Nonsub Change (sample reduction)

Quarterly Financial Report (QFR) Nonsubstantive Change – Sample Improvement

The Quarterly Financial Report (QFR) plans to implement two changes to the sampling methodology.

Sampling Changes

1. Certainty Sample Improvement

The QFR sample increases every year. From 2010Q4 to 2015Q4, the total number of active companies in the sample increased by 2,033. Of those 2,033, non-certainties increased by 1,472 and certainties increased by 561. The overall sample size was 12,200 by 2018Q1. To limit the continual growth of the QFR sample, QFR plans to implement a new sampling methodology. Table 1 presents the current sample strata.

Table 1 – Current QFR Sampling Strata

Stratum	Sampling Frame Assets
03	$250,000 - $999,999
07	$1,000,000 - $4,999,999
08	$5,000,000 - $9,999,999
14	$10,000,000 - $49,999,999
16	$50,000,000 - $249,999,999
18*	$250,000,000 and over

* Certainty companies

Note: Manufacturing is the only industry that QFR samples below stratum 16. Mining, Wholesale Trade, Retail Trade, and Selected Service Industries only sample stratum 16 and stratum 18.

The new sampling methodology introduces noncertainty stratum 17 with assets between $250,000,000 and an amount that varies by industry. The methodology will also raise the asset cutoff for certainties. The certainty asset cutoff will be monitored and could be adjusted more easily in the future.

2. Removal of Small Strata

QFR also plans to remove the two smallest strata, stratum 03 and stratum 07. The average sample size for stratum 03 is 1,035 and the average sample size for stratum 07 is 1,368 (average of 6 quarters from 2016Q4 to 2018Q1). Removing both strata 03 and 07 reduces the overall QFR sample size by an average of 2,403 companies. In addition, removing both strata 03 and 07 could increase the total unit response rate (URR) by approximately 3.0% on average. According to a recent Nonresponse Bias Study, the current URR is approximately 50% for stratum 03 and is approximately 55% for stratum 07.

Change in Burden Hour estimate

The current total burden hour estimate for QFR is 107,462 hours. Implementing the Certainty Sample Improvement changes and removing stratum 03 and 07 reduces the burden hour estimate by 26,676 hours. This results in a new total burden hour estimate of 80,786 hours. Table 2 below shows the new burden.

Table 2 – QFR new Burden Hour estimate

	QFR Forms
	200-MT (Long form)	300-S (Long form)	201-MG (Short form)	Total
Number of respondents	4,374	1,233	2,813	8,420
Number of responses per respondent/year	4	4	4	4
Total annual response (line 1 times line 2)	17,496	4,932	11,252	33,680
Hours per response	3.0000	3.0000	1.2000	2.39865
Total hours (line 3 times line 4)	52,488	14,796	13,502	80,786

Summary

The QFR plans to raise the certainty asset cutoff to limit the increasing sample size. In addition, QFR will remove the two smallest strata, 03 and 07, which means the survey will sample companies with assets of $5,000,000 and over instead of $250,000 and over. These changes are necessary because while the economy has continued to grow, the asset and receipt levels used to determine stratum classification have remained fixed. In addition, budget constraints have made the increasing number of companies difficult for staff to review. Implementing the Certainty Sample Improvement will also reduce respondent burden. Removing the smaller strata will reduce respondent burden as well and could improve the overall Unit Response Rate (URR) for QFR.

High Level Milestones

June 2019: Publish an announcement about the new Sampling Methodology with the 2019Q1 release.

December 2019: First production mailout with the new Sampling Methodology will occur with the 2019Q4 mailout.

March 2020: First publication of data with the new Sampling Methodology in the 2019Q4 release. Re-release 2018q4 to 2019q3 estimates using the new methodology.