about $2.2
million.
Table 32. Annual Recordkeeping Costs of Initial Packing (2020$)
Small
Primary
Low
High
Primary
Initial Packers
(a) FTL lots per
832
364
2,600
1,456
year
(b) Hours to
0.02
0.002
0.07
0.02
capture per lot
(c) Hourly labor
$29.06
$29.06
$29.06
$29.06
cost
(d) Number of
covered
4,022
4,022
4,022
195
establishments
$2,028,02
Total = a*b*c*d
$91,961 $21,359,019 $183,872
5

Large
Low

High

1,456

9,100

0.001

0.03

$29.06

195

$5,766

$1,454,89
8

148

e. Records of First Land-Based Receiving of Seafood on the FTL (§ 1.1335)
The final rule will require entities performing the first land-based receiving of FTL food
that was obtained from a fishing vessel to maintain traceability records. First land-based
receivers of these foods will need to link certain information covered below to the traceability
lot. First land-based receiver records must include:


the traceability lot code they assign,



a description of the food,



the quantity and unit of measure of the food,



the first land-based receiver location description (i.e., the traceability lot code source) and
(if applicable) the traceability lot code source reference,



the date the food was landed,



the harvest date range,



the harvest locations, and



the reference document type and reference document number.
We estimate the total number of first land-based receivers affected based on NAICS

311710 Seafood Product Preparation and Packaging. After removing exempt and non-covered
entities, we estimate that the final rule will impose recurring costs on about 319 small
establishments and 48 large establishments that perform first land-based receiving of FTL foods
obtained from a fishing vessel.
Table 33 summarizes our estimates of the annual recordkeeping costs of the final rule to
first land-based receiving establishments owned by small and large firms. We estimate the
annual number of FTL lots processed based on input elicited by ERG from the expert panel (Ref.
[4]). We estimate the recordkeeping times below using experts’ input, adjusting for the elicited
149

degree of electronic recordkeeping and differences between the requirements of the final and
proposed rule (with additional brief definitions of some new CTEs in their draft-final state at the
time of the traceability costs elicitation).67 To estimate labor cost, we use the average wage of an
Agricultural Worker (occupation code 45-2000) in Agriculture, Forestry, Fishing, and Hunting
from the 2020 BLS Occupational Employment and Wage Statistics, which is $14.53. We double
the wage to $29.06 to account for benefits and overhead. We estimate the total recurring costs
(sum of costs to small and large firms in the table below) of recordkeeping related to first landbased receiving of FTL food to range from about $0.004 million to $1.5 million, with a primary
estimate of about $0.3 million.
Table 33. Annual Recordkeeping Costs of First Land-Based Receiving of FTL Food
(2020$)
Small
Large
Primary
Low
High
Primary
Low
High
First Land-Based Receivers of FTL food Obtained from a Fishing Vessel
(a) FTL lots per
871
187
1,560
5,460
365
13,000
establishment
(b) Hours to
0.02
0.002
0.07
0.02
0.001
0.03
capture per lot
(c) Hourly labor
$29.06
$29.06
$29.06
$29.06
$29.06
$29.06
cost
(d) Number of
covered
319
319
319
48
48
48
establishments
$1,041,1 $136,76
Total = a*b*c*d
$163,900
$3,751
$354
$456,091
29
8

150

f. Records to Be Kept and Provided When Shipping Foods on the Food Traceability List (§
1.1340)

The final rule will require entities who ship foods on the FTL to maintain and provide
traceability records. Entities shipping FTL foods must maintain and link the following
information to the traceability lot and provide it to the immediate subsequent recipient:


the traceability lot code for the food,



the quantity shipped and unit of measure,



a description of the food,



the location from which they ship the food,



the location of the immediate subsequent recipient,



the ship date,



the location description of the traceability lot code source (TLCS) or source reference,
and



the reference document type and reference document number
Shippers must additionally provide the above information to the immediate subsequent

recipient (other than a transporter), except for the reference document type and reference
document number. We note that the traceability lot code source reference can be an internet link
or other means of digitally accessing the required information. Additionally, shipping
recordkeeping requirements do not apply to shipments of food prior to initial packing.
We estimate the total number of shippers affected by identifying NAICS categories likely
to ship foods on the Food Traceability List and removing exempt and non-covered entities. We
expect most categories of supply chain entities that handle FTL foods to incur recordkeeping
costs associated with shipping. However, as previously mentioned, the final rule’s partial
151

exemption of ad hoc sales between retailers places no traceability requirements on the selling
retailer, while the purchasing retailer can satisfy minimal requirements by keeping the sales
receipt. Thus, restaurants and other retailers under the final rule are generally only receivers, but
not shippers, of FTL foods. We thus estimate that the final rule will impose recurring costs on
approximately 24,909 small and 6,524 large establishments that ship FTL foods.
We expect that entities making capital investments will be motivated by interoperability
to prioritize streamlining shipping and receiving recordkeeping, since these are the CTEs
performed by the most entities across the supply chain. In our estimates of the time burden per
record for shipping, we therefore consider the proportion of entities making capital investments,
in addition to those currently already performing electronic recordkeeping, to be capable of
keeping and sending records via scanning and barcodes (or related technology, e.g., RFID).
Table 34 summarizes our estimates of the annual recordkeeping costs of the final rule to
establishments, owned by small and large firms, that ship foods on the FTL. We estimate the
annual number of lots shipped based on input elicited from the expert panel (Ref. [4]). We
estimate the recordkeeping times below using experts’ input, adjusting for our expected degree
of electronic recordkeeping and differences between the requirements of the final and proposed
rule (with additional brief definitions of some new CTEs in their draft-final state at the time of
the traceability costs elicitation).68 To estimate the hourly labor cost to growers and other
producers of raw agricultural commodities, we use the average wage of an Agricultural Worker

As explained in the beginning of section II.F.5, we scaled experts’ estimates of manual entry times for each
activity by the proportion of industry they estimated to have electronic recordkeeping capabilities in order to account
for baseline practices in estimating the incremental burden of the rule. Our estimates of time burden per record
therefore represent averages between manual and electronic recordkeeping weighted by the baseline prevalence of
these practices. In our estimates of the time burden per record for shipping and receiving, our weighting of this
average additionally considers the proportion of entities making capital investments to be capable of keeping and
sending records via scanning versus manual entry. For our primary estimates, we thus estimate that about seven
percent of small and large businesses will keep records manually at about two minutes per record, while the
remainder will scan records at about 2.5 seconds per record.

152

(occupation code 45-2000) in Agriculture, Forestry, Fishing, and Hunting from the 2020 BLS
Occupational Employment and Wage Statistics, $14.53, which we double to $29.06 to account
for benefits and overhead. To estimate the hourly labor cost to manufacturers and distributors,
we use the average wage of a Food Processing Worker (occupation code 51-3000) in Food
Manufacturing, $15.73, which we double to $31.46 to account for benefits and overhead. We
estimate the total recurring costs (sum of costs to small and large firms in the table below) of
recordkeeping related to shipping to range from about $0.5 million to $123.8 million, with a
primary estimate of about $30.3 million.
Table 34. Annual Recordkeeping Costs of Shipping (2020$)
Small
Primary
Low
High
Primary
Produce Farms and Sprout Growers
(a) FTL lots per year
(b) Hours to capture
per lot
(c) Hours to provide
per lot
(d) Hourly labor cost
(e) Number of
covered
establishments
Subtotal =
a*(b+c)*d*e

Large
Low

High

832

364

2,600

1,456

9,100

0.003

0.001

0.013

0.003

0.0003

0.004

0.003

0.001

0.013

0.003

0.0005

0.007

$29.06

1,065

134

$151,273

$15,978

$2,129,719

$35,475

$4,588

$388,203

Shell Eggs Producers

(a) FTL lots per year
(b) Hours to capture
per lot
(c) Hours to provide
per lot
(d) Hourly labor cost
(e) Number of
covered
establishments

832

364

2,600

1,456

9,100

0.003

0.001

0.013

0.003

0.0003

0.004

0.003

0.001

0.013

0.003

0.0005

0.007

$29.06

2,500

153

Subtotal =
a*(b+c)*d*e

$355,251

$37,524

$5,001,454

$5,600

$724

$61,281

Fishing/ aquaculture

(a) FTL lots per year
(b) Hours to capture
per lot
(c) Hours to provide
per lot
(d) Hourly labor cost
(e) Number of
covered
establishments
Subtotal =
a*(b+c)*d*e

1,040

364

2,080

3,120

1,483

16,900

0.003

0.001

0.013

0.003

0.0003

0.004

0.003

0.001

0.013

0.003

0.0005

0.007

$29.06

457

$81,221

$6,863

$731,833

$23,141

$1,422

$219,464

Manufacturing/ processing

(a) FTL lots per year
(b) Hours to capture
per lot
(c) Hours to provide
per lot
(d) Hourly labor cost
(e) Number of
covered
establishments
Subtotal =
a*(b+c)*d*e
(a) FTL lots per year
(b) Hours to capture
per lot
(c) Hours to provide
per lot
(d) Hourly labor cost
(e) Number of
covered
establishments
Subtotal =
a*(b+c)*d*e
Total

871

187

1,560

5,460

365

13,000

0.003

0.001

0.013

0.003

0.0003

0.004

0.003

0.001

0.013

0.003

0.0005

0.007

$31.46

8,145

429

$3,999

$1,927,961

$10,583,36
$462,482
3
Wholesale/ Distribution/ Warehouse/ Storage

$1,311,614

$67,987

4,875

202

4,940

14,040

1,500

24,700

0.003

0.001

0.013

0.003

0.0003

0.004

0.003

0.001

0.013

0.003

0.0005

0.007

$31.46

12,742

5,900

$52,429,12
4
$70,875,49
4

$16,343,09
9
$16,869,79
7

$11,484,42
1
$13,383,78
0

$114,890
$243,243

$225,840
$236,573

$50,340,43
1
$52,937,34
0

154

g. Records of Receipt of Foods on the Food Traceability List (§ 1.1345)
The final rule will require entities who receive foods on the FTL to maintain traceability
records. Entities receiving FTL foods must maintain records containing:


the traceability lot code of the foods,



the quantity received and unit of measure,



a description of the food,



the location for the immediate previous source (other than a transporter),



the location where the food was received,



the date of receipt,



the location of the TLCS, or a TLCS reference, and



the reference document type and reference document number.
As noted previously, entities receiving FTL foods do not need to establish a record, but

rather only need to maintain the record provided to them by the shipper (provided that record
includes all the required KDEs). In our analysis, we have also accounted for the fact that
receivers of FTL foods do not need to retrieve TLCS information made available via a TLCS
reference (e.g., following a web address to retrieve TLCS information), but are allowed to store
whatever record the shipper provides containing the TLCS reference. Additionally, receiving
recordkeeping requirements do not apply to shipments of food prior to initial packing, or to
receipt of a food by the first land-based receiver.
We estimate the total number of receivers affected by identifying NAICS categories
likely to receive foods on the FTL and removing exempt and non-covered entities. Particularly,
since shipments prior to initial packing or first land-based receiving do not require receiver
recordkeeping, we do not expect harvesters, coolers, and initial packers—whether or not they are
155

also growers—or first land-based receivers to incur receiver costs under the final rule. We
assume that all other entities who receive FTL foods will incur some recurring cost to keep
receiver records. We estimate that about 470,580 establishments will incur recurring costs due to
this provision.
However, we expect recordkeeping requirements to require more sophisticated tasks of
some supply chain entities than others. On one hand, intermediate supply chain entities, such as
manufacturers and distributors, will need to capture information upon receipt in a way that can
link incoming and outgoing product. On the other hand, entities at the end of the supply chain,
such as restaurants and other retailers, only need to maintain receiving records. We therefore
anticipate that entities at the end of the supply chain are likely to store the records provided by
their suppliers in whatever format they are provided (e.g., receipts, labels, electronic documents,
etc.).
We expect that entities making capital investments will be motivated by interoperability
to prioritize streamlining shipping and receiving recordkeeping, since these are the CTEs
performed by the most entities across the supply chain. In our estimates of the time burden per
record for receiving, we consider the proportion of entities making capital investments, in
addition to those currently already performing electronic recordkeeping, to be capable of keeping
records via scanning and the use of barcodes (or related technology, e.g., RFID).
Table 35 summarizes our estimates of the annual recordkeeping costs of the final rule to
establishments, owned by small and large firms, that receive foods on the FTL. We estimate the
annual number of lots received based on input elicited from the expert panel (Ref. [4]). We
estimate the recordkeeping times using experts’ input, adjusting for our expected degree of
electronic recordkeeping and differences between the requirements of the final and proposed rule

156

(with additional brief definitions of some new CTEs in their draft-final state at the time of the
traceability costs elicitation).69 To estimate the hourly labor cost to manufacturers and
distributors, we use the average wage of a Food Processing Worker (occupation code 51-3000)
in Food Manufacturing from the 2020 BLS Occupational Employment and Wage Statistics,
$15.73, which we double to $31.46 to account for benefits and overhead. To estimate the hourly
labor cost to restaurants and other retailers, we use the average wage of a Retail Sales Worker
(occupation code 41-2000) in Food and Beverage Stores, $12.91, which we double to $25.82 to
account for benefits and overhead. We estimate the total recurring recordkeeping costs (sum of
costs to small and large firms in the table below) related to receiving FTL foods to range from
about $5.6 million to $681.7 million, with a primary estimate of about $220.3 million.

Table 35. Annual Recordkeeping Costs of Receiving (2020$)
Small
Primary
Low
High
Primary
Manufacturing/ processing
(a) FTL lots
871
187
1,560
4,680
per year
(b) Hours per
0.003
0.001
0.008
0.003
lot to keep
(c) Hourly
$31.46
$31.46
$31.46
$31.46
labor cost
(d) Number of
covered
8,111
8,111
8,111
426
establishments
Subtotal =
$653,058
$35,870
$3,092,591 $190,075
a*b*c*d

Large
Low

High

365

13,000

0.0003

0.004

$31.46

426

$1,683

$639,639

157

Wholesale/ Distribution/ Warehouse/ Storage
(a) FTL lots
per year
(b) Hours per
lot to keep
(c) Hourly
labor cost
(d) Number of
covered
establishments
Subtotal =
a*b*c*d
(a) FTL lots
per year
(b) Hours per
lot to keep
(c) Hourly
labor cost
(d) Number of
covered
establishments
Subtotal =
a*b*c*d
(a) FTL lots
per year
(b) Hours per
lot to keep
(c) Hourly
labor cost
(d) Number of
covered
establishments
Subtotal =
a*b*c*d
Total

4,615

202

4,940

17,420

1,500

32,500

0.003

0.001

0.008

0.003

0.0003

0.004

$31.46

12,742

5,900

$5,435,959

$60,872

$95,701

$22,122,77
1

4,550

520

7,800

28,600

2,080

28,600

0.003

0.001

0.008

0.003

0.0003

0.004

$25.82

115,451

55,929

$39,853,96
8

$1,165,260

$125,032,
354

$1,032,52
5

$151,472,8
88

1,560

520

5,200

3,120

2,080

15,600

0.003

0.001

0.008

0.003

0.0003

0.004

$25.82

215,359

56,662

$224,643,2
63
$423,762,9
60

$13,818,5
43
$148,828,
924

$1,046,04
9
$2,175,95
8

$83,703,71
8
$257,939,0
16

$25,488,85
3
$71,431,83
8

$15,384,94 $9,787,95
0
2
Non-restaurant retail

$180,642,1
66
Restaurants

$2,173,643
$3,435,645

158

h. Records of Transformation of Foods on the Food Traceability List (§ 1.1350)
The final rule will require entities who transform food on the FTL to maintain traceability
records. For covered food used in transformation (if applicable), entities must keep records that
contain:


the traceability lot code of the food transformed,



a description of food transformed to which the traceability lot code applies, and



the quantity and unit of measure of food from each TLC transformed.
For covered food that was produced through transformation, entities must keep records

that contain:


the new traceability lot code of the food,



the location description for where the food was transformed,



the date transformation was completed,



a description of the food post transformation,



the quantity and unit of measure of food post transformation, and



the reference document type and reference document number for the transformation
event.
Transformation recordkeeping requirements do not apply to retail food establishments

and other food service providers with respect to foods they do not ship (e.g., foods they sell or
send directly to consumers). Transformation recordkeeping requirements also do not apply to
transformation of a raw agricultural commodity (other than a food obtained from a fishing
vessel) on the FTL that was not initially packed prior to transformation. In that situation, initial
packing records must be kept instead.

159

As in previous sections, we estimate the total number of affected entities transforming
FTL foods by identifying NAICS categories likely to transform foods on the Food Traceability
List and removing exempt and non-covered entities. We expect that all covered manufacturers
will incur recurring costs at the establishment level to keep records of transformation under the
final rule. We estimate that about 8,574 manufacturing or processing establishments will incur
recurring costs due to this provision of the rule. We expect that entities affected by this provision
will incur annual recordkeeping costs at the establishment level.
Table 36 summarizes our estimates of the annual recordkeeping costs of the final rule to
establishments owned by small and large firms that transform foods on the FTL or that produce
FTL foods through transformation. We estimate the annual number of lots transformed based on
input elicited from the expert panel (Ref. [4]). We estimate the recordkeeping times using
experts’ input, adjusting for the elicited degree of electronic recordkeeping and differences
between the requirements of the final and proposed rule (with additional brief definitions of
some new CTEs in their draft-final state at the time of the traceability costs elicitation).70 To
estimate the hourly labor cost, we use the average wage of a Food Processing Worker
(occupation code 51-3000) in Food Manufacturing from the 2020 BLS Occupational
Employment and Wage Statistics, $15.73, which we double to $31.46 to account for benefits and
overhead. We estimate total recurring recordkeeping costs (sum of costs to small and large firms
in the table below) of transformation to range from about $0.1 million to $43 million, with a
primary estimate of about $6 million.
70

160

Table 36. Annual Recordkeeping Costs of Transforming (2020$)
Small
Primary
Low
High
Primary
Manufacturing/ processing
FTL lots per year
871
187
1,560
5,460
Hours per lot to
0.02
0.002
0.08
0.02
capture
Hourly labor cost
$31.46
$31.46
$31.46
$31.46
Number of
covered
8,145
8,145
8,145
429
establishments
$4,525,74
$1,442,65
Total
$87,453 $31,351,044
7
6

Large
Low

High

365

13,000

0.001

0.07

$31.46

429

$7,275

$12,092,811

i. Electronic Sortable Spreadsheet Upon Request
When necessary to help FDA prevent or mitigate a foodborne illness outbreak, assist in
the implementation of a recall, or to otherwise address a threat to the public health, some entities
would be required to provide FDA with the information required under subpart S in the form of
an electronic sortable spreadsheet. While some may already keep records in sortable electronic
spreadsheets, others might need to put their records for the requested lots and dates in an
electronic sortable spreadsheet format upon FDA request. The final rule exempts farms with
average annual sales of no more than $250,000 and other supply chain entities with average
annual sales of no more than $1 million dollars from having to provide information in the form
of an electronic sortable spreadsheet. As this will be a low probability event for any given
establishment, we treat our estimated number of annual requests by FDA as the number of
affected establishments per year (thus assuming that no establishment receives more than one
such request in the same year).
Table 37 summarizes our estimates of the annual cost of providing traceability
information to FDA in the form of an electronic sortable spreadsheet upon request. We estimate
161

that FDA will make between 40 and 110 such requests annually (to entities not exempt from the
spreadsheet requirement) based on internal counts of CORE assignments (information requests)
between 2016 and 2021. We estimate that such requests will entail, on average, between eight
and twenty-four total hours of formatting information into a spreadsheet. We expect that the type
of employees formatting spreadsheets will be roughly equivalent to supervisors of food
preparation (occupation code 35-1010 in NAICS 445), the mean wage of which is $20.12 per
hour, which we double to $40.24 to account for benefits and overhead. We estimate the total
annual cost of formatting responses to requests for traceability information as electronic sortable
spreadsheets to range from about $0.01 million to $0.1 million, with a primary estimate of about
$0.05 million.
Table 37. Annual Cost of Providing Electronic Sortable Spreadsheets Upon Request

(a) Average hours to generate
spreadsheet
(b) Hourly labor cost
(c) Expected annual requests to
establishments not exempt from
spreadsheet requirement
Total cost = a*b*c

Primary

Low

High

16
$40.24

8
$40.24

24
$40.24

110

$48,288.00

$12,876.80

$106,233.60

6. Non-Quantified Costs
The information flows brought about by the final rule may prompt new protective
actions—for example, in farming, manufacturing or cooking processes—that themselves would
have costs. These costs have not been quantified due to lack of data; however, there is a likely
correlation between these costs’ occurrence and the realization of health and longevity benefits
attributable to this rule. One of the challenges of such attribution, for both health and longevity

162

benefits and this category of costs, is the lag in data availability as other FSMA regulations
continue to take effect.71
More generally, provisions of the final rule might generate costs that we cannot quantify,
as explained below. FDA might incur costs to review petitions requesting modified requirements
or exemptions (§1.1380), adopt modified requirements or grant exemptions on our own initiative
(§1.1385), decide that modified requirements or exemptions should be revised or revoked
(§1.1400), receive and respond to waiver petitions (§1.1435), waive requirements on our own
initiative (§1.1440), determine that a waiver should be modified or revoked (§1.1450), and
respond to a failure to comply72 with the provisions of the final rule (§1.1460).
For provisions concerning petitions, costs to FDA might include time spent reviewing
and responding, as well as publishing notices of decisions in the Federal Register. For provisions
that allow FDA to modify requirements and grant exemptions and waivers on our own initiative,
costs could include time spent making these determinations and publishing notices of decisions
in the Federal Register. Because we cannot estimate the number of petitions FDA will receive,
we cannot estimate the costs of these provisions.
Other one-time costs would result from time spent completing and submitting petitions
for modified requirements or exemptions (§ 1.1370), petitions for waiver for a type of entity (§

As noted previously, the outcomes of earlier FSMA regulations should be taken into account in the
characterization of this final rule’s regulatory baseline.
72
Enforcing the final rule on entities that are not in compliance may generate costs to the FDA. As explained in the
preamble, the FDA does not have the authority to impose fines for violations of section 204 of FSMA or subpart S.
We also note that the compliance strategy for the FDA is still in development, and that we plan to work with our
State, Local, Tribal, and Territorial (SLTT) and other regulatory partners to implement efficient enforcement of the
rule. Depending on the nature of the violation, it is generally FDA’s practice to give individuals and firms an
opportunity to take prompt and voluntary corrective action before we initiate an enforcement action. We may issue
advisory action letters, which include Untitled and Warning Letters, to notify firms of violations and to prompt
voluntary compliance. When voluntary compliance is not forthcoming, the Federal government may bring an action
in Federal court. We believe noncompliance will be a relatively uncommon event and when it does occur, entities
will generally take voluntary action to correct the noncompliance. Further, we expect coordination with SLTT
partners to minimize costs to the FDA.

163

1.1425), or waiver requests for an individual entity (§ 1.1415). Because we cannot estimate the
number of persons or entities that will submit petitions and request waivers, we cannot estimate
the costs associated with these actions. However, these potential costs will likely not increase the
net costs of this rule. First, since petitions are voluntary, firms will only submit petitions if the
cost of submitting a petition is lower than the cost of compliance. Our cost estimates do not
account for petitions, so in the case of a petition submission, the lower cost of submitting a
petition would replace the higher cost of compliance. Second, petitions must either demonstrate
that “application of the requirements requested to be modified or from which exemption is
requested is not necessary to protect the public health” (in the case of a request for modified
requirements or an exemption) or – in the case of a waiver – that “[t]he waiver will not
significantly impair [FDA’s] ability to rapidly and effectively identify recipients of a food to
prevent or mitigate a foodborne illness outbreak or to address credible threats of serious adverse
health consequences or death.” Thus, the existence of these mechanisms is not likely to interfere
with FDA’s ability to conduct traceback investigations. The cost of submitting petitions and
waiver requests would therefore replace the higher cost of compliance without compromising
benefits as estimated in this analysis.
Finally, we note that a request from FDA to produce an electronic sortable spreadsheet
(under the circumstances described above) will be withdrawn when necessary to accommodate a
religious belief of a person asked to provide such a spreadsheet. Because this does not require
any further action from persons or entities requesting a waiver of this requirement other than
stating a religious reason, we believe any such additional costs to be negligible.

164

7. Summary of Costs
Table 38 summarizes our estimates of the one-time and recurring costs of the final rule.
We estimate that the total one-time costs of the final rule will be approximately $1,684 million,
with a lower bound of $509 million and an upper bound of $5,875 million. We estimate that the
total recurring costs of the final rule will be approximately $490 million per year, with a lower
bound of $22 million and an upper bound of $2,092 million.
Table 38. Total Costs of the Final Rule (millions 2020$)
One-time Costs
Primary
Low
High
Reading and Understanding the Rule
$225.64
$203.08
$253.85
Capital Investment
$1,139.12
$278.01
$4,866.82
Training
$240.67
$12.97
$409.19
§ 1.1315 Traceability Plan
$78.55
$15.13
$345.41
Total One-time Costs
$1,683.98
$509.19
$5,875.27
Annually Recurring Costs
Capital Operation and Maintenance
$185.37
$14.91
$979.66
Recurring Training
$40.43
$0.98
$202.31
Seed lot records (Growers of sprouts) (1)
$0.10
$0.004
$0.84
§ 1.1325 Records of Harvesting
$3.11
$0.12
$27.00
§ 1.1325 Records of Cooling
$1.76
$0.10
$9.21
§ 1.1330 Records of Initial Packing
$2.21
$0.10
$22.81
§ 1.1335 Records of First Land-Based
$0.30
$0.004
$1.50
Receiving
§ 1.1340 Records of Shipping
$30.25
$0.48
$123.81
§ 1.1345 Records of Receiving
$220.26
$5.61
$681.70
§ 1.1350 Records of Transformation
$5.97
$0.09
$43.44
§ 1.1455(c)(3)(ii) Electronic Sortable
$0.05
$0.01
$0.11
Spreadsheet Upon Request
Total Recurring Costs
$489.82
$22.41
$2,092.40
(1)
Although seed lot records fall under §1.1330 Records of Initial Packing, we assume the
incidence of these costs will fall on growers of sprouts.
We present a summary of the estimated twenty-year stream of costs of the final rule in
Table 39. We expect that one-time costs of the final rule will occur evenly over the first two
years after the rule becomes effective. We expect that recurring costs will begin in the second
year, though at only half the estimated amount, lagging by one year behind the half of one-time
costs occurring in year one. We estimate that in the first year after the final rule becomes
165

effective, total costs will be approximately $842 million dollars, with a lower bound of $255
million and an upper bound of $2,938 million. In the second year, total costs will be
approximately $1,087 million dollars, with a lower bound of $266 million and an upper bound of
$3,984 million. In subsequent years, the annual cost of the final rule will decrease to $490
million, with a lower bound of $22 million and an upper bound of approximately $2,092 million.
We estimate that the total costs of the rule over 20 years will be approximately $10.7 billion,
ranging from a lower bound of $0.9 billion and an upper bound of approximately $44.6 billion.
The present value of total estimated costs of the rule is approximately $6 billion at a
seven percent discount rate and $8.2 billion at a three percent discount rate over 20 years. The
twenty-year annualized value of costs is $570.12 million at a seven percent discount rate and
$550.63 million at a three percent discount rate.
Table 39. Twenty-Year Timing of the Costs of the Final Rule (millions 2020$)
Primary
Low
High
Year 1
$841.99
$254.59
$2,937.63
Year 2
$1,086.90
$265.80
$3,983.83
Years 3-20
$489.82
$22.41
$2,092.40
Total Costs of the Final Rule
$10,745.71
$923.84
$44,584.63
Present Value of Total Costs (3%)
$8,192.05
$788.29
$33,733.08
Present Value of Total Costs (7%)
$6,039.83
$667.02
$24,608.89
Annualized Value of Costs (3%)
$550.63
$52.99
$2,267.39
Annualized Value of Costs (7%)
$570.12
$62.96
$2,322.90

G. Distributional Effects
The final rule will generate benefits and costs that may accrue unequally to
establishments depending on their industry sector and size and may also accrue unequally to
various segments of society. In this section, we discuss differential effects for consumers and
broad differences across industry sectors. We address differential effects for small entities by
industry sector in Section III of this analysis.
166

As described in Section II.F, we expect that the costs of this final rule will mainly arise
from recordkeeping requirements. Currently, entities have different baseline business practices
and therefore may face different costs depending on their size, industry, and position in the food
supply chain. Wholesalers/distributors and manufacturers are generally expected to bear the
highest per-firm costs associated with additional recordkeeping requirements, while retailers and
farms are expected to bear lower per-firm costs overall. As discussed throughout sections II.F
and III, we therefore expect the costs of the rule to be more concentrated on those industries in
the middle of the food supply chain.
The rule’s effect on traceback time and avoidance of overly broad recalls of FTL foods
will result in health benefits for consumers (estimated in section II.E), but existing inequities in
healthcare access, quality of care, and local FTL food availability and variety may result in more
benefits for some groups than others. As described in section II.E, we estimate the value of
averted illnesses and deaths by estimating 1) the cost burden of an illness on a typical individual
and 2) the number of averted illnesses through improved traceback time. The cost burden of an
illness consists of the medical care costs and the monetized value of the loss in health status.
There are significant differences in health status, healthcare access, and healthcare affordability
across sociodemographic groups (Ref. [42]). Thus, the cost burden of an illness may be unequal
across sociodemographic groups. The effect of an illness may also be unequal across groups as
differences in accessing healthcare may result in different recovery times, additional illnesses,
change in employment and income status, or other associated effects. Additionally, the risk of
contracting an illness depends on exposure, which in turn depends on the volume of FTL foods
consumed by an individual. Unfortunately, although we have information on average quality of
life measures across racial, gender, and income groups (Ref. [43, 44]), we do not have

167

information on quality-of-life loss under different illnesses for various sociodemographic groups.
We therefore cannot identify the difference in averted costs for different sociodemographic
groups associated with better traceback time and assume these averted costs to be the same
across these groups. We nevertheless acknowledge the rule could yield differential benefits from
averted illnesses for some sociodemographic groups.
The concentration of retail food establishments (RFEs) and therefore availability of
covered foods near consumers’ homes may be correlated with sociodemographic characteristics.
There is evidence that sociodemographic characteristics are correlated with the location of food
deserts (Ref. [45]). We use 2020 data from Dun & Bradstreet73 on RFE locations74 and data from
the USDA Economic Research Service on zip code-level rural/urban classifications75 to
understand the distribution of covered retail entities across the country. We find that rural areas
have the highest number of covered RFEs per 1,000 people, but in general there are few
significant differences of RFE concentrations across the country. The geographic distribution of
covered RFEs suggests that most people have a similar amount of retail food availability. It
should be noted, however, that access to transportation and purchasing behavior can also affect
food availability. For example, consumers in rural areas may have to travel far to access food and
therefore buy more foods in bulk to use for longer periods. Similarly, consumers in urban areas
without easy access to transportation may exhibit similar bulk purchasing behavior. The rate of
eating in restaurants may also differ across geographic areas. We do not, however, have
information on access to transportation and associated purchasing behavior for consumers.

Dun & Bradstreet, Dun & Bradstreet Global Business Database. 2020.
Only covered retail food establishments are analyzed; establishments that do not handle FTL foods or are exempt
based on annual sales are excluded from the analysis.
75
https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes.aspx
74

168

Using 2017 – 2018 dietary survey data from the National Health and Nutritional
Examination Survey (NHANES)76, we observed food items consumed by individuals from
across the country. We observed the diet of each respondent over two days and searched for any
item they consumed that would be on the FTL. We found some significant differences in FTL
food consumption across demographic characteristics. However, despite the observed
differences, we cannot quantify differences in benefit accrual to various socio-demographic
groups. First, diets may have changed since the data were gathered.77 Second, the final rule may
have spillover effects that affect foods not currently on the FTL. The difference in consumption
rates across sociodemographic groups may not extend to other foods. Third, not all FTL foods
carry the same risk of contamination, so the differences in consumption do not imply differences
in the risk of disease contraction.
Covered entities may pass increased costs on to consumers, raising the price of FTL
foods. We elicited input from a panel of experts, who agreed that some of the costs of the final
rule will likely be passed on through the supply chain to consumers (Ref. [4]). If the difference in
observed FTL consumption rates is due to price concerns, increased prices passed down the
supply chain may exacerbate current FTL consumption rate differences across groups. We
discuss cost pass-throughs to consumers in Section II.G and provide some evidence of increased
prices for covered foods. However, we have no evidence of the magnitude of spillover cost passthroughs to non-covered food items, the quality of diet of demographic groups, or the
substitution patterns of demographic groups as a result of price changes to FTL foods. We

The most recent dietary data from NHANES (also referred to as What We Eat In America) is from 2017 – 2018,
and can be found at: https://www.ars.usda.gov/northeast-area/beltsville-md-bhnrc/beltsville-human-nutritionresearch-center/food-surveys-research-group/docs/wweia-documentation-and-data-sets/
77
The most recent data was gathered prior to the COVID-19 pandemic and might not reflect the current diet pattern
after the significant disruption of eating patterns.

169

recognize that current differences in consumption rates of FTL foods suggests some differential
accrual of benefits to different demographic groups, but without more information, we cannot
quantify the distribution of welfare impacts across demographic groups.
Using data78 on the distribution of covered entities and consumption of certain food items
linked to demographic data, we observe some differences across demographic characteristics.
However, potential correlations between demographic characteristics and other outcomes
(healthcare access, employment opportunities, access to transportation, etc.) suggest the observed
differences could be potentially misleading. For example, the difference in FTL food
consumption rates could be because of budgetary constraints, dietary preferences, dietary
restrictions, local food availability, or several other reasons. Each of the potential causes of the
observed differences in FTL consumption rates may also be correlated with demographic
characteristics. Thus, a dedicated causal study is needed to fully understand what is causing the
differences. Because of the novelty of the FTL, there are naturally no causal studies available.
We therefore have limited knowledge on what is causing differences in FTL consumption rates,
so we also cannot determine how the final rule will affect the difference in FTL consumption
rates, and subsequently health outcomes for consumers.
In sum, we expect costs of the rule to be concentrated in industries in the middle of the
food supply chain (manufacturers, distributors, etc.), and while we recognize the potential for
benefits to accrue to some consumer segments more than others, we lack the information
necessary to quantitively estimate the distribution of benefits across sociodemographic
characteristics.

We specifically use Dun & Bradstreet (2020), USDA ERS rural/urban classification (2019), and NHANES (20172019) data.

170

H. International Effects
This section estimates costs for foreign entities (firms or establishments) who
manufacture, process, pack, or hold covered foods that are exported to the U.S. market. We
estimate foreign impacts using data from FDA’s FFRM to estimate compliance costs to foreign
entities.
In estimating the compliance costs to foreign entities, we use the average cost of the rule
per domestic entity with adjustments described below. These costs by activity type include onetime costs of reading and understanding the rule, capital investment, training, and developing
traceability plan, and recurring costs of operation and capital maintenance, training, and
recordkeeping as described in detail in section II.F of this analysis. The per entity cost is then
multiplied by the total number of foreign entities affected by each provision to get the total
compliance costs to foreign entities.
To estimate compliance costs to foreign entities, we use costs for domestic entities and
introduce adjustments to account for the number of foreign entities, foreign employee wages,
internet access, and English language proficiency as detailed below. We use 2019 FDA’s FFRM
data which contains 212,404 foreign and domestic facilities. FDA FFRM does not include data
on RFEs because they are not covered by the food facility registration regulation.79 While it is
possible that there might be a small number of foreign entities that offer covered food for sale in
the United States and meet the definition of “retail food establishment,” we assume that the
number of such entities affected by this rule is negligible. From FDA’s FFRM data, there is a
total of 212,404 foreign and domestic registered facilities of which about 75 percent (or 159,482)

https://www.regulations.gov/document/FDA-2002-N-0323-0163

171

are facilities that manufacture, process, pack, or hold covered foods. Of the 159,482 facilities, a
total of 68,566 (43 percent) are foreign facilities.
We assume that the same proportion of registered foreign establishments80 to firms is
affected by the rule (i.e., manufacture, process, pack, or hold FTL foods) as the proportion of
domestic establishments to firms. We use the ratio of 1.19 of covered domestic non-retail and
non-restaurant establishments to firms (40,754 establishments / 34,389 firms = 1.19) to estimate
that the number of foreign firms affected by the rule is 57,618 firms (68,566 foreign
establishments / 1.19 = 57,618). These estimates enable us to calculate costs of the rule to
foreign entities. We make the same assumptions in estimating compliance costs by foreign
entities as for domestic entities. Similarly to estimates in Section II.F, some costs such as costs of
reading and understanding the rule and establishing traceability plan, will occur at the firm level,
while all other costs such as those related to capital investments, training and recordkeeping are
assumed to occur at the facility level. Since the FFRM data does not contain information on
farms including firm sizes or annual receipts, we assume that the share of small foreign entities is
the same as for domestic entities.
We make two important adjustments to our estimates of compliance costs for foreign
establishments. First, since foreign wages are generally lower than domestic wages, we make
adjustment to account for this variation. To estimate wages for foreign employees and
supervisors, we take the weighted average of general wages for the top twenty foreign countries
by value of their covered foods exported to the United States and adjust this weighted average
wage to 2020 U.S. dollar (Ref. [46]). For example, this yields an average foreign general
employee wage of $3.45 per hour, mid-level supervisor wage of $7.81 per hour, and supervisor

We use ‘establishments’ and ‘facilities’ interchangeably.

172

wage of $10.77 per hour, etc. We double the wage rates to account for employees benefits and
overhead. This yields $6.91 per hour (= $3.45 x 2) for general employees, $15.63 per hour (=
7.81 x 2) for mid-level supervisor employees, etc.
The second adjustment we make is related to the time it takes for employees of foreign
entities to read and understand the rule to account for varying levels of both English proficiency
and internet accessibility. In learning about the requirements of this rule, we assume that entities
from countries with both high English proficiency and high internet access will spend a
comparable amount of time as domestic entities. However, entities from countries with lower
English proficiency but with high internet access may spend more time learning about the rule
than domestic entities because they may need to have internet access to translate the rule. Entities
from countries with both low English proficiency and low internet access may spend even more
time learning about the rule (and incur higher costs) than entities from countries with high
English proficiency or internet access.
To account for language proficiency differences, we use information from the 2020
“Education First English Proficiency Index” (EF EPI) report (Ref. [47]). This report ranks
countries by the average level of English language skills amongst adults using data collected via
English tests available over the internet. To account for country differences in internet
accessibility, we use 2022 internet user percentage estimates by country (Ref. [48]). We estimate
that on average, foreign establishments will spend 1.41 hours for every hour a domestic
establishment spends on reading the rule.
The weights are based on English language proficiency and internet access for foreign
facilities currently registered with FDA representing 114 countries who export FDA regulated
food to the U.S. The average of both weighted sums of 1.65 hours to account for differences in

173

English proficiency and of 1.17 hours to account for differences in internet usage give us a single
estimate of 1.41 (=(1.65 hours +1.17 hours)/2) for foreign establishments as equivalent to one
hour for domestic establishments.
As explained in Appendix H, we take the average between proficiency and internet
weighted hours because internet access is positively correlated with English proficiency and also
because high English proficiency alone is not enough to account for the amount of time that an
entity would require to learn about the rule. Even entities from countries with high English
proficiency would rely on using the internet insofar as to only download the rule to save it or email it to a device, whereas an entity in a country with low internet access would need to spend
more time finding internet access in order to download the rule from the internet.
Section II.F of this analysis calculates that the burden for domestic employees of reading
this rule is 16.22 hours. Assuming one foreign mid-level supervisor would be responsible for
reading and understanding the rule for small firm and three for large firms, we estimate the
burden of reading and understanding this rule per supervisor is 26.11 hours (= 16.22 hours x
1.41).

Table 40. Total Costs to Foreign Entities (Millions 2020$)
One-time Costs
Primary
$15.15
Reading and Understanding the Rule

Low
$13.63

High
$17.04

Capital Investment

$161.33

$19.71

$876.75

Training

$25.51

$1.67

$37.16

§ 1.1315 Traceability Plan
Total One-time Costs
Annually Recurring Costs

$3.74

$0.72

$16.46

$205.74

$35.74

$947.41

Capital Operation and Maintenance

$26.25

$0.89

$191.23

Recurring Training

$1.54

$0.05

$10.26

174

Seed lot records (Growers of sprouts)1

$0.00

$0.03

§ 1.1325 Records of Harvesting

$0.12

$0.00

$1.02

§ 1.1325 Records of Cooling

$0.07

$0.00

$0.35

§ 1.1330 Records of Initial Packing

$0.08

$0.00

$0.87

§ 1.1335 Records of First Land-Based Receiving

$0.01

$0.00

$0.06

§ 1.1340 Records of Shipping

$1.15

$0.02

$4.69

§ 1.1345 Records of Receiving

$8.35

$0.21

$25.84

§ 1.1350 Records of Transformation
§ 1.1455(c)(3)(ii) Electronic Sortable Spreadsheet
Upon Request
Total Recurring Costs

$0.23

$0.00

$1.65

$0.01

$0.00

$0.02

$37.80

$1.18

$236.01

Although seed lot records fall under § 1.1330 Records of Initial Packing, we assume the incidence of these costs
will fall on growers of sprouts.

As summarized in Table 40, we estimate that one-time costs to foreign entities range
from approximately $35.74 million to $947.41 million, with a primary estimate of $205.74
million. Just like for domestic entities, we expect that these one-time costs of the final rule to
foreign entities will occur evenly over the first two years after the rule becomes effective.
Recurring costs to foreign entities range from approximately $1.18 million to $236.01 million,
with a primary estimate of $37.80. million. Again, just like for domestic entities, we expect that
recurring costs to foreign entities will begin in the second year, though at only half the estimated
amount, lagging by one year behind the half of one-time costs occurring in year one.
At a seven percent discount rate, our primary estimate of the present value of costs to
foreign entities over twenty years is approximately $534.64 million, ranging from $43.22 million
to $3.03 billion. At a three percent discount rate, our primary estimate of the present value of
costs to foreign entities over twenty years is approximately $704.74 million, ranging from $50.1
million to $4.08 billion. The primary estimate of the annualized costs at a seven percent discount
rate to foreign entities is approximately $50.47 million, ranging from $4.08 million to $286.31

175

million. At a three percent discount rate, the primary estimate of the annualized costs to foreign
entities is approximately $47.37 million, ranging from $3.37 million to $274.06 million.
The costs presented in Table 40 are costs to foreign entities only. To the extent that these
costs are passed on to U.S. entities, U.S. consumers and firms that purchase covered foods from
foreign entities may experience higher costs. We assume that the requirements of this rule will
affect domestic entities in the same manner regardless of whether their suppliers are domestic or
foreign. We lack information to determine the portion of foreign producers’ compliance costs
that may be passed on to U.S. consumers.
Overall gains or losses from this rule would likely be caused by price increase or
reductions for covered varieties of foods in foreign markets. Gains to foreign consumers may
likely result from an increase in supply of domestic and imported varieties of covered foods from
other foreign markets.

Uncertainty and Sensitivity Analysis

The prospective nature of this analysis means that all our estimates have a varying degree
of uncertainty. This is the reason we present ranges to our estimates throughout this document.
1. Coverage
We derive the number of covered farms using raw USDA NASS data. Due to lack of
information on the percentage of farms producing FTL foods, we assume that all farms in
corresponding covered NAICS categories, excluding farms exempt because of their low annual
sales or direct sales to consumers, are covered entities. To that extent, we may overestimate the
number of covered farms.

176

For egg farms and aquaculture farms, we derive the number of covered farms from the
summary reports of the 2017 Census of Agriculture and the 2018 Census of Aquaculture (Ref.
[13], [14]). We may overestimate the number of covered egg farms as we assume all of the egg
farms with greater than 3,000 layers are packers. Similarly, we assume all of the aquaculture
farms with annual sales greater than $100,000 are packers, which means that we may
overestimate the number of covered egg farms and aquaculture farms.
To the extent that the U.S. Census data only cover primary NAICS codes, we potentially
exclude non-primary NAICS and may underestimate the number of total covered firms and
establishments. Due to lack of data counts specific to only entities that handle covered foods, we
modify the U.S. Census data to approximate the number of covered entities in each NAICS
category. In estimating the number of entities that handle covered foods, we only include the
numbers that were available from the NAPCS data. Hence, we may underestimate the number of
entities that we believe manufacture, process, pack, or hold foods currently covered on the FTL
and foods that contain them as ingredients.
2. Costs
In estimating costs, we are mainly uncertain of baseline trends in traceability investment.
As explained in section II.F.3 “Costs of Capital Investment,” as demand for traceability increases
and technology advances, we expect that lower-cost traceability solutions will become available
on the market. While ERG’s traceability costs elicitation provided some information on current
industry practices, we are generally unsure to what extent the rule itself drives estimated future
expansion of traceability. Although ERG’s expert elicitation asked for estimates of expected
capital costs due specifically to compliance with the proposed traceability requirements, some
experts suggested that their estimates included capital investments likely to occur for reasons

177

other than regulatory compliance, such as obsolescence. Furthermore, longstanding widespread
awareness of FSMA complicates extricating baseline traceability investment using past trends.
Past trends likely reflect, at least in part, business’ expectations of coming traceability regulation
and might change once said regulation materializes. Since experts did not separately quantify the
extent to which regulatory compliance drives their estimates, we nonetheless attributed all
estimated capital investment to the proposed traceability requirements when using the elicitation
to inform our analysis of the final rule.
Additionally, some entities that we count as distributors might also manufacture and
thereby perform the transformation CTE. We do not know the number of such entities nor the
extent to which they already factor into our manufacturing category, since, as previously stated,
Census data counts entities under only their primary activities. Some entities that we count as
distributors might also transform FTL foods via repacking (e.g., mixing peppers of different
colors). We do not know how often this would specifically be done by distributors with respect
to foods on the FTL.
Finally, throughout the cost section, we present ranges of estimates for a number of
variables to account for uncertainty stemming from yearly variability as well as imprecise
knowledge. We base these ranges mostly on ranges provided by experts in ERG’s panel, who
were asked for low, most likely, and high values. For example, low and high counts of FTL lots
handled per entity with respect to various CTEs, as well as counts of employees involved in costincurring tasks, reflect variability that we expect in yearly averages. Low and high estimates of
the proportion of industry engaged in various traceability practices, or expected to invest in
capital, account for imprecision in experts’ knowledge of these variables.

178

3. Benefits from Avoiding Overly Broad Recalls
In estimating benefits from avoiding overly broad recalls, our main source of uncertainty
was the number of firms typically impacted by an FDA advisory. While ERG’s recall expert
elicitation provided us with information about per-firm costs of dealing with an overly broad
recall, even the experts were uncertain about the scope of such an event, should one occur (Ref.
[4]). To characterize our uncertainty about the scope and variability in experts’ cost information,
we laid out a series of calculations to run a Monte Carlo simulation. Estimated benefits from the
reduction in overly broad recalls required that we assign parameters to corresponding probability
density functions to characterize the variability inherent in the costs estimates. We also assigned
parameter estimates and probability density functions to characterize the inherent uncertainty in
the estimates for the number of firms according to their respective cost category. Probability
density functions and their parameter estimates along with results of the sensitivity analysis in
the simulation showed that our estimates were mostly sensitive to the number of firms affected,
which was also our most uncertain estimate (see section II. E and Appendix G).

J. Analysis of Regulatory Alternatives to the Rule
We considered four different regulatory alternatives as described below.


Alternative a: No new regulatory action.



Alternative b: Broader exemption for retail food establishments and restaurants.



Alternative c: Reduce compliance date to two years.



Alternative d: Extend compliance date to four years.

179

Table 41 below shows a detailed summary of the costs and benefits associated with each
regulatory alternative (annualized using a seven percent discount rate over 20 years), the change
in the estimated costs and benefits relative to the final rule, the net health benefits of each
alternative, and the number of covered establishments under each alternative.
Table 41. Summary of Costs and Benefits of Regulatory Alternatives (millions $).
No. Covered
establishments

Annualized
Total Costs
(7%)

Annualized
Benefits (7%)

Net Benefit
(7%)

Final Rule

484,124

$570

$780

$210

Alt A: No action
Change from FR
Alt B. Fully exempt RFEs below
$1M
Change from FR
Alt C: Reduce compliance date
to two years
Change from FR
Alt D: Extend compliance date
to four years
Change from FR

0
-484,124

$0
($570)

$0
($780)

$0
($210)

306,680
-177,444

$508
($62)

$761
($19)

$253
$43

484,124
0

$595
$25

$857
$77

$262
$51

484,124
0

$546
($24)

$709
($71)

$163
($48)

Alternative a. No Action
We treat the alternative of taking no new regulatory action as the baseline for determining
the costs and benefits of other alternatives. In choosing an appropriate baseline, OMB Circular
A-4 recommends considering a wide range of factors, including market evolution, changes in
external factors affecting expected benefits and costs, changes in regulations promulgated by the
agency, and the degree of compliance by regulated entities with other regulations. In choosing a
baseline, we assume costs and benefits of the BT Act food tracing requirements are already
accounted for (although benefits have been either overestimated or not fully realized). As such,
if FDA pursued Alternative a, there would be no additional costs or benefits under this
alternative.
180

Alternative b. Broader exemption for retail food establishments and restaurants
Under this alternative, retail food establishments and restaurants with an annual monetary
value of food sold or provided during the previous 3-year period below $1 million (on a rolling
basis) would be fully exempt from the rule. Using SUSB data we approximate that these entities
are responsible for less than 5 percent of sales of covered foods.
As a result of exempting these entities, the Alternative b will exempt an additional
177,444 establishments throughout the entire supply chain of covered foods. The full exemption
of these entities will thus decrease the number of covered establishments from 484,124 to
306,680 (Table 41). The annualized costs of the rule will decrease from $570 million to $508
million, which is $62 million less than the estimated costs of the rule. While this will provide
relief to many smaller entities in the food supply chain, we estimate that health benefits
associated with Alternative b will decrease by $19 million, from $780 million to $761 million.
The benefits will decrease because in case of an outbreak of covered food, FDA and industry
may have less information available to them if these newly exempt entities will not be able to
provide the same traceability records in the same timely manner as covered entities. It is possible
that we underestimate the change in health benefits for this regulatory alternative because we
base it only the share of sales of covered foods and assume that the share of illnesses is
proportional to the share of sales.

Alternative c. Reduce compliance date to two years
This alternative reduces the compliance date of the rule to two years following the
effective date of the final regulation. Under this alternative, we assume that one-time costs of the
rule will occur in the first year and the recurring costs will begin in the second year without lag.

181

The number of covered establishments under this alternative will be the same as under the rule
(Table 41). The estimated annualized costs of this Alternative c will be $595 million, which is
$25 million higher than the estimated costs of the rule. The estimated annualized health benefits
will be $857 million, which is $77 million higher than the estimated benefits of the final rule.
The shorter compliance period will result in the higher annualized benefits because they would
begin year two, which is one year earlier than under the final rule. However, a shorter
compliance period means that covered entities including small entities would have less time to
prepare for implementation of the rule, especially if the supply chain is affected by the COVID19 pandemic. And it might not be feasible for small establishments to come into compliance.
Hence, our estimated benefits of this alternative could be overstated if some establishment
including small establishments would not receive capital equipment in time, have the sufficient
time to adequately train employees to use it, and have their traceability system in place before
the compliance date.

Alternative d. Extend compliance date to four years
This alternative extends the compliance date of the rule to four years following the
effective date of the final regulation. We assume that one-time costs of the rule under this
alternative will occur evenly over the first three years after the rule becomes effective. And we
assume that recurring costs will begin in the second year, though at only one-third of the
estimated amount, lagging by one year behind the third of one-time costs occurring in year one.
The number of covered establishments under this alternative will be the same as under the rule
(Table 41). Compared to the final rule, delaying the compliance date would reduce the burden on
the covered entities by shifting costs into the future as they would have additional time to comply
182

with the rule. The estimated annualized costs of this alternative will be $546 million, which is
$24 million less than the estimated costs of the rule. However, the estimated health benefits will
also decrease from $780 million to $709 million, which is $71 million less than the estimated
benefits of the final rule because delaying the compliance date of one year would result in
illnesses and death that could have been avoided.

III. Final Small Entity Analysis
The Regulatory Flexibility Act requires Agencies to analyze regulatory options that
would minimize any significant impact of a rule on small entities. Because some small covered
entities might have annualized costs (over 20 years at a seven percent discount rate) that exceed
one percent of their annual revenue, we find that the final rule will have a significant economic
impact on a substantial number of small entities. This analysis, as well as other sections in this
document, serves as the Final Regulatory Flexibility Analysis, as required under the Regulatory
Flexibility Act.81
A. Description and Number of Affected Small Entities
The entities in this small entity analysis are firms. The Small Business Administration
(SBA) publishes size standards for industry categories of firms defined by NAICS codes. SBA
defines each NAICS code’s small business threshhold either in terms of sales revenue or number
of employees. Using the 2019 SBA size standards82 in conjunction with the SUSB counts of

For descriptions of the steps that FDA has taken to minimize the significant economic impact on small entities
consistent with the stated objectives of applicable statutes, including extending the compliance period from three to
four years for all firms, please see Sections I.D comment 30, I.E, II.D.2, and Appendix D.
82
Small Business Association. Table of Size Standards. Aug 19, 2019. Available from:
https://www.sba.gov/document/support--table-size-standards

183

firms in each NAICS code by revenue and employment size,83 we estimate the numbers of
covered small firms by industry sector.84 Overall, we estimate that about 98 percent of firms
covered by this rule are small businesses by SBA standards. Table 42 shows estimated counts of
covered small firms by NAICS code.
Table 42. Small Entities Affected by the Final Rule
2017
NAICS
Code

NAICS
Industry
Description

111219

Other
Vegetable
(except Potato)
and Melon
Farming85
Sprouts (under
"Other
Vegetable
(except Potato)
and Melon
Farming")86

111339

Other
Noncitrus Fruit
Farming85

111419

Other Food
Crops Grown
Under Cover85

111219

Firm type

Revenue
per Firm
in
Millions$

SBA Size
Standard
(Millions$
or Number
of
employees)

Number of
Covered
Small
Firms

Annual
Revenue in
Millions$

3889

$791

$0.20

$47

$0.54

2249

$583

$0.26

380

$76

$0.20

Farms/Growers

We use the 2017 SUSB, the last release that contained revenue data, and inflate revenues to 2020-dollar values
using the GDP deflator.
Census Bureau. 2017 SUSB Annual Datasets by Establishment Industry. Updated 2021. Available from:
https://www.census.gov/data/datasets/2017/econ/susb/2017-susb.html
84
This rule exempts a number of firms based on size, which hence do not factor into this analysis. For discussion of
exemption thresholds and related calculations, please see Sections I.E.7, II.D.2, and Appendix D.
85
We base the small entity count and revenue estimate for this industry on the USDA National Agricultural
Statistics Service data. The SBA defines produce farms small if their total revenues are lee than $1 million.
86
We base the revenue estimate for sprout growers on Table 25 of the regulatory impact analysis for the prior
Produce Rule (Ref. [53]). We take the average of the “Average Sales Volumes” in the third and fourth columns
weighted by number of sprouting operations in each column. For our estimate of the number of covered small sprout
growers under this final rule, we use the inventory of sprout farms and operations used by the FDA’s Office of
Regulatory Affairs. Excluding very small sprout growers, this internal inventory counts 95 sprout growers. We then
assume that the same proportion of sprout growers are small as among other growers of produce.

184

112511

Chicken Egg
Production87
Finfish
Farming and
Fish
Hatcheries88

112512
114111

Shellfish
Farming89
Finfish Fishing

Aquaculture
Fishing

114112

Shellfish
Fishing

Fishing

311340

Nonchocolate
Confectionery
Manufacturing

Manufacturing/
Processing

311351

Chocolate and
Confectionery
Manufacturing
from Cacao
Beans

Manufacturing/
Processing

311352

Confectionery
Manufacturing
from Purchased
Chocolate

Manufacturing/
Processing

311411

Frozen Fruit,
Juice and
Vegetable
Manufacturing

Manufacturing/
Processing

112310

Farms (Eggs)

2500

$7,720

$3.09

$16.5

310

$83

$0.27

147

$35

$0.24

767

$555

$0.72

$22

771

$382

$0.50

290

$4,155

$14.30

1,000

109

$1,880

$17.20

1,250

498

$3,824

$7.67

1,000

$2,956

$64.52

1,000

Aquaculture

The SBA defines chicken and egg producers to be small if their total revenues are less than $16.5 million. A
producer that receives $0.85 per dozen eggs (the midpoint of seasonally adjusted December 2018 and December
2019 market egg prices) and has layers that produce 265 eggs per year would have to have over 879,000 layers in
production to earn revenues of over $16.5 million. Because only about 320 farms fall into the category of 100,000 or
more layers, more than 99 percent of the farms with more than 3,000 layers are considered small by SBA standards.
We use Table 75 from the summary report of the 2017 USDA Census of Agricultural to estimate the number of shell
egg farms. Out of total shell egg farms, we first remove shell egg farms with less than 3,000 layers who are exempt.
Out of the shell egg farms with more than 3,000 layers, we then estimate that the 99 percent of the farms are
considered small by SBA standards.
88
We use Table 9 from the 2018 USDA Census of Aquaculture to compute the weighted average revenue of small
farms (less than $1 million in sales) by fish type (baitfish, food fish, crustaceans, and mollusks). We then combine
categories by weighted average of types of small farms (Finfish includes food fish and Shellfish includes
Crustaceans and Mollusks). Finally, we multiply the average revenue by the total number of small farms to obtain
the total revenue for all small farms.
89
We use Table 9 from the 2018 USDA Census of Aquaculture to compute the weighted average revenue of small
farms (less than $1 million in sales) by fish type (baitfish, food fish, crustaceans, and mollusks). We then combine
categories by weighted average of types of small farms (Finfish includes food fish and Shellfish includes
Crustaceans and Mollusks). We assume 75% of mollusks are exempt as the raw bivalve molluscan shellfish under
the NSSP are not covered by the rule, so we include 25% of mollusks. Finally, we multiply the average revenue by
the total number of small farms to obtain the total revenue for all small farms.

185

238

$18,091

$75.89

1,250

293

$6,953

$23.76

750

4392

$2,866

$0.65

500

Retail Bakeries

Manufacturing/
Processing
Manufacturing/
Processing

1143

$9,258

$8.10

1,000

311812

Commercial
Bakeries

Manufacturing/
Processing

$1,343

$23.52

750

311813

Frozen Cakes,
Pies, and Other
Pastries
Manufacturing

Manufacturing/
Processing

114

$2,377

$20.83

1,250

311821

Cookie and
Cracker
Manufacturing

Manufacturing/
Processing

$4,533

$52.19

750

311911

Roasted Nuts
and Peanut
Butter
Manufacturing

123

$2,995

$24.33

750

311941

Mayonnaise,
Dressing, and
Other Prepared
Sauce
Manufacturing

Manufacturing/
Processing
Manufacturing/
Processing

455

$3,279

$7.20

500

311991

Perishable
Prepared Food
Manufacturing

Wholesalers/
Distributors

1273

$15,063

$11.83

250

424410

General Line
Grocery
Merchant
Wholesalers

Wholesalers/
Distributors

1260

$19,761

$15.69

200

424420

Packaged
Frozen Food
Merchant
Wholesalers

311513

Cheese
Manufacturing

Manufacturing/
Processing

311710

Seafood
Product
Preparation and
Packaging

Manufacturing/
Processing

311811

186

424430

Dairy Product
(except Dried
or Canned)
Merchant
Wholesalers

Wholesalers/
Distributors

424450

Confectionery
Merchant
Wholesalers

Wholesalers/
Distributors

424460

Fish and
Seafood
Merchant
Wholesalers

Wholesalers/
Distributors

424480

Fresh Fruit and
Vegetable
Merchant
Wholesalers

Wholesalers/
Distributors

627

$7,183

$11.46

200

690

$3,558

$5.16

200

1239

$9,085

$7.33

100

2783

$30,826

$11.08

100

3645

$21,360

$5.86

250

21597

$63,463

$2.94

$35

445110

Other Grocery
and Related
Products
Merchant
Wholesalers
Supermarkets
and Other
Grocery
(except
Convenience)
Stores

Retail Food
Establishments

13161

$15,051

$1.14

$32

445120

Convenience
Stores

Retail Food
Establishments

976

$1,206

$1.24

445220

Fish and
Seafood
Markets

Retail Food
Establishments

679

$503

$0.74

445292

Confectionery
and Nut Stores

Retail Food
Establishments

938

$1,200

$1.28

445230

Fruit and
Vegetable
Markets

Retail Food
Establishments

1913

$1,636

$0.85

445299

All Other
Specialty Food
Stores

424490

Wholesalers/
Distributors

Retail Food
Establishments

187

447110

Gasoline
Stations with
Convenience
Stores

Retail Food
Establishments

452311

Warehouse
Clubs and
Supercenters

18182

$52,743

$2.90

$32

Retail Food
Establishments

$0.1

$0.08

$32

Retail Food
Establishments

1003

$1,310

$1.31

$12

454210

Vending
Machine
Operators

Warehouses
and Storage

1520

$3,230

$2.12

$30

493110

General
Warehousing
and Storage

Warehouses
and Storage

194

$648

$3.35

$30

493120

Refrigerated
Warehousing
and Storage

Warehouses
and Storage

104

$194

$1.86

$30

493130

Farm Product
Warehousing
and Storage

Retail Food
Establishments

1513

$3,330

$2.20

$41.5

722310

Food Service
Contractors

4776

$5,222

$1.09

Caterers

Retail Food
Establishments
Retail Food
Establishments

1360

$683

$0.50

14082

$11,041

$0.78

722410

Mobile Food
Services
Drinking
Places
(Alcoholic
Beverages)

Retail Food
Establishments

110150

$129,223

$1.17

722511

Full-Service
Restaurants

Retail Food
Establishments

74655

$85,238

$1.14

$12

722513

LimitedService
Restaurants

Retail Food
Establishments

2350

$3,046

$1.30

$30

722514

Cafeterias,
Grill Buffets,
and Buffets

722320

722330

Retail Food
Establishments

188

722515

Snack and
Nonalcoholic
Beverage Bars

Retail Food
Establishments

18630

$13,345

$0.72

B. Description of the Potential Impacts of the Rule on Small Entities
ERG’s panel of industry experts, via the traceability costs elicitation, informs our
estimates of small firm compliance costs reported in this section as well as those throughout this
analysis (Ref. [4]). Whereas section II.F presented costs attributable to provisions of the final
rule, this section breaks down costs small firms across different broad industry categories are
facing. Though aggregate totals are displayed for these broad categories, the underlying analysis
in this section accounts for applicable provisions at the level of each NAICS code.
We assume that all covered small entities will incur one-time costs to read and
understand the rule. Depending on business activities and baseline practices, some but not all
covered small entities will also incur one-time and recurring capital investment and training costs
and a one-time cost to plan for compliance with the rule, in addition to various recurring annual
recordkeeping costs.
Note therefore that the following primary, low, and high per-entity estimates throughout
this section represent industry averages. Depending on business activities and baseline practices,
individual entities will likely incur costs outside our range of estimates of industry averages for
small businesses. For example, in forming our low and high estimates, we use low and high
expert estimates of the proportion of small entities in various industries that will purchase
equipment or software to comply with this rule.
Table 43 presents our estimates of the one-time cost per covered small entity. Among
small firms, we expect one-time per firm compliance costs of about $2,975 for growers of

189

produce other than sprouts, $11,122 for growers of sprouts, $13,172 for shell egg farms, $17,496
for fishing and aquaculture producers, $18,780 for manufacturers, $30,609 for wholesalers,
distributors, and warehouses, $693 for non-restaurant retailers, and $704 for restaurants.
Table 43. One-Time per Firm Compliance Costs of the Final Rule
Primary
Low
$2,975
$1,151
Farms/Growers (Produce, non-sprouts)
$11,122
$3,507
Farms/Growers (Sprouts)
$13,172
$5,641
Farms (Shell Eggs)
$17,496
$5,158
Fishing/Aquaculture
$18,780
$2,673
Manufacturing/Processing
$30,609
$2,816
Wholesalers/Distributors/Warehouses and Storage
$693
$595
Retail - Not Restaurants
$704
$598
Retail - Full and Limited-Service Restaurants

High
$11,528
$49,307
$41,351
$42,446
$50,694
$67,947
$946
$946

As previously mentioned in section II.F “Costs of the Rule,” growers of produce are the
only category of growers among which we were able to estimate counts of establishments that do
not initially pack. In all other categories of growers, we assumed that all establishments initially
pack and therefore face the possibility of capital investment costs, which we estimate in II.F.3
“Costs of Capital Investment.”
Using the same breakdown, Table 44 shows estimated cost per covered small entities,
anualized over 20 years at a seven percent discount rate. Among small firms, we expect
annualized compliance costs of about $849 for growers of produce other than sprouts, $4,295 for
growers of sprouts, $3,801 for shell eggs farms, $3,941 for fishing and aquaculture producers,
$4,625 for manufacturers, $8,027 for wholesalers, distributors, and warehouses, $402 for nonrestaurant retailers, and $180 for restaurants.
Table 44. Annualized per Firm Compliance Costs of the Final Rule (Over 20 Years, Seven
Percent Discount Rate)
Primary
Low
High
$849
$144
$5,700
Farms/Growers (Produce, non-sprouts)
$4,295
$581
$29,950
Farms/Growers (Sprouts)
190

Farms (Shell Eggs)
Fishing/Aquaculture
Manufacturing/Processing
Wholesalers/Distributors/Warehouses and Storage
Retail - Not Restaurants
Retail - Full and Limited-Service Restaurants

$3,801
$3,941
$4,625
$8,027
$402
$180

$674
$684
$314
$349
$61
$61

$22,007
$14,197
$20,668
$26,751
$1,636
$729

We use the SUSB90 data to estimate the magnitude of costs as a percent of the revenues
of covered small firms. We consider costs per firm exceeding one percent of annual revenues to
be a substantial impact. Table 45 shows our estimate of the one-time compliance costs as a
percentage of revenue for small firms, broken down by broad industry categories. Among small
firms, we expect one-time costs, as a percentage of annual revenue, of about 1.34% for growers
of produce other than sprouts, 2.05% for growers of sprouts, 0.43% for shell egg farms, 3.31%
for fishing and aquaculture producers, 0.23% for manufacturers, 0.37% for wholesalers,
distributors, and warehouses, 0.04% for non-restaurant retailers, and 0.06% for restaurants.
Table 45. One-time per Firm Compliance Costs as a Percentage of Small Firm Annual
Revenue
Primary
Low
High
1.34%
0.52%
5.18%
Farms/Growers (Produce, non-sprouts)
2.05%
0.65%
9.08%
Farms/Growers (Sprouts)
0.43%
0.18%
1.34%
Farms (Shell Eggs)
3.31%
0.98%
8.03%
Fishing/Aquaculture
0.23%
0.03%
0.62%
Manufacturing/Processing
0.37%
0.03%
0.82%
Wholesalers/Distributors/Warehouses and Storage
0.04%
0.03%
0.06%
Retail - Not Restaurants
0.06%
0.05%
0.08%
Retail - Full and Limited-Service Restaurants
Using the same categorical breakdown, Table 46 shows the annualized values of our
estimates of compliance costs over 20 years at a seven percent discount rate, again as a

For small farms and producers of raw agricultural commodities, we estimate revenues based on the FSMA
Produce Rule economic impacts analysis (Ref. [53]), the USDA National Agricultural Statistics Service, and the
USDA Census of Aquaculture. We describe these estimates in the footnotes of Table 42.

191

percentage of the revenues of covered small firms. Over 20 years at a seven percent discount
rate, we expect annualized costs, as a percentage of annual revenue, of about 0.38% for growers
of produce other than sprouts, 0.79% for growers of sprouts, 0.12% for shell egg farms, 0.75%
for fishing and aquaculture producers, 0.06% for manufacturers, 0.10% for wholesalers,
distributors, and warehouses, 0.02% for non-restaurant retailers, and 0.02% for restaurants.
Table 46. Annualized per Firm Compliance Costs as a Percentage of Annual Revenue (20
Years, Seven Percent Discount Rate)
Primary
Low
High
0.38%
0.06%
2.56%
Farms/Growers (Produce, non-sprouts)
0.79%
0.11%
5.52%
Farms/Growers (Sprouts)
0.12%
0.02%
0.71%
Farms (Eggs)
0.75%
0.13%
2.69%
Fishing/Aquaculture
0.06%
0.00%
0.25%
Manufacturing/Processing
0.10%
0.00%
0.32%
Wholesalers/Distributors/Warehouses and Storage
0.02%
0.00%
0.10%
Retail - Not Restaurants
0.02%
0.01%
0.06%
Retail - Full and Limited-Service Restaurants
In Table 47, we estimate that, on average, the total costs of the final rule per covered
small firm over 20 years will be about $13,911. At a seven percent discount rate, our estimate of
the present value of the average total costs of the final rule per covered small firm is about
$8,010. Discounted at three percent, our estimate of the present value of the average total costs
of the final rule per covered small firm is about $10,714. Over 20 years, the estimated annualized
value of average costs of the final rule per small firm is about $756 when discounting at seven
percent and $720 whan discounting at three percent.
Table 47. Costs of the Final Rule per Small Firm (Over 20 Years)
Low
Primary
$13,911 $1,244
Total Costs of the Final Rule
$8,010
$953
Present Value of Total Costs (7%)
$10,714 $1,093
Present Value of Total Costs (3%)
$756
$90
Annualized Value of Costs (7%)
$720
$73
Annualized Value of Costs (3%)

High
$56,211
$33,889
$47,284
$3,199
$3,178
192

C. Alternatives to Minimize the Burden on Small Entities
As the vast majority (roughly 98 percent) of covered firms qualify as small entities, the
analysis of regulatory alternatives for covered firms in Section II.J above effectively describes
the effects of the alternatives on small entities. In particular, Alternative B in Section II.J would
extend full exemption to retail food establishments and restaurants with an annual monetary
value of food sold or provided during the previous 3-year period below $1 million (on a rolling
basis), up from $250,000 in the rule as written. Alternative D would extend the compliance
period for all firms from three to four years. Due to traceability relying on linkages throughout
the supply chain, delaying the compliance date even for just small entities would delay the
implementation of the final rule for the vast majority of FTL products. While the postponement
of capital investments and labor expenses for compliance would reduce the present value of costs
of the final rule, it would also reduce the present value of the health benefits.

193

IV. References
[1]

[2]

[3]

[4]
[5]
[6]

[7]

[8]

[9]
[10]
[11]

[12]

[13]
[14]
[15]

M. D. Ehmke, A. Bonanno, K. Boys and T. G. Smith, "Food fraud: economic insights into
the dark side of incentives," Australian Journal of Agricultural and Resource Economics,
vol. 63, no. 4, pp. 685 - 700, 2019.
T. Bhatt, G. Buckley, J. C. McEntire, P. Lothian, B. Sterling and C. Hickey, "Making
Traceability Work across the Entire Food Supply Chain," Journal of Food Science, vol.
78, no. S2, pp. B21-B27, 2013.
U.S. Food and Drug Administration (FDA), "Requirements for Additional Traceability
Records for Certain Foods: Preliminary Regulatory Impact Analysis," FDA, Silver Spring,
MD, 2020.
Eastern Research Group (ERG), "Expert Elicitations to Estimate Traceability Costs and
Costs Savings from Avoiding Overly Broad Recalls," Lexington, MA, 2022.
J. McEntire and T. Bhatt, "Pilot Projects for Improving Product Tracing along the Food
Supply System – Final Report," Institute of Food Technologists, Illinois, Chicago, 2012.
Centers for Disease Control and Prevention, "Multistate Outbreaks of Shiga toxinproducing Escherichia coli O26 Infections Linked to Chipotle Mexican Grill Restaurants
(Final Update)," 1 February 2016. [Online]. Available:
https://www.cdc.gov/ecoli/2015/o26-11-15/index.html.
Public Health Agency of Canada, "Public Health Notice Update - Outbreak of E.coli
infections with possible link to leafy greens," 25 May 2015. [Online]. Available:
https://www.canada.ca/en/public-health/services/public-health-notices/2015/public-healthnotice-update-outbreak-e-coli-infections-possible-link-leafy-greens.html.
GS1 US,, ""Frontera Produce: Traceability from Farm to Store"," 2010. [Online].
Available:
https://www.producetraceability.org/documents/Frontera%20Produce%20Traceability%20
Case%20Study.pdf.
S. Pouliot and D. A. Sumner, "Traceability, Liability and Incentives for Food Safety and
Quality," American Journal of Agricultural Economics, vol. 90, no. 1, pp. 15-27, 2008.
J. Goldenberg, B. Libai and E. Muller, "The chilling effects of network externalities,"
International Journal of Research in Marketing, vol. 27, no. 1, pp. 4-15, 2010.
N. Mai, S. Gretar Bogason, S. Arason, S. Vikingur Arnason and T. Geir Matthiasson,
"Benefits of traceability in fish supply chains - case studies," British Food Journal, vol.
112, no. 9, pp. 976 - 1002, 2010.
K. Appold, "Challenges of Preventing Leafy Green Outbreaks," 2019. [Online]. Available:
https://www.foodqualityandsafety.com/article/fresh-produce-outbreaks-challengesprevention/?singlepage=1. [Accessed 31 August 2020].
U.S. Department of Agriculture, "National Agriculture Statistics Service, "2017 Census of
agriculture"," USDA, Washington, DC, 2019.
U.S. Department of Agriculture National Agricultural Statistics Service, "2018 Census of
Aquaculture," 2019.
GS1 US, "Foodservice GS1 US Standards Initiative," 2020. [Online]. Available:
https://www.gs1us.org/industries/foodservice/initiative.
194

[16] GS1 US, "Improving Traceability and Food Safety with GS1 Standards in Fresh Foods,"
2012. [Online]. Available:
https://www.gs1us.org/DesktopModules/Bring2mind/DMX/Download.aspx?Command=C
ore_Download&EntryId=593&language=en-US&PortalId=0&TabId=134.
[17] E. Golan, B. Krissoff, F. Kuchler, L. Calvin, K. Nelson and G. Price, "Traceability in the
U.S. Food Supply: Economic Theory and Industry Studies," Agricultural Economic
Report, vol. 830, 2004.
[18] Grocery Manufacturers Association, "Recall execution effectiveness:Collaborative
approaches to improving consumer safety and confiedence," GMA, Washington, DC,
2010.
[19] Centers for Disease Control and Prevention, "Multistate Outbreak of Human Salmonella
Enteritidis Infections Associated with Shell Eggs," 9th October 2010. [Online]. Available:
https://www.cdc.gov/salmonella/2010/shell-eggs-12-2-10.html.
[20] H. Roberts and S. R. Veil, "Health literacy and crisis: Public relations in the 2010 egg
recall," Public Relations Review, pp. 214-218, 2016.
[21] Grocery Manufacturers Association, "Capturing recall costs:measuring and recovering the
losses," GMA, Washington DC, 2011.
[22] Y. Motarjemi and M. Adams, "Emerging Foodborne Pathogens," Cambridge, the U.K. ,
Elsevier Science, 2006.
[23] J. Blankenship, T. Cloyd, K. Irvin, K. Blickenstaff and A. Fields, Memorandum of Record.
FDA Foodborne Illness Outbreak Data Used to Inform the "Requirements for Additional
Traceability Records for Certain Foods" Rule Final Regulatory Impact Analsysis, U.S.
Food and Drug Administration - Center for Food Safety and Applied Nutrition Coordinated Outbreak Response and Evaluation Network (CORE), 2022.
[24] E. Scallan, R. M. Hoekstra,, F. J. Angulo, R. V. Tauxe, M. A. Widdowson, S. L. Roy, J. L.
Jones and P. M. Griffins, "Foodborne illness acquired in the United States--major
pathogens.," Emerging Infectious Disease, pp. 7-15, 2011a.
[25] R. Pennotti, E. Scallan, L. Backer, J. Thomas and F. Angulo, "Ciguatera and Scombroid
Fish Poisoning in the United States," Foodborne Pathogens and Disease, vol. 10, no. 12,
pp. 1059-1066, 2013.
[26] E. Scallan, P. M. Griffins, F. J. Angulo, R. V. Tauxe and R. M. Hoekstra, "Foodborne
illness acquired in the United States -- unspecified agents.," Emerging Infectious Disease,
pp. 16-22, 2011b.
[27] T. Minor, A. Lasher, K. Klontz, B. Brown, C. Nardinelli and D. Zorn, "The per case and
total annual costs of foodborne illness in the United States," Risk Analysis, pp. 1125-39,
2015.
[28] L. R. Scharff, "The Economic Burden of Foodborne Illness in the United States," in Food
Safety Economics: Incentives for a Safer Food Supply, Gewerbestrasse 11, 6330 Cham,
Switzerland, Springer International Publishing AG,, 2018, pp. 123-142.
[29] U.S. Department of Health and Human Services, "Guidelines for Regulatory Impact
Analysis," 2016.
[30] A. Fields, M. Salter and k. Irvin, Memorandum of Record, U.S. Food and Drug
Administration - Center for Food Safety and Applied Nutrition, 2020.
195

[31] K. Kiesel, R. E. Goodhue, R. J. Sexton, and A. Spalding, "E. Coli in the Romaine Lettuce
Industry: Economic Impacts from the November 2018 Outbreak," Department of
Agricultural and Resources Economics, UC Davis, California, 2021.
[32] A. Meyerson, "An Analysis of the First-Order Economic Costs of the 2008 FDA Tomato
Warning," Leonard N. Stern School of Business, New York University, New York, 2009.
[33] Kawamura, The Recent Salmonella Outbreak: Lessons Learned and Consequences to
Industry and Public Health, Washington DC: US Government Printing Office, 2008.
[34] M. A. Palma, L. A. Ribera, D. Bessler, M. Paggi and R. D. Knutson, "Potential Impacts of
Foodborne Illness Incidences on Market Movements and Prices of fresh Produce in the
U.S," Journal of Agricultural and Applied Economics, vol. 42, no. 4, pp. 731 - 741, 2010.
[35] M. A. Hussain and C. Dawson, "Economic Impact of Food Safety Outbreaks on Food
Businesses," Foods, vol. 2, no. 4, pp. 585 - 589, 2013.
[36] Tyco Integrated Security, "Recall: The Food Industry's Biggest Threat to Profitability,"
Food Safety Magazine, p. 7, 11 October 2012.
[37] Food and Agriculture Organization of the United Nations, "Food wastage footprint: Fullcost accounting-Final Report," FAO, Geneva, 2014.
[38] Eastern Research Group (ERG), "Recall Cost Case Studies," Eastern Research Group, Inc,
Massachussettes, 2020.
[39] Food and Drug Administration,, Fiscal Year 2022: Justification of Estimates for
Appropriation Committees, Silver Spring, Maryland: Department of Health and Human
Services (DHHS), 2021, p. 410.
[40] J. C. McEntire, S. Arens, M. Bernstein, B. Bugusu, F. F. Busta, M. Cole, A. Davis, W.
Fisher, S. Geisert, H. Jensen and B. Kenah, "Traceability (product tracing) in food
systems: An IFT report submitted to the FDA, Volume 1: Technical aspects and
recommendations," Comprehensive Reviews in Food Science and Food Safety, vol. 1, no.
9, pp. 92-158, 2010.
[41] U.S. Food and Drug Administration, "Draft Guidance for Industry: Compliance with and
Recommendations for Implementation of the Standards for the Growing, Harvesting,
Packing, and Holding of Produce for Human Consumption for Sprout Operations," 2017.
[42] S. Mahajan, C. Caraballo, Y. Lu, J. Valero-Elizondo, D. Massey, A. Annapureddy, B.
Roy, C. Riley, K. Murugiah, O. Onuma, M. Nunez-Smith, H. Forman, K. Nasir, J. Herrin
and H. Krumholz, "Trends in Differences in Health Status and Health Care Access and
Affordability by Race and Ethnicity in the United States, 1999-2018," Journal of the
American Medical Association, vol. 326, no. 7, pp. 637-648, 2021.
[43] J. Jia, M. Zack, D. Moriarty and D. Fryback, "Predicting the EuroQol Group's EQ-5D
index from CDC's "Healthy Days" in a US sample," Medical Decision Making, vol. 31,
no. 1, pp. 174-185, 2011.
[44] P. Sullivan, W. Lawrence and V. Ghushchyan, "A national catalog of preference-based
scores for chronic conditions in the United States," Medical Care, vol. 43, no. 7, pp. 736749, 2005.
[45] S. Hamidi, "Urban sprawl and the emergence of food deserts in the USA," Urban Studies,
vol. 57, no. 8, pp. 1660-1675, 2019.

196

[46] R. B. Freeman and R. H. Oostendorp, Occupational Wages around the World (OWW)
Database: Update for 1983-2008 [Data file and documentation]., Available from NBER
website: https://www.nber.org/oww/, 2012.
[47] Education First(EF) English Proficiency Index (EPI), A Ranking of 100 Countries and
REgions by English Skills, Miami, Florida, 2020.
[48] Miniwatts Marketing Group, "www.internetworldstats.com/list2.html," Miniwatts
Marketing Group, 11 June 2022. [Online]. Available:
https:internetworldstats.com/list2.htm. [Accessed 07 September, 2022].
[49] C. Behravesh, R. K. Mody, J. Jungk and et.al., "2008 Outbreak of Salmonella Saintpaul
Infections Associated with Raw Produce," N Engl J Med 2011, vol. 364, pp. 918-27, 2011.
[50] L. A. Ribera, M. A. Palma, M. Paggi, R. Knutson, J. G. Masabni and J. Anciso,
"Economic Analysis of Food safety Compliance Costs and Fodborne Illness Outbreaks in
the United States," Hort Technology 22 (2), pp. 150-156, 2012.
[51] K. Smith, B. Miller, K. Vierk, I. Williams and C. Hedberg, "Product tracing in
epidemiologic investigations of outbreaks due to commercially distributed food items -utility, application, and considerations," 2015.
[52] D. M. Tack, L. Ray, P. M. Griffin, P. R. Cieslak, J. Dunn, T. Rissman, R. Jervis, S.
Lathrop, A. Muse, M. Duwell, K. Smith, M. Tobin-D'Angelo, D. Vugia, J. Z. Kufel, B. J.
Wolpert, R. Tauxe and D. C. Payne, "Preliminary Incidence and Trends of Infections with
Pathogens Transmitted Commonly Through Food- Foodborne Disease Active Surveillance
Network, 10 U.S. Sites, 2016-2019," Morbidity and Mortality Weekly Report (MMWR),
pp. 509-514, 2020.
[53] U.S. Food and Drug Administration, "Regulatory Impact Anaysis: Standards for the
Growing, Harvesting, Packing, and Holding of Produce for Human Consumption Final
Rule," 2015.

197

V. Appendices
A. Food Traceability List (FTL)
As of this writing, the Food Traceability List includes the following foods (Table A.1).
After publication of the final rule, the list can be updated using the procedure set forth in
§ 1.1465.
Table A.1. Food Traceability List
Foods
Cheeses, other than hard cheeses, specifically:


Cheese (made from pasteurized milk),
fresh soft or soft unripened



Cheese (made from pasteurized milk),
soft ripened or semi-soft



Cheese (made from unpasteurized
milk), other than hard cheese91

Shell eggs

Nut butters

Description

Includes soft unripened/fresh soft cheeses.
Examples include, but are not limited to, cottage,
chevre, cream cheese, mascarpone, ricotta,
queso blanco, queso fresco, queso de crema, and
queso de puna. Does not include cheeses that are
frozen, shelf stable at ambient temperature, or
aseptically processed and packaged.
Includes soft ripened/semi-soft cheeses.
Examples include, but are not limited to, brie,
camembert, feta, mozzarella, taleggio, blue,
brick, fontina, monterey jack, and muenster.
Does not include cheeses that are frozen, shelf
stable at ambient temperature, or aseptically
processed and packaged.
Includes all cheeses made with unpasteurized
milk, other than hard cheeses. Does not include
cheeses that are frozen, shelf stable at ambient
temperature, or aseptically processed and
packaged.
Shell egg means the egg of the domesticated
chicken.
Includes all types of tree nut and peanut butters.
Examples include, but are not limited to,
almond, cashew, chestnut, coconut, hazelnut,
peanut, pistachio, and walnut butters. Does not
include soy or seed butters.

“Hard cheese” includes hard cheeses as defined in 21 CFR 133.150, colby cheese as defined in 21 CFR 133.118
and caciocavallo siciliano as defined in 21 CFR 133.111. Examples of hard cheese include, but are not limited to,
cheddar, romano, and parmesan.

198

Cucumbers (fresh)

Includes all varieties of fresh cucumbers.

Herbs (fresh)

Includes all types of fresh herbs. Examples
include, but are not limited to, parsley, cilantro,
and basil. Herbs listed in 21 CFR 112.2(a)(1),
such as dill, are exempt from the requirements of
the rule under 21 CFR 1.1305(e).
Includes all types of fresh leafy greens.
Examples include, but are not limited to,
arugula, baby leaf, butter lettuce, chard, chicory,
endive, escarole, green leaf, iceberg lettuce, kale,
red leaf, pak choi, Romaine, sorrel, spinach, and
watercress. Does not include whole head
cabbages such as green cabbage, red cabbage or
savoy cabbage. Does not include banana leaf,
grape leaf and leaves that are grown on trees.
Leafy greens listed in § 112.2(a)(1), such as
collards, are exempt from the requirements of
the rule under § 1.1305(e).
Includes all types of fresh-cut leafy greens,
including single and mixed greens.

Leafy greens (fresh)

Leafy greens (fresh-cut)
Melons (fresh)
Peppers (fresh)
Sprouts (fresh)

Tomatoes (fresh)
Tropical tree fruits (fresh)

Includes all types of fresh melons. Examples
include, but are not limited to, cantaloupe,
honeydew, muskmelon, and watermelon.
Includes all varieties of fresh peppers.
Includes all varieties of fresh sprouts
(irrespective of seed source), including single
and mixed sprouts. Examples include, but are
not limited to, alfalfa sprouts, allium sprouts,
bean sprouts, broccoli sprouts, clover sprouts,
radish sprouts, alfalfa & radish sprouts, and
other fresh sprouted grains, nuts, and seeds.
Includes all varieties of fresh tomatoes.
Includes all types of fresh tropical tree fruit.
Examples include, but are not limited to, mango,
papaya, mamey, guava, lychee, jackfruit, and
starfruit. Does not include non-tree fruits such as
bananas, pineapple, dates, soursop, jujube,
passionfruit, Loquat, pomegranate, sapodilla,
and figs. Does not include tree nuts such as
coconut. Does not include pit fruit such as
avocado. Does not include citrus, such as
orange, clementine, tangerine, mandarins,
lemon, lime, citron, grapefruit, kumquat, and
pomelo.
199

Fruits (fresh-cut)

Includes all types of fresh-cut fruits. Fruits listed
in § 112.2(a)(1) are exempt from the
requirements of the rule under § 1.1305(e).

Vegetables other than leafy greens (fresh-cut)

Includes all types of fresh-cut vegetables other
than leafy greens. Vegetables listed in §
112.2(a)(1) are exempt from the requirements of
the rule under § 1.1305(e).

Finfish (fresh and frozen), specifically:


Finfish, histamine-producing species



Finfish, species potentially
contaminated with ciguatoxin



Finfish, species not associated with
histamine or ciguatoxin

Smoked finfish (refrigerated and frozen)

Includes all histamine-producing species of
finfish. Examples include, but are not limited to,
tuna, mahi mahi, mackerel, amberjack, jack,
swordfish, and yellowtail.
Includes all finfish species potentially
contaminated with ciguatoxin. Examples
include, but are not limited to, grouper,
barracuda, and snapper.
Includes all species of finfish not associated with
histamine or ciguatoxin. Examples include, but
are not limited to, cod, haddock, Alaska pollock,
salmon, tilapia, and trout.92 Siluriformes fish,
such as catfish, are not included.93
Includes all types of smoked finfish, including
cold smoked finfish and hot smoked finfish.94

Crustaceans (fresh and frozen)

Includes all crustacean species. Examples
include, but are not limited to, shrimp, crab,
lobster, and crayfish.

Molluscan shellfish, bivalves (fresh and
frozen)95

Includes all species of bivalve mollusks.
Examples include, but are not limited to, oysters,
clams, and mussels. Does not include scallop
adductor muscle. Raw bivalve molluscan
shellfish that are (1) covered by the requirements

For a more comprehensive list see Chapter 3 of the Fish and Fishery Products Hazards and Controls Guidance at
https://www.fda.gov/food/seafood-guidance-documents-regulatory-information/fish-and-fishery-products-hazardsand-controls.
93
Data for catfish were excluded from the Risk-Ranking Model because Siluriformes fish (such as catfish) are
primarily regulated by the U.S. Department of Agriculture.
94
“Smoked finfish” refers to a finfish product that meets the definition of a smoked or smoke-flavored fishery
product in 21 CFR 123.3(s).
95
Per 21 CFR 123.3(h) molluscan shellfish means any edible species of fresh or frozen oysters, clams, mussels, or
scallops, or edible portions of such species, except when the product consists entirely of the shucked adductor
muscle.

200

Ready-to-eat deli salads (refrigerated)

of the National Shellfish Sanitation Program; (2)
subject to the requirements of 21 CFR part 123,
subpart C, and 21 CFR 1240.60; or (3) covered
by a final equivalence determination by FDA for
raw bivalve molluscan shellfish are exempt from
the requirements of the rule under § 1.1305(f).
Includes all types of refrigerated ready-to-eat
deli salads. Examples include, but are not
limited to, egg salad, potato salad, pasta salad,
and seafood salad. Does not include meat salads.

201

B. Methodology Used to Estimate the Number of Illnesses
To obtain the number of illnesses, hospitalizations, and deaths reported in
Table 5, we rely on FDA Coordinated Outbreak Response and Evaluation (CORE) data
(Ref. [23]). We report the summary of these data, covering the 12-year period from January 2009
through December 2020, in columns 1-4 of Table B.1 in this appendix. We do not use CDC
outbreak data for our estimates because CDC data include illnesses resulting from improper food
handling, as well as illnesses associated with foods not regulated by FDA. We note that CORE
data include more illnesses than those attributable to covered food products. These include
adverse reactions and fungus-related illnesses. We therefore use only a subset of CORE data on
foodborne illness outbreaks associated with covered foods.1 Based on these data, of 31 known
foodborne illness-causing pathogens, 14 are commonly associated with foods currently
designated by FDA on the FTL. We list these pathogens in Table B.1.
Due to the sparsity of outbreak data on unspecified agents, as well as on underreporting
and underdiagnosis of foodborne illnesses, our estimates are subject to assumptions described
below. To account for underreporting as well as underdiagnosing of foodborne illnesses, we
apply Scallan et al. (2011a) multipliers for each pathogen, presented in columns 5, 6, and 7 of
Table B.1 (Ref. [24]). Column 5 contains Scallan et al.’s underdiagnosis multipliers specifically
for hospitalizations and deaths, taken from the authors’ Technical Appendix 3 (Ref.
[24]). Columns 6 and 7 contain the multipliers for underreporting and underdiagnosis of
illnesses, also taken from Table 3.5 of the authors’ Technical Appendix 3. Columns 8, 9, and 10
show the resulting estimates of the total number of illnesses, hospitalizations, and deaths from
covered foods over the 12-year period.

202

Scallan et al. (2011a) does not provide underreporting multipliers for three pathogens
(Ciguatoxins, Listeria, Norovirus and Salmonella ), and two contaminants (Ciguatoxin or
Scombrotoxin), and for Norovirus, Ciguatoxins, and Scombrotoxins the authors do not provide
underdiagnosis multipliers. Furthermore, at least two foodborne illness-causing agents
(Ciguatoxins and Scombrotoxin) do not have hospitalization and death underdiagnosis
multipliers, for which we assume a value of one. We obtain underdiagnosis multipliers for
illnesses from Ciguatoxins and Scombrotoxin from Pennotti et al. (2013) (Ref. [25]).
Following Scallan et al.’s (2011a) treatment of other pathogens/contaminants, we assume
100 percent reporting for pathogens/contaminants without underreporting multipliers and 100
percent diagnosis for those without underdiagnosis multipliers (Ref. [24]).2 To estimate the
number of annual illnesses, hospitalizations, and deaths caused by each pathogen/contaminant,
we divide columns 8, 9, and 10 of Table B.1 by 12 years. Columns 11, 12, and 13 provide the
resulting annual estimates. We estimate that these pathogens/contaminants cause 153,807
illnesses 364, hospitalizations, and 14 deaths annually via consumption of covered foods.
According to Scallan et al. (2011b) and CDC3, nearly 80 percent of foodborne illnesses,
53 percent of hospitalized foodborne illnesses, and 58 percent of deaths from foodborne illnesses
result from unspecified or unknown pathogens/contaminants (Ref. [26]). We multiply the
number of hospitalizations and deaths by 2.13 (= 1/ (1 - 0.53)) and 2.28 (= 1/ (1 - 0.58))
respectively to account for unspecified agents. Not considering the burden caused by unspecified
and unknown pathogens/contaminants could result in underestimation of covered foods caused
illnesses. Following Scallan et al. (2011b), the estimated number of annual illnesses in
Table 5 (153,807) before adjusting it for the number of illnesses from unspecified and
unknown agents constitute only 20 percent of all illnesses [26].

203

In our estimates we assume the same ratios for unidentified to identified illnesses due to
covered foods. The assumption is consistent with FDA’s past regulatory impact analyses,
including the RIA for the 2015 final rule on Standards for the Growing, Harvesting, Packing, and
Holding of Produce for Human Consumption. We made this assumption because outbreak data
on unidentified pathogens, specifically their associated food commodity, is extremely sparse.
The approach presumes that the percentage of identified illnesses, across all
pathogens/contaminants, attributable to FTL products and ingredients would be lower than the
percentage of illnesses from unidentified pathogens/contaminants attributable to same products.
The last row of Table B.1 present estimates of total illnesses, hospitalizations, and deaths
from covered foods after scaling for unspecified and/or unknown agents. We estimate that the
total number of illnesses from known toxins and pathogens is 153,807 and from unspecified and
unknown agents is 615,226 (= 153,807 x (1-0.2)/0.2), which yields a total of 769,033 illnesses (=
153,807 + 615,226).96 We used these estimates in our Final Regulatory Impact Analysis. We also
scale up the number of hospitalizations and deaths to account for unspecified agents and estimate
that in total 775 hospitalizations (= 364 / (1 - 0.53)) and 33 deaths (= 14 / (1 - 0.58)) are caused
by covered foods annually. Table B2. Shows the estimated range of the burden of foodborne
illnesses and cost per illness associated with covered foods.

Figures are rounded to the nearest dollar.

204

Table B.1. Illnesses, Hospitalizations, and Deaths Attributable to Illness-Causing Pathogens Associated with Covered Foods

Campylobacter

(1)
Number
of FTL
Related
Outbreaks

(2)
Illnesses

(3)
Hospitalizations

(4)
Deaths

(5)
Hospitalization
and Death
Multipliers due
to
Underdiagnosi
s

Raw Count of FTL Illnesses

(6)
Scallan
Underreport
ing
Multiplier

(7)
Scallan/
Pennotti
Underdiagn
osis
Multiplier

30.3

Estimated Total (2009-2020)
(9)
(8)
Hospitali
(10)
Total Illnesses
zation
Deaths
758

Estimated Annual Total (2009-2020)
(11)
(12)
Total
Hospitali
(13)
Illnesses
zation
Deaths
63

Ciguatoxin

9.91

377

Cyclospora

3658

122

83.1

303,980

244

25,332

E. coli (STEC) O157
E-Coli (STEC) nonO157

993

396

25.5

26.1

660,891

792

55,074

221

25.5

106.8

601,871

124

50,156

Hepatitis A Virus

1.1

9.1

160

Listeria

325

310

2.3

748

620

116

Norovirus

329

1.5

29.3

9,640

803

Salmonella typhoidal
Salmonella nontyphoidal

180

13.3

2,394

200

8193

1240

29.3

240,055

2480

20,005

207

Scombrotoxin

105

12.21

1,282

107

Vibrio-para

1.1

142.4

15,351

1,279

Vibrio-Cholerae
Yersinia
enterocolitica
Total from specified
pathogens
Total including
unspecified/unidenti
fied pathogens

1.1

33.1

437

122.8

7,736

645

205

14,256

2,192

1,845,679

4,374

168

153,807

364

9,228,394

9,261

400

769,033

775

Table B2. Estimated Range of the Baseline Economic Burden of Foodborne Illnesses Associated with Covered Foods (2020$)
Estimated
Annual
Cases

Cost of Illness:
Primary

Cost of Illness:
Low

Cost of Illness:
High

Total: Primary

Total: Low

Total: High

Campylobacter

$4,748

$2,487

$6,971

$299,718

$156,992

$440,044

Ciguatoxin

$31,402

$15,333

$47,208

$985,447

$481,175

$1,481,466

$4,451

$2,092

$6,771

$112,751,174

$52,993,812

$171,520,602

$13,757

$8,122

$19,299

$757,656,629

$447,313,160

$1,062,878,192

$2,506

$1,305

$3,687

$125,690,811

$65,453,515

$184,924,988

Cyclospora cayentanensis
E. coli (STEC) O157
E-Coli (STEC) non-O157

25,332
55,074
50,156

Hepatitis A Virus

$58,440

$28,640

$87,752

$779,979

$382,249

$1,171,197

Listeria Monocytogenes

$1,987,005

$991,975

$2,965,723

$123,773,853

$61,791,776

$184,739,829

Norovirus

803

$487

$281

$690

$391,211

$225,730

$554,283

Salmonella typhoidal

200

$7,116

$5,380

$8,824

$1,419,642

$1,073,310

$1,760,388

$7,248

$3,800

$10,639

$144,993,160

$76,017,385

$212,828,673

$548

$485

$611

$58,547

$51,816

$65,278

$2,636

$1,300

$3,951

$3,372,041

$1,662,995

$5,054,225

$1,675

$971

$2,367

$60,987

$35,354

$86,182

645

$6,255

$3,000

$9,457

$4,032,599

$1,934,100

$6,096,928

$1,276,265,797

$709,573,368

$1,833,602,274

Salmonella non-typhoidal
Scombrotoxin
Vibrio-parahaemolyticus
Vibrio-Cholerae
Yersinia enterocolitica
(i) Subtotal/Weighted
Average (rounded):
Known Pathogens

20,005
107
1,279

153,807

Average Cost per Illness
(ii) Unidentified/
Unspecified Pathogens
(iii) Total cases (i) & (ii)
(rounded)

$8,298

$4,613

$11,921

615,226
769,033

$8,298

$4,613

$11,921

$5,105,063,190

$2,838,293,471

$7,334,409,096

$6,381,328,987

$3,547,866,839

$9,168,011,370

206

C. Outbreak Case Studies Used in Estimation of Public Health Benefits
The dataset (Table C.1) represents 23 foodborne outbreaks from 2008 – 2019 coordinated
by FDA’s Emergency Coordination Response Team (ECRT) (2008 – 2010) and FDA’s
Coordinated Outbreak Response and Evaluation (CORE) Network (2011 – 2019), yielding 23
Public Health Benefit Case Studies97. Each outbreak included in this analysis:


involved a major pathogen/contaminant (Salmonella, Escherichia coli (STEC), Listeria
monocytogenes, Cyclospora cayetanensis, Vibrio parahemolyticus and Scombrotoxin);



involved FDA-regulated food(s) from the Food Traceability List (FTL) that was
identified as the outbreak vehicle and/or contaminated product;



involved a formal traceback investigation that was coordinated by FDA;



resulted in voluntary or enforced product interventions;



and resulted in public communications issued by FDA and/or CDC.

Definitions


Year – The reported year that an outbreak was evaluated/investigated by FDA.



Pathogen/Contaminant – The identified pathogen or contaminant associated with an
outbreak according to the case definition, as defined by CDC.



Species/Serotype – The species/serotype(s) that corresponds to the reported pathogen as
determined by CDC.



Commodity – The item(s) identified by FDA as the outbreak vehicle and/or
contaminated product.

Outbreaks numbered 1-15 in Table C.1 were used in the PRIA appendix C. We update this analysis with 6
additional outbreaks numbered 17 through 23 (2018-2019) from FDA’s Coordinated Outbreak Response and
Evaluation (CORE) Network.



Response Start Date – The date that a given outbreak was transferred to a CORE
Response Team.



Traceback Initiation Date – The date that represents when FDA’s traceback
investigation began.
o This date represents when the first traceback information request was issued
for record collection.



Traceback Completion Date – The date that represents when FDA’s traceback
investigation (including the review of collected records and documentation of
findings) ended.
o This date represents when the last record was received by FDA for the
traceback investigation.



Response End Date – The date that a given outbreak was closed by a CORE
Response Team.



Final CDC Publication Date – The date of publication for the final outbreak web
posting or corresponding update issued by CDC.



Final CDC Web Post Link – The link to the final outbreak web posting or
corresponding update issued by CDC.
o For these case studies, the epidemiologic data that was used for the analysis
included the final case count, hospitalization, and death totals that were
publicly reported for a given outbreak.

208

Limitations
The outbreaks used for the Public Health Benefit Case Studies were selected to represent
significant outbreaks involving some of the covered foods. It should be noted that these cases
studies do not represent all foodborne outbreaks investigated or traceback investigations
conducted from 2008 to 2019, nor do they represent all occurrences of an FTL food product
being implicated as the cause of an outbreak during that timeframe.
Another limitation of this analysis is the use of publicly reported epidemiologic data
(case count, hospitalizations, deaths). For some outbreaks, the publicly reported values may
differ from the final values internally reported by FDA and/or CDC, including the data that was
used for the risk ranking analysis. Specifically, for outbreaks associated with Cyclospora
cayetanensis, lack of a validated molecular subtyping methodology made it difficult to
differentiate historical outbreaks that may have been occurring concurrently but were associated
with different products. This led to challenges regarding the attribution of epidemiologic data to
those distinct outbreaks and/or commodities.
Additionally, the level of FDA documentation that was readily available for the outbreak
investigations included in this analysis varied, especially for those outbreaks that occurred before
CORE was established in 2011. The Traceback Initiation and Completion Dates are estimations
that best represent when the traceback investigations started and ended based on data pulled from
varying sources documenting each outbreak (e.g., email correspondence, outbreak summary
documents, etc.). The values in the dataset represent the best data currently available for
comparing the investigational elements of interest across these outbreaks.

209

Table C.1. Outbreak Case Studies Used for Estimation of Public Health Benefits
FDA
Traceba
ck
Initiatio
n Date
3/5/2008

FDA
Traceba
ck
Completi
on Date
4/10/200
8

5/28/200
8

6/1/2008

Alfalfa
Sprouts

2/26/200
9

Enteritidis

Shell
Eggs

E. coli

O145

2011

Listeria

2012

N
o.

Yea
r

Pathogen/
Contamin
ant

2008

Salmonell
a

Litchfield

Cantalou
pe

3/4/2008

2008

Salmonell
a

Saintpaul

Hot
Peppers

2009

Salmonell
a

Saintpaul

2010

Salmonell
a

2010

Species/
Serotype

Respons
e Start
Date

Respons
e End
Date

Final
CDC
Publicati
on Date

Final CDC Web-post Link

4/15/200
8

4/2/2008

https://www.cdc.gov/salmonella/2008/cantaloupes-4-22008.html

7/17/200
8

8/5/2008

8/28/200
8

https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5734a1.ht
m

3/2/2009

4/30/200
9

5/1/2009

5/8/2009

https://www.cdc.gov/salmonella/2009/raw-alfalfa-sprouts-5-82009.html

7/26/201
0

8/4/2010

8/31/201
0

9/3/2010

12/2/201
0

https://www.cdc.gov/salmonella/enteritidis/se_timeline_092010
.pdf

Romaine
Lettuce

4/21/201
0

4/27/201
0

5/11/201
0

5/21/201
0

https://www.cdc.gov/ecoli/2010/shredded-romaine-5-2110.html

monocytogen
es

Cantalou
pe

9/7/2011

9/11/201
1

11/23/20
11

12/16/20
11

8/27/201
2

https://www.cdc.gov/listeria/pdf/listeriosis-timeline102711.pdf

E. coli

O26

Clover
Sprouts

2/3/2012

2/7/2012

2/17/201
2

5/22/201
2

4/3/2012

https://www.cdc.gov/ecoli/2012/O26-02-12/index.html

E. coli

O157:H7

Spinach

11/1/201
2

11/29/20
12

12/21/20
12

12/20/20
12

https://www.cdc.gov/ecoli/2012/o157h7-11-12/adviceconsumers.html

Leafy
Greens

7/11/201
3

9/26/201
3

8/13/201
3

11/7/201
3

12/16/20
13

12/2/201
3

https://www.cdc.gov/parasites/cyclosporiasis/outbreaks/investi
gation-2013.html

9
2013

Commod
ity

Cyclospo
ra

cayetanensis

(1)

Cilantro
(1)

2016

E. coli

O157:NM

Alfalfa
Sprouts

2/18/201
6

2/19/201
6

3/22/201
6

4/5/2016

3/5/2016

https://www.cdc.gov/ecoli/2016/o157-02-16/index.html

2018

E. coli

O157:H7

Romaine
Lettuce

11/9/201
8

11/15/20
18

12/17/20
18

3/25/201
9

1/9/2019

https://www.cdc.gov/ecoli/2018/o157h7-11-18/index.html

2019

Cyclospo
ra

cayetanensis

Basil

7/11/201
9

7/15/201
9

7/30/201
9

2/3/2020

9/30/201
9

https://www.cdc.gov/parasites/cyclosporiasis/outbreaks/2019/w
eekly/index.html

2019

E. coli

O157:H7

Romaine
Lettuce

11/12/20
19

11/18/20
19

12/13/20
19

3/16/202
0

1/15/202
0

https://www.cdc.gov/ecoli/2019/o157h7-11-19/index.html

2019

Salmonell
a

Javiana

Cantalou
pe

12/6/201
9

1/8/2020

3/18/202
0

2/18/202
0

https://www.cdc.gov/salmonella/javiana-12-19/index.html

N
o.

Yea
r

Pathogen/
Contamin
ant

2018

Salmonell
a

2018

FDA
Traceba
ck
Initiatio
n Date
3/22/201
8

FDA
Traceba
ck
Completi
on Date
4/3/2018

Respons
e End
Date

Final
CDC
Publicati
on Date

6/13/201
8

7/26/201
9

https://www.cdc.gov/salmonella/Braenderup-04-18/index.html

8/3/2018

10/24/20
18

9/27/201
8

https://www.cdc.gov/vibrio/investigations/vibriop-0718/index.html

11/14/20
18

11/23/20
18

2/21/201
9

2/27/201
9

https://www.cdc.gov/salmonella/concord-11-18/index.html

2/26/201
9

4/9/2019

5/16/201
9

5/22/201
9

https://www.cdc.gov/salmonella/newport-04-19/epi.html

Tahini

4/22/201
9

4/25/201
9

5/1/2019

6/27/201
9

6/26/201
9

https://www.cdc.gov/salmonella/concord-05-19/index.html

Ground
Bison

7/1/2019

7/2/2019

8/19/201
9

9/11/201
9

9/13/201
9

https://www.cdc.gov/ecoli/2019/bison-07-19/index.html

Tuna

10/1/201
9

10/4/201
9

11/14/20
19

1/15/202
0

1/24/202
0 (2)

Hard
Boiled
Eggs

12/10/20
19

1/9/2020

2/20/202
0

3/4/2020

https://www.fda.gov/food/outbreaks-foodborneillness/outbreak-investigation-scombrotoxin-fish-poisoningyellowfinahi-tuna-november-2019
https://www.cdc.gov/listeria/outbreaks/eggs-12-19/index.html

Commod
ity

Respons
e Start
Date

Braenderup

Shell
Eggs

3/19/201
8

Vibrio

parahaemolyt
icus

Crab
Meat

7/2/2018

7/10/201
8

2018

Salmonell
a

Concord

Tahini

11/9/201
8

2019

Salmonell
a

Newport

Ground
Tuna

2019

Salmonell
a

Concord

2019

E. coli

O121:H19;
O103:H2

2019

Scombrot
oxin

2019

Listeria

Species/
Serotype

monocytogen
es

Final CDC Web-post Link

(1)

Outbreaks number 9 and 10 represent two concurrent outbreak investigations attributed to the same pathogen/contaminant but
different commodities.
(2)
Outbreak number 22 only shows FDA posting - no CDC post for this outbreak.

211

D. Estimation of the Number of Covered Entities
To obtain the number of covered entities by the traceability rule, we use several sources.
These include the 2017 SUSB and the 2017 North American Product Classification System
(NAPCS) data from the U.S. Census, data and summary reports from the 2017 Census of
Agriculture, and the 2018 Census of Aquaculture from the 2017 USDA National Agricultural
Statistics Service (NASS) (Ref. [13] [14]). All datasets used in this analysis were the latest
available to us as of January 2022.
Produce Farms
For the number of produce farms and sales of covered foods, we use the raw 2017 USDA
NASS data. We first derive total number of produce farms by NAICS code (111219, 111339,
and 111419). We also derive the number of produce farms that are fully exempt by their annual
sales with less than $25,000, and by selling directly to consumers. We then derive the number of
farms eligible for both exemption criteria. We first subtract the estimates of the fully exempt
farms (if their sales are below $25,000 or they sell their covered foods directly to consumers)
from the total number of produce farms. We separate out farms that are packers from the total
number of produce farms as they are not exempt by the rule98 and still exclude packers who are
eligible for the sales exemption (annual sales less than $25,000) or sell covered food directly to
consumers. To estimate the number of produce farms covered by the rule, we then multiply the
remaining number of farms in NAICS codes 111219, 111339, and 111419 by the FTL share to

To account for farms with packing operations, we use an indicator variable of on-farm packing facilities from the
2017 USDA NASS data.

212

account only for farms producing FTL foods. Due to lack of information on how many farms
produce covered foods, we assume that all covered produce farms produce foods on the FTL.99
For sprout growers, we use the inventory of sprout farms and operations used by the
FDA’s Office of Regulatory Affairs. Excluding very small sprout growers, this internal inventory
counts 95 sprout growers that we believe to be covered by this rule.
Shell Egg Farms Estimates
We derive the number of shell egg farms and sales of shell eggs sold directly to
consumers, and number of layers less than 3,000 from 2017 USDA NASS summary report.100
Out of 232,500 shell egg farms, most of the farms (227,340 shell egg farms) have less than 3,000
layers that approximately 98 percent of the total shell egg farms will be exempt by the final rule.
According to the 2017 NASS summary report, 8,107 shell egg farms sell directly to consumers,
which is about 3 percent of the total shell egg farms. We assume that the remainder shell egg
farms with 3,000 layers or more do not sell directly to consumers as most of the covered farms
are large farms. We therefore estimate that 3,782 shell egg farms (2 percent of the total shell egg
farms) are covered by the rule. Because we do not have additional information on the number of
packers out of the remainder shell egg farms (2 percent), we assume that all of these farms are
packers. As mentioned in the section II. I, we may overestimate the number of shell egg farms
covered by the rule as we assume direct to consumer sales among the farms with more than
3,000 layers are zero percent. Similar with the produce farms, we multiply the number of
covered shell egg farms by the FTL share to account for shell egg farms producing covered

Because we are unsure to the exact the number of produce farms who manufacture, process, pack, or hold covered
foods, we use probable estimates using triangular distribution parameter estimates assuming 0,1,1, resulting in an
expected value of 0.67.
100
We use Table 75 from the summary report of the 2017 USDA NASS to get the estimates (Ref. [13]).
https://www.nass.usda.gov/Publications/AgCensus/2017/Full_Report/Volume_1,_Chapter_1_US/usv1.pdf

213

foods. Due to lack of information on how many farms produce covered foods, we assume that all
of the covered shell egg farms produce covered foods.101
Aquaculture Farms Estimates
We derive total number of finfish and shellfish farms from the summary report of the
2018 Census of Aquaculture.102 We consider finfish and shellfish for covered aquaculture
categories. As bivalve molluscan shellfish under the NSSP are not covered in the traceability
rule (75 percent of mollusks), we assume 25 percent of mollusks will be covered by the final
rule. According to the summary report of the 2018 Census of Aquaculture, 2 percent of finfish
and 5 percent of shellfish by point of first sales are direct sales to consumers. We are unable to
separate the aquaculture farms sell directly to consumers from those fully exempt by sales (less
than $25,000) from the summary report of the 2018 Census of Aquaculture. Given that 2 percent
of finfish and 5 percent of shellfish farms total sales fall under less than $100,000 category of
farms, we only include finfish and shellfish farms with annual sales greater than $100,000 to
calculate the total number of covered aquaculture farms (percentage of total sales of finfish and
shellfish less than $100,000 are 2.5 percent and 7 percent, respectively).
We are uncertain of how many aquaculture farms are packers, so we assume that all of
covered aquaculture farms are packers. Similar with the produce and shell egg farms, we
multiply the number of covered aquaculture farms by the FTL share to account for aquaculture

101

Because we are unsure to the exact the number of shell egg farms who manufacture, process, pack, or hold
covered foods, we use probable estimates using triangular distribution parameter estimates assuming 0,1,1, resulting
in an expected value of 0.67.
102
We use Table 9 for value of sales by sales category and Table 21 for direct sales to consumers from the summary
report of the 2018 Census of Aquaculture (Ref. [14]):
https://www.nass.usda.gov/Publications/AgCensus/2017/Online_Resources/Aquaculture/Aqua.pdf

214

farms producing foods on the FTL. Additionally, we assume 100 percent of the covered
aquaculture farms produce covered foods.103
Manufacturers, Wholesalers, Warehouses, and Retailers
For manufacturers, wholesalers, warehouses, and retailers, we use the 2017 Census data
to estimate the total number of firms and establishments by North American Industry
Classification System (NAICS) industry category.104 We first calculate the number of retail food
establishments and restaurants that are fully exempt by their average annual food sales below
$250,000. Then we subtract the number of the exempt entities from the total number of entities
by each NAICS industry category. We use the 2017 Census data to approximate the number of
entities who manufacture, process, pack, or hold covered foods of each NAICS industry category
then divide them by the total entities by NAICS category to get a FTL share of each industry.105
Then we multiply the FTL share by the number of the entities after removing the fully exempt
entities (both FTL and non-FTL food products) to estimate the final number of covered entities
with foods on the FTL. To estimate the total number of covered establishments, we use
triangular distribution of the FTL share.106 We then multiply the ratio of total number of firms to
total number of establishments to estimate the total number of firms covered by the rule.

103
Because we are unsure to the exact the number of aquaculture farms who manufacture, process, pack, or hold
covered foods, we use probable estimates using triangular distribution parameter estimates assuming 0,1,1, resulting
in an expected value of 0.67.
104
As the Census data only cover primary NAICS, we may underestimate the total number of covered entities by not
counting non-primary NAICS.
105
The share is based on the ratio of FTL NAPCS establishments sum over total NAICS establishments. The share
of the FTL food products ranges from 0 (non-FTL foods) to 1 (100 percent of FTL foods). When the share of the
FTL foods related to NAPCS code is larger than total per NAICS, we truncated to 1. The shares of the FTL using
2017 Census data may be underestimated because we only include the numbers that were available from the 2017
Census data.
106
Because we are unsure of the exact number of establishments who manufacture, process, pack, or hold covered
foods, we use probable estimates using triangular distribution to estimate the FTL shares. We assume a minimum of
0, a maximum of 1, and a mode equal to the FTL share of each NAICS industry category (0; non-FTL foods to 1;
100 percent of FTL foods), resulting in expected values between 0 and 0.67.

215

Difference Between Proposed and Final RIA Estimates
The overall estimates of the number of covered entities in this RIA are lower than our
previous Option 2 estimates (98,272 fewer firms and 82,325 fewer establishments, see Table
D.1). For example, compared to our previous Option 2 estimates, the estimated numbers of
covered produce farms, egg farms, and aquaculture farms in this RIA are lower by 11,152 fewer
firms and 11,151 fewer establishments. Coverage estimates in this RIA are higher than our
previous Option 1 estimates (135,581 more firms and 152,270 more establishments, see Table
D.1).
The first reason for this difference is that for the final analysis, we use different data
sources than those used in our preliminary analysis. We use 2017 USDA NASS data in this RIA
while we previously used 2012 USDA NASS data. For egg and aquaculture farms, we derive
estimates from the 2017 USDA NASS summary report and use the 2017 USDA NASS data for
the produce farms in this RIA. We also use 2017 SUSB data from the Census instead of the
2012 SUSB data.
Other reasons for this difference are explained by changes to the proposed requirements.
While the proposed rule exemptions for retail food establishments and restaurants were based on
the threshold of 10 FTEs, the final rule exempts retail food establishments and restaurants with
annual sales below $250,000. The final rule also adds several exemptions, such as the exemption
for molluscan shellfish. Furthermore, the final rule specifies that a multi-ingredient food is
covered only if the FTL ingredient it contains is in the same form in which it appears on the FTL
(e.g., frozen pizza with a spinach topping is not covered because only fresh spinach is on the
FTL, not frozen spinach). Consequently, some businesses under certain NAICS codes that were

216

previously included in PRIA estimates are now excluded in our final analysis (NAICS 311412,
311421-311423, 311520, 311824, 311942, 445292, and 454110).
Additionally, we calculate the share of the covered entities handling foods on the FTL by
counting the number of entities that we believe manufacture, process, pack, or hold covered
foods by each NAICS industry category. In this analysis, we added the number of these entities
by NAICS category then divide the sum by the total number of establishments by NAICS
category to get the ratio of FTL establishments. Hence, the FTL shares in this analysis are lower
than the preliminary analysis. Table D.1 below contains a summary of the changes in the
estimates of the preliminary and final analyses by industry.
Table D.1. Number of Affected Entities of the Proposed and Final Rule by Industry Sector
Preliminary RIA
(Option 1)

Type
Farms /Aquaculture
/ Growers
Manufacturers /
Processors / Packers
Wholesalers /
Distributors
Warehouse and
Storage
Retail Food
Establishments
and Restaurants
Total

Preliminary RIA
(Option 2)

Final RIA

Number
of Firms

Number of
Establishm
ents

Number
of Firms

Number of
Establishm
ents

Number
of Firms

Number of
Establishme
nts

22,912

22,947

22,912

22,947

11,760

11,796

10,623

11,557

10,623

11,557

7,991

8,650

18,686

24,224

18,686

24,224

12,007

15,101

3,519

6,880

3,519

6,880

2,504

5,176

132,551

266,246

366,404

500,841

289,609

443,401

188,291

331,854

422,144

566,449

323,872

484,124

E. Changes to Cost Estimation from the Preliminary Analysis
To inform our analysis of the costs of this final rule, ERG completed an elicitation of
industry experts in December 2021 and January 2022 (Ref. [4]). Experts provided both
217

qualitative and quantitative input based on the proposed version of the rule, with additional brief
definitions of some new CTEs in their draft-final state at the time of the traceability costs
elicitation. Input included describing anticipated cost-incurring compliance activities and
expenditures, estimating variables related to cost calculations, and further commenting on factors
likely to influence costs of the rule.
Among changes to one-time costs, our estimates of costs to read and understand the rule,
in section II.F.2 “Costs of Reading and Understanding the Rule,” now account for three
employees reading the rule at large firms, versus only one employee regardless of firm size in the
preliminary analysis. We consider reading costs in this section to be separate from the costs to
identify FTL products and plan for compliance, which we estimate in section II.F.5.a
“Traceability Plan.” Additionally, we note that the final rule and preamble are about four times
as long in total number of words as the proposed versions, nearly quadrupling reading time per
affected employee.
Unlike in the preliminary analysis, in which we only considered one-time capital costs,
we now also consider recurring capital costs for those cases where capital investments made
towards compliance with the rule result in higher operation and maintenance expenses than
covered entities would otherwise face (section II.F.3 “Costs of Capital Investment”). Whereas
we previously based estimates per entity on equipment price information extrapolated from
literature, we now base cost per entity on expert estimates elicited by ERG (Ref. [4]). We now
also explicitly account for the proportion of establishments requiring additional capital using
estimates provided by the expert panel.
Unlike in the preliminary analysis, in which we considered only one-time training costs,
we now also consider recurring training costs for those cases where new training is more time

218

consuming than what covered entities would otherwise have implemented as a refresher for
continuing employees and because of turnover (section II.F.4 “Costs of Training”). Whereas we
previously proxied for the cost of a training program based on pricing offered for a single online
training course, we now use expert estimates elicited by ERG. For estimating employee labor
costs, we now use estimates by the expert panel to inform numbers of employees and training
hours in place of our previous assumptions. We now also explicitly account for the proportion of
establishments requiring additional training using estimates by the expert panel.
The final rule replaces the proposed requirement for traceability program records with the
requirement for a traceability plan. We previously estimated a recurring cost of this provision in
addition to a one-time cost because the proposed program records required operational
information that we expected to require frequent updates. In particular, these requirements
included the traceability product identifiers and product descriptions of each FTL food shipped
by an establishment. Since the traceability plan instead requires more general descriptions of
procedures, which we do not expect will change often at a typical establishment, we consider the
traceability plan to impose a one-time cost (section II.F.5.a “Traceability Plan”) and the routine
as needed updates to take de minimis time. While the final rule requires firms to update their
traceability plans “as needed,” possible future updates to the FTL, which might require some
firms to identify additional products, will only take effect two years after publication in the
Federal Register. We expect that this delay will allow firms to make necessary updates within the
scope of routine updates to standard operating procedures in the normal course of business.
Additionally, we now incorporate estimates by the expert panel on the number of employees who
will work on planning for traceability and the baseline proportion of covered entities engaging in
traceability practices.

219

The final rule also makes several changes to CTEs that affect our estimates of recurring
recordkeeping costs (section II.F.5 “Costs of Recordkeeping”). The final rule introduces
harvesting, cooling, initial packing, and first land-based receiving as CTEs, while removing
growing and first receipt CTEs and redefining transformation to include events previously
referred to as creation. As a result, we now estimate recordkeeping costs for each of these new
CTEs. Whereas in the preliminary analysis we assumed counts of traceability lots handled per
entity with regard to each CTE, we now base these counts on estimates, by the expert panel, of
the number of FTL lots handled by establishments in different categories of industries (Ref. [4]).
Additionally, the final rule changes each CTE’s corresponding set of KDEs, which we
describe in the subsections under II.F.5 “Costs of Recordkeeping.” While in this analysis we
newly base estimates of recordkeeping time on input from multiple experts, we reconcile expert
input with changes between the final and proposed KDEs, as well as baseline electronic
recordkeeping and improved efficiency from expected capital investments (described in section
II.F.3 “Costs of Capital Investment”).
Notably, as covered entities under the final rule do not assign traceability lot codes prior
to initial packing of raw agricultural commodities or first land-based receiving of food obtained
from a fishing vessel, we no longer estimate costs on a lot-level basis prior to these steps. We
now also exclude entities upstream of these new CTEs in the supply chain from our counts of
entities affected by shipping and receiving requirements, which do not apply prior to initial
packing and first land-based receiving. However, we account for information that the final rule
requires the newly defined harvesters and coolers to provide.

220

F. Case Studies Considered by Experts in Estimating Reduced Costs from Avoiding Overly
Broad Recalls Following an FDA Issued Public Health Advisory.
We present case studies of overly broad recalls used to help experts in their estimates
considering how more-targeted recalls might affect some of the identified case study costs. The
following case studies include 2008 tomato recall and the more recent 2018 and 2019 leafy green
recalls.
a) Multistate Outbreak of Salmonella Saintpaul Infections Linked to Raw Produce
This 2008 outbreak caused almost 1,500 illnesses and was initially attributed
to tomatoes, leading to a recall. The warning for the 2008 tomato recall covered all
red Roma, red plum and red round tomatoes and any other products containing these
raw, red tomatoes.107 Consumers began to avoid not only the tomatoes included in the
warning, but also all other varietals of tomatoes as well. Even though the FDA
explicitly stated that some varieties were safe, many stores removed them from their
shelves and customers began ordering their customary dishes at restaurants without
tomatoes.108 Of the sixteen traceback investigations initiated by FDA, four were
discontinued due to lack of records and the remaining 12 tracebacks resulted in no
common growing region, grower, or supplier. Challenges to the tracebacks included
lack of standardized product documentation throughout the supply chain, difficulty in
linking incoming to outgoing shipments, repacking of product, and comingling of
tomatoes. Standardized traceability documentation and linking of shipments
throughout the distribution chain would have decreased the time to complete the

107

Press Release, FDA, FDA Warns Consumers Nationwide Not to Eat Certain Types of Raw Red Tomatoes (June
7, 2008) [hereinafter FDA Recall June 7].
108
Salmonella scare hold the tomato, Chicago Tribune (Illinois), June 10, 2008.
https://www.chicagotribune.com/news/ct-xpm-2008-06-10-0806090798-story.html

221

tracebacks and provided timely information that there was no common source of
tomatoes (Ref. [49]). Ultimately the source of the outbreak was later attributed to
jalapeño and serrano peppers produced in Mexico (Ref. [50]). The investigation
showed that jalapeño peppers were a major source of contamination and that serrano
peppers also were a source.109 Although the recall of red tomatoes and tomato
products was later lifted, the negative impact on red tomatoes and tomato products
significantly affected their sales volumes at the time. In fact, costs to the Florida
tomato industry alone were estimated to be more than $100 million. In Georgia, the
costs to the tomato industry came close to $14 million.110
b) Outbreak Investigation of E. coli: Romaine (November 2018- February 2019)
On November 20, 2018, FDA issued a public advisory in response to a multistate outbreak of E. coli O157:H7 linked to romaine lettuce and advised against eating
any romaine lettuce on the market at that time. As a result, producers and distributors
voluntarily withdrew the product from the market. FDA performed a traceback
investigation to determine the source of the romaine lettuce; however, at the time of
the public advisory, FDA did not have enough traceback information to identify the
source of the contamination that would allow conducting a targeted recall. The most
efficient way to ensure keeping contaminated romaine off the market was for industry

109

Multistate Outbreak of Salmonella Saintpaul Infections Linked to Raw Produce (FINAL UPDATE)
Centers for Disease Control and Prevention, National Center for Emerging and Zoonotic Infectious Diseases
(NCEZID), Division of Foodborne, Waterborne, and Environmental Diseases (DFWED) Posted August 28, 2008
https://www.cdc.gov/salmonella/2008/raw-produce-8-28-2008.html
110
Reginald L. Brown testifying before the House Committee on Energy and Commerce, Subcommittee on
Oversight and Investigations, The Recent Salmonella Outbreak: Lessons Learned and Consequences to Industry and
Public Health, 110th Cong. 2nd sess., July 31, 2008, http://energycommerce.house.gov/cmte_mtgs/110-oihrg.073108. Brown-Testimony.pdf; “FDA tomato alert costly to Georgia producers”
Southeast Farm Press, September 4, 2008, http://southeastfarmpress.com/vegetables-tobacco/salmonella-warning0905/index.html.

222

to voluntarily withdraw product from the market, and to withhold distribution of
romaine while FDA and state partners conducted a traceback investigation to
determine whether a common supplier or source of contamination could be identified.
By December 13, 2018, FDA was able to refine the traceback investigation
implicating one farm in Santa Barbara which promptly recalled red leaf lettuce, green
leaf lettuce and cauliflower harvested on November 27, 2018, through November 30,
2018.111 On February 13, 2019, FDA completed its investigation. At the conclusion of
the outbreak, a total of 62 cases (with 25 hospitalizations and no deaths) in 16 states
and Washington DC were associated with this outbreak.112 Better traceback data
would have allowed FDA to identify the implicated farm in Santa Barbara more
quickly, such that a broad market withdrawal of all romaine lettuce might not have
been necessary.
c) Outbreak Investigation of E. coli: Romaine (November 2019- January 15, 2020)
In November 2019, FDA, along with CDC, U.S. Department of Agriculture’s
Food Safety Inspection Service (FSIS) and state health authorities, investigated an
outbreak of 167 illnesses of E.coli O157:H7 associated with salads containing
romaine lettuce. The Maryland Department of Health identified a positive sample of
romaine lettuce used in a chicken Caesar salad kit. The contaminated romaine lettuce
was supplied by farms in Salinas, CA. As a result of this positive sample, all chicken
Caesar salad kits containing the positive lot were recalled. Simultaneously, FDA was
investigating two additional outbreaks of E. coli O157:H7 associated with romaine

111

https://www.fda.gov/food/outbreaks-foodborne-illness/outbreak-investigation-e-coli-romaine-november-2018
People Infected with the outbreak strain of E. Coli O157:H7, by date of illness onset.
https://www.cdc.gov/ecoli/2018/o157h7-11-18/epi.htmlhttps://www.cdc.gov/ecoli/2018/o157h7-11-18/epi.html

112

223

lettuce and a chopped salad kit containing romaine lettuce. Based on traceback data
available initially, FDA requested that industry voluntarily withdraw romaine grown
in Salinas from the market and requested that industry withhold distribution of
Salinas romaine for the remainder of the growing season in Salinas. This was a broad
market withdrawal, because a significant portion of the romaine lettuce consumed in
the United States is grown in Salinas; however, due to a lack of more specific
traceback information, this was the most efficient way to ensure that contaminated
romaine was off the market (Ref. [30]).
d) Outbreak Investigation of Salmonella: Shell eggs (May 2010- November
2010). The 2010 shell eggs Salmonella contamination illustrates how conducting a
food recall can be a complex process. According to CDC, the shell eggs outbreak was
first reported in May 2010 and the recall was issued in August 2010. However, the
outbreak continued until October 2010 when all contaminated food vehicles were
identified and recalled (Ref. [19]). Because of the length of time it took to identify the
food vehicle, this outbreak was the largest reported foodborne disease outbreak since
the early 1970s when outbreak surveillance was established (Ref. [20] [21]). The
outbreak resulted in many illnesses and proved to be costly to businesses, something
which could have otherwise been mitigated or avoided with better tracing tools and
standardized records.
At the start of the outbreak investigation, there was a lack of clusters of illness
with an epidemiologic association to shell eggs. The clusters required more
epidemiologic evidence to be obtained on the consumption of eggs and egg-containing
foods in order to have enough information to begin a traceback investigation.

224

Whenever epidemiological investigations fail to clearly implicate possible ingredients
or foods that are causing an outbreak, federal and state authorities often elect to
identify the most likely one or two ingredients or foods in two to three of the clusters
in which the best epidemiological information and case histories are available and
trace those foods to see if there are common suppliers. These are frequently referred to
as “epidemiological tracebacks” and are meant to help inform the epidemiological
investigations (Ref. [51]). This outbreak required an epidemiological traceback to
verify that the food vehicle was in fact shell eggs, and further traceback of additional
clusters to verify the common supplier. Such a process is labor-intensive and time
consuming.
In absence of standardized records, linking shipments through the supply
chain and back to their sources, the time taken to identify implicated foods can be
unacceptably long leading to larger and more costly disease outbreak. In the shell
egg-related salmonella outbreak, some district offices had to return some clusters to
the firm for additional clarification on linking incoming to outgoing product
shipments, which prolonged and complicated the traceback investigation. Also, many
firms did not record this information at all, causing traceback to rely on analysis of
shipment dates alone. This analysis slowed the process for identification of the food
vehicle, common supplier, and affected lots. Ultimately thousands of people were
sickened because of this outbreak. The number of illnesses could have been
meaningfully reduced if a better tracing system had been in place.

225

G. Detailed Calculations Used for Estimating Benefits from Avoiding Overly Broad Recalls
Following an FDA Issued Public Health Advisory

The final step in the series of calculations used in estimating the benefits from reduction
in overly broad recalls, involves assigning parameters to corresponding probability density
functions to characterize the variability inherent in the costs estimates and also for the inherent
uncertainty in the estimates for the number of firms characterized by their respective cost
category. Column L for firms in Table G.1 shows the minimum, central and maximum estimate
for the range of affected firms from Table 20. We characterize variables in Column L as a
lognormal distribution with parameters mean and standard deviation. For Column K, or Cost per
firm we characterize variability from expert cost estimates using a Beta Subjective distribution
which uses parameters, minimum, median, average, and maximum values as shown in Table
G1.113 Column M shows all the intermediate calculations allowing us to estimate a weighted sum
of $862 million as a central estimate of total averted annual costs of overly broad recalls.

Table G1.- Estimate for Averting Costs due to Overly Broad Recalls (by Industry)
Cost per
firm
($1,000)

Firms
(L)
Distrib
ution

(K)

Stand
ard
deviat
ion

Distributi
on

Product
(Cost per
firm x
Number of
firms)
($1,000)
(M)

min

avg

max

lognormal

$68,760

lognormal

$193,090

lognormal

$122,197

Calculati
on

Producer
Low estimate
Most likely
estimate
High
estimate

$8,915
$22,041
$61,098

riskbeta
subj
riskbeta
subj
lognorm
al

(a1) low
(b1)
middle
(c1) high

113

For producer high estimate the median is larger than the mean, and for this reason we characterize it as a
lognormal distribution.

226

Shipper or Distributor
Low estimate
Most likely
estimate
High
estimate

$1,223
$3,231
$7,552

riskbeta
subj
riskbeta
subj
riskbeta
subj

lognormal

$13,732

lognormal

$33,320

lognormal

$8,886

193

311

136

lognormal

$13,969

329

156

lognormal

$240,326

lognormal

$7,650

143

lognormal

$31,359

lognormal

$119,897

lognormal

$8,323

(a2) low
(b2)
middle
(c2) high

Restaurant
Low estimate
Most likely
estimate
High
estimate

$73
$731
$3,057

riskbeta
subj
riskbeta
subj
riskbeta
subj

(a3) low
(b3)
middle
(c3) high

Non-Restaurant Retailer
Low estimate
Most likely
estimate
High
estimate

$354
$2,172
$9,134

riskbeta
subj
riskbeta
subj
riskbeta
subj

Cost of Advisory / Event -Sum, using
‐ Number of firms
incurring low estimate
61
301 545
‐ Number of firms incurring most
likely estimate
81
404 242
‐ Number of firms incurring high
estimate
3
7
11
Central Cost Value
(Sum)
145 711 798
Annual Benefit from avoiding overly broad recalls following FDA advisories
(Central Value)

$127,820
$586,633
$147,055
$861,508

(a4) low
(b4)
middle
(c4) high
A = sum
(a) low
B = sum
(b) middle
C = sum
(c) high
Total =A
+ B+ C

$861,508

227

Table G2. Simulation and Sensitivity Analysis Results for Averted Costs of an Overly
Broad Recall.

228

H. Accounting for English Proficiency and Internet Access
For estimating costs associated with learning about the requirements of this rule, we adjust
time estimates for foreign establishments to account for potential differences in time due to
differences in English language proficiency and internet accessibility for foreign facilities
representing 114 countries with valid registrations who offer FDA regulated food for sale in the
U.S. from FDA’s Food Facility Registration Module. We create a combined multiplier for each
country to simultaneously account for the two adjustment factors. We estimate this combined
multiplier for all foreign establishments as 1.41 hours. This estimate is the average of an estimated
English proficiency multiplier of 1.65 hours and an estimated internet usage multiplier of 1.17
hours. To account for language proficiency differences, we use information from a report titled
“Education First English Proficiency Index” (EF EPI) published in November of 2020.114 This
report ranks countries by the average level of English language skills amongst adults using data
collected via English tests available over the internet.
To account for country differences in internet accessibility, we use 2022 internet user
percentage estimates by country.115 Table H.1 shows establishment counts from countries scaled
according to the country’s English proficiency index (EPI). We estimate that for every hour used
by an establishment with high to very high EPI score, that establishments in countries with low
and very low EPI score would require two to three hours respectively. Similarly, we estimate that
for each hour used by an establishment with high to very high EPI score, that establishments in
countries with a moderate EPI score would require one and a half hours. We estimate 1.65

114

2020 EF English Proficiency Index – Comparing English skills between countries – EF EPI. Ef.com. Retrieved
on 2021-9-1. http://www.ef.edu/epi/.
115
Internet Usage and world populations Statistics Estimates. www.internetworldstats.com. Copyright © 2018,
Miniwatts Marketing Group. All rights reserved worldwide. Last accessed on Jun 11, 2022.

229

proficiency weighted hours as the sum of the product of the establishment percentage (A) and the
hourly equivalent (B) within each proficiency index range.
Table H.1.— Number of Foreign Establishments from FDA’s FFRM in Countries
According to their English Proficiency
English Proficiency
Index
(Range)

Percent
Number of
Establishments
Establishments
(A)

Hourly
Multiplier
(B)

Proficiency
Weighted
Hours
(A X B)

Very High
(63.2 - 100)

18,579

16%

1.00

0.16

High
(58.14 - 63.1)

6,890

1.00

0.06

Moderate
(52.82 - 58.13)

43,254

37%

1.50

0.55

Low
(48.78 - 52.81)

42,497

36%

2.00

0.72

Very Low
(0 - 48.77)

6,294

3.00

0.16

Sum

117,514

100%

1.65

Table H.2 shows the estimated number of establishments from countries scaled according
to their country’s internet access. We estimate that for every hour used by an establishment with
high to very high internet access, establishments with moderate, low, and very low internet usage
would require 1.5, two and three hours respectively. We estimate that for every hour used by an
establishment with high to very high internet usage, that establishment in countries with low and
very low internet usage would require two to three hours respectively. Similarly, we estimate that
for every single hour used by an establishment with high to very high internet usage, that
establishments in countries with moderate internet accessibility would require one and a half hours.
We estimate 1.17 internet weighted hours as the sum of the product of the establishment percentage
(A) and the hourly equivalent (B) within each accessibility index range.
230

Table H.2.— Number of Foreign Establishments from FDA’s FFRM in Countries
According to Percent Internet Access
Internet Use
(Scale)

Number of
Establishments

Percent
Establishments
(A)

Hourly Multiplier
(B)

Internet Weighted
Hours (A X B)

Very High (90%100%)

67,715

58%

1.00

0.58

High (71% - 89%)

21,196

18%

1.00

0.18

Moderate (59% 70%)

24,154

21%

1.50

0.31

Low
(57% - 58%)

1,200

2.00

0.02

Very Low
(0 - 56%)

3,249

3.00

0.08

Sum

117,514

100%

1.17

The average of both weighted sums of 1.65 hours to account for differences in English
proficiency and of 1.17 hours to account for differences in internet usage give us a single multiplier
estimate of 1.41 (= (1.65 hours +1.17 hours)/2) for foreign establishments. We take the average
between proficiency and internet multipliers because internet access is positively correlated with
English proficiency and also because high English proficiency alone is not enough to account for
the amount of time that an entity would require to learn about the rule.116 Even entities from
countries with high English proficiency would rely on using the internet insofar as to only

116

Source: https://www.visualcapitalist.com/the-most-used-languages-on-theinternet/#:~:text=English%20is%20by%20far%20the,with%20over%201.13%20billion%20speakers viewed on
September-3-2021. Based on the top 10 million websites by traffic rankings from Alexa.com Source W3Techs,
Ethnologue and the United Nations via Hootsuite.

231

download the rule and print it or e-mail it to a device, whereas an entity in a country with low
internet access would need to spend more time finding internet access in order to download the
rule from the internet.
The following three examples may help to better illustrate the relationship between
proficiency and internet adjusted hours.
1) For an entity in an English speaking country with high internet access we assign an hourly
multiplier of 1 for proficiency (Table H.1, column B) and 1 for internet access (Table H.2,
column B). The amount of time needed by an entity from a highly proficient English
speaking country with high internet access would be the same as a domestic entity,
therefore we estimate that one hour is the average of 1 hour for proficiency and 1 hour for
internet access ((1 hour from proficiency + 1 hour from internet)/2 = 1 domestic hour).
2) For entities in a country with moderate English proficiency assigned with a multiplier of
1.5 (Table H.1, column B) but assigned with the number 1 for high internet hours (Table
H.2 column B) we estimate the average of 1.25 hours would be needed for every domestic
hour spent learning and reading the rule ((1.5 hours +1 hour)/2) = 1.25).
3) Entities from countries with very low English proficiency, (assigned a multiplier of 3) and
with low internet (assigned a multiplier of 2.5) would require 2.5 more hours for every
domestic hour used in learning and reading the rule ((3 hours + 2 hours/2) = 2.5).

232

File Type	application/pdf
File Title	Microsoft Word - FDA - Food Traceability FRIA Final.docx
Author	DHC
File Modified	2022-11-20
File Created	2022-11-20