UW minimum wage study: disconcerting, distorted and biased

July 10, 2017 | John Burbank

This past month, researchers at the Evans School of Public Policy at the University to Washington released a study funded by the Arnold Foundation[1] of the minimum wage increases in the city of Seattle.[2]  They asserted that the minimum wage increases resulted in losses of hourly work and total income for workers making less than $19 an hour, causing quite a stir in the national conversation about the minimum wage.  The study has been particularly heartening for organizations and people ideologically opposed to governmental interventions to address the disequilibria of income, power, and compensation between workers and employers.[3]

The study is most disconcerting because the authors seemed to have foregone academic due diligence in their eagerness to publish this new narrative.  Distortions, discrepancies and lack of objective contextual analysis undermine their data and the claims they promulgate.  In this brief overview, I will discuss several of these data omissions and biases.

  1. The lack of a comprehensive data base: This problem of gaining actual data undermines the assertions of the UW research team.  They claim to have a new comprehensive data base from which to develop actual hours and wages, but at the same time they omit all employees of multi-place employers.  In Washington State, this omits 38% of the waged workforce.[4]  The proportion of workers in multi-place employers is most likely higher in Seattle than outside of Seattle (think Starbucks).  The authors make no effort to determine what the proportion for the city actually is, instead relying on employer surveys to assert that this omission is not a confounding factor for their study.  Without including the data from multi-place employers, the study is automatically and significantly flawed.

1a. Conflating one minimum wage level with other minimum wage levels:  The authors repeatedly refer to the minimum wage levels of $11 and $13.  But those wages are mandated only for one sector of employers in Seattle, those employers with over 500 employees who do not provide health coverage.  The authors note that there are four different tiers for minimum wages, but they fail to segregate out the impacts of each of these tiers.  The authors do not even determine what numbers of workers are covered by each of the four tiers.  As a result they are asserting impacts from a wage increase which applies only to employees of firms with over 500 employees which also do not contribute to health benefits.[5]  There is a high probability that all such firms are also multi-place firms.  As the researchers have no data on employees of multi-place firms, they are hypothesizing the impact of a raise to $13 with a data set of employees whose minimum wage is less than $13.

Absent a delineation of minimum wage levels by employer status, the researchers are simply examining wage and hour statistics and hypothesizing causation to the highest level of minimum wage increases, which do not pertain to the data and statistics which they are utilizing.  Regardless of this research fault, the effects which the authors ascribe to the increase in the minimum wage could come from a shift in the industry structure in Seattle to information technology, the catering of businesses to this high tech workforce, or the construction boom for housing and corporate offices.

  1. No accounting for contractual work: Further, the researchers completely omit the impact of the growing number of workers participating in contractual work, such as Uber drivers.  By disregarding this sector, the researchers block off any possible analysis of the impact of the increase in the minimum wage on these workers and the movement of workers into and out of this employment sector.  The authors may have a difficult time estimating compensation data and hours worked, but rather than ignoring this sector, this omission should be acknowledged as indeed jeopardizing any findings based on significantly partial data and research.
  2. The $19 an hour border:  The study describes loss of jobs below an arbitrary line of $19 an hour, as well as job gains above that line that more than offset that loss.  This is described as a problematic consequence of the increase in the minimum wage.  But even if this data is accurate and confirmed in later rigorous analyses, this is a good thing.  The object of raising minimum wages is to eliminate jobs below that minimum, that is, to eliminate low-wage jobs.  As this moves more and more workers into higher-wage jobs, workers in totality do better, the economy does better, and the public sector is better funded[6] … a win-win-win.

3a. While the border of $19 an hour as a ceiling for low-wage estimates is discussed, it is not clearly defended.  That wage level implies some level of experience, responsibility, and education.  Those are the factors, at least $4 removed from the minimum wage, for movement up the wage ladder, independent of the dynamics of the minimum wage.  Full-time work at $19 comes out to almost $40,000 a year, more than the starting salaries of K-12 public school teachers.  It is quite likely that reduction in hours for jobs with wages under $19 may have been more influenced by movement of workers in $17- and $18-an-hour jobs into $19- and $20-an-hour jobs with increased responsibilities.  To imply that job loss at the $18-an-hour range is associated with the increase of the minimum wage to $15 an hour obscures other more relevant dynamics on the labor market.

  1. The wide spread of elasticities: The authors cite an elasticity estimate for disemployment of -3 created by the increase of the minimum wage from $11 to $13.  That has been used to causally tie minimum wage increases to loss of work hours.  But in reading the footnote to this claim, the authors state that “…these estimates are imprecise. For instance, the 95% confidence intervals for the elasticities associated with a $13 minimum wage range from -2.8 to 0.3.”  To be clear, that means that by the authors’ own methodology, we could also claim a small but positive effect on employment as a result of the minimum wage increase.

4a.  Synthetic Seattle:  The authors develop these elasticity estimates comparing Seattle to a set of localities which have not increased (at that time) minimum wages.  They look at 1) King County outside of Seattle, 2) the counties in the Puget Sound area outside of King County (Snohomish, Pierce, and Kitsap counties), and 3) Washington state outside of King County.[7]  None of these comparisons are close to mirroring the economic disparities, vitality, and performance of the City of Seattle.  These measures create a false sense of “objective” comparison.

5. Failure to consider the impact of wages on public benefits: In the assertion that low-wage workers have experienced a significant reduction of hours due to the minimum wage increase, the researchers fail to appreciate the interaction of public benefits from the viewpoint of actual minimum-wage workers.  This is a rather unpleasant surprise, as the UW team is composed of public policy academics who are trained to consider the interactions of policy interventions, yet did not explore interactions between minimum wage increases and public benefit reductions and withdrawals.

Low-wage workers face benefit cliffs as their income increases – many public goods upon which low-wage workers rely are cut off, including child care subsidies, early childhood education assistance, rental assistance, food stamps, and health coverage.  There are rational incentives for workers to reduce hours and therefore overall income in order to maintain these benefits.

For example, let’s consider health coverage.  Imagine a Seattle community college student, who was working 28 hours a week at the 2015 minimum wage of $11 (for large employers who did not provide health coverage).  Her annual income would have been just over $16,000 and she would have qualified for no-cost Apple Health coverage.  The minimum wage went up to $13 for employees in these same corporate firms in 2016.  If this student maintained her hours of work, her annual income would have been almost $19,000.  She would have exceeded the income limit for Apple Health and have had to pay for her own insurance in the health care exchange.  That could cost $240 a month, when we consider premiums, co-pays, deductibles and out-of-pocket costs.[8]   But if this student-worker drops her work hours from 28 hours a week to 24 hours a week, her annual income drops below $16,400, making her again eligible for Apple Health.  Her health care costs are zero.  Her imputed annual income, just in consideration of foregone health costs, is actually $200 more than if she had maintained employment at 28 hours a week.  And she has half a day more to study.

5a. Failure to consider opportunity costs:  This latter point may seem amusing, but it speaks to the opportunity costs of work, which the researchers have also disregarded.  Four hours more for study may result in more rapid and more thoughtful academic advancement, which increases the economic prospects for such student-workers.

  1. Redistribution of income: What is missing from this minimum-wage study (as well as other minimum-wage studies) is a macro-economic analysis of the redistributional effect of raising the minimum wage.  Such analyses would study the shift of income to low wage workers from consumers and employers, as well as the shift in total proportion of income to workers, either within total compensation or from capital to labor.  This research could be of value in analyzing the impact of minimum wage increases as components for slightly retarding the polarization of wealth and income among citizens, as well as slightly shifting the total allocation of income from business to workers, that is, from capital to earned compensation, thereby rebuilding the American middle class.
  1. The researchers begin their discussion with the standard theory that an increase in price will result in a lessening of demand, in this case hypothesizing that an increase in wages will lead to fewer jobs.[9]  We know that in the real world of economics, this simplistic notion is confounded by many other factors entering into economic decision-making, including:
  • Increased worker affiliation and attachment to the workplace as a result of better compensation
  • Decreased recruitment and training costs
  • Increased overall demand and thus jobs as a result of increased consumption from increased compensation
  • Shift in the proportion of total corporate income to workers and a concomitant decrease in capital and corporate profit
  • The ability to increase prices to consumers to increase wages — a redistributional effect from consumers to workers, which is most likely from higher-income consumers (i.e. eating out in restaurants) to lower income workers, and
  • Added incentives to increase productivity as a result of increased wages.

Combined, these factors create a much more nuanced context for evaluating statutory minimum wage increases.  But by introducing an analysis with only traditional economic theory, the UW researchers set up an analysis to prove or disprove a theory instead of seeking actual data in advance of or in the place of theoretical constructs.[10]


Seattle is a booming metropolis, experiencing unprecedented three-percent population growth, making it the fastest growing major city in the United States and pushing our population over 700,000.  Housing and commercial construction is booming.  High-tech employment is ramping up and up.  Service-industry jobs are multiplying to cater to the needs and tastes of high-tech workers.[11]  While the nation is experiencing modest but steady employment increases, Seattle’s growth stands out.  In this context, to assert that the increase in the minimum wage has resulted in job loss just doesn’t hold water.

What does make sense is that increasing the minimum wage can be one of many factors contributing to economic prosperity, robust consumption, and, indeed, employment increases.   Indeed, the economy and the polity benefit with increased prosperity and enhanced quality of life when low-wage jobs are replaced with even more high-wage jobs.   That is exactly what is happening in Seattle.

The UW/Arnold Foundation study has too many caveats, absences of intersecting economic factors and incongruities of data to be considered a research study with fidelity.


[1] http://www.arnoldfoundation.org/laura-and-john-arnold-foundation-awards-research-grants-to-study-minimum-wage-increases/

[2] https://evans.uw.edu/sites/default/files/NBER%20Working%20Paper.pdf

[3] http://marginalrevolution.com/marginalrevolution/2017/06/seattle-minimum-wage-study.html

[4] https://evans.uw.edu/sites/default/files/NBER%20Working%20Paper.pdf, Table 2, p. 44

[5] Such firms are also subject to penalties under the Affordable Care Act of about $200 per month per employee for failure to provide health coverage (a penalty exceeding $1,00 an hour for each employee).  There is no discussion of the interaction of such penalties with minimum wage increases.  See http://www.kff.org/infographic/employer-responsibility-under-the-affordable-care-act/

[6] This latter observation is thanks to regressive tax structures at the city and state levels, through which public revenues are disproportionately dependent on sales taxes.  See http://www.eoionline.org/wp/wp-content/uploads/Seattle-Creating-Progressive-Funding-May-2017.pdf, page 2.

[7] https://evans.uw.edu/sites/default/files/NBER%20Working%20Paper.pdf, Table 4

[8] http://www.eoionline.org/wp/wp-content/uploads/Health-Coverage-in-King-County-Aug-16.pdf, page 7.

[9] “Economic theory suggests that binding price floor policies, including minimum wages, should lead to a disequilibrium marked by excess supply and diminished demand.”  https://evans.uw.edu/sites/default/files/NBER%20Working%20Paper.pdf, p. 1

[10] Further, they avoid any discussion of real world behavioral economics, as opposed to neoclassical economic theory based on the actions of a “rational” person.

[11] http://www.seattletimes.com/pacific-nw-magazine/money-changes-everything-will-seattle-figure-out-how-to-deal-with-its-new-wealth/

Posted in Minimum Wage

Leave a Reply

Search the blog

Subscribe to the blog

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Like what you’re reading?
Reader support helps preserve our independent voice for the middle class - please chip in to help out!