1995 Report

multifamily 1995.pdf

Study of Capital Needs in the Public Housing Program

1995 Report

OMB: 2577-0260

Document [pdf]

Download: pdf | pdf

Status of HUDInsured (or Held)
Multifamily Rental
Housing in 1995
HC-5964
Task Order #7

Final Report

December 10, 1997

Prepared for
US Department of Housing
and Urban Development
451 Seventh Street, SW
Washington, DC 20410-3000

Prepared by
Meryl Finkel
Donna DeMarco
Deborah Morse
Sandra Nolden
Karen Rich

Internal Review

Project Director

Technical Reviewer

Management Reviewer

December 12, 1997

Mr. Laurent V. Hodes
U.S. Department of Housing and Urban Development
451 Seventh Street SW, Room 8154
Washington, DC 20410

Dear Larry,

Enclosed is the revised Final Report for Task Order 7 under Contract DU100C000005964,
Status of the HUD-Insured (or Held) Multifamily Stock in 1995, in fulfillment of the
requirements under Task 7 of the Task Order. The Final Report includes two volumes: the
main report including four appendixes, and a volume with four supplementary analyses. In
addition to the Final Report, memoranda on the comparisons of Abt’s data with E&Y’s data
were submitted, as were the model and model documentation.
We have addressed the comments to the October draft as best as possible. Please let me
know if you have any further questions or comments.
I look forward to hearing from you.

Sincerely,

Meryl Finkel
Project Director

TABLE OF CONTENTS
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

1.0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1
1.2

Attributes of the HUD-Insured (or Held) Multifamily Housing Stock . . . . . . 3
Neighborhoods of HUD-Insured (or Held) Properties . . . . . . . . . . . . . . . . . 7

2.0 Tenant Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1

Tenants in Assisted Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.0 Physical Condition of the Stock of Multifamily Housing . . . . . . . . . . . . . 24
3.1
3.2

Current Backlog of Physical Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Projected Future Physical Needs - Physical Needs Accrual Costs . . . . . 36

4.0 Financial Condition of HUD-Insured and Held Properties . . . . . . . . . . . 39
4.1
4.2
4.3

Annual Net Cash Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Resources Available to Cover the Current Backlog of Physical Needs and
Annual Accrual of Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Section 8 Assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

5.0 Measures that Combine Physical and Financial Condition . . . . . . . . . . 64

6.0 Market Position of Insured Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.1
6.2
6.3

Property Finances Relative to Neighborhood and Industry Norms . . . . . . 78
Market Scenario Distress Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Market Potential of Insured Properties . . . . . . . . . . . . . . . . . . . . . . . . . . 90

APPENDIX A - Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
APPENDIX B - Data Collection Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
APPENDIX C - System for Estimating Physical Needs
Backlog and Accrual Costs From Inspections . . . . . . . . . . . . . . . . . . 121
APPENDIX D - Characteristics of Units in the Stock . . . . . . . . . . . . . . . . . . . . . . . . . . 156

Abt Associates Inc.

Table of Contents

1995 Status of the HUD-Insured (or Held)
Multifamily Stock
Executive Summary
Overview
Two of the Department of Housing and Urban Development’s (HUD’s) main responsibilities
are to enhance the availability and affordability of housing to lower-income households and to
ensure the fiscal soundness of the Federal Housing Administration (FHA) insurance fund.
These responsibilities are closely linked in that a major portion of HUD-assisted multifamily
properties has mortgages insured or held by the FHA. (There are also many properties
receiving subsidies through Section 8, Section 236, or Rental Assistance Payments (RAP)
whose mortgages are not FHA-insured or held.) This close link between HUD assistance and
insurance means that any reduction in assistance could increase claims on the insurance fund.
HUD’s involvement in FHA-insured multifamily housing frames the policy context for this
study of the HUD-Insured multifamily housing stock.
•

HUD provides mortgage insurance for this stock of over 12,000 properties. Over
$30 billion of the original principal balance on these mortgages is still outstanding,
representing a substantial contingent Federal liability.

•

HUD is responsible for providing various forms of project-based assistance to over
10,000 of these properties, housing over one million families.

•

Most of the over 1.4 million families living in the HUD-insured stock have low
income.

•

Many of the long-run, project-based Section 8 rental assistance contracts have
recently been renewed, or will come up for renewal over the next five years.

•

Current gross rents on a large portion of the assisted stock are above estimated
market rents for comparable properties in their local areas. These properties are
the focus of HUD’s portfolio reengineering efforts that are aimed at bringing rents
in line with market rent levels.

This study describes the current (1995) physical, financial, and market condition of these
properties and changes in condition that have occurred since 1989. The study universe
includes nearly all properties with mortgages insured as of 1989 that were still insured (or
Abt Associates Inc.

Executive Summary

held) in 1995. It does not include any properties that were insured after 1989, nor does it
include properties outside the contiguous states, properties in remote rural locations, nonresidential, non-rental, or single family properties, or HUD-acquired properties. The universe
includes 10,019 assisted properties (none were insured after 1989), and 2,224 unassisted
properties. To simplify presentation, the report discusses findings in terms of three categories
of insured multifamily properties:
Unassisted properties are insured under any HUD mortgage insurance program and
receive no HUD subsidy (no rental assistance and no mortgage interest subsidy).
Most unassisted properties have mortgages insured under the Section 221(d)(4)
program. This category includes 2,224 properties housing about 354,000 families.
Older assisted properties are insured under any HUD mortgage insurance program
and receive either mortgage interest subsidies (under Section 236 or 221(d)(3) Below
Market Interest Rate insurance programs) or rental assistance under the Section 8
Loan Management Set Aside, Rent Supplement, Rental Assistance Payment, Section 8
Property Disposition, or Preservation programs. This category includes 5,943
properties housing about 686,000 families.
Newer assisted properties are insured under any HUD mortgage insurance program
and receive rental assistance under one of the following Section 8 programs: New
Construction, Substantial Rehabilitation, or Moderate Rehabilitation. Most newer
assisted properties have mortgages insured under the Section 221(d)(4) program. This
category includes 4,076 properties housing about 365,000 families.
This study is based on physical inspections, market rent assessments, and secondary data
collected for a representative national sample of 621 multifamily properties. To facilitate
comparing physical needs and financial variables across properties having different numbers of
units and different sized units, all property costs were expressed per “2-bedroom equivalent”
unit. To facilitate comparing costs over time, all costs were expressed in 1995 dollars.

Study Findings
The condition of the HUD-insured stock is complex and has many dimensions. The study’s
principal findings presented below, first focus on single dimensions including: (1)
characteristics of tenants in assisted properties; (2) properties’ physical condition; and (3)
properties’ financial condition. The next set of findings focus on measures that combine
aspects of physical and financial condition, including: (4) properties’ ability to cover current
physical needs and future annual accruals of needs using internal resources; and (5) property
risk profiles based on their backlogs of physical needs and annual net cash flow. The last set
of findings, (6) shows how current property rents compare with local market rents and how
Abt Associates Inc.

Executive Summary

cash flow in assisted properties would change if rents and operating costs reverted to market
levels.

Characteristics of Tenants in Assisted Properties

The assisted portion of the HUD-insured (or held) stock serves a very low income population,
including many families with elderly and disabled members. Compared with older assisted
properties, newer assisted properties tended to have a greater portion of units assisted through
Section 8, lower tenant incomes, and a higher portion of elderly households.
•

Income Distribution: Nearly all (95 percent) residents in newer assisted
properties had very low incomes (defined as incomes below 50 percent of the local
median for their household size) as did 67 percent of residents in older assisted
properties. All the remaining residents in newer assisted properties, and most of
the remaining residents in older assisted properties had low incomes (defined as
incomes below 80 percent of the local median for their household size).
Annual income of nearly three-fourths of the households in newer assisted
properties and two-thirds of households in older assisted properties was below
$10,000.

•

Race and Ethnicity: Newer assisted and older assisted properties had similar
racial and ethnic compositions. In the average property, most (58 percent)
residents were white, while 37 percent were black. On average, 11 percent of
residents were Hispanic, regardless of race.

•

Other Demographic Characteristics: A higher proportion of households in newer
assisted properties were classified as elderly -- 40 percent compared with 28
percent in older assisted properties. In both types of properties about 11 percent
of households were classified as handicapped. Consistent with the high
concentration of elderly households in newer assisted properties, nearly half the
households in these properties (49 percent) included only 1 person.

Physical Condition-- Backlog of Physical Needs.

Each property was inspected to assess its total backlog of physical needs, which was defined
as the cost of repairs and replacements, beyond ordinary maintenance, required to restore all
property systems to original working condition. The physical needs backlog of the stock has
increased substantially since 1989. In 1989, we found that many properties did not have
sufficient funds to correct the full backlog of physical needs that existed at that time.
Abt Associates Inc.

Executive Summary

iii

Properties were also not putting sufficient funds into their reserve for replacement accounts to
cover annual accrual of future needs. Thus, while it is not surprising that the backlog of
physical needs has increased between 1989 and 1995, it is nevertheless cause for concern.
•

The mean backlog of physical needs across the whole stock of insured (and held)
properties was $3,236 per unit, with a median of $1,452. The total estimated
backlog for the stock was $4.17 billion, $3.5 billion of which was in assisted
properties.

•

The mean backlog was lowest in unassisted properties ($1,427 per unit) and
highest in older assisted properties ($3,929 per unit). The mean in newer assisted
properties, $3,214, was closer to that of older assisted properties.

•

Consistent with their lower average backlogs, nearly two thirds of unassisted
properties had backlogs of physical needs within the “normal” range of under
$1,500 per unit. This range is considered normal because, on average, a property
accrues about $1,500 per unit in repairs and replacements beyond ordinary
maintenance each year, so that a backlog of less than this amount indicates little
carryover from prior years. In contrast, only 42 percent of older assisted properties
and 55 percent of newer assisted properties had backlogs in this range.
At the other extreme, 30 percent of properties had serious physical backlogs of
over $3,000 per unit -- twice the normal annual accrual of repairs and
replacements. Older assisted properties were most likely to have serious backlogs
of physical needs. Forty-one percent of older assisted properties had backlogs of
over $3,000 compared with 25 percent of newer assisted and only 11 percent of
unassisted properties.

•

Physical needs backlogs have increased between 1989 and 1995. Even after
controlling for inflation, the mean backlog rose by 50 percent in unassisted
properties, by 40 percent in older assisted properties, and by 162 percent in newer
assisted properties. Newer assisted properties had extremely low backlogs in
1989. It is not clear to what extent the increase in backlogs of newer assisted
properties reflects their aging, and need for first time replacement of long lived
systems such as roofs or boilers. It will be important for HUD to monitor their
backlogs to be determine whether the backlog continues to increase as rapidly.

•

The physical deterioration of newer assisted properties is also evident in the
distribution of properties. In 1989, nearly three quarters of the newer assisted
stock had backlogs below $1,500 (in 1995 dollars). By 1995, only slightly more
than half did. At the other extreme, in 1989, only 10 percent of newer assisted
properties had backlogs greater than $3,000 (in 1995 dollars), while by 1995, one
quarter of the newer assisted stock did.

Abt Associates Inc.

Executive Summary

While in the short term, backlogs of physical needs may not always impinge directly on
tenants or on property viability (e.g., a property may continue for years with a heavily patched
old roof), continued deterioration could affect both tenants and property viability.

Financial Condition -- Annual Net Cash Flow

The study computed each property’s annual net cash flow per unit, defined as total annual
revenue (primarily from tenant paid rents and from subsidies) less annual expenses to cover
operations and maintenance, mortgage payments, and deposits to the reserve for replacement
account.
•

•

The stock’s financial condition of the stock has improved since 1989.
-

In 1995, annual net cash flow averaged $593 per unit, with a median of $388.
Three quarters of all insured (or held) properties had positive annual net cash
flows.

Across the full stock, average annual net cash flow improved or remained
unchanged from 1989 to 1995.

Unassisted properties experienced the largest improvement in annual net cash
flow, which increased from a mean of $158 to $487, all in 1995 dollars.
Furthermore, in 1989, 44 percent of unassisted properties had negative annual net
cash flows, while by 1995 only 25 percent did.
-

•

Improvements in cash flow in unassisted properties resulted from increases in
revenues that more than offset increases in expenses. Revenues increased by 6
percent as a result of substantial decreases in vacancy losses and small
increases in tenant paid rents. Expenses in unassisted properties increased by
only 2 percent on average, resulting from an increase in operating and
maintenance expenses that was nearly offset by a decrease in the real value of
mortgage payments. (Mortgage payments generally remained constant in
nominal dollars.)

Older assisted properties had the weakest annual net cash flows in both 1989 and
1995. While average annual net cash flow in older assisted properties increased by
6 percent from 1989 to 1995 (to $283 in constant 1995 dollars), the portion of
properties with positive net cash flow declined from 59 percent to only 35 percent.
- Because many older assisted properties receive budget-based rents, the
increase in average revenues of 5 percent was just slightly higher than the

Abt Associates Inc.

Executive Summary

increase in expenses. Expenses increased as a result of increases in deposits to
the reserve for replacement and in operating and maintenance expenses that
more than offset the decrease in real debt service payments.
•

•

Newer assisted properties had the strongest annual net cash flows in both 1989
and 1995, with an average of $1,105 per unit and 87 percent having positive cash
flows.
-

The proportion of newer assisted properties with positive annual net cash flow
stayed fairly stable between 1989 and 1995 (going from 90 percent to 87
percent), while the mean annual net cash flow per unit increased by 29 percent
during this period. This increase in mean cash flow reflects an increase in the
number of properties with very high annual net cash flows. While in 1989 one
third of all newer assisted properties had annual net cash flows of over $1,000
per unit (expressed in 1995 dollars), by 1995, 44 percent of newer assisted
properties did.

Improvements in cash flow in newer assisted properties resulted from increases
in revenues and decreases in expenses. Revenues (in constant 1995 dollars)
increased on average by 2 percent as a result of the annual adjustment factor
applied to property rents. The percentage increase in revenues in newer
assisted properties was smaller than in older assisted and unassisted properties.
Expenses in newer assisted properties decreased by 1 percent because the
increase in operating and maintenance expenses was more than offset by the
decrease in the real value of mortgage payments (though they remained
constant in nominal dollars).

The components of revenue changed substantially between 1989 and 1995,
particularly in assisted properties. In both older assisted and newer assisted
properties higher assistance payments were needed to offset decreases in tenant
paid rents. It is not clear whether the decreases in tenant paid rents reflect lower
incomes of existing tenants or poorer residents moving into the properties upon
turnover.
-

In older assisted properties tenant paid rents decreased by 10 percent. A 67
percent increase in assistance payments was required to cover this decrease in
tenant paid rents and the increase in expenses.

In newer assisted properties, a 22 percent rise in assistance payments was
required to offset the 26 percent decrease in tenant paid rents.

Abt Associates Inc.

Executive Summary

Properties’ Ability to Cover Current Backlog of Physical Needs and
Future Accruals of Needs Using Internal Resources

The study examined properties’ financial ability to cover physical backlogs and ongoing
accrual of physical needs using internal funds and annual net cash flow. As was the case in
1989, most properties did not have sufficient internal resources to cover their current backlogs
of physical needs, nor were they depositing sufficient funds into their replacement reserve
accounts to cover future needs.
Ability to Cover Current Backlog of Physical Needs. Properties generally have internal
funds that may be used to cover their physical needs backlogs. These funds may be in any or
all of the following accounts: reserve for replacement, other special purpose reserves such as
painting reserves, and in some cases residual receipts accounts, which although not intended
as repair funds, may be used for that purpose.
•

A little over one third (35 percent) of the stock had sufficient resources to cover
physical needs backlogs, including properties no physical needs backlog and others
with backlogs of physical needs less than or equal to available resources.

•

The remaining 65 percent of properties across all assistance categories lacked
sufficient resources to cover their backlog of physical needs. This included 13
percent of properties with no available resources, and 30 another percent with
insufficient resources to cover even one quarter of their backlogs.

•

The problem was most severe in the older assisted properties where only 30
percent had sufficient resources to cover their backlogs of physical needs, and least
severe in the newer assisted properties, where 42 percent had sufficient resources.

•

Ability to cover backlogs has declined since 1989. While in 1989, 45 percent of
could cover their backlogs, by 1995, only 35 percent of the stock could. Driving
this decrease in the ability to cover backlog of physical needs was the increase in
backlogs, rather than a decrease in resources. On average, available resources
increased by over 40 percent, but backlogs increased by over 60 percent. This
result held for each of the three assistance categories.

One measure of ability to cover backlog is the unfunded backlog of physical needs. This is the
total backlog reduced by available resources for properties whose total backlogs exceeds
available resources, and $0 for properties with sufficient resources. A minority of properties
is responsible for a substantial portion of the unfunded backlog.

Abt Associates Inc.

Executive Summary

vii

•

While the mean unfunded backlog was $2,630, (or 81 percent of the total
backlog), the median was only $684. This shows that a small portion of the stock
was responsible for a large portion of the unfunded backlog.

•

Almost a third of the stock (32 percent) had unfunded backlogs exceeding $2,000
per unit.

•

As with most other resource problems, high unfunded backlogs were most
common in older assisted properties.
-

The mean unfunded backlog for older assisted properties was $3,323
compared with $1,134 for unassisted and $2,437 for newer assisted.

Forty-four percent of older assisted properties had over $2,000 of unfunded
backlog of physical needs per unit, compared with 15 percent of unassisted and
25 percent of newer assisted properties.

Ability to Cover Ongoing Annual Accrual of Physical Needs. A property’s physical
needs accruals are estimates of the average annual costs needed to cover repairs and
replacements beyond ordinary maintenance for all systems over each of the next 20 years. A
property’s ability to cover the ongoing accrual of physical need is another important factor in
its long-term viability. Accrual for the stock averaged $1,437 per unit per year, with little
difference across assistance categories. There are two potential sources of funds available to
cover these accrual costs: annual deposits to the reserve for replacement accounts and positive
annual net cash flow. Properties that have positive net cash flow after covering operation and
maintenance, mortgage debt service, and reserve fund deposits could use remaining funds to
cover ongoing accruals.
•

It appears that even if the current backlog of physical needs were addressed, only
about one fourth of insured properties would be able to keep up with their ongoing
accrual of physical needs.

•

The average unfunded annual accrual was $610 per unit, and the median was
$586.

•

Compared with other two groups, older assisted properties contributed more on
average to the reserve for replacement account, but had much lower annual net
cash flow. Eighty-six percent of older assisted properties will not be able to cover
ongoing accruals of physical needs with current resources.

•

Many newer assisted and unassisted properties, too, will not be fully able to cover
ongoing accrual. Fifty-nine percent of newer assisted properties have unfunded

Abt Associates Inc.

Executive Summary

viii

accruals despite their high cash flows, and nearly three quarters of unassisted
properties also lack sufficient resources to cover ongoing accruals.

Property Risk Profile

A second way the study incorporated both physical and financial measures into a single
indicator of property condition was to place properties into four different risk profiles based
on their annual net cash flow and their backlog of physical needs. As compared with the
above measures that looked at the financial capacity of properties to cover their backlogs
and accruals, this indicator focuses on what properties are actually doing. It provides
valuable information by highlighting properties that have significant backlogs of physical needs
and at the same time also have positive annual net cash flow that is not being used to address
these needs.
•

Forty percent of all properties were classified as “minimally risky -- they were both
meeting current expenses and addressing their physical needs backlogs. These
properties had both positive annual net cash flows and “normal” physical needs
backlogs -- less than $1,500, approximating a year’s accrual. About half of
unassisted and newer assisted properties fell into this “minimal risk” category, as
did 30 percent of older assisted properties.

•

Another quarter of the stock (26 percent) was classified as “moderately risky.
These properties either had low positive annual net cash flows of under $500 per
unit along with above “normal” backlogs (over $1,500 per unit), or else they had
negative cash flows but normal backlogs. Seventeen to eighteen percent of
unassisted and newer assisted properties fell into this category, as did 37 percent
of older assisted properties.

•

Fifteen percent of the stock was classified as “high risk” -- they were both not
meeting current expenses and not addressing their backlog of physical needs.
These properties had both negative cash flows and above normal backlogs. Nine
percent of unassisted and newer assisted properties fell into this category, as did
20 percent of older assisted properties.

•

The final group of 18 percent of properties was categorized as “management risk”
-- while these properties appeared to have resources available to address at least a
portion of their backlogs, they were not doing so. They had high positive cash
flows of over $500 and yet high backlogs of over $1,500. Over one quarter (26
percent) of newer assisted properties fell into this category as did 18 percent of
unassisted properties and 13 percent of older assisted properties. While there may
be valid reasons for an owner’s temporarily deferring needed repairs, it will be

Abt Associates Inc.

Executive Summary

important for HUD’s asset managers to assess these properties and, as necessary,
take needed action.

Market Position- Property Rents Relative to Local Market Rents

Property Gross Rent Relative to Unrestricted Market Rent
For each property, the study’s market analysts estimated local market rent for comparable,
unrestricted properties. In identifying comparable properties, the analysts assumed that the
current backlog of physical needs was repaired. Actual property gross potential rents (which
equal tenant paid rents, plus utilities, subsidies, and vacancy losses added back in) were
compared with these unrestricted rent estimates.
•

Overall, the stock was evenly divided between properties with gross rents above or
below estimated market levels.

•

Nearly half the stock had gross rents close to their estimated market rents
(between 75 percent and 120 percent of the properties’ estimated market rent).
As expected, a large majority (75 percent) of unassisted properties had rents in this
range. (Not all unassisted properties had rents within this range for several
potential reasons. The market rent estimates assume the backlog of physical
needs is repaired. In addition, in the conventional market some people get better
“deals” and some worse).

•

Most (78 percent) older assisted properties had rents below their estimated market
level, including 38 percent with rents below 75 percent of their estimated market
level. Most older assisted properties have subsidized mortgages, which required
owners to maintain low, affordable, rents.

•

In contrast, the vast majority (86 percent) of newer assisted properties had rents
above estimated market level, including 40 percent with rents above 140 percent
of their estimated market level. When these properties were constructed, assisted
rents were often set above their market levels as a way to promote housing
development in locations where affordable housing was not being developed.
Rents continued to rise annually based on HUD’s Annual Adjustment Factor,
which was often higher than inflation in expenses. These properties are the focus
of the portfolio reengineering efforts that aim to bring property rents in line with
rents in their surrounding markets.

Abt Associates Inc.

Executive Summary

Market Based Cash Flow Scenario
A modified cash flow measure was developed to assess the impact of a more “market-based”
scenario on property finances. This alternative cash flow measure used estimated market rents
and operating costs instead of actuals, and assumed that deposits to reserve for replacement
accounts equal average annual accruals of capital needs. The full existing HUD-insured
mortgage was assumed to continue. This market-based scenario provides a useful baseline
comparison for the current situation.
•

Under this market-based scenario, only half of the assisted stock would have
positive cash flows.

•

Only 10 percent of newer assisted properties would have positive annual net cash
flows under this scenario, due to the decrease in revenues owners would receive at
market rents.

•

In contrast, over three quarters (78 percent) of older assisted properties would
have positive annual net cash flows at market rents, due to increases in revenues
owners would receive under a market rent scenario.

This scenario highlights the importance of the careful attention that needs to be paid to the
process of adjusting rents and expenses in the portion of the stock that currently has abovemarket rents. Clearly, reductions in rents must be accompanied by reductions in expenses or
else most of these properties would no longer remain financially viable. Similarly, careful
attention would also need to be paid to the process of raising revenues in below-market
properties. Raising revenues by allowing rents to increase to market levels would either
increase tenant rent burdens or would require additional HUD subsidies.

Summary
Overall, between 1989 and 1995 the physical condition of the HUD-insured (or held) stock
has worsened, while its financial condition has improved. Changes were most notable in the
newer assisted and unassisted properties.
For newer assisted properties, mean backlog of physical needs increased by 162 percent, the
proportion with low backlogs decreased from three quarters of the properties to just over half,
and the proportion with high backlogs rose from 10 percent to 25 percent. At the same time,
financial conditions of these properties improved, with average annual net cash flow rising by
nearly 30 percent from $859 per unit to $1,105. Contributing to their strong financial
condition is that fact that at present, the majority of these properties (78 percent) have rents
above the estimated market levels for their local markets. As is clear from the “market-based”
Abt Associates Inc.

Executive Summary

cash flow scenario, reducing rents in these properties to market level will lead to very negative
financial outcomes for these properties. Thus, the current portfolio reengineering efforts are
aimed at reducing rents to market levels, while at the same time adjusting debt service
payments downward so that properties can remain viable at market rent levels.
Older assisted properties remained in the weakest financial and physical condition over this
period. For older assisted properties, mean backlog rose by 50 percent to $3,929 per unit,
and the proportion with high backlogs increased from 34 to 42 percent of properties. Older
assisted properties experienced very slight improvements in their financial condition, with
average annual net cash flow rising by 6 percent from $265 to $281 per unit. These
properties continue to have rents substantially below estimated market levels for their local
markets, in large part due to their subsidized mortgages. Steps to address the financial and
physical condition of older assisted properties will need to be balanced against costs to the
FHA insurance fund should these properties fail.
For unassisted properties, mean backlog of physical needs rose by 40 percent from $960 to
$1,427 between 1989 and 1995. However, the major change was the strong improvement in
financial condition: average annual net cash flow more than tripled (from $158 to $487), and
the proportion with positive cash flows increased from 56 percent to 75 percent.

Abt Associates Inc.

Executive Summary

xii

1.0 Introduction
Two of HUD’s main responsibilities are to enhance the availability and affordability of housing
to lower-income households and to ensure the fiscal soundness of the FHA insurance fund.
These responsibilities are closely linked in that a major portion of HUD-assisted multifamily
properties has mortgages insured or held by the FHA1. This close link between HUD
assistance and insurance means that any reduction in assistance could increase claims on the
insurance fund. HUD’s involvement in FHA-insured multifamily housing frames the policy
context for this study of the HUD-Insured multifamily housing stock.
•

•

HUD is responsible for providing various forms of project-based assistance to over
10,000 of these properties, housing over one million families.

•

Most families living in the HUD-insured stock have low income.

•

Many of the long-run, project-based Section 8 contracts have recently been
renewed, or will come up for renewal over the next five years.

•

Current gross rents on a large portion of the assisted stock are above estimated
market rents for comparable properties in their local areas.

This study includes analysis of several aspects of the HUD-insured multifamily stock,
including its physical and financial condition, its market position and value, and tenant
characteristics. The current study updates an earlier study that collected and analyzed data on
similar aspects of the stock.2
The specific goals of the study are to provide HUD with a series of analyses and supporting
data including:
•

Assessing the current (1995) status of the FHA-insured (or held) multifamily
housing stock, including physical and financial condition, tenancy, and market
context.

There are also many properties receiving subsidies through Section 8, Section 236, or Rental Assistance Payments (RAP) whose mortgages
are not FHA-insured or held.

See Assessment of the HUD-Insured Multifamily Housing Stock. Final Report Volume I, Current Status of HUD-Insured (or Held)
Multifamily Rental Housing, HUD-1412-PD&R, September 1993.

Abt Associates Inc.

1.0 Introduction

•

Providing measures of change in the stock’s physical and financial condition over
the period since 1989 when a similar study was conducted.

•

Assessing the likely future status of the stock under alternative policy scenarios
regarding Section 8 restructuring, preservation strategies, property disposition
strategies, and asset management strategies generally.

This report focuses on the first two objectives, namely describing the status of the stock in
1995, and detailing changes that have occurred since 1989. Following an overview of the
study, the remainder of this chapter provides some basic descriptors of the properties and their
neighborhoods. Characteristics of tenants in the assisted portion of the stock are presented in
Chapter 2. The physical condition of the stock is described in Chapter 3. Chapter 4
describes the financial condition of the stock, and Chapter 5 provides overall measures of
property condition that take into account both the physical and financial condition. Finally,
Chapter 6 describes its market position. Four appendices are attached, describing the
sampling procedures used to estimate the current stock (Appendix A), the data collection
methodology (Appendix B), the system used for estimating the physical needs backlog and
accrual costs (Appendix C), and supplementary tables (Appendix D).
The study universe includes nearly all properties that were insured as of 1989 and were still
insured in 1995. It does not include any properties that were insured after 1989, nor does it
include properties outside the contiguous states, properties in remote rural locations, nonresidential, non-rental or single family properties, or HUD-acquired properties. The universe
includes 10,019 assisted properties (none were insured after 1989), and 2,224 unassisted
properties. (Additional assisted properties have been insured since 1989. These properties
are not covered by this study). As was done in the earlier study, this report discusses findings
in terms of three categories of insured (or held) multifamily properties:
Unassisted properties are insured under any HUD mortgage insurance program and
receive no HUD subsidy (no rental assistance and no mortgage interest subsidy).
Most unassisted properties have mortgages insured under the Section 221(d)(4)
program. This category includes 2,224 properties housing about 354,000 families.
Older assisted properties are insured under any HUD mortgage insurance program
and also receive either mortgage interest subsidies (under Section 236 or 221(d)(3)
Below Market Interest Rate insurance programs) or rental assistance under the
Section 8 Loan Management Set Aside, Rent Supplement, Rental Assistance Payment,
Section 8 Property Disposition, or Preservation programs. This category includes
5,943 properties housing about 686,000 families.

Abt Associates Inc.

1.0 Introduction

Newer assisted properties are insured under any HUD mortgage insurance program
and also receive rental assistance under one of the following Section 8 programs: New
Construction, Substantial Rehabilitation, or Moderate Rehabilitation. Most newer
assisted properties have mortgages insured under the Section 221(d)(4) program. This
category includes 4,076 properties housing about 365,000 families.
The findings are based on a combination of primary and secondary data collected for a
representative national sample of 621 multifamily properties. This sample includes 504
properties that were also included in the 1990 Study and 117 additional properties. Appendix
A describes the study sampling procedures. Data were extracted from existing HUD and
Census computerized data systems whenever possible. These data were supplemented with
primary data collected from on-site physical inspections and a series of telephone surveys
aimed at assessing the unrestricted market rent and value of each property. Appendix B
describes the study’s data collection.

1.1

Attributes of the HUD-Insured Multifamily Housing Stock

Exhibit 1-1 provides descriptors of the stock collected from HUD computerized data
systems.3
•

Assistance Category: Overall, 82 percent of insured properties received some sort
of HUD assistance beyond their mortgage insurance. Fifty-nine percent of the
assisted properties, defined here as “older assisted”, were assisted through Section
236, Section 221(d)(3)BMIR, Section 8 loan management set-aside (LMSA), Rent
Supplement/Rental Assistance Payments (RAP), or property disposition Section 8.
Forty-one percent of assisted properties, defined here as “newer assisted” were
assisted through the Section 8 New Construction or Substantial Rehabilitation
programs.4

Tests were conducted to determine whether differences between unassisted/assisted and older/newer were statistically significant. In all the
tables we denote where significance tests were conducted with an “a”. Where distributions were reported, we conducted tests of specific
ranges. For example, looking at the distribution of property sizes we compared the proportion that were below 50 units, and the proportion
above 200 units. In tests of a two-way variable (e.g. family/elderly occupancy) significance of the tested variable also means significance
of the other option. Differences that were significant at the 95 percent confidence level were noted with “**” and those that were
significant at the 90 percent level were noted with “*”. Variables that were tested, but not found to be different, have an “a” next to the
variable name or range, but no “*” or “**”. The formula used for calculating significance between the mean values between two groups
“x” and “y” was the was t =(Mean x - Meany)/square root [(Sex)2+(SEy)2]. If t>1.645 the difference is significant at the 90 percent
confidence level. If t >1.96 the difference is statistically significant at the 95 percent confidence level.

All tables show the stock total; then unassisted/assisted which total to the stock total; then older assisted/newer assisted which total to the
assisted stock total.

Abt Associates Inc.

1.0 Introduction

Exhibit 1-1
ATTRIBUTES OF THE HUD-INSURED MULTIFAMILY HOUSING STOCK
Total
Characteristic
Sample Properties
Total Properties
Percent of Total Properties
Number of Units
Percent of Total Units

Assisted

Total
621
12,243
100%
1,405,240
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

81
2,224
18%
354,083
25%

540
10,019
82%
1,051,157
75%

364
5,943
59%
686,309
65%b

176
4,076
41%
364,848
35%b

Property Size
<50 Unitsa

17%

50-99 Units

35%

100-199 Units
>=200 Unitsa

6%**

19%

17%

22%

31%

36%

32%

43%

36%

42%

35%

39%

29%

12%

21%*

10%

13%*

Average # of Unitsa

115

159**

Median # of Units

120

105
88

115**

100

Unit Size
<2.25 bra

80%

98%*

76%

73%*

80%

>=2.25br

20%

24%

27%

20%

Average Unit Size (brs)a

1.7

1.6

1.8

1.7

Designated Occupancy Type
Familya

75%

88%**

72%

80%**

61%

Elderly/disabled

25%

12%

28%

20%

39%

Sponsor Type
Non-profit/coopa

18%

4%**

21%

35%**

Limited Dividend

40%

48%

62%

For Profita

42%

90%**

31%

3%**

2%
26%
72%

Production Method
New Construction/ Subrehaba

87%

86%

87%

89%

84%

Existing

13%

14%

13%

11%

16%

Abt Associates Inc.

1.0 Introduction

Exhibit 1-1, continued
Total
Characteristic

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

Building Type
High Risea

26%

28%

25%

18%**

35%

Walk-up

44%

56%

42%

47%

33%

Single-Family Attacheda

31%

16%*

33%

35%

32%

Mortgage Start Year
Pre-1970a

1970-1979

55%

43%

57%

1980 or Latera

40%

52%*

37%

9%**
86%
5%**

0%
15%
85%

* Difference between unassisted/assisted or older/newer assisted significant at the 90% level.
**Difference between unassisted/assisted or older/newer assisted significant at the 95% level.
a
Significance test conducted.
b
Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.
Note:

Sums may not add to 100% due to rounding.

Source:

HUD FOMNS and MIDLIS systems, 1995 Physical Inspections.

•

Property Size: The unassisted stock tended to consist of larger properties, with an
average of 159 units; 21 percent had 200 or more units, and only 6 percent had
fewer than 50 units. Assisted properties had 105 units on average; 10 percent had
200 or more units, and 19 percent had fewer than 50 units. Older assisted
properties tend to be larger than newer assisted properties, but smaller than
unassisted properties. Older assisted properties had 115 units on average
compared with 90 units in the newer assisted portion of the stock on average.

•

Unit Size: The units in unassisted properties tended to be smaller (fewer
bedrooms) than those in assisted properties. Ninety-eight percent of unassisted
properties had an average unit size of under 2.25 bedrooms, with an overall
average unit size of 1.6 bedrooms.5 In contrast only 76 percent of assisted
properties had fewer than 2.25 bedrooms on average, and the average size was 1.8
bedrooms. Older assisted properties had the largest units on average (1.8
bedrooms) and more properties with an average size of at least 2.25 bedrooms (27
percent). Newer assisted properties averaged 1.7 bedrooms per unit, with 80
percent of properties having average unit sizes below 2.25 bedrooms.

For this study we defined properties with at least 2.25 bedrooms on average per unit as being able to house large families.

Abt Associates Inc.

1.0 Introduction

•

Occupancy type: Unassisted properties were more likely to be designated as
“family occupancy” at origination compared with assisted properties (88 percent
versus 72 percent). The newer assisted stock has the highest concentration of
properties designated as elderly/disabled, with 39 percent of properties being at
least partially occupied by elderly or disabled residents. Twenty percent of older
assisted properties were designated as wholly or partially elderly or disabled.6

•

Sponsor type: Nearly all of the unassisted (96 percent) and newer assisted (98
percent) properties had profit-motivated or limited-dividend owners. In contrast,
35 percent of the older assisted stock was owned by non-profit entities.

•

Production Method: The predominant production method across all assistance
categories was new construction or substantial rehabilitation (87 percent). The
remaining 13 percent were insured as part of the purchase of an existing property.

•

Building Type: In both assisted and unassisted properties the predominant
building type was walk-up. For fifty-six percent of unassisted properties and 42
percent of assisted properties the predominant building type was walk-ups.
Newer assisted properties had the highest concentration of high-rise buildings (35
percent).

•

Mortgage Start Year: Most of the HUD-insured stock was insured in 1970 or
later. The mortgages of the unassisted properties are roughly equally spread
across the 1970s and 1980s. The vast majority (86 percent) of older assisted
properties were insured between 1970 and 1979, while the mortgages of the newer
assisted properties—those insured in conjunction with property-based Section 8
New Construction or Substantial Rehabilitation assistance—date primarily from
1980 onward (85 percent).

The figures regarding designated occupancy type at origination are from HUD’s MIDLIS system. They differ from the actual tenant
characteristics reported in TRACS which was the source for Exhibit 2-1.

Abt Associates Inc.

1.0 Introduction

1.2

Neighborhoods of HUD-Insured (or Held) Properties)

Aside from the actual physical property itself, the local neighborhood plays an important role
in the overall quality of life of the residents. This section describes several dimensions of the
neighborhoods where the HUD-insured multi-family properties are located:
•
•
•
•
•

Characteristics of the neighborhoods
Conditions of the neighborhoods
Neighborhood demographics
Changes in the neighborhoods
Vacancies in the neighborhoods

Neighborhood Characteristics
Data used to describe neighborhood characteristics were obtained from two sources: an
Inspection Windshield Survey and the U.S. Census. The same inspectors who inspected the
sample properties also rated the property neighborhoods using the Inspector Windshield
Survey. Inspectors were instructed to drive around the neighborhood of the inspected
property and describe the neighborhood on several measures such as land use, type of
residential structure, age of most of the residential structures, and type of construction.
Central city status was obtained from U.S. Census files on Central City tracts and MSA status.
The types of neighborhoods where insured properties were located were similar across
assistance categories (Exhibit 1-2). The exhibit presents the following information on insured
properties neighborhood characteristics:
•

Land Use: Across all assistance categories, neighborhoods were primarily
residential. At least 66 percent of the land was used for residential purposes.

•

Residential Structure Type: The predominant structure type of residential
properties in neighborhoods of HUD-Insured properties was single-family homes
(49 percent on average). On average, twenty-two percent of residential structures
were large multi-family structures, making them the second most common type of
building found in the neighborhoods.

•

Construction Type: The construction type was also similar across assistance
categories, with wood structures being the dominant building type (46 percent).

Abt Associates Inc.

1.0 Introduction

•

Residential Property Age: Unassisted properties tended to be located in
neighborhoods with newer buildings compared with the assisted groups. On
average, 70 percent of properties in the neighborhoods of unassisted properties
were built after 1961, compared with 50 percent in the neighborhoods of assisted
properties. In addition, one quarter of the properties in the neighborhoods of
assisted properties were built prior to 1945 compared with 11 percent in
neighborhoods of unassisted properties.

•

Central City Status: Overall most of the properties (90 percent) were located in
metropolitan statistical areas (MSAs). Unassisted properties were more likely to
be located in non-central city portions of MSAs (59 percent) compared with
assisted properties (46 percent). Unassisted properties were less likely to be
located in non-MSA locations (5 percent) compared with newer assisted
properties (14 percent) and older assisted properties (9 percent).7

The distribution of properties by MSA/non-MSA status understates the true universe of insured properties in non-MSA areas because the
population for this study was properties in MSAs and in adjacent non-MSA counties.

Abt Associates Inc.

1.0 Introduction

Exhibit 1-2
NEIGHBORHOOD CHARACTERISTICS FOR HUD-INSURED PROPERTIES
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Land Use
a

Residential

66%

69%

66%

Commercial

21%

22%

21%

19%

22%

Industrial

Institutional

Other

Residential Structure Type
a

Single-Family Detached

49%

47%

49%

50%

Garden/Row/Townhouse

14%

16%

14%

16%

12%

Multifamily 2-4 Units

Multifamily 5-10 Units

Multifamily > 11 Units

22%

26%

21%

19%

23%

Construction Type
Wood Frame

46%

43%

47%

48%

44%

Masonry

34%

30%

34%

32%

38%

Mixed

20%

27%

19%

20%

18%

Residential Structure Age
Pre-1945
1946-1960
a

1961- Present

23%

11%

25%

23%

28%

24%

19%

25%

53%

70%**

50%

52%

47%

Central City Status
a

MSA - Central City

42%

36%

43%

46%*

38%

MSA - Not Central City

48%

59%

46%

45%

47%

Non-MSAa

10%

5%*

11%

9%*

14%

Sums may not add to 100% due to rounding.

Source:

Inspector Windshield Survey and Special Census Tabulation on Central City Status.

Abt Associates Inc.

1.0 Introduction

Neighborhood Conditions
The condition of the neighborhoods for HUD-insured stock are summarized in Exhibit 1-3.
Data on neighborhood conditions were compiled from the Inspector Windshield Survey and
from the Market Valuation Summaries. For the Market Valuation Summary, market analysts
obtained a description of the neighborhood economy from discussions with local real estate
professionals. For the Inspector Windshield Survey, inspectors rated each neighborhood on
several dimensions such as condition of streets and curbs, street maintenance, owner
housekeeping, and general condition of housing.
•

Unassisted properties were much more likely to be located in neighborhoods that
the study’s market analysts described as having an economy that was “high” (43
percent). Only 11 percent of neighborhoods of assisted properties reportedly had
a “high” economy. Similarly the analysts were far less likely to describe
neighborhoods of unassisted properties as having a “depressed economy” (2
percent) compared with neighborhoods of assisted properties (20 percent).

•

Across all the dimensions, the inspectors rated the neighborhoods of the unassisted
properties as “better” than the neighborhoods of the assisted properties. Across
most quality characteristics, the inspectors rated about ninety percent of the
neighborhoods of unassisted properties as “good” or “excellent”, whereas
approximately one fourth of assisted properties neighborhoods were rated as “fair”
or “poor”.

•

More than half of the neighborhoods of unassisted properties were considered
“better quality” or “much better quality” than other residential neighborhoods in
the local housing market. Assisted properties were more likely to be in
neighborhoods considered of “poorer quality” or “much poorer quality” than other
residential neighborhoods in the local housing market (38 percent).

There were no reported differences in the conditions of newer and older assisted property
neighborhoods.

Abt Associates Inc.

1.0 Introduction

Exhibit 1-3
NEIGHBORHOOD CONDITIONS FOR HUD-INSURED PROPERTIES
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

20%

21%

18%

Description of Neighborhood Economy
Depresseda

17%

Average

38%

35%

39%

40%

Higha

17%

43%**

11%

Mixed

28%

20%

30%

29%

31%

2%**

General Condition of Housing
Sound Conditiona

68%

83%**

64%

65%

Minor Deterioration

24%

14%

26%

27%

26%

Major Deterioration

Dilapidated/Abandoneda

1%**

Condition of Streets/Curbs
Excellent/Gooda

80%

93%**

77%

76%

78%

Fair/Poor

20%

23%

24%

22%

Street Maintenance
Excellent/Gooda

80%

96%**

77%

78%

76%

Fair/Poor

20%

23%

22%

24%

Owner Housekeeping
Excellent/Gooda

76%

93%**

72%

71%

74%

Fair/Poor

24%

28%

29%

26%

Quality as Residential Neighborhood
Excellent/Gooda

67%

90%**

62%

63%

61%

Fair/Poor

33%

10%

38%

37%

39%

Comparison of Neighborhood with Other Residential Areas in Local Housing Market
Better or Much Better Qualitya

24%

51%**

19%

17%

20%

About Average Quality

43%

38%

43%

44%

43%

Poor or Much Poorer Qualitya

33%

11%**

38%

39%

37%

Sums may not add to 100% due to rounding.

Abt Associates Inc.

1.0 Introduction

Source:

Inspector Windshield Survey and Market Valuation Summaries.

Demographic Characteristics
Demographic data for the neighborhoods of HUD-insured properties were obtained from the
1990 Census at the census tract level (Exhibit 1-4).8 While neighborhoods of unassisted
properties differed from those of assisted properties, they did not differ between newer and
older assisted properties. Exhibit 1-4 presents the following demographic information for
property neighborhoods:9

•

Race/Ethnicity: On average, insured properties were located in neighborhoods
that were 63 percent white non-Hispanic, 24 percent black non-Hispanic, and 4
percent in other non-Hispanic groups. The average neighborhood percentage of
Hispanic households regardless of race was 9 percent.

•

Unassisted properties were generally in neighborhoods with higher concentrations
of white non-Hispanic residents. Neighborhoods of unassisted properties were 77
percent white on average, compared with 60 percent in the neighborhoods of
assisted properties.

•

Income Distribution: Unassisted properties tended to be located in higher income
areas compared with assisted properties, both in terms of absolute income levels,
and in terms of income relative to the area median. The average median annual
income (1990 Census figures reported in 1995 dollars) reported in neighborhoods
where unassisted properties were located was much higher than the income in
neighborhoods of assisted properties ($40,492 compared to $28,273).

•

Twenty-eight percent of households in neighborhoods where assisted properties
were located earned more than the local area median (as reported by HUD),
compared with 39 percent in the unassisted properties’ neighborhoods. At the
other end of the income distribution, 45 percent of households in neighborhoods
where assisted properties were located had incomes below fifty percent of the local
median, compared with only 33 percent of households in unassisted property
neighborhoods.

All Census dollar values were inflated to 1995 dollars using the CPI for Urban Consumers. (i.e. multiplied by 153.4/126.1= 1.2173, the
CPI-U for the end of 1995 divided by the CPI-U for the end of 1989) and then rounded to the nearest $1000. For example, the 1990
Census ranges for income distribution include “0 - $9,999". On average across the stock neighborhoods 25 percent of households fell into
that income category. “0 - $9,999) translates into $12,172 in 1995 dollars (9999 x 1.2173), which gets rounded to $12,000. Thus the
bottom income range in the table was “0-$12,000) and included 25% of households in the neighborhoods of the insured stock.
In later chapters of this report, we use the Housing Component of the CPI as the inflation factor, because we are dealing with housing costs
and rents. Here, since the overall income distribution is used, the overall CPI is preferred.

The source for the neighborhood characteristics in the 1990 Study was the 1980 Census, thus there are reported differences in
neighborhood characteristics across the two studies.

Abt Associates Inc.

1.0 Introduction

•

The portion of the population that had incomes below the poverty level was higher
in neighborhoods of assisted properties compared with unassisted properties. On
average, 23 percent of the population in the neighborhoods of the assisted
properties was in poverty compared with 13 percent of the population in the
neighborhoods of unassisted properties.

•

Other Demographic Characteristics: Average household size was slightly higher
in neighborhoods where assisted properties were located (2.6 versus 2.5).

•

Households in neighborhoods of assisted properties were also more likely to be
headed by women (19 percent versus 12 percent) or the elderly (22 percent versus
19 percent) than unassisted properties’ neighborhoods. Insured properties were
located in neighborhoods that averaged 45 percent owner occupants.
Neighborhoods of unassisted properties had the highest proportion of owner
occupants (49 percent) and neighborhoods of older assisted properties had the
lowest proportion (42 percent).
Exhibit 1-4
DEMOGRAPHIC CHARACTERISTICS OF PROPERTY NEIGHBORHOODS
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Neighborhood Race/Ethnicity
White, non-Hispanica

63%

77%**

60%

58%

61%

Black, non-Hispanic

24%

13%

26%

Other, non-Hispanic

Hispanic
(regardless of race)

10%

11%

Neighborhood Income Distribution
$0 - $11,999a

24%

16%**

26%

$12,000 - $17,999

10%

11%

$18,000 - $29,999

18%

16%

19%

18%

$30,000 - $42,999

15%

16%

14%

$43,000 - $60,999

16%

19%

15%

$61,000 +a

17%

24%**

15%

$40,492

$28,273

$27,569

$29,299

Average Median Income

Abt Associates Inc.

$30,492

1.0 Introduction

Exhibit 1-4, continued
Total

Assisted

Total
Average Median Renter
Income

$21,236

Unassisted

Assisted

Older

Newer

$28,645

$20,001

$20,096

$19,863

Neighborhood Income Distribution
Percent < 50% Median a

43%

33%**

45%*

46%

44%

Percent 50-80% Median

12%

11%

12%

13%

Percent 80-100% Median

15%

17%

15%

14%

Percent > Median

30%

39%

28%

27%

29%

Other Neighborhood Demographic Characteristics
Average Household Size

2.5

% Population in Povertya

21%

2.5

2.6

13%**

23%

Other Demographic Characteristics:
% Households Headed by
Womena

18%

12%**

19%

% Households Headed by
Elderlya

22%

19%**

22%

21%*

% Single Person Householdsa

30%

32%**

30%

% Owner Occupieda

45%

49%*

44%

42%*

47%

18%

23%

Sums may not add to 100% due to rounding.

Source:

1990 U.S. Census Data, HUD Median Income Data for 1995.

Trends in the Neighborhoods
As part of the process of determining the market value of the HUD-insured properties, the
study’s market analysts discussed with local real estate professionals the kinds of changes that
were occurring in the neighborhoods. These neighborhood changes are presented in Exhibit
1-5.
•

Most neighborhoods of both assisted and unassisted properties were expected to
either “stay the same” (58 percent) or “improve” (37 percent). Five percent of

Abt Associates Inc.

1.0 Introduction

neighborhoods with assisted properties were expected to “decline” compared with
1 percent of unassisted properties.
•

Fifty-eight percent of the neighborhoods of the unassisted properties were
experiencing new construction, compared with only 32 percent of the assisted
properties’ neighborhoods. On the other hand, significant rehabilitation was
reported more often in neighborhoods with assisted properties (32 percent) than in
neighborhoods with unassisted properties (26 percent).

•

Neighborhoods of assisted properties were more likely to be showing evidence of
disinvestment (17 percent) compared with neighborhoods of unassisted properties
(1 percent).
Exhibit 1-5
NEIGHBORHOOD TRENDS FOR HUD-INSURED PROPERTIES
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Expected Neighborhood Change
Improvea

37%

35%

38%

37%

39%

Stay the Same

58%

64%

57%

59%

54%

Decline

Occurrence of Significant New Construction Evident
Yes a

37%

58%**

32%

30%

35%

63%

42%

68%

70%

65%

Occurrence of Significant Rehabilitation Evident
Yesa

31%

26%

32%

33%

31%

69%

74%

68%

67%

69%

Occurrence of Disinvestment Evident
Yesa

14%

1%**

17%

86%

99%

83%

Sums may not add to 100% due to rounding.

Abt Associates Inc.

1.0 Introduction

Source:

Market Valuation Summaries.

Neighborhood Vacancy Rates
Exhibit 1-6 provides indicators of neighborhood vacancy rates that were collected by the
study’s market analysts through their discussions with local real estate professionals. The
exhibit shows:
•

Forty-six percent of properties were in low-vacancy neighborhoods (vacancy rates
below 4 percent), 36 percent were in moderate-vacancy neighborhoods (vacancy
rates between 4 and 7 percent), and 18 percent were in higher-vacancy
neighborhoods (vacancy rates over 7 percent).

•

Newer assisted properties were most likely to be located in low-vacancy
neighborhoods (52 percent compared with 43 percent for both older assisted and
unassisted).

•

Older assisted properties were most likely to be located in higher-vacancy
neighborhoods (21 percent compared with 14-16 percent for newer assisted and
unassisted properties).

•

Across all assistance categories the market analysts judged that most (82 percent)
properties were in neighborhoods that had vacancy rates similar to or lower than
rates in other parts of their jurisdictions. However, 23 percent of older assisted
properties were in neighborhoods that had higher vacancy rates than other parts of
their local housing markets compared with only 12 percent of unassisted, and 15
percent of newer assisted properties.

•

Across all assistance categories, the market analysts felt that in the future vacancy
rates in the neighborhoods would likely to remain the same (78 percent). Vacancy
rates in unassisted property neighborhoods were considered more likely to increase
(15 percent) compared with rates in neighborhoods of assisted properties (9
percent).

Abt Associates Inc.

1.0 Introduction

Exhibit 1-6
NEIGHBORHOOD VACANCY INDICATORS
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Neighborhood Vacancy Rates
< 4%a

46%

43%

47%

43%*

52%

4% - 7%

36%

41%

35%

36%

34%

7% +a

18%

16%

18%

21%*

14%

Vacancy Rates in Neighborhood vs. Other Areas in Local Housing Market
Above Averagea

18%

12%*

20%

23%**

15%

Average

63%

62%

63%

61%

66%

Below Averagea

19%

26%*

17%

16%

19%

Future Vacancy Trends
Likely to Increasea

10%

15%

10%

Likely to Remain the Same

78%

74%

79%

80%

78%

Likely to Decreasea

12%

11%

12%

Market Valuation Summaries

Abt Associates Inc.

1.0 Introduction

2.0 Tenant Characteristics
This chapter first describes characteristics of tenants in assisted properties, and then presents
several of the specific housing concerns facing low-income tenants in the stock as a whole.
The assisted stock provides an important source of housing for low-income families and
elderly households. Virtually all of the households living in these properties had very-low or
low-incomes and on average one third of households were headed by the elderly. HUD
assistance plays a crucial role in maintaining these units as affordable housing.

2.1

Tenants in Assisted Properties

Tenant characteristics for residents in the assisted portion of the stock were obtained from
HUD’s Tenant Rental Assistance Certification System (TRACS), which tracks tenant
characteristics in assisted HUD-insured properties.1
Exhibit 2-1 presents characteristics of tenants by assistance category.

•

Race and Ethnicity: The racial composition of properties was similar in newer
and older assisted properties—on average 58 percent of residents were white and
37 percent were black. On average 11 percent of residents were Hispanic,
regardless of race.

•

Household Size: Nearly half the households in newer assisted properties (49
percent) included only 1 person, compared with 39 percent in older assisted
properties.

•

Income Distribution: Nearly all (95 percent) residents in newer assisted
properties had very low incomes (below 50 percent of local median for their
household size) compared with 67 percent of residents in older assisted properties.
Virtually all remaining residents had low incomes (below 80 percent of local
median for their household size).

•

Other Income Characteristics: Sources of income also varied in the two types of
properties. Although about one fourth of households in each type of property
received some public assistance, residents in newer assisted properties were more

Data were received on 460 of the 540 assisted properties (150 newer assisted and 310 older assisted). These represent 8,536 of the 10,019
assisted properties in the study universe. Income, household size and elderly/handicap status of the remaining properties were imputed
based on assistance category and occupancy type. Race and ethnicity were imputed based on assistance category, occupancy type and
characteristics of the property’s neighborhood.

Abt Associates Inc.

2.0 Tenant Characteristics

likely to receive SSI income (55 percent versus 42 percent), and less likely to have
wage income (23 percent versus 39 percent).
•

Other Demographic Characteristics: A higher proportion of households in newer
assisted properties were classified as elderly—40 percent compared with 28
percent in older assisted properties. In both types of properties about 11 percent
of households were classified as handicapped.
Exhibit 2-1
TENANT CHARACTERISTICS IN ASSISTED PROPERTIES

Total Properties
Percent of Total Properties

All Assisted

Older

Newer

10,019
100%

5,943
59%

4,076
41%

Race
White a

58%

57%

60%

Black

37%

Native American

Asian

Ethnicity
Hispanic a

11%

12%

Non-Hispanic

89%

88%

91%

Household Size
1 Person a

43%

39%*

49%

2 Person

24%

26%

22%

3 Person

17%

16%

4 + Person

16%

17%

13%

Income Distribution
Under $5,000

18%

17%

18%

$5,000-<$10,000

48%

43%

55%

$10,000-<$15,000

19%

20%

17%

$15,000-<$20,000

$20,000-<$25,000

$25,000+

Abt Associates Inc.

2.0 Tenant Characteristics

Exhibit 2-1, Continued
All Assisted

Older

Newer

Other Income Characteristics
Very low incomea
(<50% of median)

78%

67%**

95%

Low income
(<80% of median)

20%

31%

Not low income
(>80% of median)

Percent with some public assistance
income

24%

25%

24%

Percent with some SS/SSI income a

47%

42%**

55%

Percent with some wage income a

32%

39%**

23%

Other Demographic Characteristics
Percent Elderly a

33%

28%**

40%

Percent Disabled a

11%

10%

12%

Significance test conducted.
* Difference between older/newer assisted significant at the 90% level.
**Difference between older/newer assisted significant at the 95% level.
Note:

Column sums may not add to 100% due to rounding.

Source:

HUD TRACS

Housing Options for Low-Income Tenants
Exhibit 2-2 examines several specific issues facing low-income tenants across all assistance
categories.
The exhibit first reports on the difficulty in finding Section 8 qualified units in neighborhoods
where HUD-insured properties are located. This indicates how difficult it would be for lowincome tenants in insured sample properties to find alternative housing if property-based
subsidies were converted to tenant-based vouchers or certificates and they were required to
seek Section 8 qualified housing in the private sector. This information was obtained by the
study’s market analysts through conversations with local Public Housing Agency (PHA)
contacts.
The exhibit then compares the rents currently paid by tenants with the local Section 8 Fair
Market Rent (FMR). This provides an indication of how housing costs would change if

Abt Associates Inc.

2.0 Tenant Characteristics

tenants did not receive any housing assistance and rented units at the prevailing market rent
levels. The FMR, which is set at the 40th percentile of market rents, is an indicator of rents
for moderate priced units in the market.
Finally, the exhibit compares the optimal unrestricted market rent with the FMR, as an
indication of whether the units in the insured stock would be affordable with tenant-based
Section 8 assistance if owners were able to convert all units to optimal use. The optimal
unrestricted market rent was computed as the rent that, after making required repairs and
upgrades, would yield the highest net market value.2
Exhibit 2-2
OPTIONS FOR LOW-INCOME TENANTS
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Difficulty in Finding Section 8 Qualified Units
Easy/Very Easya

52%

50%

52%

55%

49%

Difficult/Very Difficult

48%

50%

48%

45%

51%

Change in Ease of Finding Units Over Last 5 Years
Easiera

18%

21%**

13%

Same

41%

38%

42%

38%

48%

More Difficulta

41%

44%

40%

41%

39%

Tenant Paid Rents/ Local FMR

< 0.25a

23%

0%**

28%

16%**

45%

0.25 - < 0.5

39%

47%

53%

38%

0.5 - < 0.75

18%

21%

27%

11%

0.75 - < 0.9

22%

0.9 - < 1

21%

1 - < 1.2

19%

1.2 - 1.4

1.4 +

15%

See Section 6.3 for details on computing optimal unrestricted market rent.

Abt Associates Inc.

2.0 Tenant Characteristics

Exhibit 2-2, continued
Total

Assisted

Total
Unassisted

Assisted

Older

Newer

Optimal Unrestricted Market Rent/Local FMR
< 0.75a

12%**

1%**

14%

15%

13%

0.75 - < 0.9

24%

27%

31%

21%

0.9 - < 1

20%

12%

22%

21%

1 - < 1.2

27%

35%

25%

1.2 - 1.4

11%

22%

15%

1.4 - < 1.75

10%

1.75 +a

2%*

9%*

Sums may not add to 100% due to rounding.

Source:

PHA Context Guide, Financial Data, FMR Data, Market Value Summary.

Section 8 Indicators
•

Properties were split about evenly between neighborhoods where the PHA contact
thought it was very easy or easy to find Section 8 qualified units and
neighborhoods where it was thought be difficult or very difficult to find Section 8
units. The ratings were similar across assistance categories.

•

Forty-one percent reported that it had become more difficult to find Section 8
qualified units over the past five years and another 41 percent reported no change.
Older assisted properties were more likely to be in neighborhoods where it was
easier now to find units (21 percent) compared with newer assisted (13 percent)
and unassisted (18 percent) properties.

Tenant Paid Rents Relative to FMR
•

In unassisted properties, rents paid by tenants were generally less than or close to
the Section 8 FMR. Nine percent were below 75 percent of FMR while 62
percent were between 75 and 120 percent of the local FMRs. Nineteen percent of
rents were over 120 percent of FMR, and therefore not easily affordable to

Abt Associates Inc.

2.0 Tenant Characteristics

moderate income households or households with Section 8 vouchers or
certificates.
•

In assisted properties, rents paid by tenants were almost always below 75 percent
of the local Section 8 FMR (95 percent). Thus, it would be expected that these
households would find it financially very difficult to rent market-rate units in their
local neighborhoods without some sort of assistance.

•

All tenant-paid rents were below the local FMR in older assisted properties, as
they were in 95 percent of newer assisted properties.

•

At the very low end of the distribution, tenant paid rents were below 25 percent of
the FMR in 45 percent of newer assisted properties and in 16 percent of older
assisted properties. Tenants in this group of 28 percent of assisted properties
would face particular hardship without their assistance.

Optimal Unrestricted Use Rents Relative to FMR
•

If rents in properties were set at their highest and best use, rents in 45 percent of
all insured properties would be above the local FMR. Seventy-one percent would
be between 75 and 120 percent of FMR.

•

Rents in unassisted properties would be higher relative to FMR compared with
rents in assisted properties. Seventy-six percent of unassisted properties would
have rents above the FMR including 41 percent with rents above 120 percent of
FMR. Only 38 percent of assisted properties would have rents above the FMR in
their optimal market position, including 13 percent with rents above 120 percent
of FMR.

•

Optimal rents relative to FMR would be lower in older assisted properties
compared with newer assisted properties. Sixty-nine percent of older assisted
properties would have rents below the FMR, and another 25 percent would have
rents between 100 and 120 percent of FMR. In contrast, 55 percent of newer
assisted properties would have rents below the FMR, and another 25 percent
would have rents between 100 and 120 percent of FMR.

•

This means that if properties were allowed to operate at their optimal market level
more than a third (38 percent) of assisted properties would not be affordable even
with Section 8 tenant-based assistance provided at the FMR and 13 percent would
not be affordable with rents allowed at 120 percent of the FMR.

Abt Associates Inc.

2.0 Tenant Characteristics

3.0 Physical Condition of the Stock of Multifamily
Housing
This chapter presents measures of the physical condition of the stock of HUD-insured (or
held) multifamily properties. Two primary measures are discussed. Section 3.1 describes the
current backlog of physical needs in the stock, and Section 3.2 describes the annual accrual of
capital needs.

3.1

Current Backlog of Physical Needs

A property’s current physical condition (physical needs backlog) was measured by the cost of
repairs and replacements beyond ordinary maintenance required to restore a property to its
original working condition. Property systems still in good working order (requiring “no
action”) by definition, had no associated costs.
The backlog of physical needs was computed using a standardized set of unit costs that were
multiplied by the quantity and action level appropriate to a particular property (e.g., the
number of windows in the property that the inspector determined needed replacement was
multiplied by the cost of replacing a window of the appropriate size). The inspection protocol
included observing conditions of 119 mechanical, electrical, and architectural systems,
organized by major property elements (site, building, or unit). The 119 specific systems were
combined into 17 major system groups for costing and reporting purposes. The systems and
their groupings are presented in Exhibit 3-1. For each system, the inspector judged and
recorded the level of remedial action. As with the 1990 Study, the A.M. Fogarty Company
supplied the per-unit costs for each of the repair and replacement items. Property costs were
then multiplied by location-specific adjustment factors to obtain the local cost of repair needs
for each property. Using the same data sources (described in Appendix B) and costing
procedures (described in Appendix C) helped to assure the consistency of cost comparisons
between the 1989 and 1995 outcomes.
To compare costs across properties having different numbers of units, each property’s costs
were expressed on a “per-unit basis”. Furthermore, to permit comparisons across properties
having different sized units (since a property of predominantly efficiency units will have lower
costs per unit than an otherwise identical property of 3-bedroom units), all property-level
costs were normalized on the basis of unit square footage, using each property’s

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Exhibit 3-1
SYSTEM GROUPS AND KEY SYSTEMS INSPECTED
Site Areas- landscaping, roadways, parking, paved pedestrian areas, curbing, fencing, retaining wall, site
drainage, pole mounted site lighting.
Site Amenities- site furniture, yards and enclosures, dumpsters, pool, tennis courts, basketball courts.
Site Distribution Systems- emergency generator, site electrical distribution, hot water distribution, domestic
hot& cold water lines, main water service, gas lines, site sanitary lines, septic system, sewage ejectors, hydrants.
Building Mechanical & Electrical - heating risers, gas distribution, sanitary distribution, fire sprinkler system,
sump pump, compactors, switchgear, building wiring, emergency lights, building smoke detector,
communication system.
Building Heating & Cooling - central vent/exhaust, central air conditioning, furnace, boiler, boiler room
piping, boiler room equipment, boiler room controls.
Building Elevators - shaftways, shaftway doors, cabs, machinery.
Building Exterior Closure - foundation, slab, exterior wall, insulation.
Building Roofs- roof covering, parapet wall, chimney, roof hatches, skylight, roof drainage.
Building Windows & Doors- windows, window security grates, exterior common doors, unit entry doors,
storm/screen doors.
Building Exterior Features - canopies, exterior stairs, building mounted site lights, fire escapes, balconies,
porches, decks, sheds.
Building Common Areas - vestibules, corridors, stairways, interior lights, mail facilities.
Unit Interior Construction - interior walls-partitions (excluding kitchen and bathroom), floor sub-base.
Unit Interior Finishes - interior walls-surface, floor covering, interior doors & frames, kitchen walls, kitchen
floor, bathroom walls, bathroom floor.
Unit Kitchen Fixtures - kitchen cabinet/counter, range and hood, refrigerator, garbage disposal, dishwasher,
microwave, trash compactor.
Unit Bathroom Fixtures - bathroom fixtures, bathroom accessories, vanities.
Unit Heating & Cooling - HVAC units, radiation, boiler (unit level), furnace (unit level), temperature control,
wall air conditioner.
Unit Electrical - electrical panel, electrical wiring, bell/intercom, smoke detector.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

“2-bedroom equivalent” rather than its actual number of units.1 This serves to normalize data
for comparisons across assistance categories that have different unit-size mixes. Throughout
this report when per-unit values are presented, they are actually per 2-bedroom equivalent
unit.
For ease of reference, the level of physical needs backlog has been divided into four
categories. These categories are defined in relation to average annual normal physical needs
accrual, which was estimated at $1,437 per unit per year.2
•
•
•
•

No backlog: less than $10 per unit
Normal backlog: $10 to $1,500 per unit
Moderate backlog: $1,500 to $3,000 per unit
Serious backlog: over $3,000 per unit

Exhibit 3-2 shows the backlog of physical needs for the full multifamily housing stock by
assistance category.3 Exhibit 3-3 shows this information graphically. The exhibits show that:

•

The mean backlog of physical needs was $3,236 with a median of $1,452. The
mean backlog was highest in assisted properties, averaging $3,638 (median
$1,661) and lowest in unassisted properties, averaging $1,427 (median $545). The
mean backlog was not statistically different between older and newer assisted
properties ($3,929 and $3,214 respectively).

•

The total estimated backlog for the stock was $4.17 billion, $3.5 billion of which
was in assisted properties.

The number of “2 bedroom equivalent” units was calculated by dividing the total square feet of living space by 844. This was the national
average square footage of a 2 bedroom unit in the 1990 Study. For consistency we used the same number in the current study. The
estimated numbers of 2BR units and actual units by assistance category are as follows:

Unassisted
Older Assisted
Newer Assisted
Total

Number of 2-bedroom
Equivalent Units
350,815
643,468
319,714
1,314,026

Number of
Actual Units
354,083
686,309
364,848
1,405,240

Appendix Exhibit D-1 presents key statistics on a “per 2-bedroom unit” basis for the stock as a whole, while the tables in the body of the
report present all data on a “per 2 BR” basis at the property level. In other words, the text tables answer questions such as what is the
average backlog per 2-bedroom unit across all properties, while the appendix exhibit answers questions such as what is the average per 2bedroom unit backlog across all units in the stock.
2

These categories are clearly somewhat arbitrary and other breaks could have been used.

In addition to comparing backlog to average accrual across the stock, we compared each property’s backlog of physical needs to its own
average annual accrual. Because average annual accruals were fairly similar across all properties, this did not yield very different results
from comparisons with accruals stock-wide.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

•

Thirteen percent of the stock had virtually no backlog of physical needs (under
$10 per unit). Unassisted properties were more likely to have no backlog.
Twenty-one percent of the unassisted stock had no backlog. Fewer older assisted
properties had no backlog (8 percent) than did newer assisted properties (15
percent).

•

Thirty-eight percent of the properties had backlogs of physical needs that can be
considered normal, in the range of about one year’s worth of average accrual of
physical needs. This level of physical need does not seem problematic, since it is
within the normal cycle of accumulation of physical needs in a property. Again,
more of the unassisted stock (45 percent) had normal backlogs of physical needs
than did the assisted stock (36 percent). Newer assisted properties were more
likely to have normal backlogs (40 percent) compared with older assisted
properties (34 percent).

•

Nineteen percent of the properties had moderate backlogs of physical needs, which
is the equivalent of about one to two years’ worth of average accrual of needs.
This level of backlog may be a cause of concern, since it appears to exceed a
normal annual accumulation of need.

•

Thirty percent of properties had serious backlogs of physical needs (over $3,000
per unit). This is at least two years’ worth of accrual of needs, and seems likely to
indicate problems that will affect residents and the marketability of the property,
and may ultimately threaten the financial viability of the property. The percentage
of older assisted properties in this category far exceeded the percentage of
unassisted and newer assisted properties in this category. Forty-one percent of
older assisted properties had serious backlogs of physical needs. In contrast only
11 percent of unassisted properties had such high backlogs. Newer assisted
properties were between these two extremes: one quarter had backlogs of physical
needs of over $3,000 per unit.

•

Fifty-eight percent of older assisted properties had moderate or serious backlogs
(of over $1,500 per unit) compared with 46 percent of newer assisted properties
and 33 percent of unassisted properties.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Exhibit 3-2
DISTRIBUTION OF BACKLOG OF PHYSICAL NEEDS BY ASSISTANCE CATEGORY
(IN 1995 DOLLARS PER 2BR EQUIVALENT UNIT)
Total
Characteristic
Total Properties
Percent of Total Properties

Assisted

Total
12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Distribution of Backlog of Physical Needs
No Backloga

13%

21%*

11%

8%*

15%

<$10

13%

21%

11%

15%

Normal Backlog

38%

45%

36%

34%

40%

$10 to <$500

15%

27%

13%

14%

11%

$500 to <$1,000

12%

13%

11%

16%

$1,000 to <$1,500

11%

12%

10%

13%

Moderate Backlog

19%

22%

19%

17%

21%

$1,500 to <$2,000

11%

$2,000 to <$2,500

$2,500 to <$3,000

Serious Backloga

30%

$3,000 to <$4,000

$4,000 to <$5000

$5,000 to <$7,500

10%

>= $7,500

11%

13%

15%

11%**

35%

41%*

25%

Statistics on Backlog of Physical Needs
Meana
Standard Error
Median

$3,236

$1,427**

$3,638

$3,929

$3,214

203

255

240

276

430

$1,452

$545

$1,661

$2,096

$1,324

Column sums may not add to 100% due to rounding.

Source:

1995 Physical Inspection Data and Costing Programs.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Exhibit 3-3
BACKLOG DISTRIBUTION BY LEVEL

Source:

Exhibit 3-2

The medians backlogs of physical needs were far lower than the means, which indicates that
some properties had extremely high backlogs.
•

The median backlog for unassisted properties ($545) was well below one year’s
average accrual, meaning that most properties are keeping up with their repair
needs.

•

The median for older assisted properties was $2,096 -- much lower than the mean,
but still well over the average annual accrual of physical needs.

•

The median for newer assisted properties ($1,324) was approximately equal to one
year’s average accrual of physical needs.

Components of Backlog of Physical Needs
The inspection protocol included observing conditions of 119 mechanical, electrical, and
architectural systems, organized by major property elements (site, building, or unit). The 119
specific systems were combined into 17 major system groups for costing and reporting
purposes. (See Exhibit 3-1 above.) As can be seen in Exhibit 3-4, most of the repair costs

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

(54 percent) were attributed to unit-level systems. Building systems were the next largest
component of physical needs backlog, accounting for 31 percent of needs overall. The
smallest costs were those associated with sites, accounting for only 14 percent of all physical
needs. The distribution of physical needs across property elements was similar across all
assistance categories.
•

The largest component of need was Kitchen Fixtures (23 percent of total backlog
costs), which includes items such as cabinets, counters, ranges, and refrigerators.
Kitchen Fixtures are subject to a high level of resident use and wear out relatively
quickly.

•

The second largest component of need, unit Interior Finishes, accounted for 21
percent of the mean physical need backlog. Interior Finishes are largely surface
elements such as wall and ceiling surfaces and interior doors. These elements are
also subject to a high level of resident use and generally wear out more quickly
than most systems.

•

The third largest component of backlog need was site areas (12 percent of backlog
costs). Site areas include items such as landscaping, roadways, and other paved
areas.
Exhibit 3-4
DISTRIBUTION OF BACKLOG OF PHYSICAL NEEDS BY SYSTEM GROUPa
Total

Characteristic

Assisted

Total
Unassisted

Assisted

Older Assisted

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

Site Costs

14%

15%

14%

15%

14%

Site Areas

12%

13%

11%

12%

11%

Site Amenities

Site Distribution

Building Costs

31%

36%

31%

30%

Mechanical & Electric

Heating & Cool

Elevators

Exterior Closure

10%

Roofs

10%

Total Properties
Percent of Total Properties

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Newer Assisted
4,076
41%b

Exhibit 3-4, continued
Total
Characteristic

Unassisted

Assisted

Older Assisted

Newer Assisted

Windows & Doors

10%

Exterior Features

Common Areas

54%

49%

55%

54%

56%

Interior Construction

Interior Finishes

21%

18%

22%

21%

23%

Kitchen Fixtures

23%

26%

23%

21%

25%

Bath Fixtures

Heating & Cooling

Electrical

Total

100%

Mean

$3,236

$1,427

$3,638

$3,929

$3,214

Unit Costs

Assisted

Total

Percents are calculated only for properties with backlog greater than 0.
Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Note:

Column sums may not add to 100% due to rounding.

Source: 1995 Physical Inspection Data and Costing Program.

Unit interior finishes and kitchen fixtures were also the two largest cost components in 1989
(their shares were 37 percent and 14 percent respectively).
While the level of repair needs was higher in the older assisted properties than in either
unassisted or newer assisted properties, the distribution of costs by system group was similar
across assistance groups.

Costs Associated with Health and Safety Systems
The effect of a given backlog of physical needs depends on the types of systems affected and
the repairs and replacements that are required. Backlogs that are in systems that directly
affect resident health and safety—such as interior construction, heating and cooling, and
building mechanical systems—are of special concern. The immediate threat to resident safety

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

is diminished to the extent that the needed repairs are in systems that are more cosmetic, such
as site amenities or interior finishes.
For the purposes of this study, a subset of the systems groups was identified as most relevant
to health and safety:
•
•
•
•
•
•

Unit Interior Construction
Unit Bathroom Fixtures
Unit Heating and Cooling
Unit Electrical
Building Heating and Cooling
Building Mechanical and Electrical

The mean costs for these Health and Safety Systems are shown in Exhibit 3-5. For all
properties, 20 percent of all physical needs (an average of $659 and a median of $65 per 2BR
equivalent unit) were in Health and Safety Systems.
This share of Health and Safety systems in the overall needs estimate was relatively constant
across assistance categories, but as with other repair costs, older assisted properties had much
higher Health and Safety needs (mean $850 and median of $150 per 2BR unit). Most
unassisted properties had no backlog needs in Health and Safety Systems. (The median for
this assistance category was $0.)
Exhibit 3-5
BACKLOG OF PHYSICAL NEEDS FOR HEALTH & SAFETY SYSTEMS
(IN 1995 DOLLARS PER 2BR EQUIVALENT UNIT)
Total
Characteristic

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

Total Properties
Percent of Total Properties

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Total Backlog
(All Property Systems)
Mean

$3,236

$1,427

$3,638

$3,929

$3,214

Backlog for Health and Safety Systems
Meana

$659

$273**

$745

$850*

$592

Standard Error

101

119

Median

$65

$105

$150

$72

Health & Safety as a % of
Total Backlog

20%

19%

20%

22%

18%

* Difference between unassisted/assisted or older/newer assisted significant at the 90% level.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

**Difference between unassisted/assisted or older/newer assisted significant at the 95% level.
a
Significance test conducted.
b
Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.
Source: 1995 Physical Inspection Data and Costing Program.

While the dollar amount attributed to Health and Safety systems increased from $377 in 1989
to $659 in 1995, the percent of all physical needs that were attributed to Health and Safety
Systems fell from 25 percent in 1989 to 20 percent in 1995. This decrease in the percent of all
physical needs that are attributed to Health and Safety Systems can primarily be attributed to
the change in the relative distribution of backlog of physical needs between unit-level and non
unit-level systems from 1989 to 1995. Most of the Health and Safety systems are unit level
systems, and in 1989 unit level systems accounted for 59 percent of the total backlog, whereas
in 1995 unit level systems accounted for only 55 percent of the total backlog.
Changes in the Physical Condition of the Stock
One of the most significant findings of this study is that the physical condition of many HUDinsured properties has deteriorated between 1989 and 1995. The properties were aging, and
apparently property owners and managers were not keeping up with repair needs. After
controlling for inflation, the mean backlog of physical needs across the stock rose by over 60
percent between 1989 and 1995. As described below, across all categories of properties the
average backlog increased, but the increase was largest by far in newer assisted properties.4

The study was not designed to provide reliable property-specific estimates of physical needs, but rather to provide reliable estimates at the
level of assistance category. Thus, comparing changes in the backlog of physical needs of specific properties may be misleading and is not
done here.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Exhibit 3-6
COMPARISON OF BACKLOG OF PHYSICAL NEEDS BETWEEN 1989 AND 1995
(IN 1995 DOLLARS PER 2BR EQUIVALENT UNIT)
(BASED ON OVERLAP SAMPLE)
Total

Assisted

Total
Total Properties
% of Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Mean Backlog of Physical Needs
1995

$3,058

$1,427

$3,420

$3,845

$2,800

1989

$1,882

$960

$2,086

$2,769

$1,091

Median Backlog of Physical Needs
1995

$1,551

$545

$1,823

$2,280

$1,390

1989

$787

$117

$982

$1,580

$373

Percentage of Properties with Backlog of Physical Needs <$10
1995

12%

21%

10%

13%

1989

20%

37%

17%

11%

26%

Percentage of Properties with Backlog of Physical Needs between $10 and $1,500
1995

37%

46%

35%

38%

39%

1989

43%

44%

42%

45%

49%

Percentage of Properties with Backlog of Physical Needs between $1,500 and $3,000
1995

21%

22%

21%

19%

24%

1989

15%

17%

18%

15%

Percentage of Properties with Backlog of Physical Needs over $3,000

1995

31%

11%

35%

42%

25%

1989

22%

10%

24%

34%

10%

Older assisted properties represent 59% of assisted properties and 49% the universe. Newer assisted properties represent 41%
of assisted properties and 33% of the universe.

Source:

1995 Data: 1995 Physical Inspection Data and Costing Program.
1989 Data: 1990 Study Analysis File.

Exhibit 3-6 presents indicators of change in the mean, median and distribution of the backlog
of physical needs in 1989 and in 1995. These numbers are slightly different from the numbers
in Exhibit 3-2 because the results are based on the overlap sample of 504 properties that were
included in both studies. Using only properties that were included in both studies eliminates
Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

the possibility that changes in the estimated backlog of physical needs results from changes in
the sample of properties used. All numbers were weighted to reflect the stock of properties
that was still insured in 1995.5
•

The mean backlog of physical needs across the entire stock in 1989 was $1,882
per unit (in 1995 dollars). The mean backlog rose to $3,058 per unit in 1995, a
62 percent increase from the 1989 mean.

•

While the absolute backlog was highest in older assisted properties in both years,
the percentage increase between 1989 and 1995 was smallest at 40 percent.

•

The absolute backlog was lowest in unassisted properties in both years. The mean
backlog use by about 50 percent between 1989 and 1995.

•

These increases in mean backlog of physical needs are certainly cause for concern,
but they are much smaller than the increase in mean backlog of physical needs in
newer assisted properties. In that portion of the stock the mean backlog rose by
over 150 percent during the same time period.

•

Across the stock as a whole, the proportion of properties with no backlog of
physical needs decreased from 20 percent in 1989 to 12 percent in 1995. At the
other extreme, the proportion of properties with backlogs of physical needs over
$3,000 per 2-bedroom unit increased from under a quarter of the stock (22
percent) in 1989 to nearly a third (31 percent) in 1995.

In all three categories of the stock, the proportion of properties with high levels of backlog
increased between 1989 and 1995.
•

Over one third (37 percent) of unassisted properties had no backlog of physical
needs in 1989, compared with only 21 percent in 1995. In contrast, the proportion
with backlogs between $1,500 and $3,000 per unit more than doubled from 9
percent to 22 percent. The proportion of unassisted properties with very high
backlogs of over $3,000 per unit stayed nearly constant, going from 10 to 11
percent.

The overall reported differences in backlog reflect both changes in actual property condition and changes in several system definitions and
costs (beyond controlling for inflation). Unit cost estimates for repairing some systems increased since 1989 and decreased for other
systems. The definitions of actions associated with some systems changed as well. As shown in a separate analytic memorandum, the
reported change in mean backlog appears to be a true result of deterioration in condition rather than a result of changes in specific cost
components. Comparing the mean backlog of physical needs in 1989 with the calculated mean using 1989 condition and the 1995 cost
files shows that the changes in costs had almost no effect on the estimated mean backlog of physical needs. The mean backlog of physical
needs in 1989 was $1,882 per unit. Using the 1995 cost file, and 1989 condition yields an overall mean estimate of $1,876 per unit.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

3.2

•

Among older assisted properties, in both years very few properties had no
backlogs of physical needs. However the proportion with backlogs between $10
and $1,500 per unit decreased from 45 percent to 38 percent of properties. At the
other extreme, in 1995, 42 percent of older assisted properties had backlogs of
over $3,000 per unit compared with about a third (34 percent) in 1989.

•

As noted above, the deterioration in physical condition was most severe in newer
assisted properties. In that portion of the stock, in 1989 three quarters of
properties had backlogs of physical needs under $1,500 per unit, including over
one quarter (26 percent) with no backlogs. Only 10 percent of newer assisted
properties had backlogs over $3,000 per unit. In 1995 the situation was very
different. Just over half (52 percent) the properties had backlogs of under $1,500
per unit, including only 13 percent with no backlog. One quarter of the properties
had backlogs of over $3,000 per unit.

Projected Future Physical Needs - Physical Needs Accrual
Costs

A property’s physical needs accruals are estimates of the average annual costs needed to
cover expected repairs and replacements beyond ordinary maintenance for all systems over
each of the next 20 years. As with backlog costs, accrual costs were computed based upon
inspectors’ examination of each Observable System. For each system a set of standardized
costs was applied, incorporating timing information based on the system’s remaining useful
life (or required action level in the case of systems, such as interior walls, which need periodic
refurbishment rather than replacement of framing and plasterboard). Estimates of future
accrual needs indicate a property’s expected need for resources in the future.
Each system was assigned an expected useful life (or required action level) and an accrual
cost. For systems requiring periodic replacement or major overhaul:
•

Useful life is the age of a system when it must be replaced (or overhauled) because
it has worn out or is approaching failure, and

•

Accrual cost is the cost of replacing or overhauling the system.

For example, boilers are expected to last 25 years (useful life) and the associated accrual cost
is the cost of a new boiler. A few items are not expected to wear out, but will need periodic
major action. For these items, the “expected life” is the action interval, and the accrual cost is
the repair cost. For example, brick chimneys or walls are not expected to wear out at any
known interval, but must have the mortar joints raked out and repointed, and be waterproofed

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

every ten years. The associated accrual cost is the cost of raking and repointing mortar joints,
as well as waterproofing.
Some systems were deemed inappropriate for accrual estimates because they were not
expected to need replacement or overhaul over the 20-year horizon used for this study. An
example is the Site-Level Domestic Hot Water Lines. Over time, a portion of the lines may
need to be replaced, but this is not an expected occurrence. (Appendix Exhibit C-4 shows the
expected useful life and accrual actions for all systems).
For most systems, the inspectors recorded system age as part of the on-site inspections; for
other systems, the study assigned age equal to the age of the buildings. To prevent doublecounting of a property’s physical needs, age was set to zero for any system that needed
replacement or overhaul as part of the physical needs backlog. In other words, for computing
accrual, the study assumes that the physical needs backlog was fully remedied in year zero.
The study’s accrual costing program determined, for each of the next 20 years, whether the
observed system would reach the end of its useful life that year (based on its expected useful
life and on the system age), and if so, added the repair/replacement cost to the accrual total
for that year.
As shown in Exhibit 3-7, the mean annual accrual cost over the next 20 years was $1,437 per
2BR equivalent unit, with a median of $1,362 (expressed in 1995 dollars). As can be seen,
three fourths of all properties are expected to have average annual accrual of needs between
$1,000 and $2,000 per 2BR unit per year. There was no significant difference in the average
annual accrual across assistance categories.6

The increase in accrual estimate is mainly due to a change in the estimation approach used to impute costs for uninspected units and
buildings. This study estimated accrual costs for uninspected units based on the ratio of inspected unit square footage to total square feet,
and accrual costs for uninspected buildings based on the proportion of the property’s units that were in the inspected buildings. In 1989,
both unit and building level accruals for uninspected units and buildings were computed based on the share of total backlog that the
inspected units (or buildings) represented. Using data on the comparison sample of properties, the stockwide estimate of average 20-year
annual accrual costs was $1,057 in 1989 (expressed in 1995 dollars) using the 1989 estimation approach. Using the 1995 methodology
and the 1989 data yields an estimate of $1,448, which is very close to the 1995 estimate of $1,437 per unit.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

Exhibit 3-7
PROJECTED AVERAGE ANNUAL ACCRUAL OF PHYSICAL NEEDS OVER 20 YEARS
(IN 1995 DOLLARS

PER 2BR EQUIVALENT UNIT)

Total
Characteristic
Total Properties
Percent of Total Properties

Assisted

Total
12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Average Annual Accrual
<$1000

16%

17%

19%

14%

$1000-$1499

45%

49%

44%

43%

45%

$1500-$1999

30%

32%

29%

26%

34%

$2,000-$2,499

$2,000 or more

Statistics on Annual Accrual
Meana

$1,437

$1,514

$1,420

$1,404

$1,443

Standard Error

19.6

54.1

20.7

28.2

29.8

Median

$1,362

$1,419

$1,331

$1,301

$1,399

Column sums may not add to 100% due to rounding.

Source:

1995 Physical Inspection Data and Costing Program.

Abt Associates Inc.

3.0 Physical Condition of HUD-Insured Properties

4.0 Financial Condition of HUD-Insured and Held
Properties
This section provides information on various aspects of property finances including annual net
cash flow (Section 4.1), ability to cover the physical needs backlog and annual accrual of
needs (Section 4.2), and Section 8 assistance (Section 4.3).

4.1

Annual Net Cash Flow

Annual net cash flow is a key indicator of a property’s viability, showing whether it can meet
its ongoing obligations. Specifically, annual net cash flow (before income taxes) shows the
degree to which a property can cover current operations and routine maintenance, mortgage
debt service, and annual deposits to its replacement reserve fund (to cover future physical
replacements). Positive annual net cash flow is also requisite for making distributions to
owners.
A property’s annual net cash flow equals its revenues less expenses. The primary revenue
source for most properties is apartment rents paid by residential tenants, and in assisted
properties, subsidies paid by HUD. Other revenue sources may include commercial rent,
financial revenue (such as interest income from reserve accounts), or forfeited tenant deposits.
Property expenses include operating and maintenance costs, debt service, and deposits to the
replacement reserve account. As was done in the section on physical needs, financial items
and reported on a per 2-bedroom equivalent unit basis to allow comparisons across properties
of different sizes (unit counts) and unit compositions (distribution of units by bedroom
counts).
For this report we define annual net cash flow as:
1. Weighted average of property revenue over most recent three years
2. Minus Weighted average of operating expenses over most recent three years
3. Minus Weighted average deposits to replacement reserve account (maximum of
actual and required deposit) over most recent two years
4. Minus Mortgage debt services (including interest and principal on mortgage and
any supplementary loans and operating loss loans) and mortgage insurance
premium.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Where:
1. The 3-year weighted average of property revenues includes actual rental income
(which equals potential rent less any vacancy losses) from tenant paid apartment
rents and tenant assistance payments (Section 8 assistance), plus any commercial
or financial income.1 A weighted average over the most recent 3 years was used,
with the more recent years receiving higher weights.2 By averaging over 3 years,
the measure focused on long-term revenue flows in a property, and reduced the
effect of one-time outliers. The application of a higher weight to more recent
years incorporated trends into the measure.
2. The 3-year weighted average operating expenses equals the sum of the cost
components reported in the project financial statements -- administrative expenses,
utility expenses, operating and maintenance expenses, and tax and insurance
expenses.3
3. The 2-year weighted average of deposits to the reserve for replacement account
equals the maximum of actual deposits as reported on the annual financial
statements and required deposits, either as reported on the annual financial
statements or as 0.5 percent of the original mortgage amount.4

Total Revenues come from HUD form 92410. The Total Revenue line from the form was adjusted when it appeared that tenants paid their
own utilities. In order to have comparable income and expense numbers across properties, we added in utility costs to both the revenue and
expense sides when it appeared, based on the value of utilities in the utility expense line that tenants paid their own utilities. Adjustments
were based on the average cost per square foot reported in the 1996 IREM Income/Expense Analysis reports by receipt of assistance,
region and building type. Financial data came from several sources. HUD supplied us with Annual Financial Statement files for 1993 1995. The 1993 and 1994 files contained more complete information (including reserves). We also obtained information from HUD’s
Data Warehouse for 1992 - 1994, and from backup HUD tapes for 1992-1994. Interest Reduction Payments (IRPs) in Section 236
properties are not reflected on form 92410. IRPs therefore, are not reflected in reported total revenues.

For properties with three years (or more) years of data, the most recent data received a weight of 0.5, the second most recent year 0.3 and
the third most recent 0.2. When four years of data were available, only the most three years were used. For properties with two years of
data, the most recent data received a weight of 0.6 and the oldest year a weight of 0.4. For properties with only one year of data, the weight
was 1. For the properties that were missing financial data, values were imputed based on median values by assistance category and
building type. HUD data files were fairly complete. For example, 354 sample properties had four years of total revenue data, 186 had
three years, 53 had two years, 16 had one year, and 12 had no financial data. For other financial variables the coverage was similar. The
most recent year of available data was 1995, for which 354 properties had at least some financial data.

Line 6200 and 6300 from HUD form 92410 for administrative expenses, Line 6400 for utilities, Line 6500 for operating and maintenance,
and Line 6700 for taxes and insurance. As discussed above, for consistency across properties, utility expenses were adjusted when it
appeared that tenants paid their own utilities for their apartments.

As of 1968 the required deposits to the replacement reserve account were 0.6 percent of the total replacement costs of structure for new
construction properties, and 0.4 percent of the mortgage amount for rehab properties. Information on deposits to the reserve for
replacement account were apparently not entered into HUD’s automated data systems. These data are only available for 1993 and 1994,
in the file provided by HUD. 563 properties had information for both years, 38 for one year, and 20 provided no information. Where two
years of data were available the weights were 0.6 for the more recent year and 0.4 for the earlier year. For about 64 percent of properties
the actual amount was used. For 9 percent, the reported required amount was used, and for 27 percent, .5 percent of the mortgage was
used.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

4. Debt service was computed from the mortgage amount, term and interest rate.
Mortgage insurance premium (0.5 percent of the outstanding principal balance)
was added to the debt service costs for all properties except Section 221(d)(3)
BMIR properties and Section 236 properties. Section 221(d)3 BMIR properties
pay no mortgage insurance premium. Section 236 properties make debt service
payments based on a 1 percent interest rate. The remaining interest payments to
the mortgagee are made by HUD. The debt service also includes payments for
supplementary loans and operating loss loans.5
Property Revenues
Exhibit 4-1 shows the components of annual revenue by assistance category. All data are
presented in 1995 dollars.6 Total revenues include rent revenues, tenant assistance payments,
commercial, financial and “other” revenues, net of vacancy losses. The exhibit shows that:
•

Mean revenues for the HUD-insured properties were $7,646 per 2-bedroom unit
per year, with a median of $6,541.

•

On average, revenues for unassisted properties ($7,978) were about midway
between revenues for older assisted ($5,868) and newer assisted ($10,057)
properties.

•

Tenant paid rents accounted for nearly all of the revenue in unassisted properties
($7,632 of the average revenue of $7,978). In assisted properties tenant paid rents
accounted for about 40 percent of revenue, with most of the remaining revenue
coming from tenant assistance payments from HUD.

•

Consistent with the higher tenant incomes in older assisted properties, tenants paid
rents in these properties were higher on average than in newer assisted properties,
averaging $3,287 per 2-bedroom unit per year, compared with $2,593 in newer
assisted properties.

Our sample included 4 properties that received operating loss loans (which are generally provided at or near the time of origination) and 21
properties that received Section 241 supplementary loans at some point after origination.

All values were converted to 1995 dollars using the CPI for Urban Consumers for Housing (1992 = 138.5, 1993 = 142.3, 1994=145.4,
and 1995 = 149.7).

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-1
COMPONENTS OF PROPERTY REVENUES
(IN 1995 DOLLARS PER 2 BR EQUIVALENT UNIT)
Total

Assisted

Total
Total Properties
Percent of Total Properties

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

12,243
100%

Total Revenues (Net of Vacancies)
Meana

$7,646

$7,978

$7,572

$5,868**

$10,057

Standard Error

156

562

144

150

279

Median

$6,541

$6,363

$6,663

$5,183

$9,128

$3,287**

$2,593

Tenant Paid Rents
Meana
Standard Error
Median

$3,846

$7,632**

121

584

$3,213

$5,976

$3,005
84

161

$2,791

$3,186

$2,124

$2,576**

$7,448

Tenant Assistance Paymentsc
Meana
Standard Error
Median

$3,730

0**

$4,558

103

127

137

239

$3,083

$3,868

$2,310

$7,106

Vacancy Loss (As a Percent of Rent Revenue)
Meana

3.12%

5.76%**

2.53%

3.37%**

1.30%

Standard Error

0.015

0.0059

0.0014

0.0022

0.001

Median

1.78%

4.50%

1.47%

2.28%

0.91%

Annual Financial Statements for 1992 - 1995.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

•

Tenant assistance payments were significantly higher in newer assisted properties
compared with older assisted properties, averaging $7,448 per 2-bedroom unit,
compared with $2,576 per 2-bedroom unit in older assisted properties. This is
because all newer assisted properties received Section 8 assistance compared with
71 percent of older assisted properties. In addition, the amount of assistance
received was also higher averaging $7,448 per unit compared with $3,250 per unit
in older assisted properties.7

•

Vacancy losses were low across all categories of insured properties, averaging
3.12 percent (with a median of 1.78 percent).

•

Vacancy losses were highest in the unassisted properties averaging 5.76 percent,
and lowest in the newer assisted properties, averaging only 1.3 percent. These
extremely low vacancy rates in assisted properties are expected because the
project-based assistance these properties receive helps them attract lower-income
renters.

Property Expenses
Exhibit 4-2 shows the components of total property expenses.
•

Total annual expenses (including operating and maintenance expenses,
replacement reserve deposits and debt service) averaged $7,052 per year.
Expenses were highest in the newer assisted properties (averaging $8,952) and
lowest in the older assisted properties (averaging $5,585).

•

The differences in total annual expenses across assistance categories result
primarily from differences in debt service costs. Debt service costs averaged
$3,760 in the newer assisted properties, $2,930 in the unassisted properties, and
only $859 in the older assisted properties. The difference in debt service costs
reflects both the timing of the loans—newer assisted properties generally have
higher mortgage principal and interest payments—and the interest subsidies and
below market interest rates provided to most of the older assisted properties.8

•

Operating and maintenance expenses were very close across all three categories of
properties, averaging $4,540 per unit per year. As in 1989, however, operating

In this analysis interest reduction payments (IRPs) for the Section 236 properties were excluded both from revenues and expenses. If IRPs
were included in tenant assistance payments the average tenant assistance payment across all older assisted properties would be $3,171 per
unit rather than $2,576. Total revenues in older assisted properties would then average $6,462 per unit.

As indicated above, IRPs were excluded from both revenues and expenses. If they were included in expenses, debt service in older assisted
properties would average $1,454 per unit instead of $859. Total expenses would average $6,179.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

and maintenance expenses of newer assisted properties ($4,928) remained higher
than those of older assisted ($4,349) or unassisted properties ($4,338).
Exhibit 4-2
COMPONENTS OF ANNUAL EXPENSES
(IN 1995 DOLLARS PER 2-BEDROOM EQUIVALENT UNIT)
Total

Total Properties
Percent of Total Properties

Assisted

Total

Unassisted

Assisted

Older
Assisted

Newer
Assisted

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Total Annual Expenses
Meana
Standard Error
Median

$7,052

$7,491

$6,955

$5,585**

$8,952

146

555

130

129

257

$6,184

$6,019

$6,265

$5,071

$8,066

$4,349**

$4,928

Operating and Maintenance Expenses
Meana
Standard Error
Median

$4,540

$4,338

$4,585

267

100

148

$4,035

$3,751

$4,114

$3,901

$4,396

Replacement Reserve Deposit [max(actual, required)]
Meana

$311

Standard Error
Median

$222**

$331

$376**

$264

$219

$160

$244

$263

$230

Total Debt Service (including MIP, supplementary loans and op loss loans)c
Mean
Standard Error
Median

$2,201

$2,930*

$2,039

$859**

$3,760

359

130

$1,512

$1,859

$1,334

$614

$3,359

Annual Financial Statements for 1992-1995.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Annual Net Cash Flow
Exhibit 4-3 brings together revenues and expenses to display annual net cash flow which
equals revenues less operating expenses, replacement reserve deposits, and debt service.

Exhibit 4-3
ANNUAL NET CASH FLOW
(IN 1995 DOLLARS PER 2-BEDROOM EQUIVALENT UNIT)
Total

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

5,943
59%b

4,076
41%b

35%**

13%

Total Properties
Percent of Total Properties

12,243
100%

2,224
18%

10,019
82%

Negative Net Cash Flowa

25%

<-$1,000

-$1,000 to <-$500

-$500 to < -$250

-$250 to <$0

Positive Net Cash Flow a

75%

66%**

87%

$0 to <$250

$250 to <$500

$500 to <$1,000

$1,000 to <$2,500

$2,500+

Statistics on Annual Net Cash Flow
Meana

$593

$487

$617

$283**

$1,105

Standard Error

341

Median

$388

$521

$347

$162

$742

Column sums may not add to 100% due to rounding.

Source:

Calculated from Financial Data.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Annual net cash flow averaged $593 per unit. However, there were substantial differences in
cash flow across assistance categories. Newer assisted properties were in the best financial
condition among the three categories of properties and older assisted properties were in the
worst financial condition. In particular:
•

Average annual net cash flow was highest in newer assisted properties averaging
$1,105 (median $742) per unit. Eighty-seven percent of newer assisted properties
had positive annual net cash flow, including over 40 percent with more than
$1,000 annual net cash flow per unit.

•

Annual net cash flow averaged $487 (median $521) per unit in unassisted
properties. Three quarters of the unassisted properties had positive annual net
cash flow.

•

The financial condition of older assisted properties was the worst among the three
categories, with cash flow averaging $283 (median $162) per unit. Over one third
(35 percent) of older assisted properties had negative annual net cash flows.

Change in Cash Flow Since 1989
In this section, we examine the changes in financial condition of the stock between 1989 and
1995, using average annual net cash flow as the indicator of financial condition. Then, we
analyze the components of revenue and expenses to identify the sources of change in cash
flow between the two periods. The numbers differ slightly from Exhibit 4-3 because the
analysis is based on the comparison sample, which includes only the 504 properties that were
inspected in both 1989 and 1995. All data were weighted to reflect the universe of properties
that were still insured in 1995, and all 1989 dollars were inflated to 1995 based on the change
in the CPI over this interval.
Exhibit 4-4 shows annual net cash flow indicators for 1995 and 1989.9 For both time periods,
the table presents mean and median cash flow per 2-bedroom unit as well as the percentage of
properties that had negative annual net cash flow, low positive annual net cash flow ($0$500), high positive annual net cash flow ($500-$1,000), and very high positive annual net
cash flow (>$1,000).

We have adjusted the 1989 data for inflation using an adjustment factor of 1.1986, which is based on changes in the housing component of
the CPI from 1989 to 1995.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-4
COMPARISON OF NET CASH FLOW
(IN 1995 DOLLARS BY ASSISTANCE CATEGORY)
BASED ON COMPARISON SAMPLE
Total

Assisted

Total
12,243
100%

Total Properties
% of Properties

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Mean Cash Flow per 2 BR
1995

$594

$487

$618

$281

$1,110

1989

$444

$158

$507

$265

$859

Median Cash Flow per 2 BR
1995

$362

$521

$338

$144

$732

1989

$232

$57

$246

$69

$708

Percentage of Properties with Negative Cash Flow (<$0 per 2 BR)
1995

25%

24%

33%

13%

1989

30%

44%

27%

39%

10%

Percentage of Properties with Low Positive Cash Flow ($0-$500 per 2 BR)
1995

32%

23%

34%

42%

23%

1989

36%

22%

38%

45%

29%

Percentage of Properties with High Positive Cash Flow ($500-$1,000 per 2 BR)
1995

17%

16%

14%

21%

1989

15%

17%

28%

Percentage of Properties with Very High Positive Cash Flow (>$1,000 per 2 BR)

1995

27%

35%

25%

12%

44%

1989

19%

25%

18%

33%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Note:

Column sums may not add to 100% due to rounding.

Source:

1995 data: Annual financial statement data for 1992-1995.
1989 data: 1990 Study Analysis File.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Overall, the stock experienced an improvement in financial condition, with the mean annual
net cash flow increasing by 34 percent from $444 to $594 and the median from $232 to $362
per unit. The percentage of properties with negative annual net cash flow dropped from 30
percent to 25 percent, and the percentage with very high positive annual net cash flow
increased from 19 percent to 27 percent.
• The biggest improvement in financial condition was among unassisted properties. Mean
annual net cash flow more than tripled from $158 to $487 and the median increased by
more than ninefold from $57 to $521 per 2-bedroom unit. The percentage of properties
with negative annual net cash flow dropped from 44 percent to 25 percent, and the
percentage of properties with high and very high positive annual net cash flow increased
from 9 percent to 17 percent and 25 percent to 35 percent, respectively.
• Older assisted properties saw a small increase (6 percent) in mean annual net cash flow
(from $265 to $281 per 2-bedroom unit), along with a moderate decrease in the number of
properties with negative annual net cash flow and substantial increases in properties with
high and very high positive annual net cash flow.
• Newer assisted properties, on the other hand, showed an increase in properties with
negative annual net cash flow (from 10 percent to 13 percent), and a decrease in properties
with low and high positive annual net cash flow, but a sizable increase in properties with
very high positive cash flow. At the same time, the mean cash flow among newer assisted
properties rose from $859 to $1,110, while the median rose only from $708 to $732 per 2bedroom unit, indicating that a small number of newer assisted properties were enjoying
dramatically higher annual net cash flows.
The following table (Exhibit 4-5) shows the property-level change in net cash flow, adjusted
for inflation.10 Overall, 37 percent of properties experienced a decrease in net cash flow in the
five-year period, while 63 percent saw an increase in net cash flow. Assisted properties were
more likely to experience a decrease in net cash flow and less likely to experience an increase
in net cash flow than were unassisted properties. Among assisted properties, older and newer
assisted properties showed similar patterns of change in cash flow.
Assisted properties (both older and newer) were more likely to experience moderate change
(either increase or decrease) in cash flow, while unassisted properties were more likely to
experience more dramatic change, with 15 percent of unassisted properties showing decreases
of over $800 per unit and 32 percent showing increases of $800 or more per unit, compared
to 7 percent and 17 percent, respectively, for assisted properties.

As noted in Chapter 3, property-level physical condition data was not intended to be measured with sufficient precision to provide reliable
property-level change estimates. In contrast, the financial data are based on actual data and thus, property-level change in financial
condition can be analyzed.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-5
CHANGE IN ANNUAL NET CASH FLOW FROM 1990 TO 1995
(PER 2 BEDROOM EQUIVALENT UNIT IN 1995 DOLLARS)
BASED ON COMPARISON SAMPLE
Total

Assisted

Total

Unassisted

Assisted

Older Assisted

Newer Assisted

Total Properties
% of Properties

12,243
100%

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Annual Net Cash Flow
Decreased

37%

30%

39%

40%

38%

By >$800 per 2 BR

15%

By $400-$800 per 2 BR

10%

By $200-$400 per 2 BR

By <$200 per 2 BR

13%

15%

16%

13%

Annual Net Cash Flow
Increased

63%

70%

61%

60%

62%

By <$200 per 2 BR

15%

17%

18%

14%

By $200-$400 per 2 BR

13%

10%

14%

11%

18%

By $400-$800 per 2 BR

15%

22%

14%

15%

13%

By $800 or more per 2 BR

29%

32%

17%

16%

18%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Note:

Column sums may not add to 100% due to rounding.

Source:

1995 data: Annual financial statement data for 1992-1995.
1989 data: 1990 Study Analysis File.

Changes in Components of Revenues and Expenses
In this section, we examine changes in the components of income and expenses to better
understand the changes in cash flow. The following table (Exhibit 4-6) shows the percentage
change in the mean values of components of revenues and expenses between 1989 and 1995.11

Percent changes were calculated using the comparison sample of properties that were included in both 1989 and 1995.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-6
CHANGES IN COMPONENTS OF REVENUE AND EXPENSES BETWEEN 1989 AND 1995
BASED ON COMPARISON SAMPLE
Total

Assisted

Total
Total Properties
% of Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

-10%

-26%

67%

22%

10%

-1%

11%

16%

Total Revenues (net of vacancies)
Percent Change

Tenant-Paid Rents
Percent Change

-10%

-16%

Tenant Assistance Payments
Percent Change

35%

N/A

35%

Vacancy Loss (As a Percent of Rent Revenue)
Percent Change

-7%

-27%

Total Annual Expenses
Percent Change

Operating and Maintenance Expenses
Percent Change

14%

19%

13%

Replacement Reserve Deposit, max (actual, required)
Percent Change

16%

18%

29%

Total Debt Service (including MIP, supplementary loans, and op loss loans) per 2 BR
Percent Change
a

-17%

-16%

-19%

-25%

-17%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Source:

1995 data: Annual financial statement data for 1992-1995.
1989 data: 1990 Study Analysis File.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Revenues
Across the entire stock average revenues rose (in real terms) by 4 percent between 1989 and
1995. This reflects a decrease in tenant-paid rents which was more than offset by increases in
tenant assistance payments and decreases in vacancy losses.12
• Unassisted properties experienced an increase of 6 percent in average revenues, resulting
from a 5 percent increase in tenant-paid rents and a 27 percent decrease in the average
vacancy loss.
• Total revenues rose by 2 percent in newer assisted properties and by 5 percent in older
assisted properties. In both cases this is a result of decreases in tenant paid rents, which
was more than offset by increases in tenant assistance payments.
• The decrease in tenant paid rents was most dramatic in newer assisted properties, where
on average they decreased by 26 percent, likely reflecting a decrease in real incomes
among residents in these properties. In older assisted properties, tenant-paid rents
decreased by 10 percent.
• The decreases in tenant-paid rents were more than offset by increases in tenant assistance
payments in both older and newer assisted properties. Although tenant assistance
payments were significantly higher in newer assisted properties in both periods, the
percentage increase was higher in older assisted properties.
Expenses
Average total expenses (in real terms) stayed nearly the same over the period, increasing by 2
percent. This is a result of increases of about 15 percent in both operating and maintenance
expenses and in deposits to the reserve for replacement account that were nearly offset by a
17 percent reduction in the real cost of debt service. Debt service payments dropped
substantially in real terms even though remaining constant in nominal dollars. That is, the
current value of debt service payments decreased in real dollars results from inflating 1989
payments to 1995 dollars even though actual debt services payments remained constant.
• Among unassisted properties, a large increase in operating and maintenance expenses (19
percent) and a small increase in deposits to the reserve for replacement (4 percent) more
than offset the decrease in debt service (of 16 percent), resulting in a 2 percent increase in
average total expenses.

For comparability to the 1995 study, interest reduction payments in the Section 236 properties were excluded in the
calculation for the 1989 values. For the 1990 Report, they were included as part of the revenue and debt service.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

• Among older assisted properties, the 11 percent increase in operating and maintenance
expenses and the 29 percent increase in replacement reserve deposits exceeded the 25
percent decline in real debt service, resulting in a 5 percent increase in average total
expenses.
• In contrast to the two other categories, among newer assisted properties, the real drop in
debt service (of 17 percent) more than offset the 16 percent increase in operating and
maintenance expenses, resulting in lower total expenses. There was no change in the
average deposit to reserve for replacement account in newer assisted properties.
In summary, the stock as a whole showed a moderate increase in cash flow, with a decrease in
the number of properties with negative cash flow and an increase in properties with high or
very high positive cash flow. Nearly two-thirds (63 percent) of properties saw an increase in
cash flow. The stronger cash flow was the result of increased total revenues (from higher
tenant assistance payments and reduced vacancy loss) which outweighed a modest increase in
expenses (primarily from higher operating and maintenance expenses).
Unassisted properties experienced the strongest increase in cash flow, with a large decrease in
the percentage of properties with negative cash flow and a large increase in properties with
high positive cash flow. Seventy percent of unassisted properties experienced increased cash
flow. The higher cash flow resulted from an increase in revenues from higher tenant-paid
rents and lower vacancy losses, which more than offset a small increase in expenses.
Older assisted properties saw a small increase in mean cash flow, with a moderate decrease in
the number of properties with negative cash flow and an increase in properties with high or
very high cash flow. The increased average cash flow was the result of higher revenues from
higher tenant assistance payments, which offset higher expenses.
Among newer assisted properties, the mean cash flow increased moderately, with 62 percent
of properties showing an increase in cash flow. However, the number of newer assisted
properties with negative cash flow also increased. The increased cash flow resulted from
increased revenues (which prevailed despite a large decrease in tenant rents because of an
offsetting increase in tenant assistance payments) and a decrease in expenses from lower real
debt service.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

4.2

Resources Available to Cover the Current Backlog of Physical
Needs and Annual Accrual of Needs

An important factor in a property’s long-term viability is its having adequate reserve funds.
This section examines the size of a property’s reserve fund balances relative to its backlog of
physical needs. There are three components of resources available to fund major repairs and
replacements:
• Reserve for Replacement: All HUD-insured or held properties are required to establish
and fund a reserve for replacements account. This is their primary resource for funding
major repairs and replacements.
• Other Reserves: Some properties have established special-purpose reserve accounts such
as painting reserves and general reserves. Few properties in our sample reported any such
reserve accounts.
• Residual Receipts Accounts: Non-profit owners and certain for-profit owners are
restricted by their mortgage regulatory (or other) contracts in their being able to take
profits from the property’s annual surplus cash after expenses. They are required to
deposit non-distributable surplus cash into a residual receipts account. Non-profit owners,
and certain owners who have received special remedial assistance or assistance under a
workout, may not distribute any profit. Limited-dividend owners may distribute only a
restricted amount, and only under stipulated conditions. While residual receipts accounts
are not reserves for the property, HUD may require owners to contribute residual receipts
funds (if any) for repairs in the case of physically deteriorated properties.
Exhibit 4-7 shows available balances in these funds.
• In most properties, the replacement reserve was the primary source of funds available to
cover physical needs backlogs, with an average balance of $1,303 per unit. Average
balances were lowest in the unassisted properties ($755 per unit) compared with the
assisted properties ($1,424).13
• Residual Receipts apply to a minority of properties, so while their overall impact may be
small (mean $206), they may be significant for particular properties.14

The values for all reserves balances were obtained using the most recent year of data available. Data were obtained from the Annual
Financial Statement file provided by HUD. Reserve for replacement balances, other reserve balances and residual receipts were not entered
into HUD’s Multifamily Data Warehouse.

Residual receipts were provided in the Annual Financial Statement file provided by HUD and were reported only for 1993 and 1994.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

• Few properties reported “other reserves” (either general reserves or painting reserves), so
this category has been excluded in Exhibit 4-7. However, for particular properties, they
may be a significant resource.15
• Total reserve balances averaged $1,643 per unit. As expected, given that most reserves
are in the reserve for replacement account, assisted properties had higher reserve balances
on average compared with unassisted properties ($1,831 compared with $797).
Exhibit 4-7
RESOURCES FOR COVERING PHYSICAL NEEDS
(IN 1995 DOLLARS PER 2-BEDROOM EQUIVALENT UNIT)
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Reserve for Replacement Balance
Meana

$1,303

$755**

$1,424

$1,327*

$1,565

Standard Error

101

Median

882

$456

$1,065

$994

$1,171

Residual Receipts
Meana

$206

0**

$252

$247

$260

Standard Error

128

Median

Percent >$0

27%

31%

38%

18%

Total Reserves
Meana

$1,643

$797**

$1,831

$1,766

$1,924

Standard Error

119

131

Median

$1,129

$479

$1,293

$1240

$1,363

Annual Financial statements.

General reserve information was provided on HUD’s FOMNS backup tapes for 1992 for 36 of the sample properties and from HUD’s
AFS file for 1995 (11 properties). Paint reserves for 1992 are from HUD’s backup FOMNS tape (for 256 properties, and from HUD’s
AFS file for 1995 (17 cases). Mean values are not provided in the exhibit, but are included in the calculation of total reserves.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Low reserve balances in themselves may not indicate problems. Reserve balances may be low,
for example, because a major repair program was recently completed. However, in properties
having both high backlogs of physical needs and low reserves, problems may be indicated. It
is important to examine both of these together. An additional source of concern is when
properties have physical needs backlogs and are not applying available reserves to address
these problems. This may be an indicator of management problems in the property.
We present two indicators of the properties’ financial ability to cover their backlogs of
physical needs:
• Backlog Coverage Ratio. This ratio compares available resources with the backlog of
physical needs. Available resources are defined as the sum of any amount by which the
reserve for replacement balance exceeds two years’ worth of annual reserve deposits,16
plus residual receipts balance, plus other reserve balances. Available resources are then
divided by the amount of the backlog.
• Unfunded Backlog. The unfunded backlog of physical needs is the total backlog reduced
by available resources (as defined above).
Exhibit 4-8 shows the backlog coverage ratio for the insured stock. The exhibit shows that a
large portion of the stock does not have sufficient resources to cover their backlogs of
physical needs.
• Thirty-six percent of the stock had sufficient resources to cover their physical needs
backlogs, including 13 percent with no backlog, and 23 percent with a positive backlog,
but with resources available to address the backlog. While the proportion of properties
that had sufficient resources was similar in unassisted and assisted properties the reasons
were different. In 21 percent of unassisted properties there was no backlog, and 15
percent had backlogs but also had resources to address the backlog. In contrast, 24
percent of assisted properties had backlogs, and also had resources to cover the backlog,
and only 11 percent had no backlog.
• The problem was most severe in the older assisted properties where only 30 percent had
sufficient resources, and least severe in the newer assisted properties where 42 percent
had. Twenty-seven percent of newer assisted properties had backlogs of physical needs,
but also had resources to cover the backlog (compared with 22 percent of older assisted

Retaining two years’ worth of deposits is in keeping with HUD’s general loan servicing practices. Had we instead assumed that properties
could use their entire reserves would have added only a small amount on average and would make little difference in the ability of most
properties to cover their backlogs.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

properties) and 15 percent had no backlog (compared with only 8 percent of older assisted
properties).
• The majority of properties (65 percent), across all assistance categories lacked sufficient
resources to cover their backlogs of physical needs. This includes 13 percent of properties
with backlogs and no available resources, and 30 percent of properties with insufficient
resources to cover one quarter of their backlogs.
• The ability to cover backlogs has declined since 1989. At that time 45 percent of the stock
either had no backlog (20 percent) or sufficient resources to cover their existing backlog
(25 percent). In 1995 only 35 percent of the stock had sufficient resources. Driving this
decreased ability to cover backlog of physical needs is the dramatic increase in backlogs
(See Section 2.1 above), rather than a decrease in resources. On average, available
resources increased by over 40 percent, but backlogs increased by over 60 percent.
Exhibit 4-8
ABILITY TO COVER REPAIR NEEDS -BACKLOG COVERAGE RATIOc
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Backlog Coverage Ratio / Available Resources/Backlog
Insufficient Resourcesa

64%

65%

69%**

57%

Backlog >0, & no available
resources

13%

15%

13%

16%

Ratio 0 to <0.25

30%

32%

30%

34%

24%

Ratio 0.25 to- <0.5

11%

10%

11%

15%

Ratio 0.5 to <1

10%

11%

Sufficient Resourcesa

36%

35%

30%**

42%

Ratio > = 1

23%

15%

24%

22%

27%

Backlog <$10

13%

21%

11%

15%

Column sums may not add to 100% due to rounding.

Source:

1995 Physical Inspection Data and Costing Program, Financial Data.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-9 shows the resulting unfunded backlog of physical needs.
•
•

•

The mean unfunded backlog of physical needs was $2,630, or 81 percent of the total
backlog.
The median unfunded backlog was only $684, indicating that a small portion of the
stock was responsible for a large portion of the unfunded backlog.
Almost one third (32 percent) of the stock had unfunded backlogs of over $2,000 per
unit. As with most other resource problems, high levels of unfunded backlog were
most common in the older assisted portion of the stock. The mean unfunded backlog
for older assisted properties was $3,323 (compared with $1,134 for unassisted and
$2,437 for newer assisted). Forty-four percent of older assisted properties had over
$2,000 of unfunded backlog per unit, compared with 15 percent of unassisted and 25
percent of newer assisted properties.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-9
ABILITY TO COVER BACKLOG - UNFUNDED BACKLOG
(IN 1995 DOLLARS PER 2-BEDROOM EQUIVALENT UNIT)
Total

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

Total Properties
Percent of Total Properties

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

$0 a

35%

36%

35%

30%*

41%

$0 - <$500

12%

22%

10%

$500 - <$1,000

13%

$1,000 - <$2,000

12%

21%

10%

12%

$2,000 - <$5,000

17%

11%

19%

22%

13%

$5,000 - <$7,500

$7,500+

11%

13%

Statistics on Unfunded Backlog
a

Mean

$2,630

$1,134**

$2,962

$3,323*

$2,437

Standard Error

194

233

231

270

410

Median

$684

$332

$817

$1,324

$540

Statistics on Backlog of Physical Needs
Meana
Standard Error
Median

$3,236

$1,427**

$3,638

$3,929

$3,214

203

255

240

276

430

$1,452

$545

$1,661

$2,096

$1,324

Sums may not add to 100% due to rounding.

Source:

1995 Physical Inspection Data, Costing Program and Financial Data.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Ability to Cover Annual Accrual
Another important factor in a property’s long-term viability is its ability to cover ongoing
accrual of physical needs. As discussed in Chapter 3 above, we have estimated the average
annual accrual of physical needs for each property. There are two potential sources of funds
available to cover these accrual costs:
•

Annual Deposits to the Reserve for Replacement Accounts: All HUD-insured or
held properties are required to make monthly deposits to the reserve for replacement
account. As discussed above, the amount assumed to be available to cover ongoing
needs is the maximum of actual deposits as reported in the annual financial statements
and the required deposits, which were approximated as 0.5 percent of the original
mortgage.

•

Positive Annual Net Cash Flow: Properties that have positive net cash flow after
covering all operating and maintenance expenses, mortgage repayment and deposits to
reserve accounts may use remaining funds to cover ongoing accruals.

It appears (see Exhibit 4-10) that even if the current backlog of physical needs were to be
addressed, only about one fourth of insured properties would be able to keep up with their
ongoing accrual needs. The average unfunded accrual was $610, and the median was $586.
Differences in the ability to cover accruals results from different resources, not from different
levels of accrual. As shown in Exhibit 3-7 (and repeated in Exhibit 4-10 for convenience),
annual accruals averaged $1,437 across the entire stock, and there were no significant
differences across assistance categories.
•

Forty-one percent of newer assisted properties had no unfunded accrual, compared
with only 14 percent of older assisted properties and 26 percent of unassisted
properties.

•

Older assisted properties contribute more on average to the reserve for replacement
account, but the other source of funds for covering accruals, ongoing net cash flow is
much lower in this group of properties compared with other types of properties.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-10
ABILITY TO COVER ACCRUAL - UNFUNDED ACCRUAL
(IN 1995 DOLLARS

PER 2-BEDROOM EQUIVALENT UNIT)

Total

Assisted

Total
Total Properties
Percent of Total Properties

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

12,243
100%

Unfunded Accrual
$0a

25%

26%

25%

14%**

41%

$0 - <$500

21%

19%

21%

20%

23%

$500 - <$1,000

31%

35%

24%

$1,000 - <$2,000

22%

25%

21%

28%

11%

$2,000+

Statistics on Unfunded Accrual
Meana

$610

$634

$604

$737**

$411

Standard Error

Median

$586

$649

$584

$766

$276

Statistics on Annual Accrual
Meana

$1,437

$1,514

$1,420

$1,404

$1,443

Standard Error

19.6

54.1

20.7

28.2

29.8

Median

$1,362

$1,419

$1,331

$1,301

$1,399

Sums may not add to 100% due to rounding.

Source:

1995 Physical Inspection Data, Costing Program and Financial Data.

In order to further explore the impact of unfunded accruals on property finances, Exhibit 4-11
presents an alternative net cash flow measure that assumes that properties fund the reserve for
replacement to a level that covers average accruals. As the exhibit shows, if properties were
to fund the reserve for replacement at a level high enough to cover average annual accrual,
three quarters would have negative net cash flow, and only one quarter percent would have
positive net cash flow. As with many other financial indicators, newer assisted properties
were in the best position and older assisted properties in the worst.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Eighty-five percent of older assisted properties would have negative cash flow if they funded
the reserve for replacement at a level that covered average annual accrual compared with 59
percent of newer assisted properties, and seventy-five percent of unassisted properties.
Exhibit 4-11
ALTERNATIVE CASH FLOW
(ASSUMES THAT RESERVE DEPOSIT EQUALS ACCRUAL)
(IN 1995 DOLLARS PER 2 BR EQUIVALENT UNIT)
Total

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

Total Properties
Percent of Total Properties

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Negative Alternative Net
Cash Flowa

75%

85%**

59%

<-$1,000

29%

28%

29%

39%

15%

-$1,000-<-$500

26%

28%

26%

28%

20%

-$500-< -$250

11%

10%

15%

-$250 - <$0

Positive Alternative Net
Cash Flow a

25%

15%**

41%

$0-<$250

11%

$250-<$500

$500 -<$1,000

$1,000-<$2,500

12%

$2,500+

Statistics on Alternative Annual Cash Flow
Meana

$-532

$-805

$-471

$-745**

$91

Standard Error

331

Median

$-628

$-649

$-615

$-843

$-284

Sums may not add to 100% due to rounding.

Source:

Calculated from Financial Data.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

4.3

Section 8 Assistance

An important source of financial assistance for HUD-insured properties is the project-based
Section 8 program, which includes the New Construction and Substantial and Moderate
rehabilitation programs, the Loan Management Set-Aside (LMSA) Program, the Property
Disposition Program, and the Preservation Program.17 Overall, 8744 properties receive some
sort of Section 8 assistance.
Exhibit 4-12 presents information on properties that receive Section 8 assistance. Newer
assisted properties generally receive assistance for nearly all property units (mean is 96
percent, median 100 percent). In contrast, LMSA assistance often covers only a portion of
units (mean is 80 percent, median 98 percent). The contract amount per assisted unit
averaged $4,833 per year in the newer assisted properties, and $2,224 in the LMSA
properties.18 Due to the annual increases in contract amount, the payments currently received
are substantially higher than the original contract amounts. Newer assisted properties
received on average $6,684 per assisted unit in tenant assistance payments, while LMSA
properties received on average $3,440.
Across all categories of Section 8, assistance contracts are being renewed for shorter periods
of time. In 1995, over half of all contracts (61 percent) were up for renewal within four years
(through 1999).

Under the Section 8 New Construction/Substantial Rehabilitation Programs private developers own and construct or rehabilitate housing
that they then rent to lower-income tenants. The maximum rents charged by owners to tenants are restricted. The difference between 30
percent of a tenant’s adjusted income and the rent being charged for the unit is paid to the owner by HUD. The other large Section 8
program, LMSA, is a form of rent supplement that was available to help troubled multifamily properties. Like the New Construction
Program, tenants pay 30 percent of their incomes towards rent, and HUD pays the owners the rest.

The contract amounts in the newer assisted stock are substantially lower than the assistance payment amount reported in the annual
financial statements. This is likely due to the fact that assistance payments escalate annually based on the AAF.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

Exhibit 4-12
SECTION 8 ASSISTANCE

Number of Properties

New
Construction/
Substantial
Rehab

LMSA
(including Rent
Sup/RAP
conversion)

Preservation

Property
Disposition

Rent
Sup/RAP19

4,076

4,011

257

134

265

Percent of Units Assisted
Mean

96%

80%

92%

100%

54%

Median

100%

99%

100%

63%

$4,563

$2,316

$3,340

$5,321

$5,527

Mean Contract Assistance Amount20
Per Assisted Unit

$4,833

$2,224

$4,113

Current Tenant Assistance Payment
Per Assisted Unit

$6,684

$3,440
Next Renewal Year

Source:

Mean

2001

1997

1998

1999

2003

1995-1996

40%

39%

34%

17%

1997-1999

25%

55%

31%

50%

46%

2000-2004

73%

30%

2005+

16%

38%

Multifamily Data Warehouse, Contracts File, 1992-1995 Annual Financial Statements.

Reflects 154 properties that were reported as rent sup/ rap in 1995 and 111 sample properties that were reported as active rent sup/rap in
1989 and had no Section 8 information in 1995.

Assistance amount was missing for 125 sample properties and was imputed per assisted unit based on the median ratio of assistance
amount/rent by Section 8 type.

Abt Associates Inc.

4.0 Financial Condition of HUD-Insured
and Held Properties

5.0 Measures that Combine Physical and Financial
Conditions
This section develops and presents comprehensive measures of property condition for
multifamily rental housing with HUD-insured (or held) mortgages that take into account both
the physical and financial condition of properties. We first look at a cross tabulation of annual
net cash flow by backlog of physical needs. Our goal in this analysis was two-fold: (1) to see
how our primary financial measure (annual net cash flow) maps with our primary physical
needs measure (backlog of physical needs); and (2) to determine the risk profile of these
properties, that is, extent to which properties appear to be either physically or financially at
risk or both.
By incorporating physical and financial measures, we also devised a combined “Distress
Index” which enables us to compare properties and assess property condition using a single
quantitative measure. We apply the Distress Index to the universe of multifamily rental
housing with HUD- insured (or held) mortgages, classify properties as sound, stressed, and
distressed, and describe the characteristics of properties in each of these categories. We then
describe changes in the multifamily stock since 1989, as measured by the Distress Index.
Each of the two indicators has advantages. The risk profile places each property into one of
four risk categories, but does not rank properties. The distress index on the other hand,
provides a single, quantifiable, numerical ranking of properties, but it focuses on properties’
financial capacity to meet current and future expenses. It does not distinguish between
properties that are, in fact, using available funds to address their backlog of physical needs and
those that are not. Thus, while the two indicators track well together, they each provide
valuable information.
Risk Profile of Properties
The summary table below (Exhibit 5-1) shows key differences in the physical and financial
condition of the housing stock and highlights areas of potential risk by assistance category.
(Exhibit 5-2 displays this information graphically.)
•

At one end of the spectrum are properties that we labeled “minimally risky”.
These are properties with positive annual net cash flow and normal backlogs of
physical needs (<$1,500). About half of unassisted (51 percent) and newer
assisted (50 percent) properties fell into this category of “minimally risky”,
compared to fewer than a third (30 percent) of older assisted properties.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

•

At the other end of the spectrum were properties that we labeled as “high risk”.
These are properties with negative annual net cash flow and high backlogs of
physical needs (>$1,500). Owners of properties in this category are both not
meeting current operating costs with current revenues and are falling behind in the
upkeep of the property. Only 9 percent of newer assisted and unassisted
properties were categorized as “high risk”, compared with 20 percent of older
assisted properties.

Exhibit 5-1
BACKLOG OF PHYSICAL NEEDS BY ANNUAL NET CASH FLOW
SUMMARY TABLE BY ASSISTANCE CATEGORY

High Positive Annual Net
Cash Flow (>$500)

Low Positive Annual Net
Cash Flow ($0-$500)

Negative Annual Net
Cash Flow
(<$0)

Low/
Normal
Backlog
<$1,500

Moderate/
Severe
Backlog
>$1,500

Low/
Normal
Backlog
<$1,500

Moderate/
Severe
Backlog
>$1,500

Low/
Normal
Backlog
<$1,500

Moderate/
Severe
Backlog
>$1,500

minimally
risky

management
risk

minimally
risky

moderately
risky

high
risk

All Categories

25%

18%

15%

17%

11%

15%

100%

Unassisted

34%

18%

17%

16%

100%

Older Assisted

12%

13%

18%

24%

13%

20%

100%

Newer Assisted

39%

26%

11%

100%

Note:

Rows may not sum to 100% due to rounding.

Source:

1992-1995 Annual Financial Statements and 1995 Physical Inspection Data and Costing Program.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Total

Exhibit 5-2
RISK PROFILE OF INSURED PROPERTIES
All Categories

Unassisted Properties

Older Assisted Properties

Newer Assisted Properties
Minimally Risky

Minimally Risky

Management Risk
13%

50%

30%

9%
37%

20%

26%
High Risk

Moderately Risky

High Risk

16%

Management Risk
Moderately Risky

Minimally Risky

Management Risk

Moderately Risky

High Risk

Derived from Exhibit 5-1.

Source:

•

In between these two extremes were properties that we labeled as “moderately
risky”. These are properties that had low positive annual net cash flow and high
backlog of physical needs, or had negative annual net cash flow and normal
backlog of physical needs. This group of properties is either not covering
operations from current revenues, or not keeping up with physical repairs. This
situation is not sustainable over time, and will likely lead to deteriorating quality of
housing for residents. Thirty-seven percent of older assisted properties were
classified as a “moderately risky”, compared with 22 percent of unassisted and 16
percent of newer assisted properties.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

•

Another category were properties we labeled as “management risk”. These are
properties with high positive annual net cash flow but also with high physical needs
backlogs. This combination of characteristics indicates a failure to address capital
needs (as evidenced by a high backlog) resulting not from lack of resources, but
potentially from poor management. Over one quarter (26 percent) of newer
assisted properties fell into the category of “management risk”, compared with 18
percent of unassisted and 13 percent of older assisted properties.

Distress Index
The second method for incorporating both physical and financial condition, involved
developing a Distress Index that reflects a property’s annual net cash flow, other financial
resources, and backlog of physical needs. The Distress Index measures a property’s financial
capacity to meet current expenses, set aside reserves for future physical needs, and undertake
a repair program to address its backlog of physical needs. The index is used to classify
properties as sound or potentially distressed, and to identify the degree of potential distress, as
measured by the extent to which properties lack required resources. In this section, we show
the development of the Distress Index based on annual net cash flow, minus unfunded backlog
of physical needs, adjusted for vacancy.1 We then compare the Distress status of properties
with the Risk Profile presented above, present characteristics of distressed properties, and
compare the results to the results of the 1990 Study on the same set of properties to gauge the
change in distress level.
Development of the Distress Index
The Distress Index is computed by taking:
(a) Net Cash Flow
(b) Minus the annual amortization of the cost of remedying the Unfunded Backlog of
Physical Needs
(c) Plus added rent from improving Vacancy Losses
The computation begins with net cash flow, which measures a property’s capacity to meet
current expenses and make deposits to its replacement reserves account. Net cash flow is

This measure has been used for comparability with the 1990 Study (presented in Wallace, et al., Assessment of the HUD-Insured
Multifamily Housing Stock, 1993). One could argue that a comprehensive measure would incorporate average annual unfunded accrual
into the Distress Index as well. This version of the Distress Index was tested and results are very similar to the version of the Index used.
As a test of the Distress Index defined in the text, we looked at the 1995 status of properties that were included in both studies, in 1989 25
percent were categorized as distressed, 15 percent as stressed and 61 percent as sound. In 1995, 15 percent of the properties that had been
categorized as distressed in 1989 were HUD-held, as were 8 percent of the stressed properties and 6.5 percent of the sound properties. The
difference in the HUD-held status in 1995 is statistically significant at the 95 percent confidence level. In other words, properties that were
classified as distressed in 1989 were significantly more likely to be HUD-held in 1995 compared with properties that had been classified as
sound.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

then reduced by the amortized cost of remedying the unfunded backlog of physical needs,
which represents the annual cost of undertaking a repair program. This simulates an owner’s
likely attempt to spread the remedial costs over time by spreading the work over time or by
spreading payments by borrowing. The final step in computing the Distress Index is to add
back a portion of a property’s excess vacancy loss to represent the higher revenues resulting
from improved operations and physical condition. These elements of the Distress Index are
discussed below.
Annual Net Cash Flow
Annual net cash flow is computed as explained in Chapter Three, taking a weighted average
over the most recent three years (expressed in 1995 dollars per 2BR unit) of both revenues
and expenses.
Annual Net Cash Flow (Weighted 3-Year Average) =
Total Revenue (Weighted 3-Year Average)
Minus Operating and Maintenance Expenses (Weighted 3-Year Average,
including expenses for administration, operations and maintenance, utilities,
taxes, and insurance)
Minus Mortgage Debt Service (Interest, Principal and Mortgage Insurance
Premium as required by mortgage)
Minus Replacement Reserve Deposit (using the greater of the property’s actual
deposit or an amount equal to 0.5 percent of the original mortgage)
Amortized Cost of Remedying the Unfunded Backlog of Physical Needs
At this step, a property’s backlog of physical needs for replacements and non-routine repairs
is taken into account. As explained in Chapter Three, a property’s unfunded backlog of
physical needs is its total backlog less available resources from the replacement reserve fund,
special reserve account, and residual receipts account. Where resources exceed the total
backlog of physical needs, there is no unfunded backlog.
Amortized cost of remedying the unfunded backlog of physical needs =
Annual debt service on a loan amount equal to the unfunded backlog cost
(20-year term at 9 percent interest)
Where
Unfunded Backlog Cost = Total Backlog Cost - Available Resources
(or 0 if resources exceed the total backlog)
and where Available Resources =
Replacement Reserve Balance in excess of 2 years’ annual deposits
Plus Residual Receipts Account Balance
Plus Other Reserve Account Balances (such as painting reserves)
Added Rent from Improving Vacancy Losses

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

In computing the Distress Index, the modified net cash flow figure is further adjusted by
adding back a portion of the property’s excess vacancy loss. This represents the income that
would result if improved management and physical condition brought a property’s excessive
vacancy loss closer to the norm for the property’s assistance category.
This computation is based on the assumption that properties whose vacancy losses rank in the
highest 25 percent (among properties in their assistance category) will be able to reduce their
vacancy losses down to the 75th percentile; that properties with vacancies between the median
and the 75th percentile will be able to reduce vacancy losses to the median level for their
assistance category; and that for all other properties, vacancy losses will remain as they are.2
Added rent from improving vacancy loss =
(1) For properties with vacancies in excess of the 75th percentile of vacancy losses
for properties in the same assistance category
Current vacancy loss - 75th Percentile Vacancy Loss (for properties in the
same assistance category); i.e., bring vacancy losses down to the 75th
percentile
(2) For properties with vacancies between the median and 75th percentile of
vacancy losses for properties in the same assistance category
Current vacancy loss - Median Vacancy Loss (for properties in the same
assistance category); i.e., bring vacancy losses down to the median
(3)
For all other properties—No adjustment
The net result of these three factors yields the Distress Index, which is a modified version of
net cash flow. A positive Distress Index indicates a sound property that can meet ongoing
operations and cover its backlog of physical needs from internal funds. A property with a
negative Distress Index is not able to cover all ongoing operations and repair of the physical
needs backlog from internal funds. This situation is not sustainable over time because the
property is either falling behind in its financial obligations (mortgage payments) or in property
maintenance and repairs. As was done in 1989, properties with a low negative index (Distress
Index of -$250 to $0) have been categorized as a “stressed”, and properties with a highly
negative index (Distress Index <-$250) have been categorized as a “distressed”.3

For unassisted properties, the top quartile of vacancy loss was 7.0 percent and the median was 4.5 percent; for older assisted properties, the
top quartile was 3.9 percent and the median was 2.2 percent; for newer assisted, the top quartile was 1.6 percent and the median was 0.1
percent.

Clearly, the cutoff between distressed and stressed was arbitrary, and could be set at another point. Essentially any negative index is an
indicator that the property will potentially have trouble in the long run.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Application of the Distress Index

Exhibit 5-3 shows the result of applying the Distress Index to the full stock of multifamily
rental housing with HUD-insured or held mortgages. The mean value of this modified net
cash flow measure was highly positive for newer assisted properties ($888), intermediate for
unassisted ($515), and close to zero ($5) for older assisted properties. In other words, after
covering ongoing operations and payments on a loan to cover the unfunded backlog of
physical needs, newer assisted properties still had, on average, $888 of cash available per unit.
Similarly after covering ongoing operations and a loan on the unfunded backlog of physical
needs, unassisted properties still had, on average, $515 of cash available per unit. In contrast,
with an average index of $5, older assisted properties were barely able to cover ongoing
operations and a loan on the unfunded backlog of physical needs.
•

Overall, about one quarter (24 percent) of properties were classified as
“distressed”—they had Distress Index deficits of more than $250 per 2BR per year
(Distress Index <-$250). These properties’ financial and physical needs
outstripped resources available from revenues and reserves. Among the three
categories of properties, older assisted properties were most likely to be classified
as distressed. Nearly one third (32 percent) of these properties were classified as
distressed compared with 15 percent of newer assisted and 19 percent of
unassisted properties.

•

At the other extreme, nearly two thirds (63 percent) of the stock was classified as
“sound” -- they had positive a Distress Index. These properties had enough
resources available from ongoing revenues and reserves to address their current
financial and physical needs. Over three quarters (78 percent) of both unassisted
and newer assisted properties were classified as sound, compared with under half
(48 percent) of older assisted properties.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Exhibit 5-3
DISTRESS INDEX BY ASSISTANCE CATEGORY
(IN 1995 DOLLARS PER 2 BEDROOM EQUIVALENT)
Total

Assisted

Total
Unassisted

Assisted

Older
Assisted

Newer
Assisted

Total Properties
Percent of Properties

12,243
100%

2,224
18%

10,019
82%

5,943
49%b

4,076
33%b

Distresseda

24%

19%

26%

32%**

15%

< -$1,000

10%

-$1,000 to <-$500

12%

-$500 to <-$250

11%

Stresseda

12%

4%**

14%

20%**

-$250 to $0

12%

14%

20%

Sounda

63%

78%**

60%

48%**

78%

$0 to < $250

12%

10%

13%

16%

$250 to <$500

12%

17%

10%

12%

$500 to <$1,000

15%

16%

15%

12%

20%

$1,000 to <$1,500

17%

10%

$1,500 to <$2,000

10%

> $2,000

10%

18%

Statistics on Distress Index
Meana

$392

$515

$364

$5**

$888

Standard Error

347

117

Median

$275

$511

$199

-$25

$712

Signifies that the differences between unassisted and assisted, or older and newer assisted, properties are statistically
significant at the 90% confidence level.
** Signifies that the differences between unassisted and assisted, or older and newer assisted, properties are statistically
significant at the 95% confidence level.
a
Significance test conducted.
b
Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.
Note:

Column sums may not add up to 100% due to rounding.

Source:

Derived from 1995 Financial and Physical Condition Data.

Exhibit 5-4 compares the two measures of property risk, the Distress Index and the Risk
Profile. The exhibit shows that the measures track fairly well, though each provides important
Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

unique information. The Distress Index provides the ability to numerically rank and compare
properties. Properties with a more negative Distress Index lack more resources, and
properties with a higher Index have more resources available. The risk profile does not rank
properties, but does it provides valuable information by highlighting whether properties are
using available resources to address problems.
•

Nearly all (93 percent) properties that were classified as distressed were also
classified as moderately or highly risky. Similarly over three quarters (79 percent)
of the properties that were classified as high risk were also classified as distressed.

•

Virtually all properties that were classified as minimally risky were also classified
as sound. However only 63 percent of the properties that were classified as sound
were also classified as minimally risky. One quarter of sound properties were
classified as management risks. This is because of the large number of properties,
especially in the newer assisted category, had available resources but were not
addressing existing physical needs backlogs.

•

Properties that were classified as moderately risky were more likely to be classified
as distressed (41 percent) or stressed (34 percent) rather than sound (25 percent).
Exhibit 5-4
COMPARISON OF DISTRESS INDEX AND RISK PROFILE OF INSURED PROPERTIES

High
Risk

Moderately
Risky

Minimally
Risky

Management
Risk

Distressed

Stressed

Sound

Total By
Risk Profile

Total Properties

1,383

256

122

1,761

Row Percent

79%

15%

14%

Column Percent

47%

17%

Total Properties

1,379

1,140

847

3,366

Row Percent

41%

34%

25%

28%

Column Percent

46%

76%

11%

Total Properties

4,870

4,892

Row Percent

100%

40%

Column Percent

63%

Total Properties

206

1,299

2,223

Row Percent

87%

18%

Column Percent

25%

2,968
24%

1,507
12%

7,768
63%

Total by Distress Category
Column Percent
Source:

12,243
100%

Derived from Exhibits 5-1 and 5-3.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Characteristics of Distressed Properties
Exhibit 5-5 below summarizes tenant characteristics and Exhibit 5-6, property, and
neighborhood characteristics of properties by distress categorization.

•

Distressed and stressed properties had proportionately fewer tenants with very low
incomes compared to sound properties. This finding reflects the high percentage
of distressed properties that are older assisted properties, which include more
moderate-income and fewer very low-income tenants than do newer assisted
properties.4

•

All categories of properties had more households headed by non-minority whites
(58 percent) than by any other racial or ethnic group. However, distressed and
stressed properties each had proportionately fewer households headed by whites
and proportionately more households headed by blacks. Distressed properties had
more households headed by Hispanics compared to either stressed or sound
properties.

•

Distressed properties had proportionately fewer single-person households or
households headed by elderly persons and more family households (again,
reflecting the difference between older and newer assisted properties).

•

Distressed and stressed properties, compared with sound properties, were more
likely to be older assisted and less likely to be newer assisted. Older assisted
properties accounted for 49 percent of the insured (or held) stock, but accounted
for 65 percent of the distressed properties and 78 percent of the stressed
properties. By contrast, newer assisted properties accounted for 33 percent of the
stock but only for 21 percent of the distressed properties and 17 percent of the
stressed properties.

•

Mean property size (number of units) was fairly even across distress categories.

•

Distressed and stressed properties had slightly larger units (higher number of
bedrooms) on average compared to sound properties. This is consistent with the
higher concentration of single and elderly households in sound properties.

•

There were few differences in the types of buildings that were distressed versus
sound. However, stress was more common in walk ups and less common in high
rises.

•

Distressed and stressed properties tended to be located in neighborhoods that were
in worse condition than those in which sound properties were located and were
more likely than sound properties to be located in central cities.

Data on tenant characteristics are available only for assisted properties.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

•

Distressed and stressed properties were more likely than sound properties to have
non-profit/cooperative or limited dividend owners and were less likely to have forprofit owners. These findings are consistent with the fact that for-profit owners
predominate among newer assisted properties and non-profits are concentrated in
older assisted properties.

•

Distressed and stressed properties were more likely than sound properties to have
rents below the local Section 8 fair market rent (FMR) levels.

•

Distressed and stressed properties were less likely than sound properties to be
located in neighborhoods with tight rental markets (vacancy rates under 4 percent).
However, the difference between property vacancy loss and neighborhood vacancy
varied little by distress category.
Exhibit 5-5
TENANT CHARACTERISTICS BY DISTRESS INDEX
FOR ASSISTED PROPERTIES
Total

Distressed
(Index <-$250)

Stressed
(Index between $250 and $0)

Sound
(Index >$0)

10,019
100%

2,556
26%

1,425
14%

6,038
60%

Race/Ethnicity
Hispanica
Non-Hispanic
Whitea
Blacka
Other

11%
89%
58%
37%
5%

14%
86%
54%*
42%*
4%

8%
92%
55%
40%
5%

10%
90%
61%
35%
4%

Household Size
1 Persona
2 People
3 People
4 People
5 People

43%
24%
17%
10%
4%

34%**
26%
19%
13%
4%

38%**
25%
19%
11%
4%

48%
23%
15%
9%
3%

Elderly Head of
Household Percenta

33%

24%**

23%**

38%

Household Income
Very low incomea
Low income
Not low income

78%
20%
1%

76%**
22%
2%

68%**
31%
1%

82%
17%
1%

Total Properties
Percent of Properties

Signifies that the differences between distressed and sound, or stressed and sound, properties are statistically significant at the
90% confidence level.
** Signifies that the differences between distressed and sound, or stressed and sound, properties are statistically significant at the
95% confidence level.
a
Significance test conducted.
Note:
Column sums may not add up to totals due to rounding.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Source:

TRACS.

Exhibit 5-6
NEIGHBORHOOD AND PROGRAM CHARACTERISTICS BY DISTRESS INDEX
MULTIFAMILY RENTAL HOUSING WITH HUD-INSURED (OR HELD) MORTGAGES

Total

Distressed
(Index Less
than -$250)

Stressed
(Index between
-$250 and $0)

Sound
(Index >$0)

12,243
100%

2,968
24%

1,507
12%

7,768
63%

Assistance Category
Unassisted
Older Assisted
Newer Assisted

18%
49%
33%

14%
65%
21%

5%
78%
17%

22%
37%
41%

Sponsor Type
Non-Profit/Coop
Limited Dividend
For-Profit

18%
40%
42%

24%
52%
24%

23%
60%
18%

15%
32%
53%

Mortgage Start Year
Before 1970
1970-1979
1980 or later

5%
55%
40%

9%
60%
32%

7%
73%
20%

4%
49%
47%

Property Size
<50 unitsa
50-99 units
100-199 units
> 200 unitsa

17%
35%
36%
12%

23%**
30%
35%
12%

18%
25%
46%
11%

14%
39%
35%
12%

Mean Unitsa
Standard Error
Median

115
3.9
96

112
7
90

117
8
100

115
5
96

Average Unit Size
<2.25 bedroomsa
>2.25 bedroomsa

80%
20%

76%**
24%**

66%**
34%**

84%
16%

Mean Unit Sizea
Standard Error
Median

1.7
0.02
1.8

1.8**
0.04
2.0

1.9**
0.07
2.0

1.6
0.03
1.7

Building Type
High risea
Walk upa
SF attached
SF detached

26%
44%
31%
0%

27%
40%
34%
0%

10%**
53%
35%
2%

28%
44%
28%
0%

Total Properties
Percent of Properties

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Total

Distressed
(Index Less
than -$250)

Stressed
(Index between
-$250 and $0)

Sound
(Index >$0)

12,243
100%

2,968
24%

1,507
12%

7,768
63%

Neighborhood Quality
Relative to City
Better than Average
Average
Worse than Average

36%
39%
23%

30%
38%
33%

35%
33%
30%

39%
41%
19%

Quality as Residential
Neighborhood
Excellent/Good
Fair/Poor

67%
33%

53%
47%

71%
29%

72%
28%

45%
36%
18%

43%
41%
16%

41%
35%
24%

48%
35%
17%

Total Properties
Percent of Properties

Neighborhood Vacancy
<4%
4-7%
>7%
*

Column sums may not add up to totals due to rounding.

Source:

Inspections, windshield survey, Census, HUD, market evaluations.

Comparison of Results with 1990 Findings
Exhibit 5-7 compares the levels of distress of properties in the current study (1995 study) with
the 1990 Study. Because the properties included in this exhibit are the same for both the 1990
measure of distress and the 1995 measure, this comparison allows us to analyze the change in
the stock over the six-year period. It shows that the level of distress in the stock as a whole
had not changed significantly. In both time periods, almost quarter (22 to 23 percent) of
properties were classified as distressed, and the percentage of properties that were classified
as “stressed” and “sound” did not change significantly.5

As was indicated in Chapter 3 on the physical condition of the stock, the study’s estimation of physical condition was not intended to
provide reliable property-level estimates, but rather reliable stratum-level estimates. Thus, we do not compare the Distress Index at a
property level between the two time periods.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

Exhibit 5-7
COMPARISON OF DISTRESS IN 1989 AND 1995
(BASED ON COMPARISON SAMPLE)

Total

Unassisted

Older
Assisted

Newer
Assisted

Total

Unassisted

1989

Older
Assisted

Newer
Assisted

1995

Distressed

22%

31%

30%

23%

19%

32%

13%

Stressed

14%

11%

21%

12%

20%

Sound

64%

58%

50%

87%

64%

78%

48%

80%

Note:

Column sums may not add up to totals due to rounding.

Source:

1995 data: Annual Financial Statements and 1995 Physical Inspection Data and Costing Program.
1989 data: 1990 Analysis File.

While overall the distribution of properties by distress level did not change, changes did occur
within the newer assisted and unassisted categories.
•

Among newer assisted properties, the percentage of distressed properties nearly
doubled from 7 percent in the 1989 to 13 percent in 1995. At the same time, the
percentage of newer assisted properties classified as sound decreased from 87
percent to 80 percent. Thus, while newer assisted properties are still the least
distressed portion of the stock, the situation in these properties has deteriorated
over the last six years, largely as a result of the deterioration in physical condition.

•

Among unassisted properties, the percentage of distressed and stressed properties
decreased substantially over the six-year period, and the percentage of sound
properties increased. In 1989, 58 percent of unassisted properties were classified
as sound and 31 percent were classified as distressed. In 1995 over three quarters
(78 percent) of the unassisted stock was classified as sound, and only 19 percent
were classified as distressed. This is largely a result of the significant improvement
in the financial condition of this portion of the stock.

•

Almost no change occurred in the older assisted portion of the stock. These
properties continued to be the most distressed (30 to 32 percent), and the least
likely to be classified as sound.

Abt Associates Inc.

5.0 Measures that Combine Physical
and Financial Conditions

6.0 Market Position of Insured Properties
This chapter describes several aspects of the market position of the stock. Section 6.1
compares property financial indicators (rents, vacancy losses, operating and maintenance
expenses, and operating ratios) with local market and industry norms. Section 6.2 presents a
modified cost flow analysis assuming properties operate at market-rate rent and operating cost
lowers. Section 6.3 reports on the potential future uses of HUD-insured properties if they
were not restricted by HUD mortgage and subsidy program requirements.

6.1

Property Finances Relative to Neighborhood and Industry Data

To help place the financial indicators presented in Chapter 4 in a broader context, we
conducted comparisons of several key financial variables with available neighborhood and
industry data. This section shows that:
•

As expected, of the three categories of properties, unassisted properties tended to
operate closest to local market and industry norms compared with the two
categories of assisted properties, although even these properties tended to have
operating and maintenance expenses that were above industry norms for
conventional properties in similar locations and similar building types.

•

The majority of older assisted properties had rents that were lower than estimates
of rents for comparable properties in their local markets. Most also had operating
and maintenance costs that were above industry norms for conventional properties
in similar locations and similar building types. The combination of lower rents and
higher operating costs meant that nearly all older assisted properties had net
operating income to rent ratios that were below industry norms for conventional
properties in similar locations and similar building types.

•

Supported by their high subsidies and very low vacancy rates, newer assisted
properties generally had rents that were significantly higher than estimates of rents
for comparable properties in their local markets. Although operating and
maintenance expenses in these properties were also higher than industry norms for
conventional properties in similar locations and similar building types, the higher
operating costs did not fully offset the higher revenues. Thus, net operating
income to rent ratios in newer assisted properties tended to be above the national
median ratios for conventional properties in similar locations and similar building
types.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Rents
Exhibit 6-1 compares property rents (gross rents paid to owners as reported on the annual
financial statements) with comparable rents in their neighborhoods. As comparable rents we
used the study’s market evaluators’ estimates of what rents would be in an unrestricted
market, assuming no upgrades were made (“as is” local market rent). Exhibit 6-2 presents
this information graphically. The stock was evenly divided between properties with gross
rents above estimated market levels and those with gross rents below estimated market levels.
•

About half (49 percent) of the stock had gross rents close to their estimated
market rent (between 75 percent and 120 percent of the properties’ estimated
market rent). As expected, a large majority (75 percent) of unassisted properties
had rents in this range.

•

Most (78 percent) older assisted properties had rents below their estimated market
level, including 38 percent with rents below 75 percent of their estimated market
level.

•

In contrast, the vast majority (86 percent) of newer assisted properties had rents
above their estimated market level, including 40 percent with rents above 140
percent of their estimated market level. These properties are the focus of the
portfolio reengineering efforts that aim to bring property rents in line with rents in
their surrounding markets.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Exhibit 6-1
CURRENT PROPERTY RENT RELATIVE TO LOCAL MARKET “AS IS”
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Property Rent Relative to Local Market “As Is”

Property Rent <75%
of Market

20%

9%**

23%

38%**

Property Rent 75 - 90%
of Market

18%

19%

18%

28%

Property Rent 90 - 100%
of Market

13%

28%

10%

12%

Property Rent 100 - 120%
of Market

18%

28%

15%

12%

20%

Property Rent 120 - 140%
of Market

13%

14%

26%

Property Rent 140 - 175%
of Market

10%

20%

Property Rent 175% +
of Market

20%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Source:

Financial Data, Market Valuation Summary.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Exhibit 6-2 displays rents relative to market graphically.
Exhibit 6-2
PROPERTY RENTS RELATIVE TO
MARKET “AS IS” RENTS

100

Percent of
Properties

Total
140-175% +

Source:

Unassisted
120-140%

Older Assisted

100-120%

90-100%

Newer Assisted
75-90%

<75%

Derived from Exhibit 6-1.

Changes in Rents Relative to Market Since 1989
In Exhibit 6-3 we compare the 1995 ratios of property rents (gross rents paid to owners as
reported on the annual financial statements) with comparable rents in their neighborhoods
with similar calculations for 1989. The 1995 figures are slightly different from those in
Exhibit 6-1 because Exhibit 6-3 relies on the comparison sample of properties that were
included in both studies. The exhibit shows that for the stock as a whole, in both time periods
about half the properties had rents above market levels (50 percent in 1989 and 48 percent in
1995) and about half had rents below. Similarly, in both time periods rents for about half the
stock were close to market levels. In 1989, 52 percent of the stock had rents between 75 and
120 percent of the estimated market rent for comparable market-rate properties in their areas,
as did 49 percent in 1995. However, within assistance categories the following changes took
place:
•

In both 1989 and in 1995, rents in about three quarters of unassisted properties
were close to estimated market rents for comparable market-rate properties in
their areas (between 75 percent and 120 percent of market). In 1995 a slightly
higher percentage of unassisted properties (43 percent) had rents that were above
market levels compared with 1989 (38 percent).

Abt Associates Inc.

6.0 Market Position of Insured Properties

•

In 1989, about two thirds (67 percent) of older assisted properties had rents that
were below estimated market rents for comparable market-rate properties in their
areas. In 1995, more than three quarters of the older assisted stock had below
market rents.

•

Rents in newer assisted properties tended to be above estimated market rents for
comparable market-rate properties in their areas in both years. The proportion of
newer assisted properties with above market rents rose from 81 percent to 86
percent during this time period.
Exhibit 6-3
PROPERTY RENTS RELATIVE TO MARKET IN 1989 AND 1995
(BASED ON COMPARISON SAMPLE)

Total

Unassisted

Ratio of Property
to “As Is”
Market Rent

Older
Assisted

Newer
Assisted

Total

Unassisted

1989

Older
Assisted

Newer
Assisted

1995

Property Rent <
Market Rent

50%

63%

67%

18%

52%

56%

77%

14%

Property <75% of
Market

18%

13%

30%

20%

38%

75% to 90% of
Market

17%

24%

23%

18%

19%

27%

90% to 100% of
Market

15%

26%

14%

12%

14%

28%

12%

Property Rent >
Market Rent

49%

38%

33%

81%

48%

43%

23%

86%

100% to 120% of
Market

20%

28%

17%

21%

17%

28%

12%

18%

120% to 140% of
Market

18%

13%

28%

140% to 175% of
Market

10%

19%

Property > 175%
of Market

10%

23%

21%

Note:

Column sums may not add up to 100% due to rounding.

Source:

1995: 1992-1995 financial data, market valuation summaries.
1989: 1990 Analysis File.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Vacancies
Exhibit 6-4 compares property vacancy losses with the vacancy rates in the surrounding
neighborhood. Vacancy losses, as reported on the annual financial statements, include both
losses from vacant units and from uncollected rents. Neighborhood vacancy rates were
collected by the market analysts as part of the market valuation process, and were provided in
ranges of under 4 percent, 4 to 7 percent, and 7 percent or above.
•

The HUD-insured stock tended to have vacancy rates that were equal to (45
percent) or lower (39 percent) than their neighborhood averages. This is as
expected, given the low mean vacancy losses reported, particularly among assisted
properties.

•

Both older and newer assisted properties were more likely to have lower vacancy
rates then their surrounding neighborhoods (40 and 47 percent respectively)
compared with unassisted properties (25 percent). Lower than average vacancy
losses are expected in assisted properties because the project-based assistance
these properties receive help them attract lower-income renters.

•

Consistent with their very low vacancy losses, only 3 percent of newer assisted
properties had vacancy rates above their local neighborhoods.
Exhibit 6-4
CURRENT PROPERTY VACANCY RELATIVE TO LOCAL MARKET
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Property Vacancy Relative to Local Market

Property Vacancy less than
Neighborhood

39%

25%

43%

40%

47%

Property Vacancy Same
Range as Neighborhood

45%

37%

47%

45%

51%

Property Vacancy Greater
than Neighborhood

15%

38%

10%

15%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Source:

Financial Data, Market Valuation Summary.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Operating and Maintenance Expenses
Exhibit 6-5 compares property operating and maintenance expenses with private sector
industry norms. Our private sector estimates relied on median values of operating and
maintenance expenses per square foot of living space by region and building type for
conventional apartments as compiled by the Institute of Real Estate Management (IREM)
data.1
Most (81 percent) HUD-insured properties had operating and maintenance expenses that were
above the medians for conventional properties in their region and building type. Operating
and maintenance expenses were above 125 percent of the median in over half the stock (52
percent).

•

Eleven percent of properties had operating and maintenance expenses that were
below 90 percent of the median expenses in conventional apartments in their
region and building type, 36 percent were between 90 and 125 percent of median,
and 52 percent had expenses that were more than 125 of the median for their
region and building type.

•

Operating and maintenance expenses in the unassisted portion of the stock were
distributed more closely around the medians of conventional properties compared
with assisted properties, although even in this portion of the stock operating and
maintenance costs tended to be above market levels. Forty-three percent of
unassisted properties had operating and maintenance costs that were between 90
percent and 125 percent of the median. However only 20 percent of unassisted
properties had operating and maintenance expenses below 90 percent of the
median for conventional properties in their region and building type, while 37
percent had operating and maintenance expenses above 125 percent of the relevant
medians.

•

Assisted properties generally had higher operating and maintenance costs
compared with conventional properties in their region and building type (85
percent).

•

Half of older assisted properties had operating and maintenance costs that were
over 125 percent of the median for their building type and region, as did nearly
two thirds (63 percent) of newer assisted properties.

1996 Income/Expense Analysis: Conventional Apartments, Institute of Real Estate Management, Chicago Il.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Exhibit 6-5
CURRENT OPERATING COSTS RELATIVE TO INDUSTRY NORMS
Total

Assisted

Total
Total Properties
Percent of Total Properties

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

12,243
100%

Operating Costs Relative to Industry Norms

Property <90%
of Industry Norm

11%

20%

11%

Property 90 - 100%
of Industry Norm

Property 100 - 110%
of Industry Norm

12%

17%

11%

12%

10%

Property 110 - 125%
of Industry Norm

17%

18%

19%

15%

Property 125 - 150%
of Industry Norm

22%

16%

24%

22%

26%

Property >150%
of Industry Norm

30%

21%

32%

28%

37%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.

Source:

Financial Data, IREM.

Net Operating Income to Rent Ratio
The last comparison (Exhibit 6-6) conducted was between the properties’ “net operating
income (NOI) to rent ratios” and “median NOI to rent ratios” for conventional properties of
the same building type and region as compiled by IREM. This provides an indication of how
the financial situation of the insured stock compares with industry norms.
The net operating income to rent ratio is defined as total revenues less operating and
maintenance expenses divided by gross potential rent. (It does not include debt service

Abt Associates Inc.

6.0 Market Position of Insured Properties

expenses or deposits to the replacement reserve account). Higher ratios indicate more funds
left for debt service and cash returns after covering operating and maintenance expenses.2

Exhibit 6-6
PROPERTY NET OPERATING INCOME TO RENT RATIO RELATIVE TO INDUSTRY NORM
Total

Assisted

Total
Total Properties
Percent of Total Properties

12,243
100%

Unassisted

Assisted

Older Assisted

Newer Assisted

2,224
18%

10,019
82%

5,943
59%a

4,076
41%a

Net Operating Income to Rent Ratio Relative to Industry Normb

Property <50%
of Industry Norm

12%

15%

12%

20%

Property 50 - 75%
of Industry Norm

24%

19%

25%

39%

Property 75 - 90%
of Industry Norm

17%

20%

13%

Property 90 - 100%
of Industry Norm

13%

18%

12%

18%

Property 100 - 110%
of Industry Norm

14%

18%

13%

23%

Property 110 - 125%
of Industry Norm

12%

13%

26%

Property >125%
of Industry Norm

14%

Older assisted properties represent 59% of assisted properties and 49% of the universe. Newer assisted properties represent
41% of assisted properties and 33% of the universe.
Net operating income to rent ratio equals (total revenue - operating and maintenance expenses)/gross potential rent.

Note:

Column sums may not add up to 100% due to rounding.

Source:

Financial Data, IREM.

Because debt service costs are very low in Section 236 properties, applying a standard NOI test to these properties would overstate their
financial hardship. Although their cash available after covering operating and maintenance costs is low, their debt service payments are
also low. We have added the IRPs to its revenues of these properties to account for this in this analysis.

Abt Associates Inc.

6.0 Market Position of Insured Properties

•

Two thirds of insured properties had net operating income to rent ratios below the
median for conventional apartments in their building type and region.

•

As expected, net operating income to rent ratios in the unassisted portion of the
stock were often closely distributed around industry norms, 58 percent were
between 75 and 125 percent of median for conventional apartments in their
building type and region. However, over a third of the unassisted stocks (34
percent) had ratios that were below industry norms.

•

In contrast, net operating income to rent ratios were below 75 percent of the
median for conventional apartments in their building type and region for more than
half (59 percent) of older assisted properties, including 20 percent with ratios
below half the median. This is a result of both the below conventional market
rents and the above conventional market operating and maintenance costs.

•

As with most financial indicators, newer assisted properties had the highest relative
net operating income to rent ratios. Sixty-three percent of newer assisted
properties had net operating income to rent ratios that were higher than those for
conventional apartments in their region and building type, including 40 percent
with ratios over 110 percent of the conventional apartment median in their building
type and region. Although operating and maintenance expenses in newer assisted
properties were higher than industry norms, rents were much higher so that overall
net operating income to rent ratios were higher than in comparable conventional
apartments.

6.2

Market Scenario Distress Analysis

All of the analyses so far that emphasized the strong financial position of newer assisted
properties focused on the “current” condition of the stock. As is obvious from the ongoing
policy initiatives, the “current” situation is changing. In particular, in Section 8-assisted
properties with above market rents, HUD is developing programs to reduce rents to market
levels and at the same time reduce debt service payments so that owners can still cover
expenses. Permanent legislation for “re-engineering” the portfolio has recently been passed.
At present, no proposals are in process to respond to the properties that have below market
rents.
In this section, we modify the Distress Index in order to project the level of distress that
would be present in the HUD insured stock if property rents were “market” rents. The
purpose of this modified Distress Index, which we call the Market Scenario Distress Index, is
to create a model to assess the level of distress that would result if rent subsidies were
removed and property rents and operating costs reverted to market levels (allowing for both
increases and decreases in rents).

Abt Associates Inc.

6.0 Market Position of Insured Properties

The Market Scenario Distress Index is computed by taking:
(a)
(b)

Net cash flow, recomputed with rents set to “as is” market rents
Minus the amortized cost of remedying the Unfunded Backlog of Physical Needs

As in the case of the original Distress Index, the computation begins with net cash flow, which
measures a property’s capacity to meet current expenses and make deposits to its replacement
reserves account. This net cash flow is different from the net cash flow used in the original
Distress Index in two respects: (1) revenues and operating costs are based on market
estimates, rather than on actual, subsidized rents and operating costs; (2) average annual
accrual is used instead of deposits to replacement reserves. Annual accrual is used instead of
replacement reserve deposits because it is assumed that the “market-oriented” property owner
will actively keep up with the property’s capital needs. Annual accrual typically is higher than
current deposits to replacement reserves.
Modified “Market Scenario” Net Cash Flow
Net Cash Flow =
Total Revenue (Weighted 3-Year Average) based on market analysts’ estimate of rent
obtainable in the market
Minus Operating and Maintenance Expenses (including expenses for
administration, operations and maintenance, utilities, taxes, insurance, and
vacancy losses, based on IREM industry norms by building type and
region)
Minus Mortgage Debt Service (interest, principal and mortgage insurance
premium as required by current mortgage)
Minus Average Annual Accrual over 20 years
As with the original Distress Index, net cash flow is reduced by the amortized cost of
remedying the unfunded backlog of physical needs, which represents the annual cost of
undertaking a repair program. No additional adjustment is made for vacancy loss, because the
IREM estimate of expenses includes vacancy losses.
Exhibit 6-7 shows the result of applying the Market Scenario Distress Index to the full stock
of multifamily rental housing with HUD-insured or HUD-held mortgages. Under the Market
Scenario Distress Index, a dramatically higher percentage of newer assisted properties would
be classified as distressed, from 15 percent under the original Distress Index to a staggering
87 percent under the Market Scenario Distress Index. The main reason for the overwhelming
difference among newer assisted properties is that a high proportion of newer assisted
properties currently receive total rents exceeding local market rents. Thus, reducing rents to
market (or forcing properties to operate without rent subsidies) would result in substantially
lower income. Further, the Market Scenario Distress Index assumes that property owners put

Abt Associates Inc.

6.0 Market Position of Insured Properties

away into reserves every year the average annual amount of accrual, not just the required
replacement reserve deposits.
Exhibit 6-7
MARKET SCENARIO DISTRESS INDEX BY ASSISTANCE CATEGORY
(ASSUMES UNSUBSIDIZED OPERATION AT “AS IS” MARKET RENTS)
(IN 1995 DOLLARS PER 2 BR EQUIVALENT UNIT)
Total

Assisted

Total
Unassisted

Assisted

Older
Assisted

Newer
Assisted

Total Properties
Percent of Properties

12,243
100%

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Distresseda

45%

49%

44%

15%**

87%

< -$1,000

30%

26%

31%

65%

-$1,000 to <-$500

10%

14%

10%

18%

-$500 to <-$250

10%

Stresseda

12%*

7%**

-$250 to $0

12%

Sounda

48%

38%**

51%

78%**

10%

$0 to < $250

$250 to <$500

$500 to <$1,000

11%

10%

11%

19%

$1,000 to <$1,500

14%

$1,500 to <$2,000

13%

> $2,000

10%

15%

Statistics on Market Scenario Distress Index
Meana

-$295

-$700*

-$206

$923**

-$1,851

Standard Error

$79

$272

$75

$81

$141

Median

-$62

-$244

$23

$848

-$1,469

Column sums may not add to 100% due to rounding.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Under the Market Scenario Distress Index, the proportion of unassisted properties classified
as distressed was also substantially higher. The main reason for the increase is the
assumption, under the market scenario that property owners are putting aside sufficient funds
to cover ongoing accruals.
In contrast, older assisted properties fared best under the Market Scenario Distress Index,
going from almost a third (32 percent) classified as distressed to only 15 percent. The reason
for this difference is that many older assisted properties are currently receiving rents that are
less than local market rents. Reverting to market rents in these cases would mean higher rents
and more revenue to deal with physical backlog and annual accrual.
6.3

Market Potential of Insured Properties

This section reports on the potential future uses of the HUD-insured properties if they were
not restricted by HUD mortgage and subsidy program requirements. First we present the
responses the study’s market analysts received to questions regarding potential uses of the
property. Next we present estimates of property value and market position, if owners were
able to convert all properties to their highest and best use.
Exhibit 6-8 shows that:
•

The overwhelming conclusion among the study’s market analysts was that most
HUD-insured properties, across all assistance categories should continue operations
“as is” with no major repairs or renovations (95 percent).

•

It was concluded that fewer than half of the properties across all assistance categories
could be physically converted or upgraded to a higher quality (44-48 percent).

•

Relatively few of the unassisted properties (16 percent) and even fewer of the assisted
properties (4 percent) were considered possible candidates for conversion to
condominiums.

The study’s market analysts estimated property rents under three alternative scenarios for
physical upgrades: “as is”, “with a moderate upgrade” and “with a major upgrade”. As part
of the physical inspection process, the inspectors identified upgrades that would be needed to
position properties for higher-end market uses. For each of the three market rent options
provided by the market analysts, we calculated a net market value based on the capitalized net
rent stream provided and the costs (repair in all cases, and upgrade in the two “upgrade”
cases, as estimated based on the physical inspection data and the costing programs) that

Abt Associates Inc.

6.0 Market Position of Insured Properties

would be required to obtain these rents.3 The “highest and best use” is defined as the market
position that provides the highest net market value. Exhibit 6-8 shows that:
•

Based on the comparison of rent streams with repair and upgrade costs, the majority
of properties (80 percent) were already situated in their highest and best use, and
would not benefit from physical upgrade projects. This includes 68 percent of
unassisted properties and 83 percent of assisted properties.4

•

If properties were positioned optimally in the market, the average value would be
about $30,000 per 2-bedroom unit, with values of unassisted properties higher than
values of assisted properties.

•

Optimal market rents in unassisted properties would also be above rents in the
assisted portion of the stock, although the differences are not statistically significant.

In particular, for each rent stream provided we calculated “net rent”, i.e. rent net of operating costs. As our estimate of operating income
after covering expenses, we used the Institute of Real Estate Management (IREM) median net operating income to rent ratios by building
type and region for 1995. The capitalization rates used were collected by the market analysts as part of their discussions with local real
estate experts.

This includes 5 sample properties, representing 68 properties in the universe where the net market value is negative even with no upgrades.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Exhibit 6-8
FUTURE MARKET POTENTIAL OF INSURED PROPERTIES
Total

Assisted

Total
Total Properties
Percent of Total
Properties

12,243
100%

Unassisted

Assisted

Older

Newer

2,224
18%

10,019
82%

5,943
59%b

4,076
41%b

Use in Current Physical Condition
Very / Somewhat Likelya

95%

97%

95%

94%

96%

Probably / Definitely Not

Physical Upgrade to Higher Quality
Very / Somewhat Likelya

45%

48%

44%

45%

44%

Probably / Definitely Not

55%

52%

56%

55%

56%

Conversion to Condominium Units
Very / Somewhat Likelya

16%**

Probably / Definitely Not

94%

84%

96%

97%

96%

Optimal Market Position
Low Marketa

80%

68%**

83%

Moderate Market

17%

25%

15%

16%

14%

High End Marketa

7%*

1%*

Optimal Net Market Value per 2BR
Mean Valuea

$30,726

$38,160**

$29,076

$28,444

$29,998

Standard Error

833

2,818

804

1,048

1,252

Median

$26,726

$31,785

$25,439

$24,923

$26,635

Monthly Rent at Optimal Position per 2BR
Mean Valuea

$631

$687

$618

$608

$633

Standard Error

Median

$565

$611

$549

$541

$521

Market Valuation Summaries, 1995 Physical Inspection Data and Costing Programs.

Abt Associates Inc.

6.0 Market Position of Insured Properties

Status of HUDInsured (or Held)
Multifamily Rental
Housing in 1995
HC-5964
Task Order #7

Final Report
Appendices

December 10, 1997

Prepared for
US Department of Housing
and Urban Development
451 Seventh Street, SW
Washington, DC 20410-3000

Prepared by
Meryl Finkel
Donna DeMarco
Deborah Morse
Sandra Nolden
Karen Rich

Appendix A
SAMPLING
This appendix describes the approach to estimating the numbers of properties insured by HUD
in 1989/90 and in 1989/95 and setting appropriate weights for five relevant samples (i.e., the
Initial, Monitoring, Analysis, Augmented, and Comparison Samples).1
Section A.1 describes the samples, sampling weights, and initial population estimates, taking
into account current information on property types. It is sequential, developing Monitoring
Sample weights from the Initial Sample weights, and Analysis Sample and Augmented Sample
weights from the Monitoring Sample weights. Section A.2 discusses revised population
estimates and Monitoring Sample weights, taking into account 1996 HUD data on properties
insured in 1989/90 and 1989/95. Revised Analysis and Augmented weights can then be
obtained by completing the steps described in Section A.1, using the revised Monitoring
Sample weights.
Thus, the methodology for developing the weights involves essentially carrying the steps in
Section A.2 first, and then using the revised monitoring sample weights to complete the steps
in Section A.1. (However, the approach is presented in the order it is because it is helpful to
understand the full approach before going into the details).
A.1

The Five Samples

Sections A.1.1 to A.1.4 describe in turn the Initial and Monitoring Samples, the Analysis
Sample, the Augmented Sample, and the Comparison Sample and their sampling weights.
A.1.1 The Initial and Monitoring Samples
In 1989, Abt developed a list of the 13,667 insured properties in HUD's MIDLIS data base
that appeared to meet certain criteria. Abt drew a two-stage sample of 1000 of these
properties, allocated over six strata. The first stage was a Probability Proportional to Size
(PPS) sample of 53 Primary Sampling Units (PSUs), using the number of properties in each
PSU as the measure of size. The second stage included samples of properties within each
sampled PSU, stratified by property type. Sampling rates differed among the strata. Within
each stratum, sampling rates for PSU properties were inversely proportional to the PSU
probability of selection, so that every property in a given stratum had the same probability of
selection.

Discussion of the sample drawn for the 1990 Study is described in detail in Appendix A of the 1992 report.

Abt Associates Inc.

Appendix A - Sampling

Data collected from HUD computer files and Field Offices for this sample early in 1990
revealed that 974 of the sampled properties were in fact eligible (met the criteria). This
sample of 974 population properties is the "Monitoring Sample". In addition, in reviewing
discrepancies between 1990 and 1996 HUD data on property types and between these data
and inspection data, we recently determined that some of the 1989 data base property type
classifications were incorrect. (One unassisted property was mistakenly coded as older
assisted non-family. Inspection results indicated that 16 older assisted family properties were
originally classified as older assisted non-family. Three older assisted non-family properties
were originally classified as older assisted family). Thus, it is useful to carry three separate
property type classification variables -- one for the original classification (which determined
initial sampling rates), one for the corrected classification, and one for the 1996 classification.
The relevant universes, sampling fractions, and population estimates are shown in Exhibit A-1.
As indicated in the exhibit, there is a choice about sample weights. Most obviously, the
original weights as shown in the third row of Exhibit A-1 can be used2, so that the weights
would be:
MO

wij

(1)

' wij *ij

where
MO

wij

= the adjusted original weight for the jth property in the ith true property

type in the Monitoring Sample;
O

These were adjusted to produce integral numbers of properties in the last row of Table 1.

Abt Associates Inc.

Appendix A - Sampling

The equal weight version for the Monitoring Sample weights would be3:
ME

wij '

(2)

1
O
wij *ij
j
nMi j

nMi ' j *ij
j

where
ME

wij

= the equal weight for the jth property in the ith true property type in the

Monitoring Sample;
nMi = the number of ith property type properties in the Monitoring Sample;
O

wij = the initial sample weight for the jth property in the ith true property type (a
function of the property's 1989 data base property type, as shown in the third row
of Table 1); and
*ij = a dummy variable indicating that the (i,j)th property in the Initial Sample was
included in the Monitoring Sample (that is, was determined to be in the study
population).
For all analyses in the report we have used the equal weight versions for each of the six
sampling strata, yielding six analytic weights. Since we are not interested in reporting NHP
and non-NHP results separately, we had hoped to collapse these strata for each of the two
older assisted categories. However key outcomes (such as 1995 cash flow) were sufficiently
different between the NHP and non-NHP strata that we decided we could not assign equal
weights to the two sub-categories of each of the two older assisted strata. Thus, we were
required to develop special programs to calculate standard errors for the variables.
It is worth noting that the HUD list appeared to be quite accurate in the sense that it included
very few properties that were not actually in the population. Not only were few listed
properties ineligible, but most of the ineligible properties reflected changes in insurance status.
From 19 to 21 of the 26 ineligible properties in the sample were properties that were no
longer insured at the time of data collection (reasons for ineligibility were not given for two
properties). We expect such discrepancies due to lags in updating data bases. Indeed, some
of these properties may have been insured when the list was drawn, but have changed status
by the time that we began data collection -- that is, our actual population is not certain types
of properties insured in 1990, but those insured in 1989 that were still insured in June of 1990.

In fact, we will never use the equal weight version of the Monitoring Sample weights, so this discussion is really to introduce the topic,
which will come up again in discussing Analysis and Augmented Sample weights.

Abt Associates Inc.

Appendix A - Sampling

A.1.2 The Analysis Sample
A sub-sample of 598 of the 974 eligible projects was drawn as the "Analysis Sample".4 The
Analysis Sample properties were allocated across PSUs in proportion to the Monitoring
Sample. Interview and inspection data collection was completed for 570 of these (95
percent). Exhibit A-2 provides details on the analysis sample. Assuming that observations
were missing at random within strata and across PSUs, the weights for completed cases reflect
the overall domain sampling fractions for completed cases. This gave the Analysis Sample
weights shown on page A-21 of the 1992 Report.
As part of the preparation for the 1995/96 study, additional information on misclassifications
in the 1989 data base was obtained. This needs to be accounted for so that the Analysis
Sample would project to our estimated totals for the true classifications. Again, we have two
obvious choices. First, we can retain the original sampling weights, inflating or deflating them
so that the sum of the weights of completed Analysis Sample cases match the population
totals for each true property category:
j wij

(AO)

wij

(3)

' wij "ij

j0MS

j wij "ij
O

j0MS

where the sums are over the Monitoring Sample, and
(AO)

wij

= the adjusted original weight for the jth completed Analysis Sample

property in the ith true property type category;
O

wij = the original sampling weight for the jth property in the ith true property type
category;
MS = the Monitoring Sample;
"ij = a dummy variable indicating that the jth property in the ith true property type
was a completed Analysis Sample property.
Alternatively, we could give equal weights to all completed Analysis Sample properties within
a given true category -- i.e.,

A sample of 600 of 976 apparently eligible projects was fielded, but it later turned out that two of these were in fact ineligible. Thus, in
fact attempts were made to collect information for a sample of 598 eligible projects.

Abt Associates Inc.

Appendix A - Sampling

(AE)

wij

(4)

1
O
j wij ,
nAi j0MS

nAi ' j "ij
j

where
(AE)

wij

= the equal weight for the jth completed Analysis Sample property in the ith

true property type category;
nAi = the number of ith true property type properties in the completed Analysis
Sample;
and other terms are as in EQ(3).

A.1.3 The Augmented Sample
The new study is intended to determine the current (1995) insurance status of the 1990
population, as well as the current characteristics of the projects in that population that were
still insured by HUD in 1995. It is based on 1995 information for an Augmented sample,
which is a sub-sample of the 1989 Monitoring sample (specifically, the 1990 Analysis sample,
less 22 that were determined to have no longer been insured in the spring of 1995, augmented
by a sub-sample of other assisted Monitoring sample projects).
The Augmented sample was drawn as follows. Using data available in early 1995, HUD
sorted the 974 monitoring properties into two classes -- those that were recorded as still
active (926 properties) - and those that were recorded as no longer active (48 properties)
Thirty of the 48 properties listed as no longer active were part of the original analysis sample.
A HUD intern called to verify the status of these 30 properties. Eight were determined to still
be active (6 had been refinanced and were active under a new FHA number and 2 properties
were acquired by HUD and resold to new owners with HUD purchase money mortgages).
Thus 22 former analysis sample properties were eliminated from the augmented sample at this
stage.
Abt drew a supplementary sample of 125 properties from among the 327 assisted monitoring
sample properties that were listed as active on HUD's file (18 monitoring sample properties
were listed as no longer active including 14 unassisted properties and 4 assisted properties.
The supplementary sample did not include any unassisted properties. HUD did not check the
true status of the 4 assisted properties listed as no longer active but they were excluded from

Abt Associates Inc.

Appendix A - Sampling

the sampling frame for the supplementary sample). Exhibit A-3 details the analysis,
supplementary and augmented samples by stratum.
Consider first the weights for estimating 1995 insurance status. For this purpose, the weights
are based on adjusting the Initial Sample weights. Thus, the weights for the "no longer active"
sample are constructed by inflating the Monitoring Sample weights for these properties so
that, for each true property type category, the sample weights sum to the weights of all the
properties in this group -- that is:
j wij

MO j0OTH

w(OTHij) ' wij

(5)

j wij

j0OTHS

where

wOTHij ' 0 if the insurance status has not been verified in the "no longer active"
sample, the numerator sum is over all Monitoring Sample properties in the "no
longer active" class, the denominator sum is over cases in the "no longer active"
sample where insurance status was verified, and
wOTHij = the weight for the jth property with insurance status verified in the ith
true property type category in the "no longer active" sample; and
MO

wij

= equals the original weight for the jth Monitoring Sample property in the ith

true property type (see EQ(1)).

The fact that the Augmented sample of properties recorded as still active consists of the
Analysis sample plus a further sample of other Monitoring sample properties is irrelevant.
Being drawn for the Augmented Sample is simply a two stage lottery. Since the sampling
rates at each stage is equal for all properties within a 1989 data base property type category,
the overall sampling rate is also equal within 1989 data base property type -- that is,
R( DBi, j0AUG * SI ) ' r(A*M) %(1 &r(A*M)) r(S,DBi)

(6)
where

R(DBi, j0AUG*SI) = the overall sampling rate for the jth Monitoring Sample property
in the ith 1989 data base property type category in selecting properties for the
Augmented Sample, given that it was recorded as still insured in early 1995;

Abt Associates Inc.

Appendix A - Sampling

r(A*M) = the sampling rate for the Analysis Sample ( r(A*M) ' 0.6 for all properties in the
Monitoring Sample); and
r(S,DBi) = the Augmented Sample sampling rate for the jth Monitoring Sample property
in the ith 1989 data base property type category, given that it was recorded as still
insured and was not in the Analysis Sample ( r(S,DBi) is the same for all properties in a
given 1989 data base property type category).
The weights can be computed in the same way as the weights for the "no longer active"
sample -- viz.
j wij

MO j0RSI

w(RSIij) ' wij

(7)

j wij

j0RSIS

where wRSIij ' 0 if the insurance status of the property was not verified in the 'recorded as
still insured" sample, the numerator sum is over all Monitoring Sample properties in
the "recorded as still insured" class, and the denominator sum is over all completed
cases in the "recorded as still insured" sample.
The estimate of the number of properties still insured in 1995 is:
NˆSIi ' j w(RSIij) 0ij % j w(OTHij) 0ij

(8)

j0RSIS

j0OTHS

where
NˆSIi = the estimated number of ith true property type properties insured in
1989/90 that are still insured in 1995;
0ij = a dummy variable indicating that the (i,j)th property was still insured in 1995;
and other terms are as defined earlier.
These weights can be retained in estimating results for properties determined to still be insured
in 1995. However, they need to be adjusted to take account of missing inspections. Again,
there are two versions that could be considered. The first simply adjusts the original weights
by true property type:

Abt Associates Inc.

Appendix A - Sampling

w(RSIij) 0ij ,ij
(9)

AUGO

wij

j w(RSIij) 0ij % j w(OTHij) 0ij
j0RSIS

j0OTHS

j w(RSIij) 0ij ,ij % j w(OTHij) 0ij ,ij
j0RSIS

'
w(OTHij) 0ij ,ij

if (i,j) 0 RSIS

j0OTHS

j w(RSIij) 0ij % j w(OTHij) 0ij
j0RSIS

j0OTHS

j w(RSIij) 0ij ,ij % j w(OTHij) 0ij ,ij
j0RSIS

if (i,j) 0OTHS

j0OTHS

where
,ij = a dummy variable indicating that inspections were completed; and
other terms are as defined earlier.
The equal weight version collapses the "no longer active" and "recorded as still insured"
strata, giving equal weight to all properties within each true property type category that are
still insured in 1995 (with the sum of the weights in each category adding up to the sum of the
weights for properties in that category in the "no longer active" and "recorded as still insured"
samples that were determined to be still insured in 1995):
AUGE

(10)

wij

1
nAUGi

j w(RSIij) 0ij % j w(OTHij) 0ij , nAUGi ' j ,ij 0ij % j ,ij 0ij
j0RSIS

j0OTHS

j0RSIS

j0OTHS

The equal weight version is the most appropriate for characterizing the still insured properties.
However, we were not able to collapse the NHP and non-NHP strata because they appear to
differ on at key measures.
A.1.4 The Comparison Sample
The Augmented Sample properties that were still insured in 1995 are a sample of 1989
properties still insured in 1995. We will often want to compare 1990 and 1995 conditions for
these properties. To do this, we would start with still insured Augmented Sample properties
for which 1995 inspections were completed, using the weights shown in EQ(9). Comparison
1990 data is available for the subset of these that were also in the Analysis Sample.
Accordingly the adjusted original weights for the Comparison sample would be:

Abt Associates Inc.

Appendix A - Sampling

100

AUGO

(ij wij
COMPO

(11)

wij

if (i, j)0OTHS

j wij

AUGO

'
AUGO

(ij wij

j0RSIS

j (ij wij

AUGO

if (i, j)0RSIS

j0RSIS

where
(ij = a dummy variable indicating that the jth property in the ith true property
category was in both the Augmented and Analysis samples, was determined to be
still insured in 1995, and was inspected in 1995.
The equal weight version would be:
COMPE

wij

(12)

A.2

(ij
nCOMPi

j wij

COMPO

nCOMPi ' j (ij
j

Taking Account of New Information on the 1989 Population

The 1989 HUD list from which the 1990 sample was drawn was no longer available when
work began on the 1995 study. HUD constructed a new listing of the 1990 population, which
we call the 1996 list. (Note that both the 1989 list and the 1996 list are lists of 1989 insured
properties -- that is, the dates refer to when the list was created, not to when the properties in
it existed). Comparison of the 1989 and 1996 lists for the initial 1990 sample suggests that
the two lists are quite similar, but that there are some differences. In particular, we estimate
that the 1996 list excludes about 3 percent of the properties in the 1989 list and that some of
these properties were in fact members of the eligible 1990 population. (Most of the
exclusions are properties that had become ineligible between the time the 1989 list was drawn
and the time insurance status was verified -- usually due to prepayment of the mortgage).
Conversely, we estimate that the 1989 list excludes about 1 percent of the properties in the
1996 list, and that almost all of these properties were in fact members of the eligible 1990
population, but were excluded to due deficiencies in the 1989 MIDLIS database, the primary
source of data used to classify properties in 1989. In addition, the 1996 list reclassifies a small

Abt Associates Inc.

Appendix A - Sampling

101

number of properties in terms of property type. We have confirmed most of these
reclassifications, though some are incorrect5.
We want to estimate the number of properties in the 1989/90 population, the number still
insured in 1995, and specify appropriate weights for the Augmented and Analysis samples in
estimating the characteristics of these two populations. The differences between the 1989 and
1996 lists create two problems. First, while the differences are quite modest, HUD would like
the population estimates to reflect the information from both lists. Second, since the sample is
based on the 1989 list, there are no direct estimates for properties not included in that list, and
we need to decide how to treat these properties in our estimates of population characteristics.
Of our initial sample of 1000 properties, 24 were not in the 1996 list - 5 original analysis
sample properties, 4 original monitoring sample properties, and another 15 properties from
among the 26 properties that were excluded from the 1989 monitoring sample. (All five of
the analysis sample properties that were not in the 1996 list were included in the group of 30
analysis sample properties recorded as "no longer active" in mid-1995. Three were
reclassified as "active" following the intern's calls. Three of the 4 monitoring sample
properties that were not in HUD's 1996 list were included in the group of 18 monitoring
sample properties that were listed as "no longer active". No follow-up occurred).
The basic approach adopted for taking the 1996 list information into account involved the
following steps:
1) Define nine 1996 strata, consisting of the four 1996 property types, each
further stratified by the 1995 insurance status recorded in the 1996 data base,
plus a ninth stratum for 1989 properties not in the 1996 data base. This allows
us to use the 1996 information on both the 1989/90 population and its 1995
insurance status.
2) Use the initial sample to estimate how many of the 1989 listed properties were
in the 1996 data base. This is primarily to reassure ourselves that using
estimates based on the 1989 sample involve extrapolation to a relatively small
number of properties that were not included in the 1989 sample frame. We can
do this by estimating:
O
Oˆ k ' j wij

(13)

i,j0 S k

where the sum is over all Initial Sample properties in the kth stratum of the 1996
data base, and

This was the source of the corrected property type categories, discussed in Section A.1.

Abt Associates Inc.

Appendix A - Sampling

102

Oˆ k = the estimated number of 1989 listed properties included in the kth
stratum of the 1996 data base; and
O

wij = the initial sample weight for the jth property in the ith true property type
(a function of the property's 1989 data base property type, as shown in the
third row of Table 1).
Exhibit A-4 presents these results.
3) We then assume that the relatively small number of properties not included in
the 1989 sample frame can be treated as a random sample of properties in each
stratum, so that we can project estimates from the 1989 sample to the entire
1996 data base. Our estimate the proportion of the properties in each of the
nine 1996 strata that were in fact in the 1989 population is:

wikj *ikj
RMO

(14)

wikj

rˆk N k

j wikj *ikj
O

if k#8

ij0S k
O

wikj *ikj

if k'9

where the sum is over Initial Sample properties listed in the kth stratum of the
1996 data base, and
O

wikj = the initial sample weight for the jth property in the ith true property
category and the kth 1996 stratum; and
*ikj = a dummy variable indicating that the (i,k,j)th property was included in
the Monitoring Sample (i.e., found to be eligible).
4) We can now create revised Monitoring Sample weights, reflecting the
information in the 1996 data base. As usual, there are two versions. The
adjusted weight version is the one that gets used; it is
RME

wikj

(15)

1
RMO
wikj ,
j
j
nMi k j

nMi ' j j *ikj
k

where

Abt Associates Inc.

Appendix A - Sampling

103

Nk = the number of properties in the kth stratum of the 1996 data base (which
is known for k < 9).
The equal weight version would be:
j wikj *ikj
O

rˆk '

(16)

ij0S k

j wikj

ij0S k

Exhibit A-5 shows the revised Monitoring Sample weights. For comparison,
Exhibit A-5a presents the Monitoring Sample weights used in the 1989 study.
Exhibit A-6 shows the revised estimate of the 1989 universe taking into account
the 1996 list as well as information from the 1989 list. For comparison, the exhibit
also shows the 1989 estimate of the 1989 universe, and the 1996 list information.
We can then use these revised Monitoring Sample weights in carrying out the
weighting for the Analysis Sample and Augmented Sample, as described in Section
A.1.
Exhibit A-7 shows the revised Analysis Sample weights. For comparison, Exhibit
A-7a presents the Analysis Sample weights used in the 1989 study.
Exhibit A-8 shows the estimate of the 1995 universe based on the estimate of the
1989 universe and the survival rates by stratum obtained for the original Analysis
Sample and the Supplementary Sample. For comparison, Exhibit A-8a also shows
the estimate of the 1995 universe using only the 1996 list information.
Exhibit A-9 shows the Augmented Sample Weights. As can be seen, within the
Unassisted and Newer Assisted categories the weights are very close, and should
certainly be combined to provide equal weights. Within the two Older Assisted
categories there is more variation in the weights. This is because within each
Older Assisted category, properties were sampled at different rates (NHP and nonNHP properties). The NHP sample was a subsample of the older assisted
properties. It included older assisted properties in the continental US, eligible to
pre-pay on their twentieth anniversary, insured prior to 1975, and with the same
SOA exclusions as the current study.
Further, several properties that were originally sampled as Non-Family properties
were reclassified as Family properties and several that were originally sampled as
Family were reclassified as Non-Family.
Abt Associates Inc.

Appendix A - Sampling

104

Exhibit A-9a shows the Augmented Sample Weights under the equal weight
option. As indicated above, key indicators for NHP and non-NHP properties were
sufficiently different that separate weights were kept.
Finally, Exhibit A-10 shows the Comparison Sample Weights.
Exhibit A-10a shows the Comparison Sample Weights under the equal weight
option.

Abt Associates Inc.

Appendix A - Sampling

105

Exhibit A- 1
INITIAL SAMPLE AND MONITORING SAMPLE:
WEIGHTS AND POPULATION ESTIMATES
1989 LISTED STRATUM
Unasst

Older Asst.
Non-Family

Older Asst.
Family

Newer
Asst.

Total

NHP

Non
NHP

NHP

Non
NHP

3,357

202

4,546

1,319

4,170

13,667

Initial Sample

205

310

180

255

1,000

Eligible Sample

188

304

178

254

974

Weight

3357
205

202
30

4546
310

73
20

1319
180

4170
255

Est.
Pop.6

Alt.
Wgt.7

Listed Universe

ELIGIBLE PROPERTIES SAMPLE

True Program Type
188

3,094

3094
189

Older Non-Fam,
NHP

195

195
29

Older Non-Fam,
Non-NHP

288

4,245

4245
291

Older Fam,
NHP

80
21

Older Fam, NonNHP

175

1,502

1502
190

Newer Asst

254

4,154

4154
254

3,079
91.7%

202
100%

4,458
98.1%

73
100%

1,304
98.9%

4,154
99.6%

13,270
97.1%

3079
188

202
30

4458
304

73
20

1304
178

4154
254

Unasst

Est Pop Prop:
Number
Percent
Pop. Wgt.8

Estimate is based on population weights from last row.

Alt. Wgt. = (Est. Pop.)/(Total Sample In Row)

Pop. Wgt. = (Listed Stratum Properties in Pop.)/(Listed Stratum Sample in Pop.).

Abt Associates Inc.

Appendix A - Sampling

106

Exhibit A-2
1990 Study Analysis Sample

Stratum

Monitoring
Sample Size

Initial
Analysis
Sample Size

Expected
Completion
Rate

Expected
Properties
with
Required
Data

Actual
Completion
Rate

Actual
Properties
with
Required
Data

Unassisted

188

123

96%

118

93%

115

Older
AssistedFamily

198

120

96%

115

96%

115

Older
Assisted
Non-Family

334

204

96%

195

95%

194

Newer
Assisted

254

153

96%

147

96%

146

Total

974

600

96%

575

95%

570

Abt Associates Inc.

Appendix A - Sampling

107

Exhibit A-3
Development of the Augmented Sample of 621 Properties

Column

Stratum
1990
Analysis
Sample

Less:
No Longer
Active

Plus:
Supplementary
Sample

Less:
Inspection Not
Complete

Plus:
Stratum
Changes

=:
Augmented
Sample

Unassisted

1159

Older
Assisted
Non-Family

194

-1010

216

Older
Assisted
Family

115

148

Newer
Assisted

146

176

TOTAL

570

125

621

1995 F47
plus phone
calls

Non-analysis
monitoring
sample still
active in 1995
F47

Includes 23
properties that
were determined
to be no longer
insured; 14
unlocateable; 9
refusals; and 6
"other"

Source:

Data
Collection

Equals
A-B+CD+E

Includes one property where the owner owned both a newer assisted and an unassisted property. The newer assisted property was sampled,
but the owner took us to the unassisted property for the inspection. In the 1989 analysis we recoded and reweighted this property as
unassisted, when in fact it should have been deleted from the sample.

Older assisted properties were classified as "family" or "non-family" based on the unit size distribution recorded on HUD's databases.
Properties with no unit size distribution were, by default, coded as non-family. Inspector counts of units resulted in reclassifying 13
properties from older assisted non-family to family, and in reclassifying 3 older assisted family properties as non-family. Thus, the net
decrease in the number of older assisted non-family properties is 10, and the net increase in the number of older assisted family properties is
9. (One unassisted property was originally miscoded as older assisted family)

Abt Associates Inc.

Appendix A - Sampling

108

Exhibit A-4
Estimated Number of 1989 Listed Properties Included in the 1996 Data Base By
Stratum

1996 Listed Stratum
1989 List of all Properties
1989 List of Universe
Unassisted
3,142
3,028
Good 1995
2,373
2,291
Not Good 1995
769
737
Older Assisted Non-Family
4,583
4,540
Good 1995
4,568
4,525
Not Good 1995
15
15
Older Assisted Family
1,474
1,466
Good 1995
1,423
1,415
Not Good 1995
51
51
Newer Assisted
4,121
4,121

Abt Associates Inc.

Appendix A - Sampling

109

Exhibit A-4, continued

Good 1995
4,072
4,072
Not Good 1995
49
49
Not on 1996 List
346
115
TOTAL
13,667
13,270

Abt Associates Inc.

Appendix A - Sampling

110

Exhibit A-5
Revised Monitoring Sample Weights Taking into Account
1996 List Information by 1996 List Category
(Including 1996 list recorded insurance status and assistance category)

Unassisted

Older Assisted
Non-Family

Older Assisted
Family

Newer
Assisted

Good
95

Goo
d 95

Not
Goo
d95

Not
Good
95

3.5714

Not
Good
95

Not
Goo
d 95

3.7822

6.8369

6.9773

7.3278

7.5932
14.4190

168
1

14.6645

14.8901

293

15.1958

15.6170

16.0791

16.1014

138

16.3655

249

16.3756

16.6275

16.9454

23.6667

56
TOTAL

Not in
1996
List

1
140

Abt Associates Inc.

324

196

249

Appendix A - Sampling

111

Exhibit A-5a
1989 Monitoring Sample Weights (Based on 1989 recorded stratum)
(Weights sum to 1989 estimate of universe)

Weight

Unassisted

Older Assisted
Non-Family

3.650
30

7.328

178

14.665

304

16.353

Abt Associates Inc.

Newer
Assisted

6.733

16.376

Older Assisted
Family

254
188

Appendix A - Sampling

112

Exhibit A-6
Revised Estimate of 1989 Universe Using Revised Monitoring Sample Weights
Takes into account 1989 and 1996 information. For comparison the estimate using only 1989
information (used in the "Blue Book", and the 1996 list-based estimate are also presented.

Stratum

Estimate of 1989
Universe
Using 1996 and 1989
Information

Estimate of 1989
Universe
Using Only 1989
Information

Estimate of 1989
Universe
Using Only 1996
List

Unassisted

3,021

3,080

3,067

Older Assisted
Non-Family

4,550

4,660

4,695

Older Assisted
Family

1,608

1,506

1,499

Newer
Assisted

4,179

4,154

4,146

TOTAL

13,358

13,270

13,407

Abt Associates Inc.

Appendix A - Sampling

113

Exhibit A-7
Revised Analysis Sample Weights Taking into Account 1996 Information
(Weights Sum to Revised Estimate of 1989 Universe)

Weight

Unassisted

Older Assisted
Family

4.1634

6.4601

7.6258

Newer
Assisted

7.6804

13.2547

13.7348

14.0431

23.7110

25.6811

26.4777

26.9286

26.9336

27.4865

27.2122

157

27.6760

28.1688

142

29.1670

40.7359

TOTAL

Older Assisted
Non-Family

115

185

123

14711

Includes one property where the owner owned two properties, one unassisted and one newer assisted. The newer assisted property was
sampled, but the owner took us to the unassisted property which was inspected. In the 1989 study, we re-assigned the weight as unassisted.
In fact, this property should not be included in the analysis sample, and is excluded from the Augmented Sample.

Abt Associates Inc.

Appendix A - Sampling

114

Exhibit A-7a
1989 Analysis Sample Weights (Based on 1989 recorded stratum)
(Weights sum to 1989 estimate of universe)

Weight

Unassisted

Older Assisted
Non-Family

4.056
27

13.446

26.694

28.450

Abt Associates Inc.

Newer
Assisted

7.481

26.782

Older Assisted
Family

167
115
146

Appendix A - Sampling

115

Exhibit A-8
Estimate of 1995 Universe
(= Properties that Were Insured in 1989 and still insured in 1995)
and Survival Rate
Stratum

Estimate of 1989
Universe

Estimate of 1995
Universe (1989
Universe still active
in 1995)

Survival Rate

Unassisted

3,021

2,224

73.62%

Older Assisted
Non-Family

4,550

4,388

96.44%

Older Assisted
Family

1,608

1,554

96.64%

Newer
Assisted

4,179

4,076

97.51%

TOTAL

13,358

12,242

91.65%

Exhibit A-8a
1996 List-Based Estimate of 1995 Universe
(= Properties that were in HUD's 1996 list as insured in 1989 and still insured in 1995
Stratum

1989 Universe based
on 1996 list

1995 Universe (1989
universe still active in
1995) based on 1996
list

Survival Rate

Unassisted

3,067

2,333

76.07%

Older Assisted
Non-Family

4,695

4,639

98.81%

Older Assisted
Family

1,499

1,474

98.33%

Newer
Assisted

4,146

4,076

98.29%

TOTAL

13,407

12,522

93.39%

Abt Associates Inc.

Appendix A - Sampling

116

Exhibit A-9
Augmented Sample Weights
(Weights Sum to Estimate of 1995 Universe)

Weight

Unassisted

Older Assisted
Non-Family

4.1634

Older Assisted
Family

Newer
Assisted

7.6258

7.6804

8.9926

10.5602

116

11.3324

20.7083

21.1334

22.2225

187

23.1527

175

23.9731

24.6561

26.7047

27.5330

28.0019

TOTAL

Abt Associates Inc.

216

148

176

Appendix A - Sampling

117

Exhibit A-9a
Augmented Sample Weights, Using Equal Weights Within Strata
(Separate Weights for NHP/Non-NHP Sub-strata)

Weight

Unassisted

Older Assisted
Non-Family

4.3684
26

11.4109

129

22.0526

190

23.1591

Abt Associates Inc.

Newer
Assisted

7.61538

27.4568

Older Assisted
Family

176
81

Appendix A - Sampling

118

Exhibit A-10
Comparison Sample Weights
(Weights Sum to Estimate of 1995 Universe)

Weight

Unassisted

Older Assisted
Non-Family

4.1634

Older Assisted
Family
18

7.6258

7.6804

12.1157

14.2278

14.6315

24.6561

26.7047

27.5330

27.9003
28.0019

Newer
Assisted

7
1

28.4730

28.6901

145

29.9605

135

31.0222

TOTAL

Abt Associates Inc.

173

114

136

Appendix A - Sampling

119

Exhibit A-10a
Comparison Sample Weights, Using Equal Weights Within Strata
(Separate Weights for NHP/Non-NHP Sub-strata)

Weight

Unassisted

Older Assisted
Non-Family

4.3684
26

15.4947

28.5034
29.9706

Abt Associates Inc.

Newer
Assisted

7.61538

27.4568

Older Assisted
Family

95
81
147
136

Appendix A - Sampling

120

Appendix B
DATA COLLECTION SUMMARY
Data collection on HUD-insured (or held) multifamily properties was conducted in three
phases, under three separate task orders plus an interagency agreement between HUD and
the US Army Corps of Engineers (USACE). Under the first task order (Task Order 5 under
contract HC-5964) and through the interagency agreement with the USACE, data on the
physical condition of properties were collected. Under a second task order (Task Order 6
under contract HC-5964) data on the market position of the properties were collected. Under
Task order 7 (the current task order), data on property finances, mortgages, tenants,
assistance contracts, and neighborhoods were assembled from a range of HUD and other data
files. For completeness, this appendix describes the data collection procedures for all three
tasks. Section B.1 describes the procedures for collecting physical condition data, Section
B.2 describes the market data collection, and Section B.3 describes the secondary data used.
B.1

Physical Condition

The physical condition of the stock was assessed on-site by the US Army Corps of Engineers
(USACE). The purpose of the on-site physical inspections was to obtain current information
on the physical condition of FHA-insured (or held) multifamily housing at a level of detail
sufficient to indicate the nature of physical deficiencies and the costs that would be required to
remedy the current backlog of physical needs as well as to estimate the ongoing accrual of
physical needs over the next 20 years.
The backlog of repair needs was estimated using the Observable Systems Method, which was
initially developed by Abt Associates for the 1985 Modernization Needs Study of Public
Housing1. Under this method, the condition of each property’s systems is observed,
evaluated, and assessed on-site; and then costed in a consistent manner off-site using a
regionalized data base of repair costs and a computerized costing program. The inspection
protocol included observing conditions of 119 mechanical, electrical, and architectural
systems. For each system, the inspector judged and recorded the level of remedial action
needed to restore the system to its original condition. The action levels were “No Action”,
“Minor Action”, “Moderate Action”, “Major Action”, and “Replace”, based on the observed
condition. Minor defects that could be corrected through routine maintenance (e.g. faucet
washer replacement) were excluded.

Dixon Bain et al., Study of the Modernization Needs of the Public and Indian Housing Stock (Cambridge, MA: Abt Associates Inc.,
March 1988). This inspection method proved sufficiently cost-effective that it has subsequently been adapted and used by at least one
commercial inspection firm.

Abt Associates Inc.

Appendix B - Data Collection Summary

121

The USACE inspectors used a standard set of seven inspection booklets (developed by Abt
Associates)—Site, Building Envelope, Building Mechanical and Electrical, Unit, Takeoff,
Property Quality Distribution (PQD) and Inspection Building Type and Quality ( IBTQ)—to
collect all relevant system-level information. For each observable system, the inspector noted
presence or absence of the system; age; type, if appropriate (e.g., battery or hard-wired smoke
detectors); number, if appropriate (e.g., the number of windows); and action level associated
with the observed condition2. Using architectural drawings, when available, or “pacing off”
when no plans were available, the inspectors calculated take-off measurements for site areas
and distribution systems, average unit square footage for all unit sizes present at the property,
and key building dimensions for up to three predominant types/sizes of buildings. These
measurements were recorded in the Takeoff booklet.
The inspectors were responsible for gathering three kinds of information on each property: 1)
current condition—observations that were used in the study to estimate the backlog of needs
(the cost to bring all systems up to their original condition); 2) upgrade feasibility—whether a
property could be physically upgraded to a higher market use, and information needed to
estimate costs of upgrading; 3) property take-offs—a measurement inventory of average units,
typical building dimensions, and certain systems, used by the study both in costing backlog
needs and estimating future accruals of repair/replacement costs. The inspectors also
conducted neighborhood windshield surveys and collected preliminary information that was
used as input for the market assessment team (discussed below).
For each system, the inspector judged and recorded the level of remedial action needed to
restore the system to its original condition. The action levels assigned to each observable
system condition were provided to all inspectors in training sessions and a series of manuals.
This uniform set of instructions assured consistency across individual inspectors. Exhibits B-1
and B-2 are samples of an inspection booklet and the action level description from the
Inspector Manual. The examples are taken from the "Full Bathroom" section of the "Unit
Inspection" booklet. (Exhibit B-1 is a page from this booklet.) Under the section labeled
"Full Bathrooms," are the seven systems observed in the bathroom inspection. Some systems
(walls and ceilings, accessories) require only an action level in order to estimate repair cost;
others require a type (i.e., the materials in use or size) as well as an action level for the repair
estimate. For example, under the Bathroom Floor Cover and Subbase System, "Type" is
necessary because replacing a ceramic tile floor would be more costly than replacing a
resilient tile floor or linoleum.

In this study, our assessment of physical needs excludes three categories of expenditures that many owners will be required to make:
modifications for accessibility for the disabled, as required by Section 504 of the Rehabilitation Act of 1973, as amended; measures taken
solely to mitigate hazards of lead paint or asbestos; and improvements for increasing energy efficiency. The only exception to this is that
the replacement of, for example, a heating system or appliance, assumes installing a standard quality replacement according to current
practice, and not simply replacing the old system.

Abt Associates Inc.

Appendix B - Data Collection Summary

122

Exhibit B-2 is taken from the Inspector's Manual of conditions and action levels. For each
system, the manual defines the system, explains where and how to observe the system, and
then describes the repair needs associated with each action level.
Two other forms—Project Quality Distribution and Inspection Building Type and Quality—
were used to obtain overall descriptions of the building stock and the relative quality of units
and buildings at the property.
In advance of the site visit, the inspector sent a Project Quality Distribution form to the site
manager. The manager completed the information on the number of units by size (bedrooms
and bathrooms) and condition, as well as the number of buildings by type (high-rise, walk-up,
etc.) and condition. A definition guide on conditions was attached to the form to make it
easier for the manager to categorize the units and buildings. When the inspectors arrived onsite, they reviewed the Project Quality Distribution form with the site manager and discussed
the general characteristics of the property, including:
•

Number, type (high rise or elevator buildings, low-rise, garden/townhouses, or
single-family detached), and age of buildings,3

•

Number of units by bedroom and bathroom size,3 and

•

The property manager's assessment of overall condition of buildings and units, i.e.,
what proportion the manager estimated were in excellent, good, fair, or poor
condition.4

From this composite of the property, inspectors selected up to three buildings and three units
to inspect, based on predominant quality categories and predominant building and unit types.
If multiple quality buildings were present, inspectors were instructed to inspect the lowest
quality building. Inspectors were also told to inspect at least one building containing an
elevator if one existed at the property, regardless of its likelihood to be inspected under the
guidelines based on predominant quality and type. For example, if the property had one high
rise building and twenty townhouse buildings, the inspector would inspect the high rise and
two townhouses.

Inspector recorded this information on the Inspector Building Type and Quality (IBTQ) form.

Manager and inspector recorded this information on the Project Quality Distribution (PQD) form.

Abt Associates Inc.

Appendix B - Data Collection Summary

123

Exhibit B-1

UNIT BOOKLET
KITCHENS (CONTINUED)
ABSENT

AGE

ACTION LEVEL
N/A

MIN

MOD

MAJ

REP

Refrigerator

102/

____

103-105/

106/

Garbage Disposal

107/

____

108-110/

111/

Dishwasher

112/

____

113-115/

116/

Microwave

117/

____

118-120/

121/

Trash Compactor

122/

____

123-125/

126/

KITCHEN UPGRADE FEASIBILITY
Is an upgrade necessary for market conversion?
NO

PARTIAL

FULL

REHAB

Rehab Feasible?
NO

Kitchen

127/

YES

128/

FULL BATHROOMS
NUMBER OF FULL BATHS PRESENT: ______
ABSENT

129/

TYPE

AGE

Walls & Ceilings - Partitions & Surfaces

G1
G2

Floor Cover & Sub-base

Ceramic

134/

Resilient

ACTION LEVEL
N/A

MIN

MOD

MAJ

REP

___% 4

138/

____
135-137/

130/
131-133/

Fixtures - Sink

____
140-142/

143/

139/

Fixtures - Toilet

____
145-147/

148/

144/

Fixtures - Tub/Shower

149/

____
151-153/

154/

____
155-157/

158/

____
161-163/

164/

G1
G2

Porcelain
Fiberglass

150/

Accessories
Vanities
(single = 24"
double = 36")

Abt Associates Inc.

159/

G1
G2

Single
Double

160/

Appendix B - Data Collection Summary

124

Exhibit B-2
103. Bathroom Floor Covering and Sub-base
Definition:

The floor sub-base refers to a rough floor, laid on joists, which serves as a base for the
finished floor. The floor covering could consist of tile, sheetgood, or carpet. There are two
types of floor covering:
G
Ceramic tile
G
Resilient sheetgoods

Where to Observe: The floor located in the inspected unit bathrooms should be observed.
Inspection Method:
G

G
Record whether the floor covering is ceramic or resilient.
Record the age of the floor.
G
The actual floor sub-base cannot be observed directly, but the inspector can note if
the floor is warped or buckled.

Rating of Repair Needs - Action Levels:
Minor Action:

Not applicable.

Moderate Action:
Major Action:

Not applicable.

The sheetgoods are severely deteriorated and need to be replaced.

Replace:

The floor is buckling, warped, or splintered, requiring the replacement of the floor covering
and sub-base.

104. Bathroom Fixtures
Definition:

There are two types of fixtures for a tub/shower (full bath):
G
Porcelain
G
Fiberglass

Common Elements:

Bathroom fixtures include the sink, toilet and tub.

Where to Observe:

These fixtures can be observed in the bathroom.

Inspection Method:

G
G
G

Each fixture is rated separately.
Record the age of the fixtures.
Record whether the tub/shower is porcelain or fiberglass (porcelain includes tile
and/or enamel on cast iron).

Rating of Repair Needs - Action Levels:
Sink:
Minor Action: The fittings need to be repaired or replaced.
Moderate Action:
Not applicable.
Major Action: Not applicable.
Replace:
The sink needs to be replaced.
Toilet:
Minor Action: The fittings need to be repaired or replaced.
Moderate Action:
Not applicable.
Major Action: Not applicable.
Replace:
The toilet needs to be replaced.

Abt Associates Inc.

Appendix B - Data Collection Summary

125

Exhibit B-2, continued
Tub/Shower:
Minor Action: The fittings need to be repaired or replaced.
Moderate Action:
Not applicable.
Major Action: Not applicable.
Replace:
The tub/shower needs to be replaced.

105. Bathroom Accessories
Common Elements:

Common bathroom accessories include a medicine cabinet, towel bar, shower rod, and a
wall-attached soap dish.

Where to Observe:

These items can be observed in the bathroom.

Inspection Method:

G
G
G

Record the age of the bathroom accessories.
Observe the condition of these items directly.
Ask the residents if the accessories are stable and operate properly.

Rating of Repair Needs - Action Levels:
Minor Action:

Not applicable.

Moderate Action:
Major Action:

Two to three accessories are broken or missing and need to be replaced (excluding the
medicine cabinet).

Replace medicine cabinet only.

Replace:

A majority of the accessories and the medicine cabinet are broken or missing and need to be
replaced.

106. Vanities
Definition:

This item refers to the vanity structure itself and not to the sink. There are two types of
vanities:
G
Single = 24" long
G
Double = 36" long

Where to Observe:

The vanity can be observed in the bathroom.

Inspection Method:

G
G
G

Record whether the vanity in the inspected unit is a single or double vanity.
Record the age of the vanity.
Observe the structure of the vanity by opening and closing the vanity doors;
observe the condition of the vanity directly.

Rating of Repair Needs - Action Levels:
Minor Action:

Not applicable.

Moderate Action:
Major Action:

Not applicable.

Replace:

Abt Associates Inc.

The vanity is beyond repair and needs to be replaced.

Appendix B - Data Collection Summary

126

For units, the inspectors were instructed to inspect units from the predominant sizes with the
provision that the inspectors select units that, in the manager's opinion, were in the worst physical
condition.5 If all the units at the property were in good condition, then the inspector made the
selection based solely on predominant unit size. If, however, there were units ranging in quality
from poor to excellent, the inspector would select poor, fair, and good units, and not inspect units
in excellent condition. This protocol was followed to obtain direct observations on the elements
most costly to repair. Adjustments to property-level repair costs for the relatively less expensive
repairs of better quality units are described in Appendix C.
In addition to assessing the current physical condition, inspectors provided information on the
physical (but not market) feasibility of upgrading certain observable systems for a market
conversion. They recorded this information in the inspection booklets, as shown in the example in
Exhibit B-1 for "Kitchen Upgrade Feasibility". (This information is needed to ascertain net
market value—that is, to subtract upgrade costs from capitalized net operating income for
market-level unassisted rents.) In some cases, upgrading meant adding a system if one did not
currently exist (e.g., a swimming pool). If the system already existed, upgrading it would involve
replacing it with better quality materials.
A total of 1,248 buildings and 1,563 units were inspected across the 621 properties of the
Augmented Sample.

B.2

Local Market Conditions

As was the case in the 1990 Study, Applied Real Estate Analysis, Inc. (AREA), a firm specializing
in market analysis, conducted local market assessments for all 621 properties in the Augmented
Sample. The local market assessments provided several types of key data:

•

Potential unrestricted market rents under current condition and with moderate and
major upgrades, and potential value as condominiums

•

Likely use of the property in an unrestricted market

•

Local market characteristics: vacancy rates, property appreciation rates, condominium absorption rates, capitalization rates

•

Local Section 8 success rates, and changes in success rates

The value to the study of the manager's rating of units and buildings by overall condition depended primarily on the manager's
consistency, rather than on the manager's use of the exact definition of excellent, good, fair, or poor. The inspector conducted quick
"walk-throughs" of units in the various categories, in addition to conducting the actual inspections, to verify the consistency of the
manager's ratings. If discrepancies existed, the inspector adjusted the distribution to reflect the differences.

Abt Associates Inc.

Appendix B - Data Collection Summary

127

For properties that were included in the 1990 Study, AREA started out with contact lists from
that study. For all properties, the inspectors provided AREA with important information
about the properties and their neighborhoods. The inspectors photographed the sample
properties, nearby potentially competitive properties, and some views of the neighborhood
surrounding the property to aid in defining the neighborhood context. They also conducted a
brief windshield survey, recording observations about the neighborhood such as age of
residential structures, density, non-residential uses, major amenities (such as a park or
shopping area), and any neighborhood elements that would detract from the market value
(such as an old, rundown industrial area). The inspectors also provided contact information
for local potentially comparable buildings. Abt provided AREA with 1990 Census data on each property’s neighborhood.
AREA staff conducted all surveys by telephone. They gathered information from realtors,
public housing and community development officials, and others knowledgeable about the
local market on possible alternative uses (such as condominiums, market-rate rental, or
nonresidential) and current rental market position (i.e., is the property currently a low-rent
property, a moderate rental, or a high-end luxury complex?), vacancy rates, and trends in
supply and demand. The final assessment of the possible rents and local market context of
each property was summarized on a Market Valuation Summary Guide form.
B.3

Data Collection from Secondary Data Sources

Data for this study were also extracted from secondary sources, including HUD automated
data bases and other existing data bases. For numerous categories of variables, multiple data
sources were available and hierarchies were developed for prioritizing sources. In addition,
when key data elements were missing, we developed procedures to impute variables. For
each category of data, the sources and imputation procedures are described below.
B.3.1.

Property Income and Expenses

Three separate data sources were used to assemble information on property finances.
Financial data are derived directly from annual statements of income and expense provided by
each property owner as required by HUD mortgage regulatory agreements. These data
include multiple years of income and expense statements for most properties in the sample.
HUD supplied us with Annual Financial Statement files for 1993 - 1995. The 1993 and 1994
files contained more complete information (including reserves). As backup to the annual
financial statement files we used the financial statement file from HUD’s Multifamily Data
Warehouse. The file contained income and expense data for 1992 through 1994. In addition
we obtained backup tapes for 1992 through 1994.
HUD financial data files were fairly complete. For example 540 of the 621 properties had
total revenue data for three or four of the possible four years, 53 properties had two years of

Abt Associates Inc.

Appendix B - Data Collection Summary

128

data, 16 had one year, and 12 had no financial dat. For other financial variables the coverage
was similar. The most recent year of available data was 1995, for which 354 properties had at
least some financial data.
When financial variables were missing, they were imputed by assigning the median value (per
unit) of the three year weighted average for all properties of the same building type and
assistance category. This method was used to impute gross revenue, tenant paid rents,
tenant assistance payments and total operating and maintenance costs. Vacancy losses were
imputed based on the median of the three year weighted average percent vacancy loss (rather
than the median dollar vacancy loss). Replacement reserve balances were imputed based on
the median per unit values by assistance category and building type.
B.3.2

Property Mortgage Information

Data on mortgages including, original mortgage amount, mortgage date, term, and interest
rate, and current status of the mortgage were obtained from files on HUD’s Multifamily Data
Warehouse. Information on additional loans (operating loss loans and Section 241
supplementary loans) was obtained from the Warehouse files and from the 1990 analysis file.
From the data we computed annual mortgage payments and outstanding principal balances.
B.3.3.

Assistance Contracts

Obtaining information on assistance contracts was especially important in order to verify the
classification of properties as unassisted or older or newer assisted, to determine the
percentage of property units that were covered by assistance contracts, the amount of
assistance, and the next (and ultimate) renewal dates of assistance contracts. Several of the
files provided by HUD were maintained at the contract level, and others were aggregated to
the project level. Data were taken from all files to calculate the total number of contracts,
contract amounts, assisted units and dates. Contract amounts were often missing and were
imputed using the median contract amount per year per assisted unit by building type and type
of Section 8 assistance. For properties that were missing Section 8 contract start dates or
renewal dates, we used the median date by Section 8 type.
B.3.4

Tenant Characteristics

Tenant Characteristics information was obtained from a HUD file that aggregated at the
project level data from HUD’s MTCS data base. The MTCS data base is a household-level
file that tracks tenant characteristics in assisted properties. Data were received on 460 of the
540 assisted properties in the sample (150 newer assisted and 310 older assisted) representing
8,536 of the 10,019 assisted properties in the study universe. No tenant characteristics data
were available for the unassisted portion of the stock.

Abt Associates Inc.

Appendix B - Data Collection Summary

129

Imputation rules were developed for each category of tenant characteristics data. Income
distributions were imputed based on assistance category. For older assisted properties we
assigned the mean income distribution based on the proportion of assisted units (under 50
percent assisted, or over 50 percent assisted) in the property. For newer assisted properties,
we assigned the overall mean income distribution, since nearly all newer assisted properties
received assistance for all units.
Race and ethnicity were imputed based on the composition of properties in racially similar
census tracts. For each property we created a flag, which indicated the dominant group in its
census tract (white non-Hispanic, Black non-Hispanic, Hispanic regardless of race, other).
Then, we calculated the mean racial and ethnic composition for all properties that provided
data by assistance by dominant group in the Census tract. For properties that were missing
racial and ethnic composition we assigned the mean distribution based on properties located in
tracts with the same predominant group, by assistance category.
Household size distribution, other income characteristics (percent with some public assistance,
percent with some SSI, percent with some wages), and percent elderly and percent
handicapped households were imputed based on the reported occupancy type (family/elderly/handicapped) and assistance category. For each of these variables we calculated
the mean distribution by assistance category and occupancy type among properties that had
data. These means were assigned to properties with missing data.
B.3.5

Neighborhood Characteristics

In addition to the information obtained from the inspectors and the market analysts,
characteristics of property neighborhood data included two additional categories of variables.
In describing characteristics of the properties’ neighborhoods, we used the characteristics of
their 1990 Census tracts (or zip codes when addresses could not be geocoded), as well as
FMRs and HUD median incomes.
All properties were geocoded using MapMarker version 2.0. Properties with very incomplete
address information could not be geocoded (e.g. addresses such as “Orleans & Illinois” and
“Kershaw St” ) and therefore, for 103 properties 1990 Census zip code-level data were used
instead of the tract-level data. All 1990 Census dollar values were inflated to 1995 dollars
using the CPI for Urban Consumers (i.e. multiplied by 153.4/126.1=1.2173), and then were
rounded to the nearest $1,000. For example, the 1990 Census ranges for income distribution
include “0- $9,999". On average across the stock neighborhoods 25 percent of households
fell into that income category. “0 - $9,999 translates into $12,172 in 1995 dollars (9999 x
1.2173), which gets rounded to $12,000. Thus the bottom income range in the table is “0$12,000 and includes 25 percent of households in the neighborhoods of the insured stock.

Abt Associates Inc.

Appendix B - Data Collection Summary

130

Data on 1996 Fair Market Rents (FMRs) and 1995 and 1996 local median incomes were
downloaded from HUD’s Web Page based on State, County and MSA.
B.3.6.

Other Sources

As backup to the data sources listed above, we also used the 1990 analysis and monitoring
files as well as HUD’s 1996 and 1989 MIDLIS files.

Abt Associates Inc.

Appendix B - Data Collection Summary

131

Appendix C
SYSTEM FOR ESTIMATING PHYSICAL NEEDS BACKLOG AND
ACCRUAL COSTS FROM INSPECTIONS
This appendix outlines the approach used to relate observations made by inspectors to costs of
repairs and replacements. The first section describes the methods for arriving at costs of the
backlog of currently needed repairs and replacements ("physical needs backlog costs").
The second section describes the method to obtain upgrade feasibility costs, that is, costs that
would have to be expended in order to convert a property to higher quality ("upgrade costs").
The third section addresses the method for estimating future accrual of major repair and
replacement costs ("accrual costs").

C.1

Estimating Physical Needs Backlog Costs from Property Inspections

The process of estimating repair costs based on the property inspections involved five steps:
•

Conducting a physical inspection of the overall site and up to 3 buildings and
units within each project in the sample;

•

Generating a system-level cost file (119 systems or groupings of physical features
were inspected in the properties);

•

Calculating system-level costs for the site and inspected units and buildings;

•

Computing property-level costs by inferring costs for uninspected units and
buildings from inspected units and buildings;

•

Regionally adjusting the property-level costs.

Physical Inspection of the Property

The physical inspection method—the Observable Systems Method—was described previously
in "Appendix B: Data Collection Summary." The inspection produces information for each
property on: the current condition and required repair action level for each of 119 systems for
the site and for the buildings and units that were inspected; upgrade feasibility to higher
market use; and property take-offs—a complete inventory of the presence, count, age, type,
and dimensions of components.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

132

System Level Cost File for Computing Physical Needs Backlog

As was discussed in Appendix B, under the Observable Systems Method, the costs of carrying
out the repair actions recorded by the inspector were computed off-site using a computerized
cost file and program. The first step in generating the cost file was developing up to five
system-specific, categorized levels of repair, ranging from no action to replacement of a
system, to correspond to action levels the inspector would use to describe the repairs needed
to bring the system up to a safe, sound, marketable condition. The action level groups are:
•
•
•
•
•

NA for no action
MIN for minor repair
MOD for moderate repair
MAJ for major repair, and
REP for replace

For any system, each action level denotes a specific repair action. For example, for Kitchen
Cabinets/Counters/Sinks (a dwelling unit system), the MIN action is to replace countertop or
sink faucet; the MOD action is to refinish existing cabinets, or repair doors and drawer hinges
as well as replacing anything covered under MIN; MAJ includes the components of MOD as
well as replacing the countertop and sink; and REP includes all MAJ components plus
replacing countertop, cabinets and sink. In the above example for cabinets/countertops/sinks,
the MIN cost is $732 for each kitchen requiring MIN action. MOD costs are $800 for each
kitchen requiring a MOD level of repair. MAJ costs are $1,532 for each kitchen. REP costs
are $2,500 per kitchen. Costs for each action level for each system are presented in Exhibit
C.1. Not all systems have 5 action levels. The Inspection Handbook for this study details
each allowable action level for each system.
As in the 1990 Study, we obtained the services of A.M. Fogarty and Associates, a firm with
extensive experience in costing for private housing construction and modernization, to review
the cost file developed for the 1990 Study and to update cost elements which corresponded to
each system and action level combination.
System Level Costs for the Site and Inspected Units and Buildings

In this step, the inspector's observations and the cost files are combined to calculate, for each
property, costs for repair actions on items that have been inspected. A mathematical
algorithm is applied to each system the inspector checked off as needing some level of repair.
The basic concept is multiplying unit cost by a quantity measure, where the quantity measure
may be scaled by a percentage of the item affected.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

133

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

ASSUMPTIONS

***SITE SYSTEMS
Landscape

0.11

0.26

0.78

1.05

Landscape-SF

Roadways

0.15

0.35

0.70

1.74

Road-SF (min 1000)

Parking Areas-Lots

0.10

0.50

1.00

1.74

Parking-SF;# of new spaces

Parking Areas-Garages

0.02

0.70

1.00

1.74

Park-SF (min 1000)

Paved Pedestrian Areas

0.30

0.76

1.89

3.73

PvdPed SF (min 1000)

Curbing-Bituminous

N/A

4.41

Curbing LF

Curbing-Concrete

N/A

7.57

15.24

Curbing LF

Curbing-Granite

N/A

2.18

7.88

N/A

Curbing LF

Fencing-Chain Link

N/A

14.50

Fencing LF

Fencing-Wrought Iron

N/A

57.84

Fencing LF

Fencing-Wood Stockade

N/A

15.27

Fencing LF

Retaining Walls-Concrete

N/A

2.99

N/A

34.04

Retain Wall-LF

Retaining Walls-RR Ties

N/A

0.67

N/A

24.17

Site Drainage-Underground

N/A

900.00

2670.00

4500.00

# Catch Basin
Landscape SF

Site Drainage-Surface

Retain Wall-LF

N/A

0.33

2.50

N/A

500.00

800.00

1350.00

3500.00

# Poles

36.00

120.00

# Units

N/A

485.00

N/A

970.00

# Yards

Dumpsters and Enclosures

1000.00

2500.00

4300.00

5700.00

Swimming Pool

5140.00

7864.00

10588.00

34475.00

# Pools

Tennis Courts

2678.00

2940.00

11655.00

24194.00

# Courts

Basketball Courts

2248.00

3094.00

N/A

10024.00

# Courts

Site Electrical Distribution-Over

N/A

95.00

130.00

Site Elec Dist-LF

Site Electrical Distribution-Under

N/A

115.00

150.00

Site Elec Dist-LF

Heating Water Distribution-Steam

N/A

325.00

Heat Water Dist-LF

Heating Water Distribution-Hot Water

N/A

175.00

Heat Water Dist-LF

Domestic Hot Water Lines

N/A

40.00

Dom Hot Water -LF

Domestic Cold Water Lines

N/A

25.00

Dom Cold Water-LF

Main Water Service

N/A

45.00

Main Water Serv-LF

Gas Lines

N/A

30.00

Gas Lines-LF

Site Sanitary Lines

N/A

40.00

Site Sanitary-LF

Septic System

N/A

350.00

8000.00

Units

800.00

1400.00

2000.00

3500.00

# Ejectors

N/A

2000.00

# Hydrants

200.00

500.00

7500.00

16000.00

Pole Mounted Site Lighting
Site Furniture
Private Yards and Enclosures

Sewage Ejectors
Hydrants
Emergency Generator

360 SF per space

Min 10% of units, Mod 25%, Maj 66%, Replace 100%

# Dumpsters

1 per project

Double Court

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

***UNIT SYSTEMS
Walls & Ceilings: Partions (not K&B)

N/A

3.00

Floor Sub-base (not K&B)

N/A

3.35

Walls & Ceilings: Surfaces (not K&B)

0.58

1.08

N/A

1.58

Floor Covering-Carpet (not K&B)

N/A

1.65

Floor Covering-Resilient (not K&B)

N/A

2.43

Interior Doors & Frames

N/A

50.00

256.00

400.00

Kitchen Walls & Ceilings:

0.70

1.25

N/A

3.00

SF
SF

# Doors need act

Partions & Surfaces
Kitchen Floor Covering & Sub-base

N/A

3.30

6.65

732.00

800.00

1532.00

2500.00

# needing replacement

50.00

N/A

500.00

N/A

# needing replacement

100.00

258.00

500.00

758.00

# needing replacement

Refrigerator

N/A

768.00

# needing replacement

Garbage Disposal

N/A

180.00

# needing replacement

Dishwasher

N/A

522.00

# needing replacement

Microwave

N/A

275.00

# needing replacement

Trash Compactor

N/A

516.00

# needing replacement

Bathroom Walls & Ceilings:

0.58

4.20

8.80

12.00

N/A

9.24

12.59

3.30

6.65

Cabinets/Counter Top/Sink
Range
Range & Hood

Partions & Surfaces
Bathroom Flr Cvr & Sub-base-Tile
Bathroom Flr Cvr & Sub-base-Resil
Bathroom Fixtures-Sink

N/A

393.00

# needing replacement

Bathroom Fixtures-Toilet

150.00

N/A

361.00

# needing replacement

Bathroom Fixtures-Tub/Shower

200.00

N/A

821.00

# needing replacement

Bathroom Accessories

N/A

100.00

160.00

230.00

# needing replacement

Bathroom Vanities-24"

N/A

24" 387.00

# needing replacement

Bathroom Vanities-36"

N/A

36" 552.00

# needing replacement

HVAC Unit-Heat Only

N/A

400.00

N/A

972.00

# needing replacement

HVAC Unit-Heat/Cool

N/A

700.00

N/A

5370.00

# needing replacement

ASSUMPTIONS

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

***UNIT SYSTEMS (Continued)
Radiation-Hydronic

N/A

9.72

N/A

19.44

Radiation-Electric

N/A

10.54

Unit Boiler

N/A

800.00

N/A

2730.00

# needing replacement

Unit Furnace

N/A

500.00

N/A

1110.00

# needing replacement

Unite Dom Hot Water Generation

N/A

150.00

N/A

450.00

# needing replacement

Temperature Controls

N/A

64.80

Wall/Window Air Conditioner

N/A

750.00

Unit Electrical Panel

N/A

1260.00

Unit Electrical Wiring

N/A

3.50

Bell/Intercom System

N/A

182.00

# needing replacement

Closed Circuit TV

N/A

100.00

# needing replacement

Emergency Call Alarm System

N/A

125.00

# needing replacement

Smoke/Fire Detection-Battery

N/A

100.00

# needing replacement

Smoke/Fire Detection-Hard Wire

N/A

153.00

# needing replacement

# Temp Controls
# Wall/Window AC’s
1 per Unit
Total Unit SF

ASSUMPTIONS

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

ASSUMPTIONS

***BUILDING ENVELOPE
Foundation-4 Foot

0.65

N/A

12.52

N/A

Perimeter-LF

Foundation-8 Foot

N/A

5.30

21.17

N/A

Perimeter-LF

Slab--Slab

N/A

0.45

2.23

5.65

Footprint-SF

Slab-Basement

N/A

0.45

2.46

6.40

Footprint-SF

Exterior Wall-Masonry

1.00

N/A

3.94

14.70

Masonry-SF

Exterior Wall-Plaster

1.00

N/A

1.44

7.20

Plaster-SF

Exterior Wall-Wood

1.00

1.25

1.17

5.83

Wood-SF

Exterior Wall-Vinyl/Aluminum

1.00

1.25

0.74

3.84

Vinyl/Aluminum-SF

Insulation-Wall

N/A

0.42

N/A

Insulation-Ceiling

N/A

0.88

N/A

Roof Covering-EDPM

0.25

1.00

1.95

4.21

Roof Covering-Shingle

0.18

0.80

1.46

2.10

Roof Covering-Built-Up

0.23

1.20

1.87

4.67

Roof Covering-Tile

0.95

1.80

7.20

8.02

Roof Covering-Metal

1.20

2.40

9.60

10.56

Parapet Wall

N/A

72.00

N/A

72.00

115.75

N/A

1064.88

Perimeter-LF

3 ft high

# Chimneys

Roof Hatches-Small

N/A

2'x2'x4' high

600.00

# Roof Hatches

Roof Hatches-Medium

N/A

< 10 SF

N/A

786.00

# Roof Hatches

Roof Hatches-Large

N/A

10-20 SF

N/A

1434.00

# Roof Hatches

Skylights-Small

20-30 SF

N/A

474.00

# Skylights

< 10 SF

Skylights-Medium

N/A

606.00

# Skylights

10-20 SF

Skylights-Large

N/A

795.00

# Skylights

20-30 SF

Penthouses-Small

N/A

2534.00

4540.00

# Penthouses

4'x10'x8'

Penthouses-Medium

N/A

5300.00

10300.00

# Penthouses

8'x14'x10'

Penthouses-Large

N/A

9600.00

19600.00

# Penthouses

20'’x20'x10'

Roof Drainage-Exterior

N/A

2.00

SF
SF

Chimney (Brick)

Roof Drainage-Interior

N/A

1.00

N/A

Windows-Small

180.00

230.00

250.00

420.00

# Windows (need act)

<15 SF

Windows-Medium

360.00

410.00

500.00

630.00

# Windows (need act)

<30 SF

Windows-Large

540.00

590.00

650.00

1080.00

# Windows (need act

>30SF

N/A

75.00

297.00

# Grates (need act)

Exterior Common Doors-Wood

200.00

N/A

605.00

747.00

# Doors (need act)

Exterior Common Doors-Metal

200.00

N/A

670.00

812.00

# Doors (need act)

Exterior Common Doors-Glass

200.00

N/A

845.00

987.00

# Doors (need act)

Window Security Grates

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

ASSUMPTIONS

***BUILDING ENVELOPE
Unity Entry Doors-Wood

200.00

N/A

605.00

747.00

# Doors (need act)

Unity Entry Doors-Metal

200.00

N/A

670.00

812.00

# Doors (need act)

Unity Entry Doors-Glass

200.00

N/A

845.00

987.00

# Doors (need act)

*See Note 1

Storm/Screen Doors

N/A

325.00

# Doors (need act)

Canopies-Small

N/A

230.00

N/A

830.00

# Canopies (need act)

6'x4'

Canopies-Medium

N/A

921.00

N/A

3321.00

# Canopies (need act)

6'x16'

Canopies-Large

N/A

2880.00

N/A

10380.00

# Canopies (need act)

10'x30'

Exterior Stairways-Wood

350.00

N/A

750.00

1980.00

# Flights need act

Exterior Stairways-Concrete

550.00

N/A

1100.00

4550.00

# Flights need act

Building Mounted Site Lights

N/A

300.00

N/A

600.00

# Lights (need act)

Fire Escapes

N/A

350.00

N/A

7260.00

Balconies-Wrought Iron

30.80

111.62

N/A

# Balconies (need act)

Balconies-Wood

38.50

50.65

1250.00

2570.00

# Balconies (need act)

Balconies-Masonry

36.00

120.00

N/A

# Balconies (need act)

Porches (w/roof)

N/A

900.00

N/A

6400.00

# Porches need act

Decks (without roof)

N/A

800.00

N/A

3840.00

# Decks need act

Attached Storage Sheds

N/A

400.00

N/A

1100.00

Vestibules

2..25

N/A

4.20

N/A

Vestibules SF

Corridors

2.25

3.25

5.75

6.75

Corridors SF

Stairways

2.75

3.75

9.00

30.85

SF2

Interior Lighting

N/A

1.00

N/A

2.75

SF3

Mail Facilities-Kiosk

N/A

1200.00

Mail Facilities-Box

N/A

70.80

Laundry Rooms

2.25

4.75

7.00

25.00

Laundry Equipment

N/A

200.00

600.00

1000.00

Common Rooms

2.25

4.20

N/A

7.00

Common Kitchens

4.20

7.00

N/A

40.00

Underground Garage

0.02

0.70

1.00

N/A

# escps*#stories

# Sheds need act

# Mail Facilities
# Mail Boxes
SF
# pieces of equip
Common Room -SF
Common Kitchen -SF

16 per mount

Exhibit C.1
SYSTEMS

MINOR

MODERATE

MAJOR

REPLACE

UNIT OF MEASURE

**BME
Heating Risers

100.00

125.00

175.00

250.00

Units

Gas Distribution

100.00

175.00

275.00

350.00

Units

Dom. Hot & Cold Water Distribution

125.00

225.00

350.00

450.00

Units

Sanitary Distribution-PVC

50.00

100.00

200.00

300.00

Units

125.00

225.00

350.00

450.00

Units

Fire Sprinkler System

0.25

0.50

2.00

35000.00

Sump Pumps-Residential

N/A

200.00

N/A

500.00

# Sump Pumps

Sump Pumps-Commercial

N/A

400.00

N/A

1600.00

# Sump Pumps

Compactors-Small

N/A

1000.00

N/A

5000.00

# Compactors

Compactors-Large

N/A

1060.00

N/A

10000.00

# Compactors

Central Vent & Exhaust

N/A

0.75

1.25

2.00

Area SF

Central Air Conditioning

N/A

1.50

N/A

4.50

Area SF

Switchgear

N/A

0.35

1.50

Area SF

Building Power Wiring

N/A

2.25

Area SF

Emergency Lights

N/A

425.00

Bldg Units / 6

Smoke/Fire Detection-Battery

N/A

100.00

Bldg Units / 4

Smoke/Fire Detection-Hardwire

N/A

153.00

Bldg Units / 4

Communication System

N/A

225.00

Bldg Units / 6

Emergency Call Alarm System

N/A

310.00

Bldg Units / 6

Master TV Antenna

N/A

2000.00

4500.00

Bldg Units / 6

Closed Circuit TV

N/A

1200.00

Bldg Units / 6

Hot Air Furnace

N/A

225.00

675.00

1500.00

Units

Boilers

N/A

350.00

550.00

1500.00

Units

Boiler Room-Piping

172.50

287.50

460.00

759.00

Units

Boiler Room-Equipment

200.00

500.00

1000.00

1500.00

Units

N/A

150.00

N/A

600.00

Units

50.00

N/A

200.00

250.00

Units

Elevator Shaftways-Hoist

2500.00

3000.00

5000.00

10000.00

Num Elevators

Elevator Shaftways-Hydraulic

3000.00

4000.00

6000.00

12000.00

Num Elevators

500.00

1500.00

2000.00

3000.00

Num Floors

Cabs

1000.00

2500.00

3500.00

Num Elevators

Machinery-Hoist

2000.00

4000.00

6000.00

20000.00

Num Elevators

Machinery-Hydraulic

5000.00

10000.00

15000.00

25000.00

Num Elevators

Sanitary Distribution-Cast Iron

Boiler Room-Controls
DHW Generation

Shaftway Doors

SF for minor-major; 1 for replace

ASSUMPTIONS

For example, for Roadways the algorithm first checks to see if the Roadway square feet (SF)
noted on the Takeoff form is larger than 1,000—the minimum SF allowed by the calculation.
The algorithm then multiplies the larger of Roadway SF or 1,000 SF by the cost element
associated with the Roadway action level noted on the Site booklet. In addition, if the action
level is MIN, then only 10 percent of the SF is used (still keeping 1,000 as a minimum
however), as the definition of the MIN action for Roadways is to "patch, pave and regravel up
to 10 percent of the roadway." Thus to calculate the MIN cost for a 25,000 SF Roadway, the
algorithm would be: $0.15 (cost per SF/minimum action) * 25,000 (# of SF) * 0.10 (% of
system affected) = $375. For a MOD action on the same system and property, the inspector
would have noted on the Site form the percentage (between 10 and 50 percent) of the
Roadway that needs to be resurfaced, regraded as well as repaved or regraveled. Thus, if the
inspector estimated that 35 percent of the roadway needed repair, then the algorithm would
be: $0.35 * 25,000 sq.ft. * 0.35 = $3,062.50.
A COBOL program was written to process the clean database by relating all the inspection
data collection instruments to each other via the HUD Project ID. A physical needs backlog
cost is then calculated for each system that required some repair or replace action. Some of
the algorithms make use of the takeoff data as in the above example on Roadways. This cost
element is on a per square foot basis. Other cost algorithms are based on the number, such as
the number of windows, that required the action. Exhibit C.1 shows the multiplier for each
cost element in addition to showing the cost for each action level.
After the per system costs are calculated, they are grouped together to form analysis groups.
For example, the Envelope system group called Windows and Doors includes the inspection
systems: Windows, Window Security Grates, Exterior Common Doors, Unit Entry Doors,
and Storm/Screen Doors. Exhibit C.2 shows which Observable systems are included in each
analysis group.
Property Level Costs

In order to generate costs for the property as a whole, costs for buildings and units that were
not inspected needed to be estimated.1
For each property, costs were generated for the buildings and units that were not inspected
based on their relationship to buildings and units that were inspected. During the inspection,
the inspector filled out an additional form—the Inspector Building Type and Quality Form
(IBTQ). For each building in the project (whether inspected or not), the inspector recorded
the age, overall building quality, the building type (High rise, Walk-up, Garden, Single family

This is not true for Site systems because all site elements were inspected.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

140

Exhibit C.2
System Groups and the Associated System Components for Backlog Repair Cost Estimates

System Group Name

System Component

Unit Interior Construction

Interior Walls-Partitions
Floors: Sub-base

Unit Interior Finish

Interior Walls-Surface
Floor Covering: Carpet
Floor Covering: Resilient
Interior Doors
Kitchen Walls
Kitchen Floor
Bathroom Walls
Bathroom Floor

Kitchen Fixtures

Kitchen Cabinet/Counter
Kitchen Range
Refrigerator
Garbage Disposal
Dishwasher
Microwave
Trash Compactor

Bathroom Fixtures

Bathroom Fixtures
Bathroom Accessories
Vanities

Unit Heating and Cooling

HVAC units
Radiation
Boiler (Unit level)
Furnace (Unit level)
DHW Generation (Unit level)
Temperature Control
Wall Air Conditioner

Unit Electric

Electrical Panel
Electrical Wiring
Bell/Intercom
CCTV
ECAS
Smoke Detector

Building Exterior Closure

Foundation
Slab
Exterior Wall
Insulation

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

141

Exhibit C.2, continued
Roofs

Roof Covering
Parapet Wall
Chimney
Roof Hatches
Skylight
Penthouse
Roof Drainage

Windows and Doors

Windows
Security Grates
Exterior Common Doors
Unit Entry Doors
Storm/Screen Doors

Exterior Features

Canopies
Exterior Stairs
Bldg Mounted Site Lights
Fire Escapes
Balconies
Porches
Decks
Sheds

Common Areas

Vestibules
Corridors
Interior Stairways
Interior Lights
Mail Facilities
Laundry Rooms
Laundry Equipment
Common Rooms
Common Kitchens
Underground Garages

Building Mechanical and Electric

Heating Risers
Gas Distribution
Domestic Hot/Cold Water Dist
Sanitary Distribution
Fire Sprinkler System
Sump Pump
Compactors
Switchgear
Building Wiring
Emergency Lights
Building Smoke Detector
Communication System
Building ECAS
Master TV Antenna

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

142

Building CCTV
Exhibit C.2, continued
Building Heating and Cooling

Central Vent/Exhaust
Central Air Conditioning
Furnace (Building level)
Boiler (Building level)
Boiler Room Piping
Boiler Room Equipment
Boiler Room Controls
DHW Generation

Elevators

Shaftways
Shaftway Doors
Cabs
Machinery

Site Areas

Landscaping
Roadways
Parking Lots
Parking Garages
Paved Pedestrian Area
Curbing
Fencing
Retaining Wall
Site Drainage
Pole Mounted Site Lighting

Site Amenities

Site Furniture
Yards and Enclosures
Dumpsters
Pool
Tennis Courts
Basketball Courts

Site Distribution

Emergency Generator
Site Electrical Distribution
Hot Water Distribution
Domestic Hot Water Lines
Domestic Cold Water Lines
Main Water Service
Gas Lines
Site Sanitary Lines
Septic System
Sewage Ejectors
Hydrants

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

143

detached), and a count of units in each size category (0BR/1Bath, 1BR/1Bath, 2BR/1Bath,
2BR/1+Baths, 3BR/1Bath, 3BR/1+Baths, 4BR/1Bath, 4BR/1+Baths) in the building.
Another form, the Project Quality Distribution Form, (PQD) was completed by the property
manager and reviewed by the inspector. The purpose of the PQD form was to collect data, at
a property level, on how many units overall (without a breakdown at the building level) in
each size category (bedrooms and baths) fell into each quality category (Excellent, Good,
Fair, and Poor).
In order to estimate the backlog cost for the uninspected units, the first step was to compute
per square foot costs for each inspected dwelling unit (the physical needs backlog costs for
the inspected units divided by the overall square feet for the particular units). The estimated
backlog costs for the uninspected units was then simply their square footage multiplied by the
average repair costs of inspected units of the same quality category. This was straightforward
because inspectors had recorded average size in square feet of each unit size.
Estimating the backlog cost for uninspected buildings was similar, but more complex because
inspectors did not collect square footages of uninspected buildings. In order to be able to
apply costs from the inspected sample to the uninspected sample, the costs for the inspected
buildings had to be normalized to account for differences in building sizes. We chose to
normalize building costs to a per 2-bedroom equivalent. The computation to normalize the
inspected building costs was as follows:

Overall national average square feet for each unit size category were calculated
as a weighted average of the square footage of all units in all buildings in the
analysis sample properties, regardless of whether the building was inspected.
The weights were the unit size distributions in each building.

The number of 2BR/1Bath equivalent units in each building was calculated as
the total square footage of living space in each building divided by the 1990
national average square footage of a 2BR/1 bath unit (843.9 sq. ft.).2 The total
square footage of living space was calculated by multiplying the national
average square feet for each unit size by the number of units of that size in the
building.

Building costs for each inspected building were normalized to a per-2BR cost
equivalent by dividing total costs by the number of 2BR equivalent units in the
building, thus generating a normalized cost for the inspected building which
could then be applied to the uninspected buildings.

For comparability, the 1990 average square footage was used.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

144

For example, Project X has 3 buildings. Building 1 is composed of 10 studio apartments, 20
1BR/1 Bath, and 10 3BR/1+ Bath. Building 2 has 20 2BR/1+ Bath. Building 3 has 10
4BR/1+ Bath. Based on the full sample of projects, the average square feet for a studio is
460.4; a 1BR/1 Bath is 640.3; a 2BR/1+ Bath is 1016.9; a 3BR/1+ Bath is 1160.3; and a
4BR/1+ Bath is 1342.7. The national average square feet for a 2BR/1 Bath used is 843.9.
Thus, the number of 2BR equivalents for Building 1 was ((10*460.4) + (20*640.3) +
(10*1160.3)) / 843.9 or 34.38. Building 2 has (20*1016.9) / 843.9 or 24.1 2BR equivalents.
Building 3 has (10*1342.7) / 843.9 or 15.91 2BR equivalents. Building 1's costs were divided
by 34.38; Building 2's by 24.1; and Building 3's by 15.91, to obtain cost per 2 bedroom
equivalent for each building.
Based on the assumption that buildings or units of the same type within the project will have
similar costs, costs for the uninspected units and properties were generated in one of three
ways:

•

Same type-same quality. If the inspection included a building of the same type
and quality as the uninspected building, the normalized 2BR equivalent cost (in
the inspected building) was multiplied by the number of 2BR equivalent units
in the uninspected building to produce the uninspected building's cost.
Similarly, if the inspection included a unit of the same size and quality as the
uninspected unit, its per square foot cost was multiplied by the total square feet
of the uninspected unit to generate the cost for that uninspected unit.

•

Same type-different quality. Ratios between quality categories within type
were calculated using the normalized costs for the inspected buildings or units.
If multiple inspected buildings (or units) of the same type but with different
quality existed for the project, the inspected building (or unit) with the closest
quality was used as a cost reference point. (Inspected buildings or units with
poorer qualities were chosen if a choice needed to be made. In other words, if
a Good high rise needed to be costed and both an Excellent and a Fair high rise
had been inspected, the Fair high rise would have been chosen as the reference
point.) Once the inspected reference point was chosen, the normalized cost
was first multiplied by the national average ratio between the costs for the
uninspected and inspected qualities for that building or unit type. In the above
example, the normalized cost for the Fair high rise would have been multiplied
by the ratio between the national average for a Good high rise to the national
average for a Fair high rise. Next, the cost was multiplied by the appropriate
factor3 for the uninspected building or unit.

For buildings, the factor is the number of 2BR equivalents discussed above. The factor for units is the total square feet for the unit.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

145

•

Different type.4 If the inspection included no building of the same type (or unit
of the same size), the ratio between the project cost and the national average
cost for inspected buildings (or units) was applied to the national average cost
for the type being costed. This ratio equals the sum of the actual inspected
costs for the project divided by the sum of the national weighted costs (i.e., the
costs for the inspected buildings using national average costs.). To cost
buildings or units with different types than those inspected in the project, the
national averages for the uninspected type and quality were multiplied first by
this project-to-national ratio, and then by the appropriate factor (either number
of 2BR equivalent units or square feet) for the uninspected building or unit
being costed. For example, if a Good high rise existed in a project for which
only Poor walk-ups had been inspected, a project-to-national ratio would have
been calculated by dividing the sum of the inspected Poor walk-up building
costs by the national average for a Poor walk-up multiplied by the number of
2BR equivalents for each inspected building in the project. The national
average for a Good high rise would then be multiplied by this project-tonational ratio, and then multiplied by the number of 2BR equivalents in the
Good high rise being costed.

Regional Adjustment to the Property Level Cost numbers

The cost element numbers created by A.M. Fogarty and Associates were based on current
costs for the Washington D.C. area. Using the R.S. Means "Location Factors" from the
Means Square Foot Costs Book for 1995, the property level physical needs backlog costs
were adjusted by multiplying them by the ratio of the R.S. Means Index for the city where the
property is located to the R.S. Means index for Washington D.C. (which is 0.95). For
example, the computed cost for a New York City property would be multiplied by 1.4105
(which is the New York-to-Washington index ratio, 1.34 / 0.95).

C.2

Upgrade Feasibility Costs

In addition to assessing the current physical condition of the properties, the inspectors
recorded in the inspection booklets, information on the physical feasibility of upgrading
certain observable systems for both a moderate and major market conversion. This
information is needed to ascertain net market value—that is, to subtract upgrade costs from
capitalized net operating income for market-level unassisted rents. The inspector rated the
feasibility of upgrading the property market level by adding amenities or improving the quality

Based on the inspection protocol, this occurrence was rare, arising only when a property contained a great diversity of building types and
quality levels. The occurrence was greater for units, however, due to the limit of 3 unit inspections per property.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

146

of materials in an existing system in order to make the property and its units marketable at a
higher rent level. Two levels of upgrading were possible: upgrading the property to a
"moderate" market quality, and upgrading the property to a high or "major" market quality. A
"moderate" market quality is defined as an average quality unit, generally in good condition,
with average amenities. A "major" market quality unit would command a high rent and may
include such amenities as tennis courts, swimming pools, and central air conditioning.
If the current condition and amenities already positioned the property into the "moderate"
market category, the upgrade feasibility rating was limited to "major" market feasibility. If the
property was already at a high-end market rent, no upgrade feasibility analysis was necessary.
In addition, if the layout or size of the buildings or units was not conducive to the upgrades
needed, the property was deemed infeasible for that upgrade level.
Upgrade actions could also be affected by physical needs backlog. In some cases, upgrading
meant adding a system if it did not currently exist (e.g., adding a swimming pool). If the
system already existed, upgrading it would involve replacing it with better quality materials
than would be used for repair. Some upgrade system costs are "additive" to the backlog
repair cost—the backlog repair would still have to occur before upgrading the system. An
example is Landscaping. If the backlog repair action requires a portion of the current
Landscaping to be reseeded, this would have to occur regardless of the Landscaping upgrade.
Other systems have "instead of" costs. This means that the backlog repair action would not
occur if the property were being upgraded. For example, there would be no reason to repair
windows that were being replaced with better quality materials. Exhibit C.3 lists for each
Upgrade system, whether its associated cost is additive to, or replaces the physical needs
backlog cost.
The method of calculating upgrade costs is similar to that used for physical needs backlog
costs. Cost elements were derived by A.M. Fogarty and Associates. Exhibit C.4 lists these
elements for each Upgrade system as well as the dimensional multiplier. For most systems
two levels of upgrade are possible. For units there is a further distinction: to Partial and Full
for Moderate Upgrade, and Rehab for Major Upgrade. Site and BME systems only allow for
moderate upgrades. If the system is present, then the upgrade is Moderate Yes; if the system
is not present, then the upgrade is Moderate Add. Envelope systems also break down
Moderate upgrade to Yes, for present systems, and Add when the system is not present. In
addition, major Yes is an upgrade option for Envelope systems.
Exhibit C.3
Upgrade Systems—Additive to Repair Backlog vs Instead of Repair Backlog

Additive Systems
Abt Associates Inc.

Instead of Systems
Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

147

Landscaping
Emergency Generator

Parking
Site Lighting
Yards and Enclosures
Swimming Pool
Tennis Court
Basketball Courts

Exterior Stairs
Bldg Mounted Site Lights
Porches
Decks
Sheds

Exterior Wall
Windows
Exterior Common Doors
Vestibules
Interior Lights

Corridors
Stairways

Mail Facilities
Laundry Facilities
Common Rooms

Central Vent/Exhaust
Central Air
Smoke Detector
Communication System
ECAS
CCTV

Electrical Service
Emergency Lights
Furnace
Boiler
DHW Generation
Elevator Cabs
Unit Interiors
Unit Kitchen
Unit Bathroom

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

148

Exhibit C.4: Upgrade System Elements
UPGRADE COSTS

MODERATE

ADD

MAJOR

REHAB

UNIT OF MEASURE

ASSUMPTIONS

***SITE UPGRADES
Landscape

0.65

Landscape-SF

Parking

1.75

735.00

1.75

Parking-SF;# of new spaces

1400.00

3500.00

1400.00

Site Furniture

180.00

# Sites/UPG; # site units divided by 6 for add

Yards-Upgrade

485.00

# Yards or # Site Units

Site Lighting-Upgrade

Yards-Add
Swimming Pool

970.00
30000.00

Tennis Courts

30000.00

20000.00

1 per project

20'x40'x6' deep

1 per project

20000.00
16000.00

400 SF

# Site Units

25000.00

Basketball Courts
Emergency Generator

65000.00

360 SF per space

# Poles for UPG; # site units/12 for add

1 per project
16000.00

1 per project

***ENVELOPE UPGRADES
Ext Wall-Plaster

7.20

Ext Wall-Plaster:SF

Ext Wall-Wood

5.83

Ext Wall-Wood:SF

Ext Wall-Vinyl

3.84

Ext Wall-Vinyl: SF

Windows-Small

483.00

546.00

# Small Windows

Windows-Medium

725.00

819.00

# Medium Windows

Windows-Large

1242.00

1404.00

# Large Windows

Common Doors

2000.00

3000.00

# Common Doors

Exterior Stairs

2000.00

# Stairs

Bldg Mtd Site Lights

400.00

800.00

400.00

# Bldg Units divided by 6

Porches

6400.00

# Bldg Units

8'x16'

Decks

3840.00

# Bldg Units

8'x16'

Storage Sheds

1100.00

# Bldg Units divided by 6

Vestibules

10.00

45.00

Vestibule-SF

10'x12'

Corridors

6.00

Corridor-SF

6' wide

10.00

Stairs
Int Lights

If Avail: Int Stair-SF; Else 160 times # Stories

2.00

Mail Facilities

2500.00

# Bldg Units/16

Laundry Room

3500.00

# Laundry Rooms

Laundry Equipment

600.00

800.00

600.00

Bldg Footprint-SF
3 washers, 3 dryers

# Pieces Laundry equipment (Add: # site units divided by
20 net # existing pieces of equipment)

Common Rooms

10.00

Common Room-SF

Exhibit C.4: Upgrade System Elements
UPGRADE COSTS

MODERATE

ADD

MAJOR

REHAB

UNIT OF MEASURE

***BME UPGRADES
Central Vent & Exhaust

1.50

Central Air Conditioning
Electrical Service

4.50

1.50

Bldg Gross Area-SF

7.50

2.50

Bldg Gross Area-SF

2.25

Bldg Gross Area-SF

2.25

Emergency Lights

159.00

Smoke Detection

153.00

Communication System

115.00

Emer Call Alarm System

425.00
340.00

159.00

Bldg Units divided by 6

153.00

Bldg Units/4

115.00

Bldg Units/6

310.00

Closed Circuit TV

Bldg Units/6

1200.00

Bldg Units/6

Hot Air Furnaces

1500.00

Bldg Units

Boilers

1500.00

Bldg Units

250.00

Bldg Units

3000.00

DHW Gneration
Elevator Cabs

Number Elevators

***UNIT UPGRADES
Interior (ex kitchen, bath)

3.00

7.00

12.00

Kitchen

3000.00

7000.00

12000.00

36.00

Total Unit SF-(kit, bath SF)

Full Baths

1500.00

2500.00

5000.00

# Full baths

Half Baths

1000.00

1500.00

3000.00

# Half baths

1 per unit

ASSUMPTIONS

After the costs are calculated for the inspected site, units, and buildings, costs are generated
for the full property (including uninspected units and buildings) using the same procedures
followed for costing physical needs backlogs:

C.3

•

Building upgrade costs for inspected buildings are normalized to a per 2
bedroom equivalent, and unit upgrade costs are normalized to a per square
foot cost;

•

Costs are generated using one of the three methods5 that were outlined above
for physical needs backlog costs; and

•

Regional adjustments are applied as discussed above.

Estimating Accrual of Repair and Replacement Costs

Accrual cost estimates are the total amount a property will need to cover expected repairs and
replacements for each Observable System over each of the next 20 years. Each system was
given either a repair or a replacement cost depending upon the standard wear of the system.
For example, boilers are expected to be replaced after a certain number of years, but
landscaping only needs periodic major maintenance. Some systems were deemed
inappropriate for accrual estimates because they generally will not need replacement or
standard maintenance over the 20-year horizon used for this study. An example is the Sitelevel Domestic Hot Water Lines. Over time, a portion of the lines may need to be replaced,
but this is not an expected occurrence. The repair or replacement system cost is based on the
same algorithm used for the physical needs backlog costings.
In addition to a repair/replacement cost, each system is assigned an average useful lifetime (or
in the case of items which will be repaired, "action-intervals" are assigned).6 For systems
requiring replacement over time, the useful life is the age the system is expected to be when it
must be replaced because it is worn-out or approaching failure. Boilers are expected to last
25 years. This is the expected life for the Boiler systems. Site Landscaping is not expected to
wear out, but will need to be reseeded and replanted every 5 years. This is the action interval
(rather than expected life) for Landscaping. Exhibit C.5 lists for each system involved in
accrual, the action level appropriate to accrual, and the useful life (or action interval).

Same type-same quality; same type-different quality; different type.

The basic reference for expected lives was Appendix B, "Accrual Actions and Expected Lives" from Future Accrual of Capital Repair
and Replacement Needs of Public Housing, Final Report, prepared for HUD by ICF, Inc., April 1989 as an update of the Abt public
housing study (Bain, 1988). Abt staff experienced in conventional residential building construction and management altered these
lifetimes for some systems.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

151

For each of the next 20 years, for each Accrual system, we test whether the system will reach
the end of its useful life (or action interval) that year. As the starting point, we used the
system ages where they were collected by the Inspector; otherwise, we estimated system age
to be the average age of the buildings in the project. We assume, however, that any system
that needed to be replaced as part of the physical needs backlog, was indeed replaced:
therefore, the age of such systems is set back to zero. The age is then increased for each
accrual year. In any year that a system's accrual age equals its expected life, then the
repair/replace cost is added into the accrual total for that year.
The accrual yearly totals are calculated on the sites, units, and buildings that were actually
inspected. These costs are then scaled up to reflect the total property. Unit level accrual
costs are scaled to property totals based on the proportion of the property’s total square
footage the inspected units represent. Building level accrual costs are scaled to property
totals based on the proportion of the property’s total units the inspected buildings represent.
The property totals are then regionally adjusted as discussed previously. Accrual costs are
based on 1995 current dollars.

Abt Associates Inc.

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

152

Exhibit C.5
Life Expectancies and Repair/Replace Action Levels for Accrual Systems

SYSTEM
Landscaping
Roadways
Parking Lots
Parking Garages
Paved Pedestrian
Curbing
Fencing
Retaining Walls
Site Drainage
Pole Mntd Lighting
Site Furniture
Yards
Dumpster
Pool
Tennis
Basketball
Domestic Hot Water Dist
Domestic Cold Water Dist
Sewage Ejector
Engineering Generator
Unit-Carpet
Unit-Floor Resilient
Kitchen Floor
Kitchen Cabinet
Kitchen Range
Refrigerator
Garbage Disposal
Dishwasher
Microwave
Trash Compactor
Bath Floor-Ceramic
Bath Floor-Resilient
Bath Fixtures
Bath Accessories
Bath Vanities
Unit HVAC
Radiation2
Unit Boiler

LIFE
EXPECTANCY
5
25
25
25
25
25
20
10
25
25
15
20
15
15
15
15
40
40
40
35
10
20
15
25
15
15
7
15
10
15
50
20
40
40
40
20
25
25

( 5)1
(15)
(10)
(20)
(10)

(15)
(25)
(25)
(25)
(20)

Numbers in parenthesis are life expectancies for family occupied units and buildings

Hydronic only.

Abt Associates Inc.

REPAIR
ACTION LEVEL
MIN
REP
REP
REP
REP
REP
REP
MOD
REP
REP
REP
REP
REP
MAJ
MAJ
MOD
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

153

Exhibit C.5, continued
Unit Furnace
Unit DHW Generation
Temperature Control
Wall Air Conditioner
Bell/Intercom
Unit CCTV
Unit ECAS
Unit Smoke Detector
Building Foundation
Exterior Wall
Roof-Membrane
Roof-Shingles
Roof-Builtup
Roof Covering-Tile
Roof Covering-Metal
Parapet Wall
Chimney
Penthouse
Roof Drainage
Windows
Security Grates
Ext Common Door
Unit Entry Door
Storm/Screen Door
Canopies
Exterior Stairs
Bldg Mtd Site Lights
Fire Escapes
Balconies
Porches
Decks
Sheds
Vestibules
Corridors
Stairways
Interior Lights
Mail Facilities
Laundry Rooms
Laundry Equipment
Common Rooms
Common Kitchen
Underground Garage
Heating Riser
Gas Distribution
Dom Hot/Cold Water

Abt Associates Inc.

25
20
25
15
30
30
30
40
10
10
40
20
40
30
30
10
10
10
25
40
40
60
20
15
20
10
10
40
40
40
25
40
10
10
10
25
20
15
10
10
20
25
15
15
50

(15)

(30)
( 7)

( 8)

(10)

REP
REP
REP
REP
REP
REP
REP
REP
MIN
MIN
REP
REP
REP
REP
REP
MOD
MIN
MAJ
REP
REP
REP
REP
MOD
REP
MOD
MIN
REP
REP
REP
REP
REP
REP
MIN
MOD
MIN
MOD
REP
MOD
REP
MOD
MAJ
MAJ
MOD
MOD
REP

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

154

Exhibit C.5, continued
Sanitary Dist
Fire Sprinkler System
Sump Pump
Compactor
Central Vent/Exhaust
Central Air
Emergency Lights
Smoke Detector
Communication System
Building ECAS
Master TV Antenna
Building CCTV
Building Furnace
Building Boiler
Boiler Room Piping
Boiler Equipment
Boiler Room Controls
DHW Generation
Elevator Shaftways
Shaftway Doors
Elevator Cabs
Elevator Machinery

Abt Associates Inc.

10
5
20
10
10
20
35
40
30
30
30
30
25
25
50
25
25
20
15
15
30
30

( 7)

(20)

(10)
(10)
(15)
(25)

MIN
MIN
REP
REP
MAJ
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP

Appendix C - System for Estimating Physical Needs
Backlog and Accrual Costs from Inspections

155

Appendix D
CHARACTERISTICS OF UNITS IN THE STOCK

Exhibit D-1
CHARACTERISTICS OF UNITS IN THE HUD-INSURED MULTIFAMILY STOCK
BY ASSISTANCE CATEGORY
(IN 1995 DOLLARS PER 2BR EQUIVALENT UNIT)
Total
Characteristic
Number of 2BR Units
Percent
Number of Units
Percent

Assisted

Total
1,314,026
1,405,240

Unassisted

Assisted

Older Assisted

Newer Assisted

350,815
27%
354,083
25%

963,211
73%
1,051,157
75%

643,468
67%
686,309
65%

319,743
33%
364,848
35%

Backlog per Unit
$<10

13%

23%

13%

$10 - <1,500

$1,500 - <3,000

$3,000 - <7,500

$7500+

Mean

$3,172

$1,488

$3,785

$3,917

$3,520

Average Annual Accrual per Unit
$<1,000

17%

21%

24%

15%

$1,000 - <1,500

$1,500 - <2,000

$2,000+

Mean

$1,389

$1,491

$1,352

$1,323

$1,410

Cash Flow per Unit
$< -500

15%

$-500 - <0

$0 - <500

Abt Associates Inc.

Appendix D - Characteristics of Units in the Stock

156

Exhibit D-1, continued
Total
Characteristic

Assisted

Total
Unassisted

Assisted

Older Assisted

Newer Assisted

Cash Flow per Unit, continued
$500 - <1,000

$1,000 - 2,500

$2,500+

Mean

$407

$72

$529

$278

$1,034

Property Size
<50 Units

50 - 99 Units

100 - 199 Units

200+ Units

Mean

115

157

105

108

Designated Occupancy Type
Family

84%

93%

81%

87%

67%

Elderly/Handicapped

Source:

1995 Physical Inspection Data and Costing Programs, 1992-1995 finacial data, MIDLIS.

Abt Associates Inc.

Appendix D - Characteristics of Units in the Stock

157

Status of HUDInsured (or Held)
Multifamily Rental
Housing in 1995
HC-5964
Task Order #7

Final Report
Supplements
December 10, 1997

Prepared for
US Department of Housing
and Urban Development
451 Seventh Street, SW
Washington, DC 20410-3000

Prepared by
Meryl Finkel
Donna DeMarco
Deborah Morse
Sandra Nolden
Karen Rich

SUPPLEMENT 1
Risk Analysis Based on Loan-to-Value and Debt-Coverage Ratios
Commercial lenders use two key ratios to evaluate loan applications: the loan-to-value ratio
(LTV) and the debt-coverage ratio (DCR), or debt service coverage ratio. LTV, defined as
the mortgage balance divided by the market value, is a measure of risk due to low equity. In
general, property owners with high LTVs are more prone to default on their mortgage
because they have less equity in the property and a higher probability that any future drops in
value could result in negative equity. Properties with LTVs greater than 1 are at greatest risk
of default, all else equal, because their loan balances are greater than the value of the property,
in other words they already have negative equity. At loan origination, underwriters often
require an LTV of .80 or lower on multifamily properties. For the 221(d)(4) and 221(d)(3)
programs LTV at origination was allowed to be higher, up to 1.0. Of course, over time, the
actual LTV will change as the owner pays down the principal on the loan and/or as the market
value of the property changes.
Estimating the value of assisted properties is difficult because of the uncertainty of continued
subsidies. For this analysis LTV is defined as the current mortgage balance divided by the
estimate of the market value of the property, assuming that both rents and operating costs
are at market levels. Thus, this measure of LTV does not reflect the current operating
conditions for the properties. Under current conditions both revenues and operating costs are
often above or below market levels due to rent restrictions and assistance payments. The
market value of the property is equal to the capitalized value of market rent net of operating
and maintenance expenses. The market rent and capitalization rate estimates were provided
by the market analysts. Operating cost estimates are based on medians for conventional
apartments by region and building type, as compiled by IREM.1 The net market value equals
the market value net of physical needs backlog, because to achieve market rents it is assumed
that all physical needs backlogs must be addressed. This approach to estimating LTV assumes
that the existing mortgage loan remains, but essentially assumes that the properties receive no
ongoing assistance, so that rents are equal to prevailing market rents, and operating expenses
equal industry medians.2
DCR, defined here as total revenue net of operating and maintenance costs and deposits to
the reserve for replacement account divided by debt payment, is a measure of whether
operating income is sufficient to cover the mortgage payments. This measure relies on actual
current revenue, operating cost, reserve deposit, and debt service. Typical multifamily

1996 Income/Expense Analysis, Conventional Apartments, Institute for Real Estate Management, Chicago IL

A more conservative measure of net market value could also subtract the present value of other costs of conversion to market such as
evictions, advertising and redecorating.

Abt Associates Inc.

Supplement 1

underwriting criteria specify a DCR of 1.25 or greater. A DCR of less than 1.0 indicates
financial distress due to cash flow difficulties -- cash flow after covering operating and
maintenance costs and deposits to the reserve account is insufficient to cover debt service.3
Both of these indicators have been incorporated into numerous loan default models.4 Most
models use LTV and DCR at origination because property values are only available from the
original appraisal. For this study, we have used an estimate of current property value, which
may provide a better indicator of current financial condition.
As shown in Exhibit S1-1, 59 percent of all insured properties had LTVs less than 0.8, and 30
percent had LTVs greater than 1.0.5 The proportions are similar in assisted and unassisted
properties. However, older and newer assisted properties had very different LTV profiles.

Alternatively a DCR based on market rents and operating costs ,and reserve deposits sufficient to cover ongoing accrual costs could be
developed.

Bogdon and Follain, An Examination of Three Sets of Indicators of Financial Risk Among Multifamily Properties, Urban Institute, April
14, 1996.

Five sample properties, representing 68 properties in the universe had negative net values. In other words, their capitalized rent stream
could not cover both operations and repairing the backlog. For these tables, the negative value properties are included with the properties
with LTVs greater than 1, though in fact their LTVs are less than zero due to the negative value.

Abt Associates Inc.

Supplement 1

Exhibit S1-1
LOAN TO VALUE RATIO AND DEBT COVERAGE RATIO
BY ASSISTANCE CATEGORY

Total Properties
Percent Properties

Total

Unassisted

Assisted

Older
Assisted

Newer
Assisted

12,243
100%

2,224
18%

10,109
82%

5,943
59%

4,076
41%

Loan to Value Ratio
< = .80 (good)

59%

64%

58%

48%

14%

.80 - 1.0

11%

15%

10%

18%

> 1.0 (bad)6

30%

21%

32%

68%

Debt Coverage Ratio

Source:

< 1.0 (bad)

25%

24%

34%

13%

1.0 - 1.25

26%

25%

27%

18%

39%

> 1.25 (good)

49%

51%

49%

48%

Financial Statements, Market Valuations.

Very few older assisted properties (7%) had negative equity (LTV>1.0), because the property
owners have had several years to pay down the principal on their mortgages over time
through regular payments. Of concern is the high percentage of newer assisted properties
(68%) that had negative equity. This may be due to the relatively few years owners of newer
assisted properties have had to pay down the principal as well as declines in property values.
But it is also due to the fact that these properties were often built in neighborhoods where the
market was not adding to the supply of affordable housing. Thus, by design, these programs
allowed LTVs above even at origination in order to promote housing development.
About half (49 percent) of properties across all assistance categories had DCRs above 1.25,
and about one fourth had DCRs below 1.0. Although across all assistance categories the
proportion of properties with DCRs above 1.25 was similar, the proportion with DCRs below
1 differed across assistance categories. Over one third of older assisted properties had DCRs
below 1.0, in spite of their low debt service payments, compared with only 13 percent of
newer assisted properties (this in spite of their high debt service payments) and 25 percent of
unassisted properties.

Includes 5 sample properties representing 68 universe properties with negative value. These properties actually have very low (negative)
calculated LTVs.

Abt Associates Inc.

Supplement 1

Exhibit S1-2 combines DCR and LTV to examine how these two indicators relate to each
other and to estimate the number of properties that both have negative equity and cannot
cover debt service. “Single Trigger” default models often assume that the property owner is
at risk to default if the property has either negative equity (LTV>1) or a debt-coverage ratio
of less than unity (DCR<1).7 Forty-eight percent of insured properties fell into at least one of
these risk categories. “Double Trigger” models of default assume that the owner is at risk to
default only if the property has both negative equity and a debt-coverage ratio of less than 1.
An estimated 701 properties, or 6 percent of all properties had both negative equity and a
DCR of less than one.
Of particular interest is the percentage of properties in each assistance category that have both
negative equity (LTV>1.0) and insufficient income to cover debt service (DCR<1.0). Among
unassisted properties, 9 percent were in this dubious position, as were 10 percent of newer
assisted properties. Because of their relatively low LTVs, only 2 percent of older assisted
properties had both negative equity and insufficient income to cover debt payments.
Looking at the strongest LTVs and DCRs, 43 percent of unassisted properties and 45 percent
of older assisted properties had DCRs of 1.25 or greater and LTVs less than or equal to 0.80.
By contrast, only 11% of newer assisted properties meet these conditions, largely due to the
high LTVs found in this group of properties.

Goldberg, Lawrence, Default Risk Models for FHA Multifamily Loans: The Effect of Terminating Project Based Rental
Assistance, Office of Federal Housing Enterprise Oversight (no date).

Abt Associates Inc.

Supplement 1

Exhibit S1-2
LOAN TO VALUE RATIO AND DEBT COVERAGE RATIO
MULTIFAMILY RENTAL HOUSING WITH HUD-INSURED (OR HELD) MORTGAGES
Debt-Coverage Ratio (DCR)
LTV

< 1.0 (bad)

1.0 - 1.25

> 1.25 (good)

Total

Total Properties
Percent Properties

3,072
25%

3,203
26%

5,963
49%

12,243
100%

Total

LTV < = .80 (good)

2,112
17%

1,163
10%

4,049
33%

7,324
60%

0.8 < LTV ˜ 1

264
2%

428
3%

634
5%

1,327
11%

LTV >1 (bad)

701
6%

1,612
13%

1,280
10%

3,592
29%

Unassisted

549
25%

1,125
51%

2,224
100%

LTV < = .80 (good)

275
12%

192
9%

961
43%

1,428
64%

0.8 < LTV ˜ 1

82
4%

165
7%

82
4%

329
15%

>1.0 (bad)

192
9%

82
4%

467
21%

Older Assisted

1,995
34%

1,056
18%

2,892
49%

5,943
100%

LTV < = .80 (good)

1,791
30%

901
15%

2,648
45%

5,430
90%

0.8 < LTV ˜ 1

89
1%

56
1%

119
2%

256
4%

>1.0 (bad)

115
2%

100
2%

132
2%

346
6%

Newer Assisted

533
13%

1,598
39%

1,945
48%

4,076
100%

LTV < = .80 (good)

46
1%

69
2%

440
11%

556
14%

0.8 < LTV ˜ 1

92
2%

208
15%

440
11%

741
18%

>1.0 (bad)

394
10%

1,320
32%

1,065
26%

2,779
68%

Source:

Financial Data and Market Valuations.

Abt Associates Inc.

Supplement 1

SUPPLEMENT 2
Comparison of Findings with RAMS
To compare our system of assessing property condition with HUD’s current practice, we
compared our measures of physical condition (physical needs backlog), financial condition
(net cash flow), and overall condition (Distress Index) with HUD’s Risk Assessment
Management System (RAMS) for the 567 properties for which we had both sets of data.
Under RAMS, each property is assigned a ranking from 0 to 40 each for physical and financial
condition, with 40 representing the highest risk. Using a weighted sample to represent the
universe of HUD-insured multifamily properties, we did a cross tabulation of physical needs
backlog by RAMS physical rankings and another cross tabulation of annual net cash flow by
RAMS financial rankings to see how our physical and financial measures mapped with RAMS.
We also compared our overall summary Distress Index with the RAMS overall system rank to
see how our overall measure compares with HUD’s current practices.
Comparison of Physical Condition. The RAMS Physical Inspection Rank is based on the
overall rating of physical condition from the latest physical inspection report. If the latest
inspection report is over three years old, additional points are assigned. The physical ranking
system is shown in Exhibit S2-1.
Exhibit S2-1
RAMS PHYSICAL RANKING SYSTEM
Inspection Report Evaluation
Superior

Satisfactory

Below Average

Unsatisfactory

No Physical Inspection
Source:

Date of Inspection Report

Points

3 years ago or less

more than 3 years ago

3 years ago or less

more than 3 years ago

3 years ago or less

more than 3 years ago

3 years ago or less

more than 3 years ago

n/a

HUD RAMS Pilot Training materials (points). October 18, 1996

Abt Associates Inc.

Supplement 2

The following table (Exhibit S2-2) shows, for each assistance category the percentage of
properties in each of three RAMS physical inspection ranking categories (superior,
satisfactory, or below average/unsatisfactory) by their level of physical needs backlog.
In general, our comparison showed that properties with RAMS scores associated with below
average or unsatisfactory inspection reports had the highest physical needs backlogs. In the
unassisted category, 48 percent of the properties in our weighted sample had a RAMS ranking
of 15 or higher, indicating a below average or unsatisfactory physical inspection report. Of
these, 42 percent had backlogs of $1,500 or greater, including 15 percent with backlogs
exceeding $3,000. Of the properties with satisfactory RAMS physical inspection rankings,
two-thirds (67 percent) had backlogs less than $1,500, including 17 percent with no backlog.
Of those unassisted properties with RAMS rankings based on superior physical inspection
reports, 100 percent had backlogs of less than $1,500, including a third (33 percent) that had
no backlog.
As with unassisted properties, nearly half (49 percent) of older assisted properties in our
weighted sample had a RAMS ranking of 15 or higher, indicating a below average or
unsatisfactory physical inspection report. Of these, 59 percent had backlogs of $1,500 or
more, including 46 percent that had backlogs exceeding $3,000. Among those older assisted
properties with satisfactory physical RAMS rankings, 40 percent had backlogs of less than
$1,500, including 10 percent with no backlog. Of those older assisted properties with
superior physical RAMS rankings, 64 percent had backlogs of less than $1,500, including 11
percent that had no backlog.

Abt Associates Inc.

Supplement 2

Exhibit S2-2
PHYSICAL BACKLOG BY RAMS PHYSICAL INSPECTION RANk
BY ASSISTANCE CATEGORY
RAMS Physical Inspection Rank (0 = Lowest Risk)

Superior
0-1

Satisfactory
2-3

Below Avg/
Unsatisfactory
15-40

Total

Total Physical
Needs Backlog
Unassisted

165
9%

824
43%

906
48%

1,895
100%

Backlog <$10

33%

17%

15%

17%

$10 - $1,500

67%

50%

42%

48%

$1,500 - $3,000

20%

27%

22%

>$3,000

13%

15%

13%

Older Assisted

317
6%

2,445
45%

2,624
49%

5,389
100%

Backlog <$10

11%

10%

$10 - $1,500

53%

30%

36%

34%

$1,500 - $3,000

12%

24%

13%

18%

>$3,000

24%

36%

46%

40%

Newer Assisted

440
11%

2,316
59%

1,158
30%

3,914
100%

Backlog <$10

26%

16%

15%

$10 - $1,500

32%

46%

30%

40%

$1,500 - $3,000

26%

19%

22%

21%

>$3,000

16%

19%

40%

25%

Note:

Column sums may not add up to totals due to rounding.

Fewer than one-third (30 percent) of newer assisted properties in our weighted sample had a
RAMS ranking of 15 or higher (indicating a below average or unsatisfactory physical
inspection report), compared to nearly half of unassisted and older assisted properties. Of
these, 62 percent had backlogs exceeding $1,500, including 40 percent with backlogs
exceeding $3,000. Of those newer assisted properties with satisfactory physical RAMS
rankings, 62 percent had backlogs less than $1,500, including 46 percent with no backlog. Of
those newer assisted properties with RAMS rankings based on superior physical inspection

Abt Associates Inc.

Supplement 2

reports, 58 percent had backlogs less than $1,500, including one-fourth (26 percent) with no
backlog.
The comparison of RAMS physical rankings with backlog by assistance categories shows the
following: (1) among properties with good RAMS scores, unassisted properties were more
likely to have no or low backlog than either older or newer assisted properties; (2) among
properties with poor RAMS scores, unassisted properties were less likely to have a high
backlog than older or newer assisted properties. The correlation coefficient of the RAMS
physical rankings, and physical needs backlog was 0.18, showing a relatively weak correlation
between the RAMS scores and physical needs backlog.
Comparison of Financial Condition. The RAMS financial ranking is based on vacancy,
percent gross rent subsidized, units, weighted average surplus cash per unit, weighted average
operating cost per total revenue, reserve for replacement per total revenue, weighted average
total margin, and weighted average current ratio1.
Overall, across all assistance categories, RAMS financial rankings and this study’s annual net
cash flow were moderately, and negatively correlated, with a correlation coefficient of -0.34.
The following table (Exhibit S2-3) shows, for each assistance category, the percentage of
properties in each of three RAMS financial ranking categories by their annual net cost flow
(negative cash flow, low positive cash flow ($0-$500), and high positive cash flow (>$500)).
As the table shows, unassisted properties had the worst RAMS financial rankings of the three
assistance categories, with more than a third (35 percent) having scores of 24 or higher. Of
these, 58 percent had negative cash flow. By the same token, only 28 percent of unassisted
properties had RAMS financial rankings of 12 or lower. Of these 84 percent had high positive
cash flow.
Among older assisted properties, 18 percent had RAMS ranks of 24 or worse, compared to
35 percent of unassisted properties. Of those older assisted properties with RAMS scores of
24 or worse, 59 percent had negative cash flow. At the other end of the spectrum, 52 percent
of older assisted properties had scores of 12 or better, of which 81 percent had positive cash
flow.
Newer assisted properties scored the best on the RAMS financial rankings, with only 4
percent having scores of 24 or worse. Of these, 50 percent had negative cash flow and 50
percent had low positive cash flow. A full 76 percent of newer assisted properties had scores
of 12 or better, of which 98 percent had positive cash flow and 76 percent had high positive
cash flow.
Exhibit S2-3

See page 4 of Appendix B of RAMS Pilot Training Materials, October 1996, for definitions of these components of RAMS financial rank.

Abt Associates Inc.

Supplement 2

NET CASH FLOW BY RAMS FINANCIAL RANK
BY ASSISTANCE CATEGORY
RAMS Financial Rank (0 = Lowest Risk)

Note:

Net Cash Flow

0-12

12-23

24+

Total

Unassisted

522
28%

714
38%

659
35%

1,895
100%

<$0

16%

58%

26%

$0-$500

46%

17%

23%

>$500

84%

50%

25%

51%

Older Assisted

2,796
52%

1,605
30%

988
18%

5,389
100%

<$0

19%

41%

59%

33%

$0-$500

48%

37%

28%

41%

>$500

33%

21%

14%

26%

Newer Assisted

2,988
76%

787
20%

139
4%

3,914
100%

<$0

47%

50%

13%

$0-$500

22%

15%

50%

21%

>$500

76%

38%

66%

Column sums may not add up to totals due to rounding.

Comparison of Overall Property Condition. The RAMS Overall System Rank combines
physical, financial, and management ranks, which have maximum points of 40, 40, and 20,
respectively, to produce a score from 0 to 100, again with zero being the lowest risk. We
compared this RAMS overall rank with Distress Index for the 567 properties for which we
had both sets of data. The RAMS physical and financial rankings were described above. The
RAMS management rank is based on the overall rating of project management from the latest
management review report, with additional points added for management reviews conducted
more than three years ago.
Using a weighted sample to represent the universe of HUD-insured multifamily properties, we
did a cross tabulation of the Distress Index by RAMS overall system rank to see how the
Distress Index mapped with RAMS.
The Distress Index correlated moderately and negatively with the RAMS Overall System
Rank, with a correlation coefficient of -0.24. As we can see from Exhibit S2-4, of the

Abt Associates Inc.

Supplement 2

S10

properties that had the RAMS ranks associated with least risk (0-40), 72 percent were sound
according to the Distress Index, compared with 45 percent of properties with more risky
ranks (71-100). At the same time, 40 percent of properties with higher-risk ranks (71-40)
were distressed, versus 16 percent of properties with less risky ranks (0-40).
Exhibit S2-4
DISTRESS INDEX BY RAMS OVERALL SYSTEM RANK
BY ASSISTANCE CATEGORY
RAMS Overall System Rank (0 = Best)
Lowest Risk
0-40

Middle Risk
41-70

Highest Risk
71-100

Total

All Categories

5,565
50%

3,464
31%

2,168
19%

11,197
100%

Distressed (<-$250)

16%

29%

40%

25%

Stressed (-$250-0)

12%

16%

12%

Sound (>$0)

72%

63%

45%

64%

Unassisted

549
29%

659
35%

686
36%

1,895
100%

Distressed (<-$250)

15%

21%

24%

20%

Stressed (-$250-0)

Sound (>$0)

85%

75%

68%

75%

Older Assisted

2,399
45%

1,786
33%

1,204
22%

5,389
100%

Distressed (<-$250)

27%

33%

44%

33%

Stressed (-$250-0)

20%

10%

23%

18%

Sound (>$0)

53%

56%

32%

49%

Newer Assisted

2,617
67%

1,019
26%

278
7%

3,914
100%

Distressed (<-$250)

25%

58%

15%

Stressed (-$250-0)

Sound (>$0)

87%

66%

42%

78%

Distress Index

Note:

Column sums may not add up to totals due to rounding.

Source:

RAMS Data, and 1995 Physical Inspection Data and Costing Program, 1992-1995 AFS.

Abt Associates Inc.

Supplement 2

S11

Among properties that were categorized as distressed, the RAMS system was more likely to
assign a low-risk rank to older assisted properties than to unassisted or newer assisted
properties. While only 15 percent of unassisted properties and 7 percent of newer assisted
properties with low-risk rankings (0-40) were classified as distressed, a full 27 percent of
older assisted properties with low-risk RAMS ranks were distressed according to the Distress
Index. This is due largely to the fact that the RAMS physical inspection rank, which
comprises 40 percent of the Overall System Rank, ranks most properties as low risk. Over
half of unassisted and older assisted properties and two-thirds of newer assisted properties
have RAMS physical scores of 3 or less (out of 40).
Among properties classified as sound by the Distress Index, the RAMS system was more
likely to assign a high-risk score to unassisted properties than to older or newer assisted
properties. While only 32 percent of older assisted and 42 percent of newer assisted
properties with high-risk rankings (71-100) were classified as sound, more than two-thirds (68
percent) of unassisted properties with RAMS scores between 71-100 were sound. This
discrepancy is largely due to the heavy weight RAMS attributes to vacancy in its ranking
system relative to the Distress Index. Under RAMS, vacancy is assigned a weight of 4 to 7,
depending on the assistance category2, while other indicators in the RAMS ranking system are
give weights of 1 to 4. Because unassisted properties had much higher vacancies than either
older or newer assisted properties, they were more likely to be given a high-risk rank under
RAMS even if they were categorized as sound under the Distress Index, which places less
emphasis on vacancy.

Vacancy weights under RAMS: 4 for Section 221(d)(4) properties, 6 for older assisted housing, and 7 for other
rental mortgages.

Abt Associates Inc.

Supplement 2

S12

SUPPLEMENT 3
Details on Change in Physical Condition
The physical condition of the HUD-Insured Multifamily Stock is measured using the cost to
remedy the backlog of physical needs. Changes in the reported physical condition between
1989 and 1995 reflect changes in the incidence and severity of reported deficiencies as well as
changes in some system definitions and costs. Chapter 3 of the report described the change in
physical condition of the stock between 1989 and 1995. This memorandum provides further
details on changes in physical condition, and the reasons for change.
The following analysis focuses on the comparison sample of properties, the subset of 504
properties that were inspected in both 1989 and 1995. All numbers were weighted to reflect
the stock of properties that was still insured in 1995.
The overall reported differences in backlog reflect both changes in actual property condition
and changes made by Abt Associates in several system definitions and costs used to compute
backlogs. Even after adjusting for inflation, unit cost estimates for repairing some systems
increased since 1989 but decreased for other systems. We changed the definitions of actions
associated with some systems as well in order to be closer to current practices in the industry.
The appendix provides details on system definitions, costs and any changes for 1989 and
1995. In order to separate the total measured change in physical condition into changes
resulting from costs and those resulting from condition changes, Exhibit S3-s 1 and 2 show
several relevant data items for the stock as a whole and for each assistance category. Exhibit
S3- 1 shows changes in mean physical needs backlogs by system group, changes in incidence
of reported backlogs, and the contribution of each system group to the overall increase in
mean backlog. Exhibit S3- 2 provides an indicator of the change in severity of needs by
system group.
Exhibit S3- 1 shows:

•

The mean physical needs backlog by system group in 1989 using the 1989
costs (inflated to 1995 dollars).1

•

The mean physical needs backlog by system group in 1989 using the 1995
costs. Comparing this column with the 1989 backlog shows the impact of
changing some system repair cost estimates on the mean backlog by system group
and overall.

As described in Chapter 1, over 115 specific systems (such as landscaping, roadways, parking, paved pedestrian areas) were inspected in
each property. These systems were combined into system groups (such as Site Areas) for property-level costing and for reporting.

Abt Associates Inc.

Supplerment 3

S13

•

The mean physical needs backlog in 1995 by system group. Comparing this
column with the means for 1989 condition using 1995 costs shows the impact of
changes in condition and system definitions on the mean backlog by system group
and overall.

•

The proportion of properties with backlogs in 1989 by system group. This
provides an indication of the incidence of problems in 1989.

•

The proportion of properties with backlogs in 1995 by system group. This
provides an indication of the incidence of problems in 1995.

•

The percentage each system group contributed to the overall change in mean
backlog. (For example, the overall change in mean backlog for the entire stock
was $1,176 per 2BR ($3,058-$1,882). Site areas, thus, contributed 6 percent to
the overall change in mean backlog (mean backlog in 1995 was $225 per 2BR, and
in 1989 it was $152. (225-152) / 1,176 = 0.06). This allows us to focus our
attention on system groups that had a significant impact on the overall change in
average backlog between 1989 and 1995.

Exhibit S3- 2 shows:
•

Stock-wide and by assistance category, the mean physical needs backlog in
1989 and in 1995 by system group only for properties that had backlogs of
needs within the system group (both using the 1995 cost file). This provides an
indication of the change in severity of needs, because it shows the average backlog
associated with the system among properties that had backlogs within the system.

Abt Associates Inc.

Supplerment 3

S14

Abt Associates Inc.

Supplerment 3

S15

24
120

Building Mechanical

Building Heating and
Cooling

Unit Electrical

Note:

$1,876

256

662

215

125

102

$149

System Mean in 1989
(using 1995 costs)

Numbers are weighted and 1989 costs were adjusted to 1995 dollars.

$1,882

Unit Heating and
Cooling

Total

Unit Bathrooms

266

Unit Construction

Unit Kitchen Fixtures

Building Common
Areas

699

Building Exterior
Features

Unit Finishes

101

182

Building Windows
and Doors

Building Roof

Building Closure

Site Distribution

Building Elevators

$152

System Mean in 1989
(using 1989 costs)

Site Amenities

Site Areas

Group Systems

$3,058

151

542

791

129

367

186

181

152

$225

System Mean in 1995
(using 1995 costs)

83%

52%

Percent of Properties
with Costs > 0 in 1989

91%

64%

Percent of Properties
with Costs > 0 in 1995

Exhibit S3- 5-1
INDICATORS OF CHANGE IN BACKLOG OF PHYSICAL NEEDS (TOTAL STOCK)

100%

-1

Percent Contribution
to Total Change

Abt Associates Inc.

Supplerment 3

S17

41
157
5

Building Mechanical

Building Heating &
Cooling

Building Elevators

Unit Construction

36
65

Unit Heating &
Cooling

Unit Electrical

Note:

$2,760

132

406

972

102

286

196

138

$216

System Mean in 1989
(using 1995 costs)

Numbers are weighted and 1989 costs were adjusted to 1995 dollars.

$2,770

124

Unit Bathrooms

Total

422

Unit Kitchen Fixtures

1,003

Building Common
Areas

Unit Finishes

140

Building Exterior
Features

Building Windows &
Doors

Building Roof

283

Site Distribution

Building Closure

$224

System Mean in 1989
(using 1989 costs)

Site Amenities

Site Areas

Group Systems

$3,845

150

197

656

968

132

496

191

231

169

$320

System Mean in 1995
(using 1995 costs)

92%

71%

Percent of Properties
with Costs > 0 in 1989

94%

70%

Percent of Properties
with Costs > 0 in 1995

Exhibit S3- 1, continued
INDICATORS OF CHANGE IN BACKLOG OF PHYSICAL NEEDS (OLDER ASSISTED PROPERTIES )

100%

-3

-5

-1

Percent Contribution
to Total Change

Abt Associates Inc.

Supplerment 3

S18

6
61
6

Building Mechanical

Building Heating &
Cooling

Building Elevators

Unit Construction

Unit Electrical

Note:

$1,016

138

246

217

$119

System Mean in 1989
(using 1995 costs)

Numbers are weighted and 1989 costs were adjusted to 1995 dollars.

$960

Unit Heating and
Cooling

Total

Unit Bathrooms

146

Building Common
Areas

Unit Kitchen Fixtures

Building Exterior
Features

269

Building Windows &
Doors

Unit Finishes

Building Roof

107

Site Distribution

Building Closure

$110

System Mean in 1989
(using 1989 costs)

Site Amenities

Site Areas

Group Systems

$1,427

272

289

158

127

132

163

$84

System Mean in 1995
(using 1995 costs)

65%

42%

Percent of Properties
with Costs > 0 in 1989

84%

49%

Percent of Properties
with Costs > 0 in 1995

Exhibit S3- 1, continued
INDICATORS OF CHANGE IN BACKLOG OF PHYSICAL NEEDS (UNASSISTED PROPERTIES )

100%

-1

-5

-5%

Percent Contribution
to Total Change

Abt Associates Inc.

Supplerment 3

S19

3
75
30
56

38
3

Building Elevators

Building Closure

Building Roof

Building Windows &
Doors

Building Exterior
Features

Building Common
Areas

Unit Construction

Unit Electrical

Note:

$1,057

102

437

110

$67

System Mean in 1989
(using 1995 costs)

Numbers are weighted and 1989 costs were adjusted to 1995 dollars.

$1,091

Unit Heating &
Cooling

Total

Unit Bathrooms

103

Building Heating &
Cooling

Unit Kitchen Fixtures

Building Mechanical

491

Site Distribution

Unit Finishes

$72

System Mean in 1989
(using 1989 costs)

Site Amenities

Site Areas

Group Systems

$2,800

145

524

806

157

294

210

135

122

$163

System Mean in 1995
(using 1995 costs)

76%

40%

Percent of Properties
with Costs > 0 in 1989

88%

61%

Percent of Properties
with Costs > 0 in 1995

Exhibit S3- 1, continued
INDICATORS OF CHANGE IN BACKLOG OF PHYSICAL NEEDS (Newer Assisted Properties)

100%

Percent Contribution
to Total Change

Abt Associates Inc.

Supplerment 3

S20

232
1,388
140
436
226
737

137

294

Building Mechanical

Building Heating &
Cooling

Building Elevators

Building Closure

Building Roof

Building Windows &
Doors

Building Exterior
Features

Building Common
Areas

415
713

Unit Heating &
cooling

Unit Electrical
$2,312

299

Unit Bathrooms

Total

713

1,268

Unit Kitchen Fixture

Unit Finishes

410

554

Site Distribution

Unit Construction

121

$271

1989

Site Amenities

Site Areas

Group Systems

Total

$3,406

426

483

368

880

1,383

426

422

180

944

567

482

203

1,070

558

201

354

1995

$1,552

150

172

337

746

741

142

152

1,034

222

370

302

889

156

282

1989

$1,700

106

244

181

537

734

317

302

799

448

485

1,469

335

170

1995

Unassisted

$3,042

573

420

349

773

1,548

522

325

155

695

233

475

133

1,493

300

554

136

310

1989

$4,134

625

641

425

1,014

1,596

459

441

210

1,044

551

550

222

1,084

634

284

453

1995

Older Assisted

$1,396

1,621

508

162

480

928

124

288

679

204

329

1,379

123

165

1989

$3,174

275

179

329

828

1,305

383

444

174

799

650

366

240

876

495

267

1995

Newer Assisted

Exhibit S3- 2
MEAN COSTS PER 2BR FOR PROPERTIES WITH REPORTED BACKLOGS IN EACH SYSTEM GROUP

These two Exhibit S3-s together, along with the appendix on changes in system definitions,
provide information on the overall change in repair backlogs and on the reason for changes,
whether due to cost, incidence or severity of problems.
As is evident from comparing the first two columns of Exhibit S3- 1, overall and within each
assistance category, changes in the repair cost estimates for specific systems had little impact on
the overall estimate of repair backlogs in 1989, although the means for certain system groups
changed. In other words, the increases in repair costs for some system groups were generally
offset by decreases in costs for other systems. Thus, although the mean backlogs associated with
some specific system group changed based on the updated cost estimates, the overall estimated
average backlog did not. For the stock as a whole, the mean backlog in 1989 using the 1989 cost
files (but inflated to 1995 dollars) was $1,882. Using the 1995 cost file, the mean backlog in
1989 (in 1995 dollars) was $1,876. (Similarly for unassisted properties the mean backlog in 1989
was $960 using the 1989 cost files and $1,016 using the 1995 cost file. For older assisted
properties the numbers are $2,770 and $2,759 respectively, and for newer assisted properties the
numbers are $1,091 and $1,057).
In contrast, comparing the second and third columns indicates that there were significant changes
in incidence and severity of problems between 1989 and 1995. The mean backlog stock-wide in
1989 was $1,876 per 2BR (using 1989 condition and 1995 costs), and it was $3,058 per 2BR in
1995. This increase is a result of change in incidence and severity of problems as well as changes
in some system definitions.
Comparing the fourth and fifth columns of Exhibit S3- 1 shows the change in incidence of
problems. The two columns show the proportion of properties that had backlogs of physical needs
overall and in each system group by year. For example, in 1989, 83 percent of all properties had
backlogs while in 1995, 91 percent of properties had needs. Across all system groups the
proportion of properties reporting backlogs increased.
As an example of how the tables can be used we look at the change in backlog for the site area
system group for the stock as a whole. Exhibit S3- 1 (column 1) shows that in 1989 insured
properties had an average of $152 of backlog per 2BR associated with site areas. In 1995
(column 3) the average rose to $225. Comparing columns 1, 2, and 3 shows that the overall
change in backlog in site areas results from changes in condition rather than changes in costs. The
mean backlog of physical needs associated with site areas in 1989 (using 1989 costs) was $152
and using the 1995 cost file with 1989 condition, the mean was slightly lower at $149.
Column 4 shows that in 1989, 52 percent of properties had backlogs in systems associated with
site areas and column 5 shows that in 1995, 62 percent of properties had backlogs in this system
group. Clearly the incidence of problems increased.

Abt Associates Inc.

Supplerment 3

S21

As indicated above, column 2 showed that the mean backlog per 2BR associated with site areas
across all properties was $149 in 1989 (using the 1995 cost file). However, Exhibit S3- 2 shows
that among the (52 percent of) properties that had backlogs associated with site areas the mean
backlog was $354. In 1995, 64 percent of properties had backlogs in systems associated with Site
Areas. The mean backlog per 2BR associated with site areas across all properties was $225,
which is higher than the 1989 average. Exhibit S3- 2 shows, however, that among the (64
percent of) properties that had backlogs associated with site areas the mean backlog was only
$282. Thus, the increase in overall mean backlog associated with site areas is a result of
increased incidence (52 to 64 percent), rather than an increase in severity of problems because
there was a decrease in mean backlog for properties that had backlogs.
System Groups Contributing to Increased Repair Backlogs

The discussion below focuses on the systems that contributed substantially (10 percent or more)
to the change in mean backlog.
•

There were three system groups that together accounted for over half the increase in
mean backlog: unit kitchen fixtures (24 percent), building windows and doors (23
percent), and building roofs (11 percent). These three systems accounted for a major
portion of the increase in mean backlog across all three assistance categories.

•

In unassisted properties two additional system groups also contributed substantially to
the change - building heat and cooling (22 percent), and building common areas (10
percent). Unit kitchen fixtures contributed 27 percent of the total change. Five
system groups - building windows and doors, building roofs, building common areas,
building heat, and unit kitchen fixtures contributed 92 percent of the increase in mean
backlog in unassisted properties.

•

In older assisted properties one system in addition to unit kitchen fixtures, building
windows and doors and building roofs also contributed substantially to the change unit heating and cooling (11 percent). Building windows and doors alone represented
33 percent of the increase in costs. Building windows, unit kitchen fixtures building
roofs and unit heating and cooling contributed 77 percent of the increase in backlog.

•

In newer assisted properties one system in addition to unit kitchen fixtures, building
windows and doors and building roofs also contributed substantially to the change unit interior finishes (18 percent). Building windows and doors, unit kitchen fixtures
building roofs and unit interior finishes contributed 68 percent of the increase in
backlog.

Abt Associates Inc.

Supplerment 3

S22

Unit Kitchen Fixtures include kitchen cabinets/counters, ranges and hoods, refrigerators,
garbage disposals, microwaves and trash compactors. The mean repair backlog in 1989
estimated using the using 1995 cost file ($256 per 2BR stock-wide) was slightly lower than the
1989 estimate using the 1989 cost file ($266 per 2BR). This is due to decreases in the cost
estimates for repairing/replacing cabinets and counters and ranges and hoods that were only
partially offset by an increase in the cost estimate for replacing refrigerators.
In 1995, backlogs associated with kitchen fixtures averaged $542 per 2BR. This reflects a
substantial increase in the proportion of properties requiring actions and an increase in the severity
of actions required. In 1989, 40 percent of properties had backlogs associated with kitchen
fixtures compared with 62 percent of properties in 1995. Across all assistance categories the
proportion of properties requiring actions in kitchens increased (from 19 to 51 percent in
unassisted properties, from 54 to 65 percent in older assisted properties, and from 21 to 63
percent in newer assisted properties).
In 1989, 24 percent of inspected units required repairs in their kitchen cabinets/counters, including
3 percent that required major repairs or total replacement. In contrast, repairs were required in
33 percent of units inspected in 1995, including 9 percent that required major repairs or total
replacement. For the other kitchen fixtures the only action level available was total replacement.
In 1989, 11 percent of ranges, 3 percent of refrigerators, and 1 percent of garbage disposals
inspected required replacement. In 1995, 17 percent of ranges, 13 percent of refrigerators, and 9
percent of garbage disposals required replacement.2
Part of the increase in reported requirements for replacing kitchen fixtures may be a result of a
change in the instructions in the training manual. In 1995, the manual said that refrigerators,
garbage disposals, dishwashers and microwaves were to be replaced if they were over-age. 3
Building Windows and Doors include windows, window security grates, exterior common
doors, unit entry doors and storm/screen doors. The increase in the average repair backlogs for
windows and doors between 1989 and 1995 reflects changes in costs, incidence and severity of
problems. Costs for repairing different size windows increased by between 200 to 300 percent in
the 1995 cost file, while costs for repairing doors decreased by about 16 percent. (See appendix
for details). The mean backlog for windows and doors increased from $101 per 2BR using the
1989 condition with 1989 costs to $215 per 2BR using the 1989 condition and 1995 costs, largely
reflecting the increase in cost estimates for repairing/replacing windows.

Information is presented on the percentage of (weighted) properties requiring some action in a system group. This is calculated based on
whether the costs associated with that system group property-wide were positive. However, reporting on action levels for specific systems is
based on the proportion of units, buildings or sites inspected that required a specific action. We are not able to inflate this to property-wide
or stock wide estimates (because, for example, each inspected unit within a property represents a different number of property units).

In fact, inspectors did not always follow this rule. In 1989, 19 percent of inspected refrigerators were over 15 years old and 12 percent of
these were said to require replacement. In 1995, 13 percent of inspected refrigerators were over 15 years old and 52 percent were said to
require replacement.

Abt Associates Inc.

Supplerment 3

S23

In 1995, backlogs associated with building windows and doors averaged $367 per 2BR. This
increase is due to increases in the reported incidence and severity of needs in older and newer
assisted properties, which was partially offset by a decrease in incidence and severity in unassisted
properties. In 1989, 41 percent of older assisted properties and 16 percent of newer assisted
properties had needs associated with windows and doors. In 1995 the proportions were 48 and
35 percent respectively. In addition the severity of needs increased. For example, in 1989, 2
percent of both small and medium windows inspected in older assisted building required total
replacement. In 1989, 7 percent needed to be replaced. Similarly, among older assisted
properties with backlogs in the windows and doors system group, the average backlog was $695
per 2BR in 1989 and $1,044 in 1995.
In contrast, in unassisted properties, the average repair need for windows decreased from $217
per 2BR in 1989 (using 1995 costs) to $158 in 1995, largely due to a decrease in the severity of
reported problems (in both years about 20 percent of unassisted properties had backlogs
associated with windows and doors, but the average backlog in properties with backlogs was
$1,034 per 2BR in 1989 and $799 in 1995).
Building Roofs include roof coverings, parapet walls, chimneys, roof hatches, skylights and roof
drainage. Comparing the mean repair backlog associated with roofs per 2BR unit using 1989
condition and 1989 cost ($57 per 2BR for the stock as a whole) with the 1989 condition using
1995 costs ($48 per 2BR), the exhibit shows a slight decrease in average backlog reflecting a
slight decrease in the cost estimates for repairing/replacing roof coverings, which is the largest
contributor to this system group. The estimate of backlogs for roofs was $186 per 2BR in 1995.
This change is due to an increase in the incidence of reported problems (33 percent of properties
required some action associated with roofs in 1995 compared with 23 percent in 1989), as well as
in the severity of needs in some roof systems. For example, in 1989, 5 percent of inspected
buildings required major repairs or replacements of roof coverings. In 1995, 11 percent of
inspected buildings required major repairs or replacements of roof coverings. Similarly, among
properties with backlogs in roofs, the average backlog was $226 per 2BR in 1989 and $567 in
1995. Thus, even though the cost estimate per square foot of roof replacement decreased slightly,
the increase in incidence and the absolute high cost of replacing roofs led to an overall increase in
the mean backlog associated with roof repairs.
Building Heating and Cooling includes central vent/exhaust, central air conditioning, furnace,
boiler, boiler room piping, boiler room equipment, boiler room controls. The per unit repair cost
estimates for each of the building heating and cooling subsystems decreased between 1989 and
1995. Thus, the estimate of 1989 backlog using 1995 costs ($102 per 2BR overall) is below the
estimate for 1989 using 1989 costs ($120 per 2BR overall). However, overall the proportion of
properties requiring action increased between 1989 and 1995 from 8 to 14 percent. In unassisted
properties the proportion more than doubled from 5 percent to 11 percent. Because these
systems are very costly ($1,070 per 2 BR on average in 1995 in the properties where any action

Abt Associates Inc.

Supplerment 3

S24

was required), even a slight increase in the number of properties requiring action can contribute
substantially to the overall mean repair backlog.
Building Common Areas include vestibules, corridors, stairways, interior lights, and mail
facilities. Most individual cost components for this system group increased moderately between
1989 and 1995, as is reflected in the comparison of mean backlog in 1989 using 1989 costs ($59
per 2BR overall) with the mean backlog in 1989 using 1995 cost estimates ($71 per 2BR).
The mean backlog associated with Building Common Areas rose to $129 per 2BR in 1995 . This
is due to increased severity of reported problems. The overall percentage of properties with
repair costs associated with building common areas was fairly constant, 25 percent in 1989 and 29
percent in 1995. However, among the properties with reported backlogs in common areas the
mean backlog rose from $293 per 2BR in 1989 (using 1989 condition and 1995 costs) to $422 in
1995.
Building common areas contributed 10 percent of the overall increase in backlog in unassisted
properties. This is only a $46 increase per unit, but given the relatively low overall increase in
mean backlog in unassisted properties this represents a large percentage. (In contrast, for
example the mean backlog associated with building common areas was $86 per 2BR in older
assisted properties in 1989 and $132 in 1995. In this case, the same $46 increase represents only
4 percent of the overall increase in mean backlogs per unit).
Unit Heating and Cooling includes unit-level HVAC units, radiation, unit-level boilers, unitlevel furnaces, temperature controls, and wall air conditioners. Overall the repair costs for this
system’s components stayed the same between 1989 and 1995, as can be seen from the
comparison of costs in 1989 using 1989 condition and costs ($23 per 2BR overall) with the 1989
condition 1995 cost estimate ($22 per 2BR). However in older assisted properties both the
incidence and severity of problems increased between 1989 and 1995. In 1989, 8 percent of older
assisted properties required actions in this system group versus 24 percent in 1995. Among the
properties that required action, the mean repair need rose from $420 per 2BR in 1989 (using 1995
costs) to $641 in 1995.
Unit Interior Finishes include interior walls and surfaces, floor covering, interior doors and
frames, kitchen walls, kitchen floors, bathroom walls, and bathroom floors. As is shown in the
appendix table, costs for some components of this system group rose and some fell, but the
overall mean backlog associated with unit interior finishes fell slightly going from the 1989 cost
file to the 1995 cost file (from $699 per 2BR overall to $662).
In older assisted properties the mean backlog associated with unit interior finishes actually
decreased between 1989 and 1995, and in unassisted properties the mean backlog rose by $20 per
2BR unit contributing 4 percent to the total increase in mean repair backlog. However, in newer
assisted properties the mean repair backlog associated with unit interior finishes rose from $491
Abt Associates Inc.

Supplerment 3

S25

per 2BR unit in 1989 to $806 in 1995, contributing 18 percent of the total change. This is due to
both changes in incidence and severity of needs. Needs were reported in 62 percent of newer
assisted properties in 1995 compared with 47 percent in 1989. The average backlog associated
with unit interior finishes in newer assisted properties that reported needs was $927 in 1989
(using 1995 costs) compared with $1,305 per unit in 1995.
In Summary, between 1989 and 1995 it appears that stock-wide the condition of HUD-insured
(or held) multifamily properties deteriorated. The properties were aging, and apparently property
owners and managers were not keeping up with backlogs. Most of the reported changes in
average backlog appear to be a result of actual changes in the condition of the stock (both
incidence and severity of problems), rather than a result of changes in the costing programs or
system definitions. The mean backlog in 1989 was $1,882 (in 1995 dollars). Using the 1995 cost
file, the mean backlog in 1989 (in 1995 dollars) was $1,876. The mean backlog in 1995 was
$3,058, a 63 percent increase over the 1989 backlog (estimated using the 1995 cost files). While
the absolute backlog was highest in older assisted properties in both years, the mean backlog in
both unassisted and older assisted properties rose by about 40 percent between 1989 and 1995.
In contrast, the mean backlog in newer assisted properties rose by 165 percent from $1,057 per
2BR to $2,800.

Abt Associates Inc.

Supplerment 3

S26

Supplement 4
System by System Comparison

Abt Associates Inc.

Supplement 4

S26

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

FOUNDATION

Minor
Moderate

4 Foot
Major

Repair Cracks

Cost
0.72

Type
Action

l.f.

Patch wall

N/A
Repair cracks and refinish. Indicate
percentage to be repaired.

1995

3.00

Repair cracks, replace interior wall
surface; %

N/A

Minor

N/A

Major

Difference
Cost
0.65

Minor

Moderate

Major

% Change
-9.62

Repair cracks and refinish. Indicate
percentage to be repaired.

1.92

Repair cracks and/or exposed upper
wall surface; waterproof. Indicate
percentage to be repaired.

3.72

l.f.

>25% of foundation cracked and
deteriorated and needs to be repaired
and refinished
Repair cracks, replace interior wall
surface; %

N/A

12.52

‘90 costs appear to
be s.f. costs
although the
calculation is in l.f.

317.82*

5.30

‘90 costs appear to
be s.f. costs
although the
calculation is in l.f.

176.36*

21.17

‘90 costs appear to
be s.f. costs
although the
calculation is in l.f.

469.75*

‘90 form allowed for
replace action
which was defined
as N/A in the
handbook.

Replace

Slab

Change

N/A

Replace

Moderate

8 Foot

Action

Unit of
Cost
Measure

Repair cracks.

0.48
0.48

Repair cracks and replace 30% of
slab.

2.40
3.00

s.f.

Slab
BMT

Repair cracks.

0.45
0.45

-6.14
-6.14

Slab
BMT

Replace up to 30% of slab.

2.23
2.46

-6.97
-17.90

1990 costs have been adjusted to 1995 using a factor of 1.1986.

* Indicates absolute change of greater than 25%.

Abt Associates Inc.

Supplement 4

S27

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Replace

Action
Replace slab completely.

Cost
5.69
6.29

Unit of
Cost
Measure

Type

1995
Action

Slab
BMT

Replace slab.

Difference
Cost

Change

5.65
6.40

% Change
-0.76
1.71

1990 costs have been adjusted to 1995 using a factor of 1.1986.

* Indicates absolute change of greater than 25%.

Abt Associates Inc.

Supplement 4

S28

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System
EXTERIOR
WALL

Level

Action

Minor

Waterproof or paint affected
percentage. Indicate percentage to be
repaired.

Cost
1.20

Unit of
Cost
Measure
s.f.

Type

1995
Action
Evidence of water penetration; %

Difference
Cost

Change

1.00

% Change
-16.67

Masonry
Moderate

Major

Replace

Minor

Moderate

N/A

Point and waterproof masonry.

4.79

Indicate percentage to receive
replacement.

17.98

Waterproof or paint affected
percentage. Indicate percentage to be
repaired.

1.20

Repair/replace up to 20% of surface
(water penetration, severe damage).
>20% of surface requires
repair/replace; %
s.f.

N/A

Evidence of water penetration; %

3.94

Major action
definition change in
‘95.

-17.75

14.70

-18.24

1.00

-16.57

N/A

Plaster
Major

Replace

Minor

Moderate

2.40

Repair/replace up to 20% of surface
(water penetration, severe damage).

1.44

Indicate percentage to receive
replacement.

7.79

>20% of surface requires
repair/replace; %

7.20

-7.58

Waterproof or paint affected
percentage. Indicate percentage to be
repaired.

0.90

Paint wood; %

1.00

11.24

Prepare, paint wood.

1.50

Extensive peeling, chipped, bubbling
paint.

1.25

-16.57

Repair/replace wood.

2.40

Repair/replace up to 20% of surface
(water penetration, severe damage).

1.17

-51.19*

Indicate percentage to receive
replacement.

4.79

>20% of surface requires
repair/replace; %

5.83

21.60

Wood
Major

Replace

Major action
definition change in
‘95.

Prepare, paint.

Abt Associates Inc.

s.f.

Supplement 4

-39.93*

S29

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

Action
Waterproof or paint affected
percentage. Indicate percentage to be
repaired.

0.90

Minor

Prepare, paint V/A.

Moderate

Cost

Unit of
Cost
Measure

Type
Action

s.f.

Replace

Insulation
(add only)

Change

% Change
11.24

1.50

Extensive peeling, chipped, bubbling
paint.

1.25

-16.57

Repair/replace V/A.

2.10

Repair/replace up to 20% of surface
(water penetration, severe damage).

0.74

-64.72*

Indicate percentage to receive
replacement.

3.00

>20% of surface requires
repair/replace; %

3.84

28.15*

0.42
0.88

-49.94*
-26.58*

N/A

Moderate

N/A

Replace

Cost
1.00

Minor

Major

Difference

Paint wood; %.

Vinyl/Aluminum
Major

1995

Add insulation.

N/A

Abt Associates Inc.

0.84
1.20

s.f.

Wall
Ceiling

Add insulation.

N/A

Supplement 4

S30

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

Action

Cost

Patch roof; replace shingles.

0.90
0.60
1.20

Repair up to 25% of roof area.
Indicate percentage.

1995
Action

Difference
Cost

Change
‘95 costs reflect
implied 25%; ‘90
calculation adjusted
cost by 25% for
minor action.
Percent change in
parenthesis
respresents 25% of
‘90 costs. Types tile
and metal aded in
‘95.

% Change

Small leaks requiring repairs to
isolated areas.

0.25
0.18
0.23
0.90
1.20

1.20
0.96
1.44

EDPM
Shingle
Built-up
Tile
Metal

Up to 25% of roof requires surface
repair; %.

1.00
0.80
1.20
1.80
2.40

-16.57
-16.57
-16.57

Resurface over existing roof covering.

2.70
1.20
3.30

EDPM
Shingle
Built-up
Tile
Metal

Resurface roof but removal of existing
roof not necessary

1.95
1.46
1.87
7.20
9.60

-27.69*
21.81
-43.27*

Remove existing roof covering
system, including insulation and install
new roof and insulation.

3.90
1.98
5.39

EDPM
Shingle
Built-up
Tile
Metal

Removal of existing roof and
installation of new roof necessary.

4.21
2.10
4.67
8.02
10.56

8.07
6.18
-13.42

Moderate

Roof Covering

Major

Abt Associates Inc.

s.f.

Type

EDPM
Shingle
Built-up
Tile
Metal

Minor

Replace

Unit of
Cost
Measure

Supplement 4

-72.19*
-69.96*
-80.81*

( 11.11)
( 20.00)
(-23.33)

S31

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Minor

Action

Cost

Unit of
Cost
Measure

Type
Action

N/A
Repair and replace 10% of parapet
wall and coping.

1995

Difference
Cost

35.96

l.f.

Replace brick & coping; %

72.00

‘95 form collects
percentage to be
replaced; ‘90 form
did not, but adjusted
cost for moderate
action by 10%.

72.00

‘95 form collects
percentage to be
replaced; ‘90 form
did not.

Parapet Wall
N/A
Replace all parapet wall and coping.

Chimney
(brick)

71.92

Rebuild/replace > 50% of parapet
wall.

239.72

Repoint joints, waterproof chimney.

Moderate

N/A

Major

N/A

Replace

Roof Hatches

Rakeout and repoint mortar joints;
waterproof.

Replace/rebuild chimney.

958.88

Extensive deterioration; rebuild
chimney.

Minor

N/A

Moderate

N/A

Major

N/A

Replace roof hatch
Replace

Abt Associates Inc.

100.23*

N/A

Replace

Minor

% Change

N/A

Moderate

Major

Change

599.30
1078.74
1797.90

Small
Medium
Large

Difficult to open/close; replace.

0.12

115.75

-51.71*

1064.88

11.05

600.00
786.00
1434.00

0.12
-27.14*
-20.24

Supplement 4

S32

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Skylights

Level

Action

Type

1995
Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace skylight.
Replace

Penthouses

Cost

Unit of
Cost
Measure

359.58
958.88
1438.32

Small
Medium
Large

Deteriorated/damaged; replace.

Minor

N/A

Moderate

N/A

Major

Moisture-proof penthouse and replace
door.

Rebuild penthouse.
Replace

Roof Drainage

671.22
1654.07
3595.80

Small
Medium
Large

Moisture penetration; replace 2
components (doors, roofing, siding).

2157.48
7551.18
26968.5

Small
Medium
Large

Replace doors, roofing, siding.

N/A

Moderate

N/A

Major

N/A

Difference
Cost

Change

474.00
606.00
795.00

2534.00
5300.00
9600.00
4540.00
10300.00
19600.00

% Change

31.82*
-36.80*
-44.73*

Major action scope
change in ‘95.

277.52*
220.42*
166.98*

‘90 costs did not
include replacing
the sides, only the
top.

110.43*
36.40*
-27.32*

‘95 form allows for
each type of
drainage to have a
separate action
level. ‘90 form
allowed one action
level.

Minor

Exterior

Replace

Replace indicated percentage of
gutters, downspouts, and fascia.

Abt Associates Inc.

2.40

s.f.

Replace > 50% of gutters and
downspouts.

2.00

Supplement 4

-16.57

S33

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

Action

Cost

Unit of
Cost
Measure

Type

1995
Action

Minor

N/A

Moderate

N/A

Interior
Major

Replace

Replace accessories.

1.20

s.f.

Accessories are damaged/missing;
replace

N/A

Difference
Cost

1.00

% Change

-16.57

N/A

Reglaze with thermopane.

59.93
89.90
119.86

Small
Medium
Large

Replace window glass; #

180.00
360.00
540.00

Costs were
underestimated in
‘90.

200.35*
300.47*
350.53*

Replace window security devices.

89.90
149.83
209.76

Small
Medium
Large

Replace hardware, balances, clips,
child guards, screens, locks, glass; #

230.00
410.00
590.00

Costs were
underestimated in
‘90. Moderate
action definition
change in ‘95.

155.85*
173.65*
181.28*

Replace storm/screen windows.

89.90
113.87
161.81

Small
Medium
Large

Replace sashes; #

250.00
500.00
650.00

Major action
definition change in
‘95.

178.10*
339.11*
301.70*

419.51
629.27
1078.74

Small
Medium
Large

Replace sash, frame, storm windows;
#

Minor

Moderate

Windows
Major

Replace entire window unit.
Replace

Window
Security Grates

Change

Minor

N/A

Moderate

N/A

Major

N/A

Repair grates; #

Replace

N/A

Replace grates; #

Abt Associates Inc.

420.00
630.00
1080.00

0.12
0.12
0.12
System added in ‘95

75.00
297.00

Supplement 4

S34

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Minor

Moderate

Exterior
Common Doors

Replace hardware, rehang,
recondition door.

Replace building entry door.

Minor

Moderate

Replace building entry door and
frame.

Replace hardware, rehang,
recondition door.

239.72
239.72
239.72

Replace unit entry door.
Major

Replace unit entry door and frame.

1995
Action

Wood
Metal
Glass

Damaged/missing hardware; #

‘95 form allowed for
all 3 types; ‘90 form
only allowed for one
type.

% Change
-16.57
-16.57
-16.57

Replace door; #

605.00
670.00
845.00

-8.23
1.63
8.46

1018.81
1018.81
1138.67

Wood
Metal
Glass

Replace door and frame; #

747.00
812.00
987.00

-26.68*
-20.30
-13.32

239.72
239.72
239.72

Wood
Metal
Glass

Damaged/missing hardware; #

200.00
200.00
200.00

‘95 form allowed for
all 3 types; ‘90 form
only allowed for one
type.

-16.57
-16.57
-16.57

N/A
659.23
659.23
779.09

Wood
Metal
Glass

Replace door; #

605.00
670.00
845.00

-8.23
1.63
8.46

898.95
898.95
1018.81

Wood
Metal
Glass

Replace door and frame; #

747.00
812.00
987.00

-16.90
-9.67
-3.12

325.00

8.46

N/A

Moderate

N/A

Major

N/A

Abt Associates Inc.

200.00
200.00
200.00

Change

Wood
Metal
Glass

N/A

Replace door.

Cost

659.23
659.23
779.09

Minor

Replace

Difference

N/A

Replace

Storm/Screen
Doors

Cost

Type

N/A

Major

Replace

Unit Entry
Doors

Action

Unit of
Cost
Measure

299.65

Replace door; #

Supplement 4

S35

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Minor

Moderate

Action

Cost

Unit of
Cost
Measure

Type
Action

N/A
Repair damaged canopy; reroof;
repair/replace columns, reanchor,
paint.

1995

Difference
Cost

Change

% Change

N/A
239.72
1018.81
2397.20

Small
Medium
Large

Replace roofing and minor repairs; #

230.00
921.00
2880.00

-4.05
-9.60
20.14

830.00
3321.00
10380.00

15.41
10.83
44.34

350.00
550.00

-2.66
-8.23

Canopies
Major

N/A
Replace canopy.

Replace

Minor

Moderate

Exterior
Stairways

Major

Replace

Minor

Building
Mounted Site
Lights

Moderate
Major
Replace

Patch stair or refinish wood.

N/A
719.16
2996.50
7191.60

Small
Medium
Large

Structurally unsound; replace canopy;
#

359.58
599.30

Wood
Concrete

Chipped nosing; missing railing; #

N/A

Patch stair; replace, refinish railing;
refinish wood.

719.16
1198.60

Wood
Concrete

Patch stairs, replace railings, refinish
wood; #

750.00
1100.00

4.29
-8.23

Remove and replace stair structure.

2157.48
4794.40

Wood
Concrete

Remove and replace stair structure; #

1980.00
4550.00

-8.23
-5.10

300.00

0.12

600.00

25.15*

N/A
Replace exterior entry lights.

N/A
299.65

N/A
Replace building mounted lights.

Abt Associates Inc.

Replace fixtures; #
N/A

479.44

Replace fixtures and wiring; #

Supplement 4

S36

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Minor
Moderate

Fire Escapes

Major

Replace

Action

Cost

Unit of
Cost
Measure

Type
Action

N/A
Refinish, repair, repaint.

359.58

story

Refinish; #

1797.90

Structurally unsound; replace; #

N/A

Wrought Iron
Wood Masonry

Scrape and paint railings; #

N/A

Wrought Iron
Wood Masonry

Replace railings; #

Moderate

Major

Replace

Minor

Porches
(w/roof)

Moderate

Major
Replace
Minor

Decks
(without roof)

Moderate

Major
Replace

Cost

350.00

7260.00

30.80
38.50
36.00

Replace floor (wood only); #

1250.00

N/A

Replace floor and railings (wood only);
#

2570.00

N/A

N/A
958.88

N/A
Replace structure.

4794.40

Abt Associates Inc.

Broken hardware, minor roof leaks; #

Costs were
underestimated in
‘90.

303.80*

System added in
‘95.

900.00

-6.14

Structurally unsound; replace; #

6400.00

33.49*

800.00

11.24

3840.00

28.15*

N/A
719.16

N/A
Replace structure.

-2.66

N/A

N/A
Repair surface materials; paint;
replace hardware.

% Change

111.62
50.65
120.00

N/A

Repair surface materials; paint;
replace hardware; reroof.

Change

N/A

Minor

Balconies

Difference

N/A

N/A
Replace fire escapes.

1995

Broken hardware, minor repairs; #

N/A
2996.50

Structurally unsound; replace; #

Supplement 4

S37

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level
Minor

Attached
Storage Sheds

Moderate

Major
Replace
Minor
Moderate

Vestibules
Major

Replace

Cost

N/A
Repair surface materials; paint;
replace hardware; reroof.

Paint interior; perform minor repairs.

419.51

s.f.

Action

Difference
Cost

Change

% Change

Broken hardware, minor roof leaks; #

400.00

-4.65

1.97

N/A
1078.74

Structurally unsound; replace; #

1100.00

1.80

Paper/paint walls; minor repairs.

2.25

25.15*

4.20

7.82

N/A
Replace flooring, paint; perform
repairs; recondition.

1995

N/A

N/A
Replace structure.

Type

N/A
3.90

N/A

s.f.

Paper/paint walls; minor repairs;
replace floor.
N/A

Paint walls and ceilings; replace
wallpaper.

1.80

Paper/paint walls and ceilings.

2.25

25.15*

Moderate

Paint walls and ceilings; replace
wallpaper; replace floor coverings.

3.60

Paint/paper surfaces and patch
plaster.

3.25

-9.62

Paint walls and ceilings; replace
wallpaper; replace floor coverings;
repair/replace railings.

5.09

Paint/paper surfaces and replace floor
covering.

5.75

12.88

Major

Replace indicated percent of walls
and ceilings; paint walls and ceilings;
replace wallpaper; replace floor
coverings; repair/replace railings.

5.99

Replace walls, ceilings, floors; %

6.75

Minor

Corridors

Action

Unit of
Cost
Measure

Replace

Abt Associates Inc.

Major cost added to
replace action in
‘90, but not in ‘95.

Supplement 4

12.63

S38

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

Action

Cost

Paint walls and ceilings; replace
wallpaper.

1.80

Moderate

Paint walls and ceilings; replace
wallpaper; replace floor coverings.

Major

Unit of
Cost
Measure
s.f.2

Type

1995
Action

Difference
Cost

Paint walls and stairs.

2.75

4.20

Paint walls and stairs and patch
plaster.

3.75

Paint walls and ceilings; replace
wallpaper; replace floor coverings;
repair/replace railings.

5.99

Paint walls and stairs and replace
finish on treads and risers.

9.00

Replace indicated percent of walls
and ceilings; paint walls and ceilings;
replace wallpaper; replace floor
coverings; repair/replace railings.

7.19

Replace stairs and patch walls; %

30.85

Minor

Stairways

Replace

Minor
Moderate

N/A
Replace light fixtures.

Change
The ‘90 action
definitions were not
correct for this
system and did not
include the costs to
replace stairs.

% Change
52.96*

-10.61

50.18*

Major cost added to
replace cost in ‘90,
but not in ‘95.
Percent change in
parenthesis
represents the
aggregate costs.

328.97*

(134.07*)

N/A
1.20

s.f.3

Replace light fixtures

1.00

-16.57

2.75

8.23

Interior Lighting
Major
Replace

2.
3.

N/A
Replace fixtures and wiring.

N/A
3.00

Replace fixtures and wiring.

The ‘90 form collected stairway s.f., the calculation used a proxy of 160 s.f. x # stories, when S.F. was missing. The ‘95 form did not collect stariway s.f. Thus, the claculation always used the
proxy.
In ‘90, s.f. was calculated as the sum of corridor and common room s.f. In ‘95, s.f. was equal to the building footprint.

Abt Associates Inc.

Supplement 4

S39

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Mail Facilities

Level

Action

Type

1995
Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace mailboxes.
Replace

2337.27
77.91

Kiosk
Box

Cost

Change

% Change

1200.00
70.80

‘95 form collected
the number to be
replaced; ‘90 did
not.

-48.66*
-9.12

Paint room; minimal repairs.

2.25

Costs were
underestimated in
‘90.

87.72*

Not secure; replace; #

1.20

Paint; replace floor coverings;
recondition.

3.60

Paint and replace floor covering.

4.75

32.10*

Major

Paint; replace floor coverings;
recondition; perform M&E repairs.

4.20

Paint, replace floor covering and M&E
repairs.

7.00

66.86*

Paint; replace floor coverings;
recondition; perform M&E repairs;
replace interior components.

5.99

Gut rehab necessary.

Replace

Moderate

Laundry Rooms

Minor

Moderate

N/A
Repair electrical service or water
supply/drain or vent.

s.f.

Difference

Paint laundry rooms.

Minor

Laundry
Equipment

Cost

Unit of
Cost
Measure

25.00

239.72

N/A

Replace all components.

Abt Associates Inc.

317.15*

N/A
Repair motors, belts, switches.

200.00

Replace equipment; #

600.00

Major action added
in ‘95 (same as
replace action in
‘90).

1000.00

Replace action
definition change in
‘95.

Major

Replace

Replace action
definition change in
‘95.

719.16

Replace equipment, electrical service,
drain and vents.

-16.57

Supplement 4

39.05*

S40

System By System Comparison
Booklet: BUILDING ENVELOPE
19901

System

Level

1.20

Paint; replace floor coverings;
recondition.

Major

Replace

1995
Action

Difference
Cost

Change
Minor action
definition change in
‘95.

% Change

2.25

3.60

Paint, replace floor covering,
additional repairs.

4.20

Paint; replace floor coverings;
recondition; perform M&E repairs.

4.20

N/A

Paint; replace floor coverings;
recondition; perform M&E repairs;
replace interior components.

5.99

Paint, replace floor coverings, M&E
repairs.

7.00

16.80

Paint; replace floor coverings;
recondition.

1.20

Paint, replace floor covering.

4.20

250.41*

Moderate

Paint; replace floor coverings;
recondition; perform M&E repairs.

3.60

Paint, replace floor coverings, M&E
repairs.

7.00

94.67*

Paint; replace floor coverings;
recondition; perform M&E repairs;
replace interior components.

5.99

N/A

Major

Minor

Common
Kitchens

s.f.

Type

Paint and minimal repairs.

Moderate

Underground
Garage

Cost

Paint common rooms.

Minor

Common
Rooms

Action

Unit of
Cost
Measure

s.f.

16.80

Major action
dropped in ‘95.

Replace

N/A

Gut rehab necessary.

40.00

Minor

N/A

Paint parking stripes.

0.02

Moderate

N/A

Replace lighting fixtures.

0.70

Major

N/A

Resurface and major repairs.

1.00

Replace

N/A

Abt Associates Inc.

87.72*

Replace action
added in ‘95.
System added in
‘95.

Supplement 4

S41

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level

Action

Cost

Replace valve stems; replace riser
gate valves; replace temp probes.

89.90

Moderate

Repair supports, patch insulation,
patch or replace leaky sections, clean
clogged return lines.

Major

Minor

Heating Risers

Replace

Minor

Unit of
Cost
Measure

Type

Action

Difference
Cost

Change

% Change

Repair/replace valve stems, riser gate
valves, temp. probes.

100.00

11.24

119.86

Replace supports, insulation piping,
<25% or return lines.

125.00

4.29

Replace indicated percentage of runs;
and insulate piping.

179.79

Replace 25% -2/3 of piping; insulate
piping; %

175.00

-2.66

Replace entire system.

239.72

Replace risers.

250.00

4.29

Repair/replace valve stems, riser gate
valves, temp. probes.

100.00

11.24

Replace valve stems; replace riser
gate valves; replace temp probes.

89.90

unit

1995

unit

Repair supports, patch insulation,
patch or replace leaky sections, clean
clogged return lines.

179.79

Replace supports, insulation piping,
<25% or return lines.

175.00

-2.66

Moderate

Major

Replace indicated percentage of runs;
and insulate piping.

299.65

Replace 25% -2/3 of piping; insulate
piping; %

275.00

-8.23

Replace entire system.

383.55

Replace risers.

350.00

-8.75

Gas Distribution

Replace

Abt Associates Inc.

Supplement 4

S42

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level
Minor

Action

Cost

Replace valve stems; repack gate
valves; replace pressure gauges.

119.86

Replace indicated percentage of
piping and insulation.

Unit of
Cost
Measure

1995
Type

unit

Action

Difference
Cost
125.00

239.72

Replace up to 25% of piping.

225.00

‘90 form collected
percentage; ‘95
assumes 25%.
Minor cost added to
moderate cost in
‘90, but not in ‘95.
Percent change in
parenthesis
represents the
aggregate cost.

-6.14

(-37.43*)

Replace hot or cold water distribution.

359.58

Replace 25% -2/3 of piping; %

350.00

Major action
definition/scope
change in ‘95.
Minor cost added to
major cost in ‘95,
but not in ‘90.
Percent change in
parenthesis
represents the
aggregate cost.

-2.66

( 32.10*)

Replace entire hot and cold water
distribution.

479.44

Systems corroded; replace.

450.00

Major

Replace

Sanitary
Distribution

Minor

Replace broken floor or wall; clean out
covers; rout and clean problem areas;
snake floor drains.

Moderate

Replace broken floor or wall clean out
covers; rout and clean problem areas;
snake floor drains. Replace 10% of
piping.

Major

Replace

% Change

Replace valve stems & pressure
gauges.

Moderate

Domestic Hot
& Cold Water
Distribution

Change

-6.14

Type added in ‘95.

119.86

PVC
Cast Iron

Backups; replace broken floor, clean
covers, snake drains.

50.00
125.00

239.72

PVC
Cast Iron

“MINOR” + pipe deterioration; replace
up to 25% of piping.

100.00
225.00

-6.14

Replace indicated percentage of
affected piping.

359.58

PVC
Cast Iron

Corrosion, cracks; replace 25% -2/3
of piping; %

200.00
350.00

-2.66

Replace entire system including vent
stacks.

599.30

PVC
Cast Iron

Drains clogged, systemic leaks;
replace system.

300.00
450.00

-24.91

Abt Associates Inc.

unit

4.29

Supplement 4

4.29

S43

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level
Minor

Action

Cost

Unit of
Cost
Measure

1995
Type

Major

Replace

Minor

Cost

Replace 10% of heads & piping.

0.25

-16.57

Replace indicated components.

0.60

“MINOR” + replace siamese twin
connection, switches, flow control
valve.

0.50

-16.57

Replace entire system, except water
pump.

2.00

Major action and
replace action
switched in ‘95.

35000.0

Major action and
replace action
switched in ‘95.

Replace fire pump.

Replace entire system including fire
pump.

41951.00

2.40

Replace entire system, including water
pump.

N/A
Replace motor.

N/A
179.79
479.44

Residential
Commercial

Replace motor.

200.00
400.00

Sump Pumps

Replace

Minor

N/A
Replace all components.

479.44
1318.46

Residential
Commercial

Replace system.

500.00
1600.00

Replace one major component.

1797.90
1797.90

Small
Large

Replace pump or motor or piston
cylinder.

Replace two major components.

2996.50
3595.80

Small
Large

N/A

Replace entire system.

5993.00
14383.2

Small
Large

Replace system.

Abt Associates Inc.

11.24
-16.57

4.29
21.35

N/A
1000.00
1060.00

Compactors

Replace

‘95 form allows both
types of sump
pump; ‘90 form
allowed only 1 type.

N/A

Moderate

Major

% Change

0.30

Moderate

Major

Change

Replace 10% of heads.

Moderate

Fire Sprinkler
System

Action

Difference

‘95 form allows both
types of
compactors; ‘90
form allows only 1
type

-44.38*
-41.04*

Major action
dropped in ‘95.

5000.00
10000.00

Supplement 4

-16.57
-30.47*

S44

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level
Minor
Moderate

Central Vent &
Exhaust

Major

Replace
Minor

Central Air
Conditioning

Moderate
Major
Replace

Switchgear

Cost

1995
Type

N/A
0.90

Replace fan; replace up to 10% of
ductwork.
Replace entire system.

s.f.

Change

% Change

Replace fan.

0.75

1.50

Replace fan & up to 25% of ductwork.

1.25

2.40

Replace ductwork.

2.00

-16.57

1.50

-16.57

4.50

-16.57

N/A

-16.57
Major action
percentage change
in ‘95.

-16.57

N/A
1.80

s.f.

Replace compressor and/or fan.

N/A
Replace entire system.

Cost

N/A

Replace fan.

Replace components.

Action

Difference

N/A
5.39

Replace system.

Type dropped in ‘95.

Minor

N/A

Moderate

N/A

Additional capacity necessary; add
panel.

0.35

Replace system.

1.50

-16.57

2.25

-6.14

Major

Replace

Building Power
Wiring

Action

Unit of
Cost
Measure

Replace entire system.

1.20
1.80

s.f.

w/out heat
with heat

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace all wiring within building.

Abt Associates Inc.

2.40

s.f.

Replace >50% of wiring.

Major action added
in ‘95.

Supplement 4

S45

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Emergency
Lights

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace unit and fixtures.

479.44

6 units

Replace light unit.

N/A

Moderate

N/A

Major

N/A

Difference
Cost

425.00

Replace

Communication
System

239.72
299.65

4 units

Battery
Hardwire

Detectors & annunciator panel
broken; replace system.

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Emergency Call
Alarm System

Replace system including annunciator
panel.

Replace central system.

239.72

6 units

Wiring & receiver broken; replace
system.

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace central system.

Abt Associates Inc.

359.58

6 units

Annunciator panel, horns, lights
broken; replace system.

% Change

-11.35
‘95 form allows for
both types; ‘90 form
allowed for only 1
type.

Minor

Smoke/Fire
Detection

Change

100.00
153.00

-58.28*
-48.94*

225.00

-6.14

310.00

-13.79

Supplement 4

S46

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Master TV
Antenna

Level

Cost

1995
Type

Action

Difference
Cost

Change

% Change

Minor

N/A

Moderate

N/A

Major

N/A

Replace antenna or master dish.

2000.00

Replace antenna/ master dish &
cables.

4500.00

25.15

1200.00

0.12

Repair/replace minor component
(burner, controls, humidifier).

225.00

-6.14

Replace burner or combustion
chamber or fan.

675.00

-6.14

1500.00

-3.73

Replace

Closed Circuit
TV

Action

Unit of
Cost
Measure

Replace antenna. Replace system.

3595.80

6 units

Minor

N/A

Moderate

N/A

Major

N/A

Replace
Minor

Moderate

Hot Air Furnace

Major

Replace

Replace CCTV central system.

1198.60

6 units

N/A

Tune-up burner, adjust/replace
controls; patch small air leaks, secure
loose mounts; clean/replace
humidifier; lube fan; clean flue, adjust
draft controls.

239.72

Replace the faulty component.

719.16

Replace system.

Abt Associates Inc.

Replace system.

Major action added
in ‘95.

1558.18

unit

Replace system.

Supplement 4

S47

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level

Action

Cost

Unit of
Cost
Measure
unit

1995
Type

Action

Difference
Cost

Moderate

Type added in ‘95.
Action shift in ‘95;
major action
dropped; minor
became moderate;
moderate became
major.

119.86

Replace major component.

359.58

Replace burner.

350.00

-2.66

Replace two major components.

719.16

Replace major component (boiler,
cast-iron sections, insulation,
combustion chamber).

550.00

-23.52

1500.00

-3.73

Replace insulation.

172.50

-4.05

Major

Replace

% Change

Replace burner.

N/A

Minor

Boilers

Change

Replace system.

1558.18

Replace system.

Boiler Room
Replace piping insulation.

179.79

Replace indicated percentage of
piping.

299.65

Replace up to 25% of piping; %

287.50

‘90 form did not
collect percentage
of piping to be
replaced; ‘95 form
does.

-4.05

Moderate

479.44

Replace 25% -2/3 of piping; %

460.00

‘90 form did not
collect percentage
of piping to be
replaced; ‘95 form
does.

-4.05

Major

Replace indicated percentage of
piping.

Replace entire system.

791.08

Corrosion, excessive leaks; replace
system.

759.00

Minor

Piping

Replace

Abt Associates Inc.

unit

Supplement 4

-4.05

S48

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level

Action
Perform minor repairs.

Cost
599.30

Unit of
Cost
Measure
unit

1995
Type

Action

Difference
Cost

Isolated damage; minor repairs.

200.00

Minor

Change

% Change

Boiler room
peripherals was
split into separate
systems of
Equipment and
Controls in ‘95.
Percent change
cannot be
calculated.

Equipment
Moderate

Major

Replace

Replace 1/3 of associated equipment.

1078.74

Repair/replace up to 25% of
equipment.

500.00

Replace 2/3 of associated equipment.

1678.04

Repair/replace 25% - 2/3 of
equipment.

1000.00

Replace all peripheral systems.

2517.06

Replace all equipment.

1500.00

N/A

Minor

Moderate

Controls

Controls included with equipment in
system called peripherals IN ‘90.

Major
Replace

Abt Associates Inc.

unit

Replace up to 25% of temp. controls
& zone valves.

150.00

N/A
Replace temp. controls & zone valves.

600.00

Supplement 4

S49

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

DHW
Generation

Level

Action

Cost

Unit of
Cost
Measure

Type

Action
Repair/replace minor component
(temp. controls, heat exchanger,
gaskets).

Difference
Cost

Change

50.00

% Change

59.93

Minor

Adjust/replace temp. controls; clean
heat exchanger; tune-up burner;
tighten fittings and valve packings;
replace gaskets; replace small
circulating pumps; patch insulation.

149.83

N/A

Moderate

Adjust/replace temp. controls; clean
heat exchanger; tune-up burner;
tighten fittings and valve packings;
replace gaskets; replace small
circulating pumps; patch insulation.

209.76

Replace major component (burner,
combustion chamber).

200.00

-4.65

Major

Replace either burner(s) or all of the
peripherals and tank insulation or
retube exchanger or replace the
combustion chamber.
Replace entire system.

281.67

Replace system.

250.00

-11.24

Replace dysfunctional items and
perform all adjustments necessary.

239.72

Repair/replace electrical item;
compensation chain rope.

2500.00

Repair/replace 50% of the listed
observations and perform all
necessary adjustments.

599.30

Repair/replace 50% of components.

3000.00

5000.00

Replace

unit

1995

-16.57

Moderate action
dropped in ‘95.

Elevator Shaftways

Minor

Hoist

Moderate

Major

Replace

Replace dysfunctional components.

2397.20

Replace hoist ropes or guide rails.

Replace all shaftway systems.

5993.00

Replace system.

Abt Associates Inc.

‘95 form allows for
both types; ‘90 form
allowed only 1 type.
Costs were
underestimated in
‘90.

942.88*

400.58*

Major action
definition change in
‘95.

10000.00

Supplement 4

108.58*

66.86*

S50

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Level

Minor

Hydraulic

Moderate

Major

Replace

Minor

Moderate

Shaftway Doors
Major

Replace

Action

Cost

Moderate

Major

Replace

1995
Type

Action

Difference
Cost

Replace dysfunctional items and
perform all adjustments necessary.

119.86

Repair/replace electrical item;
compensation chain rope.

3000.00

Repair/replace 50% of the listed
observations and perform all
necessary adjustments.

239.72

Repair/replace 50% of components.

4000.00

6000.00

Replace dysfunctional components.

1198.60

Replace piston assembly or guide
rails.

Replace all shaftway systems.

5993.00

Replace system.

Replace indicated components.

359.58

Replace indicated components.

599.30

floor

Replace small components
(interlocks, level switches, wiring).

Change
Costs were
underestimated in
‘90. Minor action
definition change in
‘95.

500.00

Replace door gibs, rollers.

1500.00

1438.32

Replace door.

2000.00

Replace doorway system.

2397.20

Replace door system (frame, door,
equipment).

3000.00

% Change
2402.90*

1538.6*

Major action
definition change in
‘95.

12000.00

Replace shaftway doors and include
both minor and moderate repairs.

400.58*

100.23*
Costs were
underestimated in
‘90.

39.05*

150.29*
Major action
definition change in
‘95.

39.05*

25.15*

Costs were
underestimated in
‘90.

Replace the indicated component(s).

359.58

Replace light fixture, fan, display
lamps, buttons, telephone, handrails,
toe guard.

1000.00

Replace indicated components and
perform all necessary adjustments.

599.30

Replace door, guides, motor & drive &
>20% of interior finish damaged.

2500.00

317.15*

Replace the components and
waterproof the floor if necessary.

839.02

Replace safety blocks, rollers, cables.

2500.00

197.97*

Replace system.

3500.00

16.80*

Minor

Cabs

Unit of
Cost
Measure

Replace entire system.

Abt Associates Inc.

2996.50

Supplement 4

178.10*

S51

System By System Comparison
Booklet: BUILDING MECHANICAL & ELECTRICAL
19901

System

Elevator
Controller /
Dispatcher

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Difference
Cost

Change

% Change

System dropped
in ‘95.

Minor

Replace indicated components.

599.30

N/A

Moderate

Replace indicated components.

1198.60

N/A

Major

Replace indicated components.

1797.90

N/A

Replaace entire system.

5993.00

N/A

Replace/repair indicated components
and perform all adjustments.

2397.20

Repair/replace worm thrust bearing,
brake pads, brake pin, governor
guard, governor power interlock.

2000.00

-16.57

Replace brake drum, turn and
undercut commutator(s), turn and
undercut drive sheave vee groove,
align worm and worm gear,
replace/repair dysfunctional
component(s) and perform all
necessary adjustments.

4794.40

Repair/replace brake drum, motor
generator, drive sheave, brake
solenoid, cooling fan.

4000.00

-16.57

Replace/repair dysfunctional
components and perform all
necessary adjustments.

7191.60

“Moderate” + replace major
component.

6000.00

-16.57

Major

Replace

Replace all hoist machinery.

23972.00

Replace >75% of system.

20000.00

-16.57

5000.00

-16.57

Replace

Machinery
Minor

Moderate

Hoist

Minor
Moderate

Hydraulic

Major

Replace

Replace hydraulic pump.

5993.00

Replace pump.

Replace electric motor.

11986.00

Replace electric motor.

10000.00

-16.57

Replace pump and motor.

17979.00

Replace pump & motor.

15000.00

-16.57

Replace all system components,
including hydraulic fluid tank.

47944.00

Replace all components, including
hydraulic fluid tank.

25000.00

-47.86*

Abt Associates Inc.

Supplement 4

S52

System By System Comparison
Booklet: SITE
19901

System

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Difference
Cost

Change

% Change

Reseed, replant selected areas.

0.12

10% s.f.

Worn areas.

0.11

Reseed, replant up to 1/3 of site.

0.24

s.f.

Up to 25% deteriorated.

0.26

Reseed, replant up to 2/3 of site.

0.90

25% - 2/3 deteriorated.

0.78

-13.23

Reseed replant all of site.

1.20

>2/3 deteriorated.

1.05

-12.40

Minor

Patch as necessary; repave, regravel
less than 10% of road .

0.17

10% s.f.

Isolated damage; up to 10% requires
patching, paving, regravel .

0.15

-10.61

Repair selected holes, curbs and
resurface 10-50% of roadway.
(Indicate percentage to be repaired.)

0.35

s.f.

10-50% requires repair; %

0.35

Moderate

Repair selected holes, curbs and
resurface 51-100% of roadway.
(Indicate percentage to be repaired.)

0.61

51-100% requires resurfacing; %

0.70

14.51

Major

Replace

Demolish existing roadway and
reconstruct new base and surface.

1.87

Removal of roadway; new base &
surface necessary

1.74

-6.94

Minor

Patch as necessary; seal coat and
stripe.

0.17

10% s.f.

Isolated damage; up to 10% requires
patching, paving, regravel.

0.10

-40.41*

Repair selected holes or ponding
resurface 10-50% of paving. (Indicate
percentage to be repaired.)

0.35

s.f

10-50% requires repair; %

0.50

43.85*

Repair selected holes, curbs and
resurface 51-100% of paving.
(Indicate percentage to be repaired.)

0.61

51-100% requires resurfacing; %

1.00

63.59*

Demolish existing roadway and
reconstruct new base and surface.

1.59

Removal of parking area; new base &
surface necessary.

1.74

9.15

Minor

Moderate

Landscaping
Major
Replace

-8.23
Moderate action
percentage change
in ‘95.

8.46

Roadways

Moderate

Parking Areas Lots
Major

Replace

Abt Associates Inc.

Supplement 4

S53

System By System Comparison
Booklet: SITE
19901

System

Level

Action

Cost

Unit of
Cost
Measure
s.f.

Minor

Moderate

Parking Areas Garages

Major

Replace

Type

Action

Difference
Cost

Paint stripes.

0.02

Replace lighting fixtures; %

0.70

Resurface garage; %

1.00

Removal of parking area; new base
and surface necessary.

1.74

Change

% Change

System added in
‘95.

Spot repair less than 10% of paved
areas.

0.24

10% s.f.

Isolated damage; up to 10% requires
resurfacing.

0.30

‘90 costs did not
include removal.

25.15*

Moderate

Resurface 10-50% of paving.
(Indicate percentage to be repaired.)

0.48

s.f.

10-50% requires resurfacing; %

0.76

‘90 costs did not
include removal.

58.52*

Major

Resurface 51-100% of paving.
(Indicate percentage to be repaired.)

0.84

51-100% requires resurfacing; %

1.89

‘90 costs did not
include removal.

125.26*

Demolish and reconstruct new
pedestrian walkway.

1.73

Removal of parking area; new base &
surface necessary.

3.73

‘90 costs did not
include removal.

116.11*

4.41

Form change in ‘95
to collect %.

Minor

Paved
Pedestrian
Areas

1995

Replace

Curbing

Bituminous

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Remove and replace curbing

3.60

l.f.

Remove and replace curb; %

Minor

N/A

Moderate

N/A

22.64

Concrete
Major
Replace

Repair, patch and realign.

11.99

Remove and replace curbing.

19.18

Abt Associates Inc.

l.f.

Repair, patch and realign curb.
Remove and replace curb.

7.57

-36.84*

15.24

-20.53

Supplement 4

S54

System By System Comparison
Booklet: SITE
19901

System

Level
Minor
Moderate

Action

Cost

Unit of
Cost
Measure

N/A

1995
Type

Action

Difference
Cost

Change

% Change

N/A

Repoint joints.

2.40

Realign and replace joints.

9.59

l.f.

Repoint joints.

2.18

-9.06

Repair, patch and realign curb.

7.88

-17.82

Granite
Major

N/A

Moderate

N/A

Major

N/A

Replace

Fencing2
Minor

Form change in ‘95 to
allow separate
measurements per type.

Chain Link

Replace

Indicate percentage of fence to be
replaced.

14.38

l.f.

Beyond repair; %

Minor

N/A

Moderate

N/A

Major

N/A

Wrought Iron

Replace

Indicate percentage of fence to be
replaced.

39.55

l.f.

Beyond repair; %

14.50

‘90 costs did not
include vinyl coating

57.84

Costs were
underestimated in
‘90.

0.81

46.23*

Additional types (concrete, tubular) costed in 1995.

Abt Associates Inc.

Supplement 4

S55

System By System Comparison
Booklet: SITE
19901

System

Wood Stockade

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Minor

Indicate percentage of fence to be
replaced

11.99

l.f.

Beyond repair; %

N/A
Repair walls, clean, fill cracks, reset
portions.

Difference
Cost

15.27

17.98

l.f.

Concrete
RR Ties

Isolated cracking, water seepage
requires resetting to portions of wall;
%

2.99
0.67

Retaining Walls

Replace

N/A
Indicate percentage of wall to be
replaced.

% Change

27.40*

N/A

Moderate

Major

Change

‘90 calculatations
assumed a 6 foot
height--the calculation was cost by l.f.;
‘95 calculates l.f. x
height. Percent
change cannot be
calculated.

N/A
179.79

Substantial heaving or settlement
problems; %

34.04
24.17

Site Drainage
Minor

Moderate

Underground

Major

N/A

Regrade and reset rims.

1066.75

Rebuild portions of basins.

4015.31

Replace catch basins.

9409.01

Replace

Abt Associates Inc.

Catch
Basin

Catch basins ineffective requiring
regrading of land & resetting of rims.

900.00

-15.63

Rebuild catch basins.

2670.00

-33.50*

Replace catch basins; #

4500.00

‘95 form collects #
catch basins to
replace; ‘90 form did
not.

Supplement 4

-52.17*

S56

System By System Comparison
Booklet: SITE
19901

System

Level
Minor

Moderate

Action

Cost

Unit of
Cost
Measure

N/A

1995
Type

Action

Difference
Cost

Change

% Change

N/A

Regrade up to 50% of site.

0.90

Regrade up to 100% of site.

1.14

landscap
e s.f.

Earthwork required to offset erosion.

0.33

More earthwork and concrete
required.

2.50

Tighter specs in ‘95.

-63.29*

Surface
Major

Replace
Minor

Pole Mounted
Site Lighting

Moderate
Major
Replace

Replace fixtures.

599.30

Replace supports.

958.88

pole

Replace fixtures.

500.00

-16.57

Replace supports.

800.00

-16.57

Replace supports and fixtures.

1558.18

Replace fixtures and supports.

1350.00

-13.36

Replace wiring, supports and fixtures.

4195.10

Replace fixtures, supports, wiring.

3500.00

-16.57
Calculation change
in ‘95. “90 costs
were per 100%
units.

35.96

10% units
(‘95 calc.)

Peeling paint, ,rust, minor damage;
repair/refinish necessary.

Moderate

Replace up to 1/3 of elements.

71.92

25% units
(‘95 calc.)

Up to 25% needs to be replaced.

120.00

66.86*

Major

Replace up to 2/3 of elements.

119.86

66% units
(‘95 calc.)

25%-2/3 needs to be replaced.

120.00

0.12

Replace all site furniture.

179.79

100% units
(‘95 calc.)

All furniture needs to be replaced.

120.00

-33.26*

485.00

15.61

970.00

7.90

Replace

Minor
Moderate

Private Yards
and Enclosures

N/A

Repair and/or refinish appropriate
elements.

Minor

Site Furniture

N/A

119.55*

Major

Replace

N/A
Repair enclosures; patch patios.

Abt Associates Inc.

0.12

N/A
419.51

N/A
Replace enclosures; remove and
rebuild patios.

36.00

yard

Up to 1/3 needs repair.
N/A

898.95

Replacement of enclosure and
removal and construction of patio
necessary.

Supplement 4

S57

System By System Comparison
Booklet: SITE

Cost

Unit of
Cost
Measure

Replace enclosure.

1198.60

dumpster

Replace dumpster.

Major

Replace

19901

System

Level
Minor

Dumpsters and
Enclosures

Moderate

Minor

Moderate

Swimming Pool
Major

Replace

Minor

Moderate

Tennis Courts
Major

Replace

Action

1995
Type

Action

Difference
Cost

Change
Definitions of all
actions changed
from ‘90 to ‘95.

% Change

Replace dumpster.

1000.00

4195.10

Replace dumpster & repair enclosure.

2500.00

-40.41*

Replace pad and enclosure.

2517.06

Replace dumpster & enclosure.

4300.00

70.83*

Replace pad, enclosure and
dumpster.

6712.16

Replace dumpster, enclosure & pad.

5700.00

-15.08

Paint pool with epoxy and install new
filters and pumps.

5513.56

Paint pool w/ epoxy & repair filters &
pumps.

5140.00

-6.78

Repair/resurface deck.

6712.16

Paint pool w/epoxy, replace filters &
pumps, patch decks.

7864.00

Drain and resurface pool interior;
replace filtration system;
repair/resurface deck.

pool

16780.4

N/A
Apply sealant.

3595.80

N/A

Replace posts and net; apply sealant.

Rebuild court, install new posts and
net.

Abt Associates Inc.

5153.98

25170.60

court

Resurface interior, repair & resurface
deck, replace filtration system.

10588.0

Replace pool.

34475.0

Surface requires sealant & relining.

2678.00

Surface requires sealant & relining &
replace posts & net.

2940.00

Replace posts & net & resurface
court.

11655.00

Rebuild court; install new posts & net.

24194.00

Moderate action
changed in ‘95.

-16.57

17.16

-36.90*

Replace action
added in ‘95.

-25.52*
Moderate action
added in ‘95.

Major action scope
change in ‘95.

Supplement 4

126.14*

-3.88

S58

System By System Comparison
Booklet: SITE

Cost

Unit of
Cost
Measure

Apply sealant.

3595.80

court

Install new backboards and supports.

2157.48

19901

System

Level
Minor

Moderate

Basketball
Courts

Major

Replace

Action

1995
Type

N/A
Rebuild court, install new backboards
and supports.

Action

Difference
Cost

Surface requires sealant & relining.

2248.00

Surface requires sealant; replace
backboards & supports.

3094.00

Moderate

Major

Replace

Heating Water
Distribution

13184.60

Rebuild court; install new backboards
& supports.

N/A

Domestic Hot
Water Lines

Replace damaged wiring, supports,
conduits.

113.87
137.84

Replace entire system.

155.82
179.79

l.f.

over
under

10024.00

‘90 form
differentiated
between with and
w/out electrical heat
for each type.

130.00
150.00

-16.57
-16.57

N/A

Major

N/A

N/A
179.79
209.76

l.f.

Steam
Hot Water

Replace lines; %

Minor

N/A

Moderate

N/A

Major

N/A

Abt Associates Inc.

29.97

l.f.

-23.97

Replace system.

Moderate

Replace indicated percentage.

43.41*

-16.57
-16.57

N/A

Replace

Moderate action
scope change in
‘95.

95.00
115.00

N/A

Replace indicated percentage.

-37.48*

Replace up to 40% of wiring,
supports, conduits.

Minor

Replace

% Change

N/A

Minor

Site Electrical
Distribution

Change

Replace lines; %

325.00
175.00

Costs were
underestimated for
steam in ‘90.

80.77*
-16.57

40.00

Costs were underestimated in ‘90.

33.49*

Supplement 4

S59

System By System Comparison
Booklet: SITE
19901

System

Domestic Cold
Water Lines

Level

Cost

1995
Type

Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Main Water
Service

Action

Unit of
Cost
Measure

Replace indicated percentage.

23.97

l.f.

Replace lines; %

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace indicated percentage.

47.94

l.f.

Replace lines; %

Minor

N/A

Moderate

N/A

Major

N/A

Difference
Cost

Change

% Change

25.00

4.29

45.00

-6.14

30.00

0.12

40.00

-4.65

350.00

-6.56

8000.00

-11.01

Gas Lines

Replace

Site Sanitary
Lines

Replace indicated percentage.

29.97

l.f.

Replace lines; %

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace indicated percentage.

41.95

l.f.

Replace lines; %

Minor

N/A

Moderate

N/A

Septic System
Major
Replace

Replace tanks.
Replace the entire system.

Abt Associates Inc.

374.56
8989.50

unit

Replace tanks.
Replace tanks and leaching fields.

Supplement 4

S60

System By System Comparison
Booklet: SITE
19901

System

Level
Minor

Sewage
Ejectors

Action
Replace pump(s).

Cost
958.88

Unit of
Cost
Measure
ejector

1995
Type

Action
Replace pumps.

Difference
Cost

Change

% Change

800.00

-16.57

Moderate

Replace pump(s) and motor(s).

1678.04

Replace pump and pump motors.

1400.00

-16.57

Major

Replace pump(s), motor(s) and
controls.

2397.20

Replace pump, pump motors and
controls.

2000.00

-16.57

Replace the entire system

4195.10

Replace system.

3500.00

-16.57

Replace
Minor

N/A

Moderate

N/A

Major

N/A

Hydrants
Replace indicated percentage of
hydrants.

1797.90

Hydrant

Replace hydrants; #

2000.00

‘90 form collected
percentage of
hydrants to replace;
‘95 form collects
number of hydrants
to replace.

11.24

Tune-up generator.

200.00

‘90 form collected
this site-level
system on the BME
form.

-58.28*

500.00

-65.24*

7500.00

25.15*

16000.00

55.50*

Replace

Tune-up generator.

479.44

Minor

Emergency
Generators

Moderate

Major

Replace

Tune-up and perform general repairs.

1438.32

Tune-up generator; repair valves.

Perform general rebuilding/repairs.

5993.00

Replace engine or generator; rebuild
control panel.

Replace entire system.

Abt Associates Inc.

35958.00

Replace system.

Supplement 4

S61

System By System Comparison
Booklet: UNIT
19901

System

Walls &
Ceilings:
Partitions (not
kitchen & bath)

Level

Type

Action

N/A

Moderate

N/A

Major

N/A

Replace wall/ceiling material and
paint. Indicate percentage to be
replaced.

3.00

s.f

Water damage, buckling; replace; %

Minor

N/A

Moderate

N/A

Major

N/A

Difference
Cost

Change

3.00

% Change

0.00

‘90 costs do not
include demolition;
‘95 costs do.

Indicate percentage to be replaced.

1.80

s.f.

Buckling, warped, splintered; replace;
%

3.35

Surface material needs to be restored
with minimal prep work; paint.

0.72

s.f.

Paint.

0.58

-19.35

Moderate

Major prep work required for surface
material restoration; paint.

1.14

Surface damage (patch, spackle) but
partition fine.

1.08

-5.15

Major prep work required for surface
material restoration including tile
replacement; paint.

1.62

N/A

Major

Replace

Minor

Walls &
Ceilings:
Surfaces (not
kitchen & bath)

Cost

1995

Minor

Replace

Floor Sub-base
(not kitchen &
bath)

Action

Unit of
Cost
Measure

Replace

N/A

Abt Associates Inc.

Replace & paint surface area.

86.33*

Major action
dropped in ‘95.

1.58

Replace action
added in ‘95.

Supplement 4

S62

System By System Comparison
Booklet: UNIT
19901

System

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Difference
Cost

Change

% Change

Floor Covering

Carpet (not
kitchen & bath)

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Resilient (not
kitchen & bath)

2.70

s.f.

Carpet worn, stained; replace; %

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Minor

Interior Doors &
Frames

Replace indicated percentage of all
floor coverings.

Moderate
Major
Replace

Replace indicated percentage of all
floor coverings.

1.80

s.f.

N/A
Replace hardware, rehang door.

Holes & cracks; replace; %

1.65

-38.82*

2.43

35.16*

50.00

19.19

N/A
41.95

door

Replace hardware; rehang door; #

Replace door.

239.72

Replace door; #

256.00

6.79

Replace door and frame.

383.55

Replace door & frame; #

400.00

4.29

Abt Associates Inc.

Supplement 4

S63

System By System Comparison
Booklet: UNIT
19901

System

Kitchen Walls
& Ceilings:
Partitions &
Surfaces

Level

Action

Minor

Surface material needs to be restored
with minimal prep work; paint.

0.78

Moderate

Major prep work required for surface
material restoration; paint.

Major

Cost

Unit of
Cost
Measure
s.f.

1995
Type

Action

Difference
Cost
0.70

-10.15

1.08

Patch & paint.

1.25

15.88

Major prep work required for surface
material restoration including tile
replacement; paint.

2.34

N/A

Replace wall/ceiling material and
paint. Indicate percentage to be
replaced.

4.55

Replace & paint surface & framing
system; %

Minor

N/A

Moderate

N/A

Major

Replace

% Change

Paint.

Major action
dropped in ‘95.

3.00

Replace

Kitchen Floor
Covering &
Sub-base

Change

Replace sheetgoods.

1.92

Replace floor covering and sub-floor.

3.12

Abt Associates Inc.

s.f.

Replace sheetgoods.

3.30

Replace floor covering & sub-base.

6.65

Cost of major action
added to replace
cost in ‘90; cost of
moderate action
added to replace
cost in ‘95. Percent
change in
parenthesis
represents
aggregate costs.

-34.13*

(-38.32*)

72.08*
‘90 costs do not
include sub-base;
‘95 costs do.

Supplement 4

113.39*

S64

System By System Comparison
Booklet: UNIT
19901

System

Level

Cost

1995
Type

Action

Difference
Cost

Change

% Change

Replace countertop, keep sink;
replace fittings (faucet).

719.16

Replace countertop & sink faucet.

732.00

1.79

Refinish existing cabinets; repair
drawers, replace door hinges; replace
fittings (faucet).

839.02

Refinish cabinets, repair drawers &
door hinges, replace sink faucet.

800.00

-4.65

Moderate

1678.04

“MINOR” + “MOD” (repair/replace
countertop, cabinets, sink).

1532.00

-8.70

Major

Refinish existing cabinets; repair
drawers, replace door hinges; replace
fittings (faucet); replace countertop
and sink.
Replace countertop and backsplash;
replace sink and fittings; remove and
replace cabinet system.

2636.92

Replace countertop, cabinets, sink.

2500.00

-5.19

Replace

Minor

Cabinets /
Counter Top /
Sink

Action

Unit of
Cost
Measure

N/A

Burner non-functional.

Moderate

N/A

Major

N/A

Replace range.

Replace

N/A

50.00

Minor

System added in
‘95. Ranges without
Hoods were
captured under
Ranges and Hoods
in ‘90.

Range

Replace burner, clean hood, perform
minor repairs.

149.83

Burner non-functional.

100.00

-33.26*

Moderate

Replace hood.

239.72

Replace hood.

258.00

7.63

Major

Replace range.

539.37

Replace range.

500.00

-7.30

Replace range & hood.

758.00

-35.14*

Minor

Range & Hood

500.00

Replace

Replace range and hood.

Abt Associates Inc.

1168.64

Supplement 4

S65

System By System Comparison
Booklet: UNIT
19901

System

Refrigerator

Level

Action
N/A

Moderate

N/A

Major

N/A

Replace refrigerator.

539.37

Nonfunctional / >useful life (15 years;
10 years / family).

Minor

N/A

Moderate

N/A

Major

N/A

Replace garbage disposal.

239.72

Nonfunctional / > useful life (7 years).

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Microwave

Type

N/A

Replace

Dishwasher

Cost

1995

Minor

Replace

Garbage
Disposal

Action

Unit of
Cost
Measure

Replace dishwasher.

539.37

Nonfunctional / > useful life (15
years).

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace microwave.

Abt Associates Inc.

299.65

Nonfunctional / > useful life (10
years).

Difference
Cost

768.00

Change

Costs were
underestimated in
‘90.

% Change

42.39*

180.00

-24.91

522.00

-3.22

275.00

-8.23

Supplement 4

S66

System By System Comparison
Booklet: UNIT
19901

System

Trash
Compactor

Level

Type

Action

N/A

Moderate

N/A

Major

N/A

Replace compactor.

509.41

Surface material needs to be restored
with minimal prep work; paint.

0.78

Moderate

Major prep work required for surface
material restoration; paint.

Major

Minor

Replace

Bathroom
Floor Cover &
Sub-base

Cost

1995

Minor

Replace

Bathroom Walls
& Ceilings:
Partitions &
Surfaces

Action

Unit of
Cost
Measure

Nonfunctional / > useful life (15
years).

Change

516.00

% Change

1.29

0.58

1.08

Patch & paint surface.

4.20

Costs were underestimated in ‘90.

289.34*

Major prep work required for surface
material restoration including tile
replacement; paint.

4.61

Tile replacement and Paint.

8.80

Costs were underestimated in ‘90.

90.70*

Replace wall/ceiling material and
paint. Indicate percentage to be
replaced.

6.89

Replace & paint surface & framing
system; %.

Cost of major action
added to replace
cost in ‘90; but not
in ‘95. Percent
change in
parenthesis
represents
aggregate costs.

74.12*

9.24
3.30

‘90 costs do not
include demolition;
‘95 costs do.

18.60
83.55*

12.59
6.65

‘90 costs do not
include demolition;
‘95 costs do.

31.30*
121.93*

N/A

Moderate

N/A

Replace

Cost

Paint.

s.f.

Minor

Major

Difference

Replace floor covering.

7.79
1.80

Replace floor covering and sub-floor.

9.59
3.00

Abt Associates Inc.

s.f.

Tile
Resilient

Replace sheetgoods.

Buckling, warped, splintered; replace
floor covering & sub-base.

12.00

-25.55*

Supplement 4

( 4.35)

S67

System By System Comparison
Booklet: UNIT
19901

System

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Difference
Cost

Change

% Change

Bathroom Fixtures
Replace sink.

359.58

Repair/replace fittings.

150.00

Minor

Sink
Moderate
Major
Replace
Minor
Moderate

N/A
Replace sink and toilet.
Replace all fixtures (sink, toilet, tub).

Fixtures treated as
one system in ‘90;
minor (repair) action
added in ‘95.
Percent change
cannot be
calculated.

N/A
719.86
1558.18

N/A

N/A
Replace sink.

393.00

Repair/replace fittings.

150.00

Replace toilet

359.58

N/A

Replace toilet and sink.

719.86

N/A

Toilet
Major
Replace

Replace all fixtures (sink, toilet, tub).

1558.18

Replace toilet.

361.00
200.00

Minor

N/A

Repair/replace fittings.

Moderate

N/A

Tub/Shower
Major
Replace
Minor

Moderate

Bathroom
Accessories

Major

Replace

Replace tub.
Replace all fixtures (sink, toilet, tub).

719.86
1558.18

N/A
Replace 2-3 broken or missing
accessories.

Abt Associates Inc.

Replace tub/shower.

821.00

N/A
89.90

N/A
Replace all accessories.

N/A

179.79

Replace 2-3 accessories (not
medicine cabinet).

100.00

Replace medicine cabinet only.

160.00

Replace accessories & medicine
cabinet.

230.00

11.24

Major action added
in ‘95.

Supplement 4

27.93*

S68

System By System Comparison
Booklet: UNIT
19901

System

Bathroom
Vanities

Level

Action

Cost

Unit of
Cost
Measure

1995
Type

Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Minor

Moderate

Replace vanity.

329.62
389.55

24"
36"

N/A
Replace the indicated component.

Replace vanity.

Difference
Cost

Change

% Change

387.00
552.00

17.41
41.70*

400.00
700.00

-7.30
-39.17*

972.00
5370.00

-15.53
-6.66

N/A
Heat only
Heat/cool

431.50
1150.66

Replace 1 component (fan, controls,
coil, cabinet).

HVAC Unit2
Major

Replace

N/A
Replace entire system.

N/A
1150.66
5753.28

Replace 2 or more components.

Radiation
Minor

Moderate

N/A
Replace or repair the indicated
component.

N/A
7.19

l.f.

Replace component (cover, valves,
traps).

9.72

35.16*

19.44

35.16*

10.54

-45.04*

Hydronic
Major

Replace

N/A
Replace entire system.

N/A
14.38

Radiator leaks, damaged
components; replace system.

Minor

N/A

Moderate

N/A

Major

N/A

Electric

Replace

Replace entire system.

Abt Associates Inc.

19.18

Nonfunctional; replace.

Supplement 4

S69

System By System Comparison
Booklet: UNIT
19901

System

Level
Minor

Moderate

Action

Cost

N/A
Major overhaul or replacement of a
major component.

Replace

958.88

Minor

Moderate

Unit Furnace
Major
Replace
Minor

Unit Domestic
Hot Water
Generation

Moderate

Major

Replace

Temperature
Controls

N/A
Replace entire system.

2397.20

Action

Replace 1 component (burner boiler,
cast-iron sections, combustion
chamber).

Replace system.

599.30

Repair/replace component (tune-up
burner, adjust controls, lube fan).

1438.32

Replace system.
N/A

N/A

Replace component (temp. control,
valves).

N/A

N/A
419.51

Nonfunctional / > useful life (20
years).

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace temperature control.

Abt Associates Inc.

Change

% Change

800.00

-16.57

2730.00

13.88

500.00

-16.57

1110.00

-22.83

N/A

Replace entire system.

Cost

N/A

N/A
Replace entire system.

Type

Difference

N/A

N/A
Tune-up burner, adjust controls,
patch air leaks, secure loose mounts,
clean humidifier, lube fan, clean flue,
adjust draft controls.

1995

N/A

Unit Boiler
Major

Unit of
Cost
Measure

119.86

Nonfunctional or missing; replace.

150.00

Moderate action
added in ‘95.

450.00

7.27

64.80

-45.94*

Supplement 4

S70

System By System Comparison
Booklet: UNIT
19901

System

Wall / Window
Air Conditioner

Level

Moderate

N/A

Major

N/A

Replace entire system.

839.02

Nonfunctional; replace.

Minor

N/A

Moderate

N/A

Major

N/A

Replace entire panel.

1078.74

Panel burned-out; replace.

Minor

N/A

Moderate

N/A

Major

N/A

Replace branch wiring, outlets and
fixtures.

3.60

s.f.

Deteriorated wiring, switches, outlets.

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Closed Circuit
TV

Action
N/A

Replace

Bell/Intercom
System

Type

N/A

Replace

Unit Electrical
Wiring

Cost

1995

Minor

Replace

Unit Electrical
Panel

Action

Unit of
Cost
Measure

Replace unit level signaling
component.

179.79

Nonfunctional; replace.

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace unit level component.

Abt Associates Inc.

119.86

Nonfunctional; replace.

Difference
Cost

Change

% Change

750.00

-10.61

1260.00

16.80

3.50

-2.66

182.00

1.23

100.00

-16.57

Supplement 4

S71

System By System Comparison
Booklet: UNIT
19901

System

Emergency Call
Alarm System

Level

Cost

1995
Type

Action

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Smoke / Fire
Detection

Action

Unit of
Cost
Measure

Replace unit level signaling
component.

149.83

Nonfunctional; replace.

Minor

N/A

Moderate

N/A

Major

N/A

Replace

Replace detector(s).

Abt Associates Inc.

119.86
149.83

Detector

Batttery
Hard Wire

Nonfunctional; replace.

Difference
Cost

Change

125.00

100.00
153.00

% Change

-16.57

‘95 form allows for
both types; ‘90 form
allowed only 1 type.

Supplement 4

-16.57
2.12

S72

File Type	application/pdf
File Title	C:WORKROBINSONAEE02E4E.PDF
Author	Unknown
File Modified	1998-03-24
File Created	0000-00-00