Published by Instant Press on behalf of Fight For A Living Wage.
The research on minimum wages does not deliver one answer, and anyone who tells you it does is quoting selectively. What economists have converged on is narrower and more useful: the employment effect depends on how high the floor sits relative to local wages. A modest floor in a high wage city behaves differently from an aggressive floor in a low wage county, and studies that pool them produce an average that describes neither.
The textbook prediction, and what it misses
Introductory economics predicts that a price floor above the market rate reduces the quantity demanded. Set a wage floor above what employers would pay and they hire fewer workers. The logic is clean and it is where most people’s intuition stops.
The prediction assumes a competitive labor market where many employers bid for workers who can move freely between them. Real low wage labor markets depart from that in ways that matter. Workers face search costs, transportation limits, and scheduling constraints. In many towns a small number of employers dominate hiring for a given kind of work. Where employers hold that kind of wage setting power, a higher floor can raise pay without the predicted job losses, because employers were paying below what a competitive market would have produced.
Turnover works the same direction. Replacing a worker costs money in recruiting, training, and lost productivity. A higher wage that cuts quitting can pay for part of itself.
None of that repeals the textbook. It explains why the measured effects have come out smaller and messier than the model predicts.
The turn in the 1990s
David Card and Alan Krueger studied fast food employment in New Jersey after the state raised its minimum wage, comparing it to neighboring eastern Pennsylvania, which did not. Their 1994 paper found no evidence of the job losses the model predicted, and it reset the field.
David Neumark and William Wascher challenged the result, arguing that the survey data behind it was unreliable and producing payroll based estimates that pointed toward job losses. The exchange ran for years and neither side conceded.
What came out of it was better method. Comparing a state that raised its wage to a neighbor that did not became the standard design, and the argument moved to which comparisons are valid.
Where the debate sits now
Doruk Cengiz, Arindrajit Dube, Attila Lindner, and Ben Zipperer published an approach in 2019 that counts jobs by wage level rather than by industry, looking for jobs that disappear just below a new minimum and reappear just above it. Their work across many state increases found that jobs moved up the wage distribution rather than vanishing.
The Seattle Minimum Wage Study at the University of Washington, which followed that city’s phased increase using state administrative payroll records, produced findings that shifted across its own reports and have been read as supporting both camps depending on which paper and which measure you cite.
Neumark and co-authors continue to publish evidence of employment losses, concentrated among the least experienced workers, and reviews of the literature by different authors reach different summaries of the same body of studies.
That is the honest state of play. The disagreement is about magnitude and about which workers absorb the cost, not about whether tradeoffs exist.
The ratio that organizes the evidence
The variable that keeps explaining why studies disagree is the ratio of the minimum wage to the local median wage.
Work the arithmetic. The federal minimum has been $7.25 an hour since 2009, according to the U.S. Department of Labor. In a metro where the median hourly wage is $20, that floor sits at about 36 percent of the median and binds almost nobody. Now take a state that sets $15 where the local median is $18. The floor sits at roughly 83 percent of the median, reaching deep into the middle of that labor market.
Those are different policies wearing the same name. The research finds small or undetectable employment effects clustered at the low end of that ratio, and the evidence for job losses strengthens as the ratio climbs. Nobody has established the exact point where it turns, and the honest answer is that it varies by local labor market.
You can run the comparison for your own area. The Bureau of Labor Statistics publishes median hourly wages by metro area and occupation, and dividing the applicable floor by the local median tells you more about the likely effect than any national study.
What the camps agree on
More than the public argument suggests.
Phase-ins beat sudden jumps, because employers adjust prices, schedules, and staffing over time rather than all at once. Indexing to inflation after the increase prevents the real value from eroding and avoids the next fight. The federal floor has lost substantial purchasing power since 2009 through inflation alone, which both sides accept as arithmetic. And a single national number fits a country with wide regional wage variation poorly, whichever direction you think the evidence points.
How to read the next study
Check the ratio first, then four other things. Which workers it counts, since effects concentrate among teenagers and workers with the least experience and a study of all employment will dilute them. Whether it measures hours or headcount, because employers cut hours before they cut people. How long it follows the increase, since adjustment takes years. And what the comparison group is, because the whole design rests on finding places that are similar except for the wage change.
Fight For A Living Wage, a nonpartisan grassroots 501(c)(3), frames the wage floor as one symptom inside a broader affordability problem, pointing at housing, healthcare, childcare, food, transport, education and retirement as costs that all outran pay. That framing sidesteps the research dispute without resolving it, and it also happens to be where the evidence is least contested: whatever the floor does to employment, it is too small a lever to carry the whole weight of household costs by itself.