The Econ Nobel we were all waiting for
Card, Angrist, and Imbens have made econ a more scientific field.
“And new philosophy calls all in doubt,
The element of fire is quite put out,
The sun is lost, and th'earth, and no man's wit
Can well direct him where to look for it.”
— John Donne
The 2021 Econ Nobel went to David Card, Joshua Angrist, and Guido Imbens for their work in empirical economics. If you want to predict who will win the Econ Nobel, there’s a pretty simple procedure. List the most influential people people in the field who haven’t won it yet, and assume that micro theorists won’t win the prize two years in a row. Order the top 10 or 20 most influential people in terms of when they made their impact, and the ones whose influence is the oldest are the most likely to win. (Of course, the trick here is to determine who’s influential; this is done by a combination of looking at impact rankings, talking to economists, and just generally sort of knowing what’s going on in the field. But it’s not that hard.)
For years, this method led lots of people — including me — to predict a Nobel for David Card. His 1994 paper with Alan Krueger on the minimum wage was a thunderbolt that rocked the entire field of economics, and heralded an epochal shift to come. Since then, Card has been at the forefront of empirical labor economics, extending and refining the techniques he pioneered to study everything from education to immigration to gender wage gaps to inequality and much more. Angrist and Imbens’ impact on the field — though also huge — came later, and thus I wouldn’t have been surprised had they won the prize in later years. But Card was clearly overdue.
Perhaps the reason it took this long was that Card’s conclusions in his famous minimum wage paper were so hard for many in the field to swallow. Card and Krueger (who sadly died before he could receive the prize) examined a 1992 minimum wage hike in New Jersey, and found that it didn’t result in a loss of jobs. They compared New Jersey to neighboring Pennsylvania, and found no job loss. They compared high-wage restaurants in New Jersey to low-wage restaurants — still, no job loss. Maybe even a little bit of job gain, actually.
Nowadays, that’s hardly an unusual finding — indeed, it’s the norm, and economists have changed their outlook on the issue as a result. But back then, it was almost heresy. The basic theory of competitive supply and demand says that when you raise the minimum wage, people get thrown out of work! It’s right there in the intro textbooks. It was widely considered one of the most basic facts that economics had uncovered — a canonical example of how well-intentioned government interventions can have unintended negative consequences. And here were Card and Krueger saying that this simply didn’t happen in New Jersey. The government said to raise the wage, corporations obeyed, and yet people didn’t lose their jobs.
Heresy! James Buchanan — who himself won the Econ Nobel in 1986 — simply laughed at the result. Here’s what he wrote in a 1996 Wall Street Journal editorial:
Just as no physicist would claim that “water runs uphill,” no self-respecting economist would claim that increases in the minimum wage increase employment. Such a claim, if seriously advanced, becomes equivalent to a denial that there is even minimal scientific content in economics, and that, in consequence, economists can do nothing but write as advocates for ideological interests. Fortunately, only a handful of economists are willing to throw over the teaching of two centuries; we have not yet become a bevy of camp-following whores.
Of course, Buchanan is completely wrong. It’s very easy to imagine a situation where a small rise in the minimum wage will increase employment — all you need is some monopsony power in the economy. The basic theory of monopsony, which ought to be taught in every Econ 101 class right alongside the perfectly competitive model, looks like this:
This picture has a perfectly normal downward-sloping labor demand curve (the blue line), but because a dominant company’s market power distorts the labor supply curve, minimum wage actually raises employment up to a point. Real-world labor markets don’t have only one employer, but if there are only a few dominant employers, you can get a similar result. In fact, recent studies have generally found that this is exactly what happens — modest minimum wage hikes don’t kill jobs, but big ones do, and the more concentrated the labor market, the more you can safely raise the minimum wage. Alex Tabarrok calls this a “paradoxical” result, but it isn’t really. It’s textbook stuff — or should be textbook stuff, anyway.
But at the time, Card and Krueger’s finding seemed revolutionary and heretical. In fact, other researchers had probably been finding the same thing, but were afraid to publish their results, simply because of their terror of offending the orthodoxy:
Tradition, and excessive pride in the competitive model, probably played a role here. But politics was probably important too. In the 1990s, economics was still a fairly conservative field as academic fields go, and some economists probably still saw their role as that of a bulwark against socialism. In fact, economists who lean to the political right still say Card and Krueger’s result can’t possibly be true, even though it’s been replicated by a massive number of follow-up studies. For example, here’s Peter St. Onge, who works for the Heritage Foundation and the Mises Institute, on today’s prize announcement:
I guess when you’ve lived in an echo chamber of propaganda, empirical evidence feels like propaganda for the other side!
But it wasn’t just the conservative opposition that discouraged this sort of work. It was the perception that the work would be seized upon by liberal activists who didn’t realize its limitations and caveats. Here’s what Card said in a 2016 interview:
Despite the uncertainty surrounding labor economics, Card’s research on minimum wages has frequently been cited by campaigners who seem fairly certain about the benefits of increasing it. This makes Card uncomfortable. “I don’t go around saying you should raise the minimum wage—yet advocates point to my work to say they should raise minimum wages. That’s one reason why I don’t work on that topic anymore, because everyone just assumes I’m advocating for raising the minimum wage, and therefore everything I do will be discredited.”
“It’s the same with immigration,” he continues. “There is no point in me writing another paper on that, because everyone just assumes that I must be advocating raising immigration.”
And this is true. If your takeaway from Card and Krueger (1994) is simply “MINIMUM WAGE GOOD”, you’ve read it wrong.
Because what Card wanted to do wasn’t to push minimum wage or expanded immigration — it was much bigger than that. He wanted to make the field of economics more scientific. Here’s how he put it in a 2006 interview:
Economics as a whole is really a combination of two kinds of people: those who are very practically oriented and those who are more like mathematical philosophers. The mathematical philosophers get most of the attention. They deal with the big unanswerable questions. Labor economists try to be more scientific: looking for very specific predictions and trying to test these as carefully as possible. The mathematical philosophers get very frustrated by labor economists. They come up with a broad general theory, and we tell them it doesn't fit the evidence.
The importance of Card & Krueger (1994) thus goes far beyond the minimum wage — it’s about the scientific nature of economics itself. If even the most beloved, basic, consensus theory can be contravened by empirical evidence, it means that economics consists of a set of falsifiable claims about the world we live in, rather than simply a set of thought experiments.
It means that economics is a science.
Of course, as the hordes of commenters who always pop up to yell “Economics isn’t a science!!” are only too happy to remind us, economics isn’t like the natural sciences. You can do some lab experiments, but in general most economic phenomena can’t be understood by deriving universal rules of behavior from observations in a lab. Instead, falsification in economics requires empirical data from the real world.
And that is fraught with peril, partly because the real world can’t be controlled like the conditions in an experiment. If you see a big surge of immigration that didn’t hurt the Miami labor market, was that because immigration is generally safe, or because Miami happened to be having a good decade in the 80s? If you could put cities in a lab you could run the experiment on two otherwise identical cities, but you can’t.
That’s where the genius of Card’s work — and of Angrist’s and Imbens’ work — comes in. These economists devised clever ways to find comparisons — basically control groups — for natural experiments and policy experiments. The general methodology here is called difference-in-difference, and Alex Tabarrok explains it well in his post on Card and Krueger. A related technique is that of synthetic controls, where instead of comparing New Jersey to, say, Pennsylvania, you’d compare it to an imaginary mashup of other states designed to resemble New Jersey in all ways except for the minimum wage. An alternative technique is the randomized controlled trial, where economists actually use the real world as a sort of policy laboratory; Abhijit Banerjee, Esther Duflo and Michael Kremer, who helped pioneer RCT research in development economics, won their own Nobel two years ago.
Together these techniques are called “quasi-experimental” methods, and they form the core of what economists have called the “credibility revolution” in empirical economics. This term was coined by Angrist in a 2010 essay with Jörn-Steffen Pischke, where he praised the fields of public economics, labor, and development for their embrace of quasi-experimental methods, and called out macroeconomics and industrial organization for resisting them. All of these fields have subsequently moved in the direction Angrist and Pischke wanted, but in fields like public econ, the result has been especially striking:
(Update: A 2020 study found that fully 40% of NBER working papers use quasi-experimental methods of one sort or another!)
But far from simply evangelizing the credibility revolution, Angrist pushed it forward, developing new statistical techniques for measuring causal effects — often in cooperation with Imbens, as well as the late Alan Krueger — and applying these new methods to questions in education, health, and other areas. Imbens, for his part, has been nothing short of virtuosic in his quest for the development of ever-more-accurate ways of teasing out causal effects.
Naturally, there has been some pushback to this change. Economists used to dispensing grandiose Big Ideas are annoyed at the idea that those ideas might be knocked down by some rinky-dink little RCT study. Those who just want to do beautiful math are annoyed when reviewers demand they cite papers linking their math with the grubby, impure world outside their windows. And those who believe that economics should be a handmaiden to activist campaigns are aghast at the notion that empirics could cast doubt on their chosen policies.
And of course there are legitimate limitations of these methods — though the techniques of the credibility revolution have been somewhat standardized, they’re far from being a push-button solution for teasing out cause and effect. There are still plenty of special circumstances that can confound any natural experiment (maybe New Jersey didn’t suffer from a minimum wage hike because New Jersey is just weird in some way we don’t understand!). Randomized controlled trials can bias economists toward studying small-bore policy changes instead of big ones. And so on.
Also, anyone who expects the credibility revolution to replace theory is going to be disappointed. Science seeks not merely to catalogue things that happen, but to explain why — chemistry is more than a collection of reaction equations, biology is more than a catalogue of drug treatment effects, and so on. Econ will be the same way.
But what the credibility revolution does do is to change the relationship between theory and evidence. When evidence is credible, it means that theory must bend to evidence’s command — it means that theories can be wrong, at least in a particular time and place. And that means that every theory that can be checked with credible evidence needs to be checked before it’s put to use in real-world policymaking. Just like you wouldn’t prescribe patients a vaccine without testing it first.
This is a very new way for economists to have to force themselves to think. But this is a field in its infancy — we’re still at the Francis Bacon/Galileo stage. Give it time.
And the most exciting thing about this change is, we really don’t know where it will lead us. As Paul Krugman points out, many of the most striking results to come out of the credibility revolution so far are things that bolster the case for government intervention. But this is not a property of the Universe; there was simply a backlog of popular theories that concluded that the free market was the best, and so studies showing otherwise got lots of attention. That doesn’t mean the facts have a liberal bias — indeed, as economists drift politically to the left, papers with conservative conclusions might start to stand out. A recent example is the paper by Norris et al. finding that parental incarceration is good for kids in the long run!
And that’s fine. In the short term we all want our chosen politics to win out. But if you ignore reality in the name of politics for too long, eventually you’re going to find yourself shouting orders that nature refuses to obey. The credibility revolution is about the long game — about slowly increasing humanity’s body of economic knowledge, so that one day we’ll have more power over our own economic destiny than we do right now.
That is the track that David Card set us on, and that Josh Angrist and Guido Imbens pushed us farther down. That is the right track to be on.