Tuesday, July 4, 2017

UW Evans School Study of Seattle's Minimum Wage Looks Like an Outlier

Two new studies of Seattle’s minimum wage have recently been published. One was done by a UW Evans School team and finds that, over the study period, wages went up an average of 3%, while total labor hours decreased by 9%.  This amounts to an elasticity of -3.o, which simply means that a 1% increase in wages results in a 3% decrease in employment hours.

By contrast, a UC Berkeley study finds that raising "minimum wages in Seattle up to $13 per hour raised wages for low-paid workers without causing disemployment. Each ten percent minimum wage increase in Seattle raised pay by nearly one percent in food services overall and by 2.3 percent in limited-service restaurants. The pay increase in full-serve restaurants was much smaller and not statistically significant, consistent in part with higher pay in full-service restaurants and the establishment of a tip credit policy. Employment effects in food services, in restaurants, in limited-service restaurants and in full-service restaurants were not statistically distinguishable from zero. These results are all consistentwith previous studies that credibly examine the causal effects of minimum wages.

In addition to these local studies, there's a new meta-study focusing on 46 events with the largest wage changes, which found that average wages of the affected earners increased significantly by 10.8%, but that employment was little changed with a statistically insignificant increase of 0.2%. This study estimates 0.02 as the implied elasticity of employment with respect to wages, with 0.307 as its standard error, which rules out elasticities smaller than -0.585 at 95 percent confidence level. Not to put too fine a point on it, and ignoring the positive sign of employment, the meta-study's estimated elasticity of 0.02 is 1/15 the size of the UW study's elasticity of 3! Something seems to have gone haywire.

Commenting on the UW study, the authors of the meta-study contend that the idea that raising the minimum wage has a much larger effect on hours than on wages strains credulity, especially since, as economists Ben Zipperer and John Schmitt have noted, Seattle’s increase “is within the range of increases that other research has found to have had little to no effect on employment.” It’s not entirely clear why the University of Washington team gets such a weird result — since their data isn’t public, we can’t check it — but it’s worth noting at least two important issues with their study.

First, the UW data exclude workers at businesses that have more than one location; in other words, while workers at a standalone mom-and-pop restaurant show up in their results, workers at Starbucks and McDonald’s don’t. Almost 40 percent of workers in Washington state work at multi-location businesses, and since Seattle’s minimum wage increase has been larger at large businesses than at small ones — right now, a worker at a company with more than 500 employees is guaranteed $13.50 an hour, while a worker at a company with fewer than 500 employees is guaranteed only $11 an hour — these workers’ exclusion from the study’s results is an especially germane problem (note that low-wage workers in Seattle have had an incentive to switch from small firms to large firms since the minimum wage started rising). In earlier work, in fact, the University of Washington team’s results were different depending on whether these workers were included in their analysis; including them made the effects of the minimum wage look more positive.
Second, the University of Washington team does not present enough data for us to assess the validity of its “synthetic control” in Washington — that is, the set of areas to which they compare the results they observe in Seattle. The Seattle labor market is not necessarily comparable to other labor markets in the state, and given some of the researchers’ implausible results, it’s hard to believe the comparison group they chose is an appropriate one.

Discussion.  Let’s consider some examples of how firms and employees might respond to increases in the legal minimum wage.

1. The Perfect Competition case. Suppose workers supply services worth  $13/hour.  If a new Miniumum Wage were to exceed $13/hour, these workers would be offered fewer hours of work. This reduced demand for labor at higher wages is, in effect, the classical mechanism (the price effect) that's at work behind the scenes in the UW study, which found that a 3% average wage increase went hand-in-hand with a 9% reduction in total labor hours.

Yet, even if these UW percentages were correct, the classical interpretation of the results doesn’t necessarily follow.  It could be that, for some low-wage employees, the greater income afforded by the 3% wage increase (the income effect) allowed these employees to take classes, travel, etc., which shows up as a reduction in total work hours.

Now, there’s nothing in this "trade-off" between higher wages and few hours worked that justifies a claim made about the UW study, which is that the “cost” of a 9% reduction in paid hours “isn’t worth the “benefit” of a 3% increase in pay per hour of work.

It is true, however, that, all else equal, a reduction in total labor income, due to fewer labor hours (for whatever reasons), will have some secondary multiplier effects, in particular less aggregate demand, less sales revenue, and lower demand for labor.

2.  The Monopsony Case.  Suppose there’s a group of workers who produce $13.50/hour in value, but are only paid $10/hour.  Let's assume the employer knows she's getting this extra $3.50/hour in "exploitation profits." Let's also assume the employee knows this, but isn't interested in looking for another job.

·   In this case, raising the minimum wage to $13/hour shifts $3/hour from employer to worker. It’s still rational for the owner to keep the employee on payroll. If we assume the worker decides to remain at the now $13/hour job, then, all else the same, wages paid go up, while profits fall a bit (but not into negative territory).

·   Insofar as low-wage workers spend a larger proportion of their marginal income than business owners typically do, the shift of $3/hour from profits to wages will increase aggregate demand and sales revenue, pushing up demand for labor.

·   This effect could be important if large low-wage employers have bargaining advantages over individual workers.

3. The wage bargain as an implicit contract. Some labor economists believe that, at higher wages, workers work harder and smarter, and that there’s less labor turnover, which affords benefits to workers and firms. Depending on the size of minimum wage increase and the other factors involved, it's conceivable that more workers could be hired at higher wages. If so, wage income rises, profits may rise, sales revenue will increase, and the demand for labor will increase.

No comments:

Post a Comment