This year's econ critics make a few good points
And some bad ones too.
When I started my blog a decade ago, I was an economics grad student eager to point out the flaws and deficiencies in my field (especially macroeconomics). Eventually I felt like I had said all I had to say, and moved on to other topics. But I still find it fun to observe — and to referee — the various attacks that other people make against the econ field. Some of these attacks are deserved, but as the profession has listened to its critics (both internal and external) and striven to improve itself over the past decade, the critics have often failed to keep up. Every few months, some writer would pen a screed — usually in a British newspaper — harrumphing that economics isn’t a science, it has no predictive power, it’s just a way of cheerleading for free markets, and so on. And every year I’d write an article for Bloomberg patiently explaining why this was mostly nonsense.
This year, in the wake of an Econ Nobel award that showcased how the field is progressing in a more empirical direction, there have been some more interesting and nuanced critiques along with the same-old-same-old. So I thought I’d go over some of these and offer my thoughts.
Critiques from other fields
Since the start of the Covid pandemic, there has been some sniping between economists and epidemiologists. This is pretty natural, as the two fields deal with relatively similar phenomena — large groups of human beings interacting to produce some sort of quantifiable outcome. In both fields, it’s difficult to do controlled experiments, so the researchers have to rely more on a combination of natural experiments, empirical correlations, and theory. So it has been interesting to watch these two groups of scholars discover each other.
In a Twitter thread, epidemiologist Ellie Murray tried to pour some cold water on the Credibility Revolution in empirical economics that won the Nobel this year. Some excerpts:
This broadside could have generated a lot of interesting discussion about specific papers, but it didn’t; Murray politely declined to name the problematic papers she was thinking about, and only a handful of economists showed up to answer her request for examples of well-done papers.
There’s probably a reason for this, beyond just the fact that Twitter arguments easily degenerate into negativity and name-calling. Empirical work will always depend on a set of assumptions in order to identify what causes what. Suppose you see a state raise its minimum wage, and afterward you see that employment doesn’t fall. Do you assume that the state’s policy was effectively just a random decision that might as well have come out of nowhere? Or do you worry that the state government simply knew it was headed for an economic boom that would raise wages anyway, and changed its policy to look like it was responsible for the wage hike? And when you’re comparing that state to other states — or to some imaginary, synthetic combination of other states — is it an appropriate comparison, or was the state that raised the minimum wage simply less likely to suffer negative employment effects?
And so on. It’s simply, flatly, utterly impossible to give definitive proof that all of the identifying assumptions in an empirical paper hold. Ed Leamer has a nice paper explaining some of the reasons why this is true, but the problem is more general; if you could prove all the assumptions, you wouldn’t have to make assumptions in the first place.
So what can you do? All you can do is argue that your assumptions are plausible to the other people in your field. When they’re satisfied, then you get published. That’s it. It’s the same in every empirical field. If you go to an empirical econ seminar, you’ll find that it consists mainly of economists challenging authors to justify their assumptions, and authors trying to convince the crowd. Epidemiologists who are really interested in the standards of rigor in economics should attend a few of these seminars!
As economists point out in their replies, there are plenty of active efforts within the profession to improve the quality of empirical research (indeed, this is a large part of what Angrist and Imbens won the prize for). Murray’s perception that economists think the Credibility Revolution is over and done is very, very wrong:
If you want to read about some really top-notch examples of economists justifying their identifying assumptions and trying to improve their methods, check out Scott Cunningham’s blog, starting with his post about recent papers on the minimum wage.
But no matter how much economists try to make things better, they could always use some help from an outside perspective. When Murray complains of “unclear” estimands, “incorrect” interpretations, and “omitted” variables, she’s saying that the authors’ arguments in the econ papers she read didn’t satisfy her. And that’s fine! The solution to that is to have more people like Murray critique empirical econ papers, and force economists to deal with those new objections.
Which, in fact, is what the Credibility Revolution is all about, and why it’s so important and good. The very fact that an epidemiologist like Murray can read econ papers and identify potential problems with their empirical strategies is a huge victory. If she read a DSGE paper, she wouldn’t even know where to start. The Credibility Revolution gave economists a language to talk to each other, and to smart people from other fields, about the hard problems of causality and identification. That’s a revolution worth celebrating, even if it’s never over.
The Macro No Good critique
The next criticism is that even if economics is becoming more empirical and more credible, macroeconomics is still unscientific and unreliable due to the difficulty of causal identification.
I spent years making a similar case, so I’m not going to argue too much with the basic premise here. In fact, I’d argue that at this point, we almost define macroeconomics as “the study of economic phenomena for which reliable causal identifications cannot be made”. I thought about this when Ellie Murray asked what separates “macro” from “micro”:
Traditionally, “macro” is the study of the business cycle and long-term growth, while “micro” is just everything else. But this has little to do with how “big” they are; development economics overlaps heavily with growth, and tax policy is arguably just as systematically important as recessions, but development and tax econ are both “micro”. The real difference isn’t bigness, it’s how good the evidence is. For development you can do RCTs, for tax policy changes you can often make good identifying assumptions (e.g. you can study differences between companies that are impacted to different degrees by a given tax). But since booms and busts and overall growth affect everything at once, it’s much harder to get a handle on what causes what.
But that doesn’t mean you shouldn’t try to get a handle on it. There are certainly some macroeconomists who use their field’s paucity of solid evidence as an excuse to disappear into unaccountable theory-world. But there are many who are doing their utmost to scrape and scrounge every last scrap of evidence and insight out of the rocky badlands.
At the forefront of these efforts, in my opinion, are Emi Nakamura and Jon Steinsson of the University of California Berkeley. If they do not eventually win a Nobel for their efforts, I will be very surprised. A couple of years ago they wrote a paper called “Identification in Macroeconomics”, laying out the challenges in their field and proposing some ways forward. A brief summary won’t do it justice, but the basic conclusion is: Since macro evidence isn’t good, use evidence from microeconomics. Make sure the pieces of your macro theory don’t contradict what we know from credible studies of things like consumer behavior.
The approach Nakamura and Steinsson suggest won’t completely solve the problem; macro isn’t due for its own Credibility Revolution. But it’ll help. It’ll allow macroeconomists to toss out large numbers of models that don’t pass the microeconomic smell test. They’ll still be left with a lot of competing hypotheses and theories, but a lot fewer than before.
So macro is hard, but at least now macroeconomists are trying to do the right thing. I much prefer that modest, incremental approach to the vapid, bloviating alternatives.
The History of Thought critique
Another interesting criticism of economics this year came from within the profession itself. It was a debate over whether modern econ is doing itself a disservice by neglecting classic works.
This could have turned into a really interesting debate about whether economics is still in what Thomas Kuhn called a “pre-paradigmatic” phase — whether it has a solid foundation of ideas to build upon, or whether economists should toss out recent ideas and go back to scrounging for scraps of inspiration from Keynes, Hayek, Marx and Smith in order to build something that works better.
Unfortunately, this is Twitter, so it did not turn into that. Instead, the people arguing for and against reading the classics never really made their reasons clear, mostly choosing to dunk on each other with some form of “you’re just ignorant”. Then Dani Rodrik came in and offered a sensible middle path. Some excerpts from his thread:
Basically, when you distill a long scholarly book into a few simple equations, a lot gets left behind. And if you decide those equations don’t quite cut it as an explanation of the economy, you can go back to the book to try to help you glean subtle insights that will let you make a better set of simple equations.
This is true. But it’s not the only reason to read the classics. History of thought is helpful because it teaches you how people at the forefront of their field struggled with questions they couldn’t quite answer yet. This is something you don’t learn from textbooks. When Newton was inventing kinematics, he didn’t just pop out the clean, efficient version that you learn in high school; he groped around in the dark, trying to formulate ideas no one had formulated before. So if you want to be a researcher at the forefront of your field, hacking your way into the dark forest of the unknown, you could probably benefit from knowing how others did the same in their day.
So yes, there are good reasons for economists to read the classics, even if you think that economics has made progress since the days of Keynes, Smith, and Marx.
The British Leftist critique
Finally, we come to this year’s version of the good old British Leftist critique of economics. This is the tired old thing that gets trotted out year after year in the pages of British newspapers like The Guardian. In a nutshell, it goes something like this: “Neoclassical” economics isn’t a science, but it pretends to be one, spinning useless but fancy-looking mathematical models that are rigged so as to further the neoliberal political ends of rich people and big business at the expense of everyone else.
This critique is powerful and enduring because it contains some nuggets of truth embedded in a thick gooey wrapping of B.S. I’m not going to go over the whole thing yet again. But this year, a version of the British Leftist critique was offered by one of my favorite science fiction writers, Cory Doctorow:
In addition to being one of my favorite authors, Cory is one of the most interesting thinkers out there today (see his interview with Brad DeLong and me for our podcast Hexapodia). But he hangs around with a lot of British leftists. They sell a powerful, simple narrative, but it just misses way too much to be very useful.
Cory’s description of economics bears little resemblance to the real thing. Imperfect information is core to modern econ; theories showing how imperfect information can cause markets to break down received a Nobel 20 years ago. Perfect rationality has been successfully challenged by behavioral economics for decades, and received Nobels in 2002, 2013, and 2017. In recent years, Econ Nobels have gone to development economists who use randomized controlled trials to study anti-poverty programs in developing countries, to empirical economists who found that minimum wages aren’t as harmful as people think, to empirical economists who found better ways to measure poverty and household welfare, to institutional economists who studies the problem of the commons, and so on.
Meanwhile, though economists generally failed to predict the 2008 financial crisis, Ben Bernanke’s theories described how financial crises could spill over into the broader economy with devastating effects. That research helped motivate Bernanke to carry out a swift and unprecedented program of quantitative easing and emergency lending that ended up saving the U.S. economy and the global economy from a far greater disaster.
That’s not to say everything is hunky-dory. Cory is absolutely right to point the finger at Industrial Organization economists who act as hired guns for big businesses in antitrust cases — a practice that, frankly, smells like a massive case of scientific malfeasance and needs to be investigated by the American Economic Association. Cory is also correct when he points out that many undergraduate-level economics textbooks — most notably Greg Mankiw’s Principles of Economics — present a simplified version of economics that is out of step with recent research, way too heavy on theory and light on empirics, and tilted toward free market ideology.
And Cory is right when he suggests that the solution to the latter problem is simply better textbooks. And he’s right to identify the CORE project as the most promising alternative:
Cory hopes that CORE will be more ideologically in step with his own beliefs. But the real advantage of CORE — besides that it’s free — is that it emphasizes empirics much more than traditional textbooks. This, in my opinion, is actually much more important than any ideological shift. By teaching kids that econ needs to be grounded in evidence, CORE will help them understand what we do and don’t know about how the economy works. This is very much in keeping with the broader evolution within the econ profession — the shift that was recognized in this year’s Nobel.
So although Cory gets the problems with econ partially wrong, he ultimately points to some of the people who are making progress toward solving the problems that do exist.
Anyway, that’s about it for this year’s edition of Econ Critics. Tune in next year for more! There will always be more.
You missed the Judea Pearl monomaniacal critique, namely: economists spend a lot of time doing causal inference but by and large refuse to use DAGs and structural causal models to represent their assumptions and determine their identification strategies, even though this would make things much clearer with no downside.
It's interesting that I think Macro has some of the most obvious examples how it's improved the world. Like, comparing the response to 08/09 to the Great depression is pretty amazing. And the role of the world bank in China in 1983 was also pretty amazing and seems to me at least played a pretty signifigant part in China's liberalisation (you guys should read Vogel's biography of Deng, it's pretty epic.) I know much less about this case, but also feels pretty signfigant for India's liberalisation in the 1990s. If macro get's get a few percent of the credit for the declines of poverty in India and China that's pretty epic.