After California’s $20 minimum wage for fast-food workers went into effect in April, some economists expected affected restaurants to cut jobs. So what actually happened? They not only added workers but did so at a faster pace than fast-food restaurants in the nation as a whole—or at least that was the claim of a research paper by two labor economists at the University of California, Berkeley, and the University of California, Davis.

If you actually read it, you’ll find that the results celebrated in the press release and echoed by the media aren’t in the paper. In fact, it barely addresses the effect of the minimum wage increase on fast-food employment in California. It offers no numbers and no models. There’s no evidence that fast-food jobs increased after the law was implemented.

The paper’s findings were trumpeted as evidence that government-mandated wage increases have no adverse effect and that we should be raising the minimum wage higher and in more places.

Only toward the end of the 25-page study is employment shown. There you’ll find a graph that represents the closest thing to an argument in the paper. It shows full-service and fast-food restaurant employment in California, represented by the red line, and in the U.S., represented by the blue line, from 2023 to 2024.

The authors state that the data are prone to sampling errors, and make an inconclusive finding that “we do not detect evidence of an adverse employment effect.” But the paper’s abstract neglected the fine print caution, boldly asserting, “We find that the policy…did not reduce employment.” The accompanying press release, which is likely all that journalists bothered to read, states that “contrary to fears expressed by restaurant groups, the wage increase did not lead to job cuts.”

But the solid red line on the chart clearly shows California fast-food employment increasing more slowly than the solid blue line showing national fast-food employment, which is the opposite of the authors’ claim. If they suggest anything, these data show that the minimum wage increase reduced California fast-food jobs.

But it’s still hard to make a precise estimate from the way the chart is presented. So I looked up the numbers, which tell a different story than the authors claim. Even though the paper was published in September, the chart ends in July 2024, when California fast-food employment was up 1.85 percent since March 2024 while national fast food was up 3.22 percent.

This is a sign that the minimum wage is having a negative impact. In 2021 and 2022, national and California fast-food employment grew at nearly identical rates: 7.7 percent over the two years nationally, and 7.8 percent in California. But in 2024, growth slowed dramatically in California, and after July, employment began to decline.

The slowdown started a couple of months before the law took effect, but that’s exactly what you’d expect because it was signed by the governor in September 2023 and management’s decisions to close, open, or rebrand their restaurants would have been made in anticipation of the law being implemented.

But there’s a big problem even with my version of the chart. The data used to draw the red solid line don’t only represent fast-food restaurants impacted by the law; they also include casual dining restaurants exempted from the law, such as buffets, Panera Breads, smaller fast-food chains, donut and snack shops, grocery store concessions, and most delis. If fast-food restaurants were negatively impacted by the law, we would expect some of the exempted establishments to expand to take their market share, thus adding jobs. By combining data from exempted establishments that were likely growing with data from restaurants impacted by the minimum wage increase, the negative effect of the law may be hidden in the data.

Looking at crude aggregates tells us little. But California possesses the information from employer job reports that would settle the issue. Every quarter, California employers submit a series of reports to the state giving details of each employee’s hours and pay by Social Security number. The state knows everyone who worked for a fast-food operation covered by the law, and what their wages and hours were before and after the law took effect.

Another study on the same subject, “Early Effects of California’s $20 Fast Food Minimum Wage” by Daniel Schneider, Kristen Harknett, and Kevin Bruey, sponsored by Harvard’s Malcolm Wiener Center for Social Policy, used data from semi-annual surveys of retail workers in Western states. In this case, the researchers focused only on fast-food workers covered by the law, excluding exempt restaurants.

Immediately after the new fast-food law became effective, California fast-food workers lost an average of two work hours per week due to the law, according to the paper. But the authors used misleading language to report this result because their margin of error for the estimate meant that the actual change could be anything from an average loss of five hours to an average gain of one hour. “We can reject large reductions in work hours,” the study reports. “We find no significant effects of the minimum wage increase on usual hours.”

Both statements are misleading. The authors can reject that the average loss in hours was greater than five per week, but five hours–or even two–is a large loss. The estimated two hours per week loss is economically significant to the low-wage workers, just not statistically significant by conventional criteria (which means it might be the result of random noise). The correct phrasing is “we failed to find statistically significant proof” of wage losses, not “we find no significant effects.” Absence of evidence is not evidence of absence.

Another deficiency is that the study only covered fast-food workers who kept their jobs after the law went into effect, excluding missing workers who were laid off and the employees of restaurants that closed. And the respondents were self-selected—people who respond to Facebook and Instagram ads to take a survey for a chance to win a $500 gift card. This is not a random selection of people, and they do not always answer the questions seriously or honestly. This is not a sample from which anyone would expect to get solid proof of anything, so failure to find it doesn’t mean much one way or the other.

Finally, the study only deals with the first few months after the law took effect. Some changes can take months or years to emerge.

The main objection to high minimum wages is not their effect on overall employment, prices, or profits—it’s the fear that they cut off the bottom rungs of the economic ladder for low-skill workers. Instead of being able to work at low wages while improving job skills and making contacts for advancement, they are forced into the underground cash economy or public assistance. Therefore studies of these laws should focus on the effect on these low-skill workers, not on economic aggregates.

The correct way to study the impact of the $20 minimum wage is to see what happened to fast-food workers who were earning less than that amount before April 1, 2024. How many had their pay raised and hours maintained? How many lost their jobs or lost hours, and what did they do afterward? Were low-skill workers able to compete for the new job openings after the law’s implementation?

If minimum wage increases were a drug, governments would have to conduct trials and monitor adverse effects afterward. That’s what happened in Seattle when it raised the minimum wage in 2014. The city called for proposals to study the impact on actual workers earning below the minimum before the law. The Evans School of Public Policy and Governance at the University of Washington was the only volunteer. Its researchers found that the law didn’t cause an increase in layoffs among workers who had previously earned below minimum wage, but it did reduce their hours by an average of 7 percent. That was partly offset by a 3 percent increase in hourly pay for the hours they did work. On net, the law cost these workers an average of $888 per year.

That amount is significant in itself, but it’s important to consider that it accounts for only the short-term effects. As mentioned above, some layoffs and hour reductions will happen immediately, but others—such as more businesses closing and fewer opening, or automation and other changes reducing employment—can take years. Another point is that the workers who benefited from higher pay were the ones most likely to have risen out of the minimum wage ranks to the middle class even without a mandated increase, while the workers who lost much more than $888 per year are more likely to be the ones blocked forever from economic advancement. In fact, the paper found that the workers who benefitted net were the most experienced and highest paid among the group–earning more than the old minimum but less than the new–while the less-experienced workers earning the old minimum or close to it, lost considerably more than the average.

Seattle legislators must have been unhappy with those findings because they cut funding for the Evans School and reached out to the same group at U.C. Berkeley that did the California minimum wage study to do its own distorted analysis, which was rushed out a week before the Evans study was made public. Eventually, Seattle raised the minimum wage again.

These studies aren’t about the search for truth with statistics; their purpose is to score political points, with little regard for the low-skill workers whose lives are directly impacted.

• Video Editor: Adani Samat
• Audio Production: Ian Keyser
• Color Correction: Cody Huff
• Camera: Jim Epstein