Indexing the White Collar Salary Test: A Look at the DOL’s Proposal
Aug 27, 2015
In its Notice of Proposed Rule Making (NPRM), the Department of Labor (DOL) proposes to increase the current level for the salary test and for that threshold to be automatically increased each year. The initial salary threshold level was the focus of our previous article (available here). This article examines the two options proposed by the DOL for automatic annual increases, which are (1) to make adjustments using the 40th percentile standard and (2) to index salary tests to the Consumer Price Index. Both proposals raise concerns: the first could cause the salary test threshold to increase by at least an additional 50 percent over the next five years, substantially increasing the number of white collar jobs that are currently exempt but will have to be re-classified to non-exempt. The second proposal raises questions about whether a formulaic adjustment to the salary test is appropriate because it limits the DOL’s ability to adapt to changing future economic conditions. Both proposals will likely result in consistent annual increases in the fraction of salaried white collar workers who fail the salary test, lose their FLSA exemption, and require re-classification to non-exempt status.
Automatic Adjustments to the Salary Test
The salary test for the white collar exemption to the FLSA was last adjusted in 2004 and prior to that was last adjusted in 1975. When the salary test remains unchanged in nominal dollars and general inflation occurs, fewer and fewer employees fail the salary test and eventually the salary test becomes irrelevant for determining exempt status. The purpose of automatic adjustments to the salary test is to neutralize the impact of inflation so that workers with similar duties and roughly equal real earnings are treated comparably in different years.
The DOL’s proposal in the NPRM to use a formula to automatically adjust the salary test for the white collar exemption is an unprecedented change in the history of the FLSA. The DOL has considered and rejected automatic adjustments to the salary test in previous reports as well as in the 2004 Final Rule, when it was again decided that the minimum salary threshold should be determined by an examination of available data on actual salary levels currently being paid. The DOL stated in 2004 that its decision to set a salary threshold after examining available salary data “reflects the Department’s long-standing tradition of avoiding the use of inflation indicators for automatic adjustments to these salary requirements.” The DOL also explained that “a mechanical adjustment for inflation could have an inflationary impact or cause job losses. We are particularly concerned about, and required to consider, the impact that an inflation adjustment could have on lower-wage sectors such as businesses in rural areas, businesses in the retail and restaurant industry, and small businesses.” The DOL’s rejection of automatic adjustments to the salary test date back to the 1949 Weiss Report which considered and rejected proposals to increase salary levels based upon the change in the cost of living from 1940.
Proposal 1: Adjustments using the 40th Percentile Standard
The first DOL proposal for automatically adjusting the salary test is to set the salary threshold as the 40th percentile of the pay distribution among workers in the Current Population Survey who respond that they are not paid hourly each year. It should be noted that the Current Population Survey asks workers whether or not they are paid hourly, but does not inquire as to whether a nonhourly worker is paid a salary, earns commissions, is paid by piece rates, or is compensated in another way.
While this proposal examines available data on actual earnings, the mechanical approach to re-setting the salary test at the 40th percentile every year is flawed for several reasons. First, and most importantly, the sample of workers who respond that they are not paid hourly will change as workers who fail the salary test are re-classified to non-exempt. For example, if the salary test is set at $970 per week in 2016 (as projected in the NPRM), all nonhourly white collar workers earning less than this threshold and not eligible for any other exemptions must now be classified as non-exempt. It is expected that many of these workers will be converted from nonhourly to hourly positions (although some may remain salaried non-exempt employees).
The top panel of the chart illustrates the pay distribution among nonhourly workers as of 2015. The 40th percentile of this distribution is $950 per week. If just one quarter of the full-time nonhourly workers earning less than $49,400 per year ($950 per week) were re-classified as hourly workers, the pay distribution among the remaining nonhourly workers would shift so that the 40th percentile of the 2016 pay distribution would be $54,184 ($1,042 per week), about 9.6 percent higher than it was in 2015. This process will continue each year as the lowest paid nonhourly workers fail the salary test and many are re-classified as non-exempt hourly workers. The bottom panel of the chart represents the pay distribution after five years of salary tests with one quarter of the nonhourly workers who failed the salary test transitioning to hourly pay each year. This graphic illustrates that, even in the absence of inflation, in five years the new 40th percentile of the nonhourly pay distribution would be $72,436 ($1,393 per week), which is about 46.6 percent more than the minimum salary threshold in 2015.
The ratcheting up of the salary test threshold becomes more pronounced if more nonhourly workers who failed the salary test are re-classified into hourly positions each year. For example, if half of the nonhourly workers below the salary test threshold are reclassified as hourly each year, the 40th percentile threshold will increase by 19.9 percent in the first year and by 94 percent over a five year period. This means that a salary threshold of $49,400 ($950 per week) in 2015 would increase to $95,836 ($1,843 per week) by 2020, even in the absence of inflation.
To put this into perspective, in 2015 about 50 percent of first line supervisors of office and administrative support workers, an occupation title that the DOL considers to have “a high probability of exemption” (in excess of 90 percent), would fail the salary test of $950 per week. If the minimum salary threshold is adjusted to the 40th percentile each year, the 94 percent increase in the salary test to $1,843 per week would be large enough that 90 percent of first line supervisors of office and administrative support workers would fail the salary test by 2020.
While the anticipated large increase in the salary test threshold from one year to the next is the primary flaw with this proposal, there are other problems with setting the salary test at the 40th percentile of the nonhourly pay distribution each year. Annual adjustments are intended to mitigate the impact of inflation, not to change the salary test as the composition of the workforce changes. A worker at the 40th percentile of the educational attainment distribution has attended some college but does not have a four-year degree. As educational attainment increases and more students receive associates degrees or degrees from two-year colleges, the worker at the 40th percentile of the salary distribution will become more educated, work in a different type of job, and perform different tasks. Consequently, the characteristics of a worker at the 40th percentile of the nonhourly pay distribution will change over time, and some of the anticipated increase in pay at the 40th percentile will be due to an increasingly educated workforce and not inflation.
Proposal 2: Adjustments using the Consumer Price Index
The second DOL proposal for automatic adjustments to the salary test is to index the test to the Consumer Price Index (CPI). The risk of indexing a salary test to the CPI, which is constructed to measure inflation in commodity prices, is that commodity price inflation and wage inflation may diverge. For example, the CPI increased by 25.3 percent from 2004 to 2014, while the median full-time salary of high school graduates increased by 16.4 percent, with corresponding median pay increases for adults with some college of 15.1 percent and college graduates of 20.2 percent over the same period. Because there has been low overall inflation over the past ten years, these differences are modest; however if Federal Reserve policy were to change, or the dollar were to weaken relative to foreign currencies, commodity price inflation could return and year-to-year changes in commodity prices may not track closely with year-to-year changes in wages and salaries. By adopting an automatic adjustment tied to the percentage change in commodity prices, the DOL proposal introduces the possibility that the adjustments to the salary test either understate or overstate the change in average wages and salaries.
There are, in fact, strong economic arguments for why the CPI will overstate inflation in wages. First, the CPI overstates changes in the cost-of-living because it suffers from “substitution bias.” The CPI assumes that all consumers will buy the same shopping basket of goods, regardless of changes in relative prices. Smart, rational consumers will economize on commodities whose relative prices have gone up and buy more of relatively lower priced goods. In this sense, the CPI consistently overstates the changes in the “true” cost-of-living faced by workers.
In addition, the BLS tracks different components of employee compensation in the Employment Cost Index and has found that the cost of benefits have increased 48 percent more than wages and salaries since 2001. Consequently, even if total compensation increased in proportion to the CPI, it could well be due to more rapidly increasing expenditures on fringe benefits such that wages/salaries are increasing, but by less than the overall cost-of-living. Therefore a salary test for the FLSA exemption that assumes that wages and salaries will increase in lockstep with the CPI is likely (based on historical trends) to cause the test threshold to increase too quickly and cause many white collar workers to lose their FLSA exemption.
The NPRM includes two alternatives for automatic adjustments to the salary test. One proposal would likely result in steep increases in the salary test threshold causing millions of white collar workers to be converted to non-exempt hourly positions. The other, a proposal to index the salary test to the CPI, might appear benign after a decade of low inflation, but would not be in a more inflationary environment. The DOL’s own arguments for not including automatic adjustments still hold. Even if it is determined that the benefits of automatic adjustments outweigh the concerns, careful thought must be given to the process by which the adjustment amount will be determined.
 Page 22167 of “Defining and Delimiting the Exemptions for Executive, Administrative, Professional, Outside Sales and Computer Employees; Final Rule; Preamble“, 29 CFR Part 541, Federal Register, Vol. 69, No. 79, Friday April 23, 2004.
 Ibid, page 22168.
 The $970 per week in NPRM is the projected threshold for 2016, but in data from the first two quarters of 2015 (the most recent data available), the 40th percentile of the pay distribution is about $950.
The DOL’s economic analysis does not predict how many employees who fail the salary test will remain salaried but non-exempt and how many will be converted to hourly employees. We assume that one quarter of salaried employees will be converted to hourly employees each year to be conservative and because some nonhourly employees are exempt due to the outside sales exemption or other FLSA exemptions (e.g. physicians, lawyers, or teachers).
 Between 2004 and 2014 the fraction of full-time employees with a high school diploma or less declined from 39.5 percent to 33.5 percent while the fraction with a college degree or more increased from 33.0 percent to 39.3 percent.
 These pay changes include workers paid salaries and those paid by the hour.