Case Study 2 — Seattle's $15 Minimum Wage Experiment

In June 2014, the Seattle City Council voted unanimously to raise the city's minimum wage to $15/hour, phased in over several years. At the time, $15 was about double the federal minimum wage of $7.25. The Seattle ordinance was the most aggressive minimum wage increase any major U.S. city had attempted, and economists, journalists, and policymakers across the country watched it closely. The question: would Seattle's experiment confirm the standard model's prediction of significant employment losses, or would it confirm the more recent Card-Krueger framework of small employment effects?

The honest answer, as of 2026, is that the evidence is mixed and the debate is still live. But the Seattle case is one of the cleanest natural experiments we have in modern minimum wage research, and what it has shown — and what it has not shown — tells us a lot about how empirical economics works in practice.

The setting

Seattle in 2014 was a booming city. The tech sector was driving rapid growth in jobs and wages. The median wage was already well above the federal minimum. Cost of living was high, particularly for housing. The pre-ordinance minimum wage in Washington State was $9.32/hour. Seattle's median hourly wage in 2014 was about $26 — meaning the proposed $15 minimum was about 58% of the median. (For comparison, the federal minimum of $7.25 was about 30% of the national median wage at the time.)

The phase-in schedule depended on employer size and benefits: - Large employers (501+ workers): $11/hour starting April 2015, with annual increases reaching $15 by January 2017 - Large employers with benefits: similar but slower - Small employers: $10/hour starting April 2015, with annual increases reaching $15 by January 2019 (or 2021 with benefits)

The phased approach meant that the full $15 wage didn't take effect immediately, giving the city's labor market time to adjust.

What the simple model predicted

The simple supply-and-demand model with a moderately elastic labor demand curve would have predicted: - Significant unemployment for low-wage workers - A reduction in hours worked for those who keep their jobs - Higher prices at restaurants and other minimum-wage-intensive businesses - Some businesses closing or relocating outside the city

The size of these effects would depend on the elasticity of labor demand. With a typical pre-Card-Krueger elasticity of −1, a 60% wage increase would predict roughly a 60% reduction in low-wage employment. This would be catastrophic. With a Card-Krueger elasticity near zero, the predicted effect would be minimal.

Most economists were somewhere in between. The pre-implementation predictions ranged from "small but real job losses" (CBO-style estimates) to "essentially no employment effects" (Card-Krueger-style estimates).

What actually happened

The Seattle Minimum Wage Study (a research collaboration including the University of Washington and other institutions) tracked employment, hours, and wages in Seattle from 2014 onward, comparing to a synthetic control of similar Washington State counties.

Wages went up for low-wage workers. This was clear and consistent with the policy's intent. Workers who kept their jobs earned substantially more.

Employment effects were mixed across studies. This is where the debate gets interesting:

• The University of Washington team (Vigdor, Jardim, et al.) initially found that hours worked by low-wage workers fell substantially after the ordinance — by enough that total earnings for the affected group actually decreased. Their interpretation: the wage gains were more than offset by hours losses.

• The Berkeley team (Reich, Allegretto, et al.) found different results using a different methodological approach (focusing on the food service sector). Their interpretation: small employment effects, large wage gains, net positive effect on workers.

• Subsequent research has tended toward middle ground. Most analysts now believe Seattle saw some hours reduction but not as severe as the UW team's initial findings, and not as small as the Berkeley team's findings. The full $15 wage (which took effect for large employers in 2017 and for small employers later) probably caused a modest but real reduction in low-wage employment hours.

Restaurant prices rose. This was consistent with the price-pass-through mechanism — restaurants absorbed some of the wage increase as higher prices.

Restaurant closures did not significantly increase. Despite predictions of widespread closures, Seattle's restaurant industry remained healthy. Some restaurants closed; others opened. The net effect was modest.

Low-wage workers continued to work in Seattle. There was no exodus of low-wage employment to neighboring jurisdictions. The labor market mostly absorbed the wage increase rather than relocating.

Why the results are contested

The Seattle study illustrates one of the hardest problems in empirical economics: identifying causal effects in a complex, changing environment. Several confounding factors complicate the analysis:

1. Seattle was booming. The city's tech-driven economic expansion was creating jobs at every wage level. Some of the resilience of low-wage employment may have been due to the strong economy rather than the elasticity being small.

2. The phase-in was gradual. Effects of a sudden $15 wage might be different from effects of a gradual transition. The empirical record only tells us about the gradual case.

3. Workers adjusted in ways that confused measurement. Some workers who would have worked in Seattle moved to higher-paying jobs as their wages rose; others took jobs in suburban areas. The data don't always cleanly capture who is affected.

4. The denominator problem. Are we counting "low-wage workers" as those earning less than some threshold (which changes when wages rise) or as a stable population? Different choices give different answers.

5. The right counterfactual. What would Seattle have looked like without the minimum wage increase? Comparing to other Washington counties involves assumptions about how representative they are.

These methodological disputes are not signs of dishonesty among the researchers. They are the kinds of disputes that arise whenever you try to estimate causal effects from observational data in a complex economy. The honest framing is that the Seattle experiment told us something — that the very large minimum wage increases produced wage gains for some workers, modest hours reductions for others, and price pass-through to consumers — but didn't settle the broader debate.

What we can confidently say

After several years of follow-up research, here is roughly what most labor economists agree on about Seattle:

1. The wage gains were real. Low-wage workers in Seattle earned more after the ordinance than they would have without it.

2. There were some hours losses. Not all of the wage gain translated into income gain because employers reduced hours for some workers. The size of this effect is contested.

3. Net earnings effect for affected workers was positive but smaller than the wage increase suggested. Workers came out ahead on net, but by less than a naive calculation would predict.

4. There was no catastrophic employment collapse. The simple supply-and-demand model with elasticity of −1 would have predicted massive job losses. They did not occur.

5. The labor demand elasticity is small but nonzero. This is consistent with the Card-Krueger framework and inconsistent with the pre-1994 consensus.

6. At very high wage floors, the elasticity may be larger. This is the open question for $20 or above.

What the Seattle experiment did not settle

• Whether a $20+ minimum wage would have substantially different effects
• Whether the same policy would work the same way in lower-wage cities
• Whether the long-term effects (over 10+ years) match the short-term effects
• Whether the employment effects are evenly distributed across worker types or concentrated on certain groups (teenagers, immigrants, those with less education)
• Whether the wage gains for affected workers eventually pass through to higher prices that hurt low-income consumers

Each of these is an open question. Each has been studied. Each has produced mixed results. The honest position is that we know more about minimum wage effects after Seattle than before, but we don't know everything.

The political reception

The Seattle experiment has been cited by both proponents and opponents of higher minimum wages, often selectively. Proponents emphasize the wage gains and the absence of catastrophic employment effects. Opponents emphasize the hours reductions and the modest negative employment effects. Both can find support in the data.

The 2020s saw a wave of $15+ minimum wage adoptions across U.S. states and cities, partially inspired by Seattle's apparent success. California, New York, Massachusetts, and several smaller jurisdictions adopted phased $15 (or higher) minimums. The federal minimum wage debate, however, has not produced a $15 federal floor — partly because of political resistance, partly because of concern about the heterogeneity of labor markets across the country (a $15 minimum is moderate in Seattle but extreme in rural Mississippi).

Lessons

Lesson 1 — Empirical economics is hard. Even the cleanest natural experiment doesn't fully resolve a contested question. The Seattle data has been analyzed by multiple research teams using multiple methodologies, and they have come to somewhat different conclusions. This is normal. The debate is honest.

Lesson 2 — Predictions matter. The pre-Seattle predictions (significant employment losses) did not match what happened. This is a real update to the field's understanding, even if the precise size of the actual effect is contested.

Lesson 3 — Heterogeneity matters. The Seattle experiment was conducted in a high-wage, booming labor market. Generalizing to other contexts requires caution.

Lesson 4 — The trade-off is real but smaller than the simple model predicted. Higher minimum wages do produce some employment loss. The loss is smaller than the wage gain for most workers. Whether the trade-off is acceptable is a values question, but it is at least a less catastrophic trade-off than the simple model suggested.

Lesson 5 — Pure "the market knows best" is not the right framing. Markets are useful but imperfect. Where minimum wages have passed without producing the predicted disasters, it's appropriate to update toward a more nuanced view of the labor market — one that includes monopsony power, search frictions, and behavioral effects that the simple model abstracts from.

Discussion questions

1. The Seattle results have been interpreted differently by different research teams. What does this tell you about empirical economics? Should researchers always converge on the same answer?

2. The Seattle experiment was conducted in a high-wage city. How might the results differ if the same $15 minimum were imposed in a low-wage rural area?

3. The case study suggests that a $20 minimum wage is more uncertain than a $15 one. What's the economic reasoning for this? At what wage floor would you expect substantial employment effects?

4. Larry Summers has argued that "demand curves slope down" — meaning that high enough minimum wages must eventually cause large employment losses. The empirical literature suggests this is true at very high wage floors but not at moderate ones. How would you reconcile these views?

5. Imagine you are advising a city council on a $15 minimum wage proposal. Based on the Seattle evidence, what would you recommend? What additional information would help you?