The dream of evidence-based policy is that as data and analysis get better and better, policymakers will be able to make informed decisions in real time. Instead of relying on judgment or on long-cherished theories when an economic crisis arises, they’ll have the facts at their fingertips. In reality, however, this dream may always remain just out of reach.
Sometimes, the barrier to evidence-based policymaking is that the evidence just isn’t all that solid. When the financial crisis struck in 2008, it was difficult to find comparable situations -- even the Great Depression, or the Japanese land bubble, were very different in terms of which financial institutions were threatened, which types of assets had gone bad and the types of responses made by political leaders. No one really knew for certain whether bailing out the banks was the right move. As for quantitative easing, it had been tried in Japan, but there was no way to know if the results would carry over. And no one was quite sure what those lessons were, in any case -- statistical analysis of macroeconomic policies is inherently limited.
A second problem is that the people gathering and evaluating the evidence are often politically biased. This is true even in academia. Back in 2014, economists Zubin Jelveh, Bruce Kogut and Suresh Naidu performed an ingenious study to detect political bias in the economics literature. Using text-mining, they classified economists into ideological groups, and then tested whether those groups found different numbers when evaluating empirical questions -- for example, how much minimum wages affect employment. They found a statistically significant correlation -- more ideologically conservative economists tended to find results that implied less government intervention was warranted, while more ideologically liberal ones found numbers that implied government needed to do more. The effect was rather modest in size, but if even the ivory tower is subject to bias, the web of think tanks, pundits and political advisers who are close to the halls of power are likely to be much worse.
But in the end, the biggest barrier to evidence-based policymaking may simply be that events tend to move faster than evidence. For example, take the local and national campaigns to raise the minimum wage to $15 an hour. Seattle enacted a law in 2014 that would raise the minimum wage to $15 over the course of three years. New York City began a similar process two years later. California adopted a law in 2016 to raise the statewide minimum wage for large businesses to $15 by 2022, and for small businesses by 2023.
These were sensible policy experiments. Although plenty of evidence shows that most past minimum wage hikes in the US didn’t substantially raise unemployment, and plenty of evidence shows that higher minimum wages are usually good for workers’ health and happiness, there‘s always the possibility that $15 is too big of a jump. By raising minimum wage in cities and states that have high productivity levels and high costs, the danger from unemployment was limited (a $15 wage in rural Kansas would have been riskier). And starting with cities and states instead of the whole country allows economists to compare places that adopted the $15 minimum with places that didn’t, in order to figure out any harms the new wage floor might cause. Several such efforts are underway, such as the Seattle Minimum Wage Study at the University of Washington.
But politics tends to outpace science. Flush with the energy of victory from their local and regional campaigns, the activists are demanding a national $15 minimum wage. They have received support from Democrats in Congress, as well as most of the Democratic presidential candidates.
This isn‘t irrational or foolish on the part of the activists. Political momentum and enthusiasm doesn’t last forever; they have to do what they can before the national discourse shifts to issues like health care and climate change. But it means that the nation might have to give a national $15 minimum wage an up or down vote before the evidence on local $15 wages is conclusive. The Seattle Minimum Wage Study has so far found a few negative effects on the labor market, but these were mild. Evidence of the effect on consumer prices is also mixed.
Meanwhile, minimum wage opponents are scrambling to marshal any evidence they can, even though this evidence is often shoddy. For example, a survey by workplace-management-software company Harri asked restaurants how they tended to respond to minimum wage hikes; most said they raised prices and reduced worker hours. Meanwhile, another survey by the NYC Hospitality Alliance, a membership association for restaurants and bars, found that most establishments said they had cut employee hours in response to the recent wage hike. But these results are not nearly as reliable as the economics papers I mentioned above -- they don‘t measure restaurants’ actual behavior, and they don‘t compare businesses or areas affected by minimum wage with others whose minimum wage stayed the same.
But opponents feel such intense pressure to blunt the drive for a national $15 minimum wage that they’ll take whatever they can get; the Harri survey, for example, got tons of attention on social media.
This episode illustrates how hard it is to base policy on solid evidence when political events are moving quickly and biased, politically motivated actors are grasping at whatever straws they can find. On balance, the weight of evidence says that a national $15 minimum wage is less frightening than many believed a few years ago. But a decision will have to be made on this, and an untold number of other policy choices, before the science is fully settled.
So although the drive for evidence-based policy will make some progress toward the ideal of rationality, it’s always going to be an uphill battle.
Noah Smith is a Bloomberg Opinion columnist. He was an assistant professor of finance at Stony Brook University. -- Ed.