I have a T-shirt that I never put on because I don’t deserve to wear it. It is written “Master of Metrics” on the back.
I got it in 2015 as a promotional link to a copy of the review of a book on econometrics called “Mastering ‘Metrics: The Path From Cause to Effect”, co-authored by Joshua Angrist, who received Monday the Nobel Prize in Economics with David Card and Guido Imbens. To wear the T-shirt, you really have to finish the book. I’m only on page 85, so the T-shirt stays in the dresser.
Having said that, I’m pretty excited about the award being given to Angrist, from the Massachusetts Institute of Technology; Card, University of California, Berkeley; and Imbens, Stanford University. Many excellent articles on the Nobel Prize have focused on how these researchers upset conventional economic wisdom on topics such as minimum wages and immigration. Rather, I want to focus on the tools that the three have developed. These tools are powerful but easily grasped, like a good pair of pliers.
The problem with econometrics is that correlation does not imply causation. Just the fact that you wore mismatched socks to a job interview and didn’t get the job doesn’t prove the hypothesis that the wardrobe malfunction was what killed your chances. And you can’t test the hypothesis by resuming the interview with matching socks.
Economists call this “the fundamental problem of causal inference.” Fortunately, there is a way around this. If it is impossible to go back to observe the two possibilities for the same individual (interview with matching socks vs interview with non-matching socks), it is possible to find the average effect by experimenting on several people. We will never know for sure if taking aspirin was what cured your headache, but we can measure the average effect of aspirin on thousands of headache sufferers who have or have not taken one. compressed.
Sometimes economists can conduct appropriate experiments, where some randomly selected people are “treated” (experienced) and the rest serve as a “control” group. The 2019 Nobel Prize in Economics went to Abhijit Banerjee, Esther Duflo and Michael Kremer for such experiments, which aimed to reduce poverty in the world. More often than not, however, appropriate experiments are impossible. You cannot randomly assign certain people the status of smoker or dropping out of school, for example. As a fallback, economists look for “natural experiments”: real-life situations which, due to some quirk of nature or government policy or some other source, resemble designed experiments.
Card, Angrist, and Imbens are adept at identifying and learning from natural experiences. Card and fellow economist Alan Krueger exploited a variation in the state minimum wage between New Jersey and Pennsylvania to see if the increase in the minimum wage is killing jobs. The fast food outlets on either side of the New Jersey-eastern Pennsylvania border were similar in all respects except how much they had to pay workers, since New Jersey had raised its minimum wage. Contrary to popular belief, economists have found “no indication that the increase in the minimum wage has reduced employment.”
If Card and Krueger had only looked at employment in New Jersey, they would have had a hard time disentangling the effect of the higher minimum wage from the effect of seasonal changes in fast food employment. So they exploited the fact that the seasonal effects in eastern Pennsylvania are similar to those in New Jersey, effectively using Pennsylvania as the “control” group.
It’s an example of an ingenious tool that this year’s Nobel laureates have developed. Here is another one:
Let’s say you want to understand the effect of military service during the Vietnam War on income later in life. It is not enough to compare the lifetime salaries of people who served and not, as they can be systematically different from each other in other ways that are difficult to detect. For example, what if the people who had not served tended to come from wealthier families?
In a 1990 article that examined the relationship between military service during the Vietnam War and income later in life, Angrist proposed a technique to get around the problem: he focused on the lottery number of a anybody. Having a low lottery number increased the likelihood of serving in the military, and there was no risk that people who drew low numbers would be systematically different from people who drew high numbers, as the lottery numbers were awarded. randomly.
Angrist admitted that this approach was not perfect. Many of those who served in the military during the Vietnam War were volunteers, meaning they would have served even if they had had a high number of lotteries. Conversely, some who had low lottery numbers were not used, in some cases because they were labeled conscientious objectors.
But Angrist, along with Imbens, figured out how to make reliable inferences even when the natural experience was muddled. In the case of the draft, Angrist has shown that he can remove the mud by focusing on how the draft count – the natural experience – influences whether a man serves, ignoring other factors. He found that it is possible to draw useful conclusions about the men who served because they were recruited, but impossible to conclude anything useful about the men who served because they were recruited. volunteered. He found that “in the early 1980s, long after their service in Vietnam ended, the earnings of white veterans were about 15% lower than those of comparable non-veterans.”
The beauty of the work of Nobelists is that it is about the real world. Finding successful natural experiments requires not only intelligence, but a deep understanding of the phenomenon being studied.
This week, I interviewed Paul Romer, who received the 2018 Nobel Prize in Economics for his work on growth theory. In a 2015 article, he severely criticized his fellow economists for what he called “mathematics,” which he defined as the use of the language of mathematics, but in a sloppy way that “leaves ample room for improvement. make way for skidding ”. That’s not a problem with this year’s laureates, who used math appropriately, he told me.
“There has been a reaction in the profession away from theory and towards more attention to facts,” he said. “If you take it seriously, you have to take the follow-up question seriously: can I interpret these correlations as telling me something about causation?” “
Romer identified this approach as “the real heart” of what this year’s winners have done in the job.
I asked Romer if he thought his 2015 review of “mathiness” could have pushed the profession and the Nobel Committee towards the type of work being honored this year. He laughed, noting that this is exactly the kind of question that the fundamental problem of causal inference says it’s impossible to answer. Nonetheless, he said he thinks the profession is on a better path.
If you are interested in learning more about this research, two good resources are the Nobel website and an online video series featuring Angrist on the online business resource Marginal Revolution University. Now I have to finish Angrist’s book so I can wear this T-shirt.
David Card got half of the 10 million Swedish kronor ($ 1.1 million at current exchange rates) awarded for this year’s Nobel Prize in economics, while Joshua Angrist and Guido Imbens each got a quarter . As far as I know, this is one of only four ways Nobel money in economics is distributed. The others are: one person gets everything; two people get half each; and three people get a third each. (There cannot be more than three recipients.) Please let me know if I missed any other division.
What do you think of this system? If you think all winners should be treated equally, then you should oppose the 50-25-25 split and prefer one-third each. At the other extreme, if you think a distribution on the basis of a proportional contribution is fair, then why not 40-30-30 or 60-30-10 or, while we are at it, 98-1 -1? (It would be a serious insult to 1.)
Quote of the day
“There are years that ask questions and years that answer. “
– Zora Neale Hurston, “Their Eyes Watched God” (1937)
Do you have any comments? Send a note to email@example.com.