Six reasons not to put too much weight on the new study of Seattle’s minimum wage
Seattle’s minimum wage increases are one of the most important local policy developments in recent years, but the new study by University of Washington researchers Jardim et al (2018) is largely uninformative about the effects of the policy because it uses the same flawed methodology that economists criticized in connection with earlier studies by the group. But, even if you believe the results of the new study, a careful reading of its actual findings shows the minimum wage benefitted all of the city’s low-wage workers who had jobs prior to the increase.
1. The new study is based on a flawed comparison between Seattle and other areas in Washington state. The comparison causes the study to measure a reduction in the number of new jobs under $15/hour, when in fact this is not a cause for concern.
By comparing workers in Seattle with workers elsewhere in Washington state, the study incorrectly assumes that the low-wage labor market in Seattle would have grown like other areas in Washington, were it not for the city’s 2015-2016 minimum wage increases. This comparison is unreasonable because, as other researchers have demonstrated (Dube 2017, Rothstein and Schanzenbach 2017, Zipperer and Schmitt 2017), Seattle experienced much faster wage gains for reasons that had nothing to do with the minimum wage. Indeed, the authors of the new study find that Seattle had faster wage growth and diverged from other regions prior to the city’s minimum wage increases (see their Table 8 for the 2012-2013 period).
The flawed comparison underlying the study causes it to mistakenly attribute negative employment changes to the minimum wage, when in reality Seattle’s economic boom simply meant that low-wage jobs were converted into higher-wage jobs. For example, the authors document a decline in newly employed workers earning less than $15/hour and argue that the minimum wage is causing “losses in employment opportunities.” Instead, as jobs in Seattle’s tightening labor market were upgraded from lower-wage to higher-wage jobs, there was a mechanical decline in the number of new entrants under any given low-wage threshold. The purported decline in new entrants is not a cause for concern. The fast wage growth in Seattle relative to comparison regions prevents the study from making credible claims about the consequences of the city’s minimum wage increases in 2015 and 2016.
2. The study’s analysis excludes newly entering workers in chains or other multi-site working establishments—these multi-site businesses are more than a third of low-wage employment in Washington.
The study uses detailed individual-level data on hours and earnings, but omits all newly employed individuals who work for an employer with multiple locations, like a chain, because the data cannot identify the location of any multi-establishment business. According to the authors, this exclusion is more than one-third of all jobs in Washington paying under $11 or $13 per hour. If larger, multi-site employers are more likely to be hiring low-wage workers after the minimum wage increase, this would cause the study to measure a reduction in the number of new low-wage entrants, even when no such reduction is taking place.
3. Although the study is not a reliable guide to the effects of the policy, if its findings are actually correct, all of the groups of incumbent workers analyzed by the authors were better off after the minimum wage increase.
According to the study, the average low-wage worker with a job before the minimum wage increase received significant gains in take-home pay as a result of the policy–about $500 per year in 2016. Even those who worked very low hours prior to the minimum wage change benefitted. Workers who were working very low hours (about eight hours a week on average) experienced a small reduction in their hours, but it was fully offset by hourly wage increases–leaving these workers with the same take-home pay from fewer hours of work. Meanwhile, the group of workers with more working hours prior to the minimum wage increase saw a jump up in take-home pay, exactly as intended by the policy.
4. The hours reduction the study attributes to the minimum wage is very small: according to the study’s findings, low-wage workers worked about 45 minutes less per week after the minimum wage increases.
The study finds that incumbent low-wage workers work fewer hours after the minimum wage increase, so much so that a group of workers with low hours during the baseline period did not see any change in total take-home pay. But the actual drop in hours estimated by the study is very small–about 45 minutes per week–the average estimated effect across all periods for those earning either under $11/hour or $13/hour prior to the minimum wage increases. This very small reduction in hours holds for those workers with both lower and higher baseline hours of work per quarter.
5. The study uses non-standard terminology about “experience” to describe those who worked different amounts of hours prior to the minimum wage increase.
What the authors call “low-experience” or “high-experience” workers are actually workers with lower or higher attachment to the workforce in the nine months before the minimum wage increases went into effect. These terms are not based on the usual definition of experience used by economists or the general public. The term experience as used in the University of Washington study does not refer to workers’ age, their total time in the labor market, or their total time with a given employer. It is more accurate to label these individuals as low-hours or high-hours workers, or having low or high attachment to the labor market.
It is also important to note that the population analyzed in the study worked very few hours per week. On average, the sample worked 18 hours per week, and 93 percent worked fewer than 40 hours per week. While a high share of part-time work is characteristic of low-wage populations, the very high share reported in the study calls into question the representativeness of the study’s data. National data from 2015, for example, shows about 74 percent of workers at or below the federal minimum wage worked fewer 40 hours per week.
As a consequence, when the study divides its sample into low-hours and high-hours groups, the low-hours group worked very few hours: about eight hours per week. Nationally, less than 5 percent of workers earning at or below the federal minimum wage work fewer than eight hours per week. Part of the reason behind the low hours worked in the study’s sample may be related to the exclusion from the sample of larger, multi-site businesses, where employees may tend to work more hours.
6. Seattle was already a relatively high-wage city before the recent minimum-wage increases. The increases enacted in Seattle in 2015 and 2016 were not particularly high relative to the pre-existing wage levels in the city. As a result, other high-quality research, based on many historical minimum wage increases, is applicable to the 2015-2016 Seattle minimum wage increases.
In 2016, at the end of the study period, the highest minimum wages in Seattle were $12.50-$13.00/hour for large employers and $10.50-$12.00/hour for small employers. Although minimum wage increases to this level were substantial by national standards, given Seattle’s already high wages, these minimum wages are well within the evidence base analyzed by high-quality research, which on balance finds large increases in take-home pay for low-wage workers, with little-to-no employment consequences.
Key relevant studies include: Rinz and Voorheis (2018), who use high-quality, administrative tax records and find that minimum wage increases led to large annual earnings increases for those at the bottom of the income distribution. Allegretto et al’s (2018) evaluation of six city-wide minimum wage increases, including Seattle’s, which concluded that the policies raised earnings in the low-wage restaurant sector without any detectable change in employment. And Cenzig et al’s (2018) analysis of 138 state-level minimum wage increases, including those as high as Seattle’s, which finds no significant evidence of reduced employment among low-wage workers, including among new entrants and incumbents. Indeed, all of these studies rely on multiple historical state- or city-wide minimum wage increases and are therefore less likely to be affected by the kinds of economic shocks such as Seattle’s rapidly tightening labor market, which can bias the results of a single case study.