Seattle's $15 Vote and the Studies That Argued Over It

On June 2, 2014, the Seattle City Council met in a packed chamber and voted, nine to zero, to raise the city's minimum wage to $15 an hour. It would phase in over several years, faster for big employers and slower for small ones, and when it was done Seattle would have one of the highest wage floors in the country. The vote had been a long time coming. A year earlier, SeaTac, the small city that surrounds the airport just to the south, had passed a $15 minimum for airport and hotel workers by ballot measure. That fall, a socialist candidate named Kshama Sawant had won a seat on the Seattle council campaigning on $15. Fast-food workers had been walking off the job in cities across the country. The unanimous vote looked like a settled question. Three years later it became the most argued-over minimum wage in America, because two teams of serious researchers looked at what happened and came back with stories that could not both be complete.

In short: In June 2017, a University of California, Berkeley team reported that Seattle's increases raised pay in food service without costing jobs, and a University of Washington team reported that the move to $13 cut low-wage workers' hours enough to lower their average earnings. They were measuring different workers in different ways. The fight over those studies taught economists a great deal, and the wider research since then points toward modest increases having small effects on jobs, with real uncertainty about very high ones.

Why Seattle went first

Seattle was well placed to try a big number. Its economy was booming on the strength of technology and aerospace, its rents and living costs were rising fast, and Washington state already had one of the highest state minimum wages in the country, indexed to inflation since a 1998 ballot measure. A $15 wage in Seattle was a large jump, but not a leap into the unknown in the way it would have been in a small town in the rural South.

The politics were also unusual. The mayor, Ed Murray, set up an advisory committee that included both business owners and labour leaders, and the ordinance that came out of it was a negotiated deal. Small businesses got a longer runway. Employers could count some tips and health benefits toward the wage for a period. The threat of a ballot measure with a faster schedule hung over the room and helped push the business side toward agreement.

The city also did something less common. It paid for researchers at the University of Washington to study what happened. That decision is where the trouble began.

The Berkeley study

In late June 2017, a team at the University of California, Berkeley led by the labour economist Michael Reich released a study of Seattle's food-service industry. Restaurants are the classic test case for minimum wage research, because so many of their workers earn at or near the minimum and because restaurants are everywhere, which makes them easy to compare.

The Berkeley team compared Seattle's restaurant sector with a statistically constructed comparison group made of other areas whose trends had matched Seattle's before the law. They found that pay in food service rose as the minimum went up, and that employment in the sector did not fall relative to the comparison areas. That fit what Reich and his colleagues had found in many earlier studies of other cities: moderate increases raised wages at the bottom without measurable job losses.

The study arrived a few days before the University of Washington's. Reporting at the time said the mayor's office had seen the Washington team's findings before the Berkeley study came out, and some read the timing as an effort to get a favourable result into the news first. Whatever the intent, it made a technical disagreement look like a political brawl.

The University of Washington study

The University of Washington team, a group of economists led by Jacob Vigdor, had access to something the Berkeley group did not: Washington is one of a handful of states that collects hours worked, not just earnings, in its unemployment insurance records. That let them look at how many hours low-wage workers actually worked, not only how many jobs existed.

Their working paper, released in June 2017 through the National Bureau of Economic Research, focused on the step from $11 to $13 in 2016. They found that hours worked in low-wage jobs fell by around 9 percent relative to comparison regions, and that the drop in hours was large enough that the average low-wage worker's monthly earnings went down, by roughly $125 a month in their main estimate. Hourly pay went up; total pay, on average, did not.

That was a striking finding, and it was everywhere within days. Opponents of $15 minimums cited it as proof that well-intended raises hurt the people they were supposed to help.

Why they disagreed

The two studies were not really measuring the same thing, and the differences explain most of the disagreement.

The first is scope. Berkeley looked at food service, where wages are low across the board. Washington looked at all jobs paying under a given hourly rate, across every industry. Those are different populations, and a result in one does not automatically carry to the other.

The second difference is the one that drew the sharpest criticism. The Washington state data identify businesses by their tax account, and a chain with locations inside and outside Seattle shows up as a single account with no way to know which workers were in the city. So the Washington team left multi-site businesses out of its analysis. Chains like large restaurant groups and retailers employ a big share of low-wage workers. Critics, including Reich and the economist Arindrajit Dube, argued that if some low-wage work shifted from single-site businesses to chains, or from inside the city to just outside it, the study could mistake a move for a loss.

The third is what counts as a decline. If the wage threshold for being counted as low-wage stays fixed at $19 an hour, say, while wages rise across the board, some workers will simply climb above it and drop out of the count. The Washington team adjusted for this in various ways, but critics argued the fix was incomplete in a city where wages at every level were rising fast.

The fourth is the comparison group. Seattle in those years was not a normal city; it was growing much faster than most of its neighbours. Finding places that looked like Seattle without its minimum wage is genuinely hard, and different reasonable choices produce different answers.

None of this means either team was careless. Both were working at the edge of what the data could show, and both said so. That is often how real research looks from the outside: not a clean verdict, but a set of careful people arguing about which imperfection matters more.

What happened next

The University of Washington team kept working, and later papers added something important. Following individual workers over time, they found that people who already had low-wage jobs when the minimum rose generally did well. Their hourly pay went up, and they worked somewhat fewer hours, but their total earnings rose on average. The losses, in their analysis, showed up mainly as fewer new entrants getting hired into low-wage jobs. That is a more nuanced picture: gains for experienced workers, possibly at some cost to people trying to get their first foot in the door.

Meanwhile, the broader literature moved too. A widely cited study by Doruk Cengiz, Arindrajit Dube, Attila Lindner and Ben Zipperer, published in 2019, looked at many state minimum wage increases over decades. Rather than looking at one industry, it counted jobs just below the new minimum before an increase and just above it afterwards. It found that the number of low-wage jobs stayed about the same: jobs below the new floor disappeared and roughly the same number reappeared at or slightly above it.

Other economists, notably David Neumark, have continued to find job losses in some settings, especially for teenagers and the least skilled. The Congressional Budget Office, reviewing the research for its 2019 analysis of a $15 federal minimum, concluded that job losses were likely but uncertain, with a wide range of plausible outcomes.

Where economists now agree and disagree

Here is a fair summary of the state of the debate, as best I can describe it, and readers should treat it as a summary rather than a ruling.

There is broad agreement that moderate minimum wage increases, relative to local wages, raise pay for low-wage workers and have small effects on total employment, small enough that many studies cannot detect them. There is broad agreement that the old textbook prediction, that any price floor on labour must reduce jobs in proportion, does not describe real labour markets well. Employers often have some power to set wages below what a competitive market would pay, which economists call monopsony power, and a minimum wage can push pay up without cutting jobs when that power exists.

There is real disagreement about very high minimums relative to local wages, about effects on teenagers and the least experienced workers, about effects on hours rather than jobs, and about longer-run effects, such as faster automation or fewer new businesses opening. Seattle's experience is relevant to all of those questions and settles none of them.

There is also near-agreement on a point that should matter to progressives: where you set the minimum relative to local wages matters more than the dollar figure itself. $15 in Seattle in 2017 and $15 in a low-wage rural county are not the same policy.

The lesson for progressives

The most useful thing to take from Seattle may be about how we argue. Supporters of $15 sometimes treated the University of Washington study as an attack to be discredited. Opponents treated it as a final verdict. Neither response was honest about what the research could show.

I think the case for strong minimum wage increases is strong enough to survive a serious study that complicates it. The question was never whether a higher wage floor has any trade-offs; most policies do. The question is whether the gains for millions of low-wage workers outweigh the costs, and whether the policy can be designed to keep the costs small: phasing increases in, indexing to inflation, adjusting for regional costs, and watching what happens to new workers. Seattle did several of those things. It also kept studying itself, which is more than most governments do.

A decade on, Seattle's minimum wage has kept rising with inflation and is among the highest in the country. The city's economy did not collapse. Restaurants opened and closed as restaurants do. And workers who had been told for years that $15 was a fantasy found that, in their city at least, it became the law and then the floor.

What a city can see that a nation cannot

Part of what made the Seattle fight so valuable is that the city became one of the most closely watched local labour markets in the world. Few places had hours data in their unemployment records, a law with a clear starting date and a team of economists paid to look. That gave researchers a view of how employers actually adjusted, not just whether total jobs went up or down.

Employers have several ways to respond when the wage floor rises. They can cut jobs, cut hours, raise prices, accept lower profits, reduce turnover by keeping workers longer, invest in equipment or find ways to get more done per hour. Most real businesses do some mix of these. The University of Washington team's hours data captured one channel unusually well. Other studies captured others. Arguments about whether a raise hurts or helps are often really arguments about which channels dominate, and the honest answer is that it depends on the industry, the size of the increase and the local labour market.

The city also built an enforcement office. Seattle created an Office of Labor Standards to investigate wage theft and enforce its minimum wage and related rules. That matters more than it sounds. A minimum wage that employers can ignore without consequence is a suggestion, and much of the gap between what a law promises and what workers receive comes down to whether anyone checks.

The national debate has kept returning to the same numbers. In 2021, when Congress considered the Raise the Wage Act, the Congressional Budget Office updated its analysis and estimated that a $15 federal minimum by 2025 would cost about 1.4 million jobs in its central estimate while lifting about 900,000 people out of poverty. Supporters disputed the job estimate, citing the newer research; opponents cited it as a reason to go slower. The question Seattle raised, how high is too high for a given place, is still the right one.

What you can do

• Compare your state with Washington. The state minimum wage table lists every state's current rate. If yours is far below, the gap tells you more than any one study will.
• Bring evidence to the next local fight. City and county minimum wage proposals keep coming up. When they do, the people who show up to council meetings with a fair account of the research, including its uncertainties, tend to be the ones who get listened to.
• Support the workers who started it. Seattle's $15 grew out of organizing by fast-food, airport and hotel workers. The labor page tracks the newest union elections the NLRB has counted, and many of the current campaigns are in the same low-wage industries.
• Read the wider debate. The minimum wage is one front in a broader argument about workers and unions and who has power to set pay. Seeing the whole picture makes it easier to answer the claim that wages are simply whatever the market says.