The Montgomery County minimum wage impact study is absurd junk science
In January, Montgomery County, Maryland County Executive Isiah Leggett vetoed an ordinance passed by the county council that would match the minimum wage in the District of Columbia, raising the county minimum to $15 by 2020. Leggett then commissioned the consulting firm PFM to analyze the likely economic effects. The firm just released their study and their findings are so implausible that they border on the absurd. The study essentially concludes that raising the minimum wage in Montgomery County—even a small amount—would be the most devastating economic shock the county has experienced in a generation, more damaging than the Great Recession. To say that the study has methodological problems would be a gross understatement. No county official, business owner, worker, or resident in Montgomery County—and certainly not editorial boards of local newspapers—should give any credence to this report.
The report posits that the proposed $3.50 minimum wage hike over 5 years will lead to massive losses in jobs, income, and county revenues. Ostensibly wanting to present both the costs and benefits, the authors do also note that “increased wages are associated with improved mental health, reduced hunger, and decreased stress for workers and their families.” Admittedly, I have only skimmed the full 145 page report, but one only needs to read the initial section on job impacts to see how flawed this “study” is. The alleged large negative outcomes for incomes and county revenues all stem from the jobs findings, so there really isn’t need to read much further.
The report’s methodology for how they calculate expected impacts on employment is completely divorced from any actual research. First, the authors go through a long discussion of other localities that have enacted higher minimum wages—such as the District of Columbia, Los Angeles, and San Jose, among others— which they refer to as “comparison jurisdictions,” implying that the impacts of minimum wage hikes in these locations might provide guidance for how a higher minimum wages might affect Montgomery County. Ironically, they note that in virtually all these “comparison jurisdictions,” studies that analyzed the resulting or likely employment effects of the local minimum wage showed that any impact on jobs was negligible. Nevertheless, the authors assert that Montgomery County is not a “twin” of any of these places, thus none of these chosen comparisons should serve as a guide.
The report then summarizes a “literature review” provided to the authors by economist David Neumark of UC- Irvine. For those not steeped in minimum wage research, Neumark is probably the most prominent critic of the minimum wage among all academic researchers. His research is in no way representative of a consensus among economists. In fact, the bulk of recent research has found that minimum wage increases have had little effect on employment. The study group can certainly consider Neumark’s input, but for him to be their only source of background research should raise a red flag.
Yet even with Neumark’s highly pessimistic take on the effects of the minimum wage in hand, the PFM report sets aside Neumark’s findings and instead makes use of their own dubious survey results to produce their entire analysis. They base their study on a survey of business owners in the county, in which they asked these business owners what percentage of their workforce they would eliminate if the minimum wage were raised to various levels. To put it bluntly, this is junk research. The closest analogy I can think of would be if the FDA was considering approval of a new drug and instead of reviewing any studies or trials, they instead simply asked the drug company “what percentage of patients do you think this drug will harm versus help?”
The PFM analysts then take the “usable” responses from their survey and “weight” their responses based upon the number of minimum wage workers they reported employing. It’s not clear what the “unusable” responses were; however, other sections of the report suggest that the “unusable” responses were from businesses that either don’t employ low-wage workers or said they would not eliminate any jobs in response to the increase. In other words, it appears that they throw out all the responses where the employer said it won’t affect jobs at their business, then among everyone who said they would reduce jobs, they give disproportionate weight to the entirely speculative and subjective opinions of employers who report having the most low-wage workers (i.e., those who would likely have the largest increases in labor costs resulting from a minimum wage increase.) Even for basic survey analysis, this is grossly improper: not only have they seemingly tossed out responses that would have made any average effect less negative, they have intentionally magnified the responses of those respondents who would, by their nature as more intensive users of low-wage labor, necessarily predict the largest negative impacts.
[A brief aside for anyone who actually reads the study: the way that they frame the whole “anticipated job loss” survey with business owners also seems highly misleading. Table 16 seems to indicate that they asked business owners to quantify the expected percentage of their workforce they would eliminate from the proposed minimum wage increases, starting with the increase from $10.75 to $11.50 scheduled to occur in July 2017. Yet they describe this as a 20.4 percent increase, using the 2015 county minimum of $9.55 as their starting point. In other words, they ask business owners to describe their anticipated staff reductions resulting from a future minimum wage increase, yet in presenting the magnitude of the increase, they include past increases which has already occurred and to which businesses have already adjusted. This makes no sense. Moreover, it’s not clear why they should be analyzing the July 2017 increase at all—it’s settled law. If PFM is supposed to be analyzing legislation that would enact new future increases, they should be making comparisons to what the minimum wage is under current law ($11.50) or at the very least what it was at the time of the survey ($10.75). There is no reason to compare the proposed minimum wage to what it was in 2015 unless you’re intentionally trying to make the proposed increase look larger than it is.]
Using the weighted survey responses from employers who said they would cut jobs, the authors produce absolutely ludicrous “job loss elasticities” that they then appear to use improperly. When using elasticities, economists will typically describe them as the change in one thing resulting from a 10 percent change in something else. Indeed, the report employs this same construction in the first paragraph on page 44:
“The weighted results show that a 20.4 cumulative percentage increase in the minimum wage rate is expected to yield a 5.0 percent loss in minimum wage jobs, a 30.9 percent increase yields a 7.7 percent loss, etc. The elastic measure is the percentage of jobs lost divided by the percentage wage increase, so that 5.0 divided by 20.4 yields an elasticity estimate of -0.24 for 2017. Stated slightly differently, a 10 percent increase in the wage rate, given the pre-July 2017 rate of $10.75 per hour, would yield a loss of 2.4 percent of local minimum wage jobs.”
Setting aside the lunacy of how they’ve come up with these “elasticities,” if you actually apply them as they themselves explicitly state (“a 10 percent increase in the wage rate…would yield a loss of 2.4 percent of local minimum wage jobs”) you get dramatically smaller impacts than they claim. According to their elasticity, the 6.98 percent increase in the minimum wage that occurred on July 1, 2017 when the county minimum rose from $10.75 to $11.50 should have reduced low-wage employment in the county by 1.67 percent. But the authors don’t use these values as “elasticities” in the way any economist typically would. Instead, they take what they are calling “elasticities,” which are actually just the total percentage of low-wage jobs that employers reported they would be cut in each step according to their bizarrely weighted survey, and they reduce jobs by that full percentage.
According to the report, the Montgomery County minimum wage increase from $10.75 to $11.50, which took place last month, should have eliminated or will eliminate 18,318 jobs or 24.4 percent of all low-wage jobs in Montgomery County. To put that in context, the authors are claiming that as a result of the $0.75 increase in the minimum wage that took effect last month, Montgomery County will suffer larger job losses than the county experienced in the worst year of the Great Recession. (From 2008 to 2009, Montgomery County lost 14,500 jobs per the Quarterly Census of Employment and Wages.) That’s right, according to the PFM study, Montgomery County just last month experienced its most damaging economic crisis in a generation.
I won’t bother to dig into their equally outrageous estimates for future years. Any objective observer—especially those that sit on the county council—should look at the current state of the Montgomery County’s economy and ask whether the sky is failing. If the answer is no, they should be very skeptical of PFM’s claims. Researchers have produced plausible assessments of proposed $15 local minimum wage increases. This is not one of them.
(8/4/17) Update: It appears that the report’s methodology was even worse than I thought. Adam Pagnucco, a researcher and journalist in Maryland, received the PFM study’s official survey in May via an email from the Montgomery County government. Adam notes that the online survey had no collection of any respondent identifying information and zero access control. This means that anyone with the link could have responded, and there would be no way to verify whether they were an actual Montgomery County business owner, or as in Adam’s case, whether they had any employees. Moreover, there would be no way to stop someone from filling out the survey multiple times. The report’s claims about job impacts were bogus—now it seems even their survey of employer attitudes is not credible.
 The report states they had 307 “usable” responses (p42) from a survey of 520 business owners (p16). Of those 520, 97 percent were located in Montgomery County (p16), 16 percent are said to not employ any minimum wage workers (p16), and 27.3 percent reported that they would not reduce employment in response to the increases (p60). 520 x 0.97 x 0.84 x (1-0.273) = 308.