Economists Are Not Plumbers

Peter G. Klein. Peter G. Klein is Carl Menger Research Fellow of the Mises Institute and W. W. Caruth Chair and Professor of Entrepreneurship at Baylor University's Hankamer School of Business.

"The American Economic Review has published Esther Duflo's Richart T. Ely Lecture, "The Economist as Plumber" (ungated version here, video version here). The essay summarizes Duflo's vision of economics. First, economics is primarily useful as a policy tool; i.e., the main role of the economist is to advise governments in designing laws and regulations. Second, economics is not a set of theoretical propositions about human action and interaction, useful for understanding history and policy, but a technique for building things, albeit imperfectly and experimentally. To do so, one must focus on the details, without worrying too much about the big picture. "Plumbers try to predict as well as possible what may work in the real world, mindful that tinkering and adjusting will be necessary since our models gives us very little theoretical guidance on what (and how) details will matter. This essay argues that economists should seriously engage with plumbing, in the interest of both society and our discipline."

I recognize Duflo's achievements (along with Alvin Roth, Abhijit Banerjee, John List, and many others) in pioneering the use of "randomized controlled trials" in economics. But I am not, in general, very sympathetic to this approach. To be sure, there is a role for applied policy analysis, particularly among private actors. Randomized controlled trials can be a useful tool for entrepreneurial forecasting, just like other forms of business analytics (1, 2). And of course, RCTs are supposed to be the gold standard for biomedical research (albeit imperfectly applied). But are they economics?

I have argued before that experiments, and RCTs in particular, are not substitutes for economic theory and more conventional forms of applied economics, because they deal with very small problems. By "small" I don't mean socially unimportant — Banerjee and Duflo became famous for their Poverty Action Lab, an attempt to alleviate poverty in the world's least developed areas — but rather, problems that don't involve much economics beyond something like, "incentives matter." RCTs have been used to study how to get students to study harder for tests, how to write fundraising letters to get more money, how to get people to eat healthier food (maybe), and other social issues. It's unclear that they can provide any insight into the core questions addressed by economic theory and policy, both Austrian and neoclassical. What is the basis of social cooperation? How does an economy grow? What causes business cycles? Should we adopt the gold standard? Does regulation protect private interests? There is nothing wrong with providing a little extra understanding, on the margins. But RCTs don't easily handle the big questions. 

Duflo recognizes this concern but thinks it's a feature, not a bug: 

[A]n economist who cares about the details of policy implementation will need to pay attention to many details and complications, some of which may appear to be far below their pay grade (e.g. the font size on posters) or far beyond their competence level (e.g. the intricacy of government budgeting in a federal system). It will sometimes appear that the extensive training they received is underused if . . . the theoretical complexities turn out to be second order. On the other hand, they will have a chance to apply their economist’s mind, since many of the details have implications for issues that are an economist’s bread and butter: incentives, information, imperfect rationality, etc. They will also need to be very observant, and keep a close eye on the impact of any change they recommend.

There is nothing wrong with being careful and observant in applying one's "economist's mind" to applied work in economic history or policy. (Mises would argue, in contrast to Duflo, that introspection and a deep understanding of the motives and beliefs of the relevant actors are more useful here than experimental methods.) But is that the best use of the profession's scarce talent? 

An additional problem, which I discussed before in the context of Alvin Roth's work, is that RCTs deal primarily with non-monetary exchange. In these models and settings, coordination takes place not through the price mechanism, but by administrative rules created by the "market designer." Again, this may be useful for entrepreneurs, managers, and technicians. But I fear for a world in which government bureaucrats see their role as making private markets "work better." As Hayek pointed out in The Counter-Revolution of Science, engineering is for, well, engineers, not for social scientists. Likewise, plumbing is fine and necessary, but an economist-plumber is acting mainly as a plumber, not as an economist. 

Update: Carmen Dorobăț offered a similar critique on an earlier version of Duflo's paper, emphasizing that economists are not entrepreneurs, economics is a vocation, and economists are not poets. I agree on all three points!"

Most of Econ 101 Is Right

Noah Smith doesn't even try to argue otherwise 

By David R. Henderson at FEE.

"Economist Noah Smith has a recent article titled “Most of What You Learned in Econ 101 Is Wrong.” He doesn’t actually make the case that would support that title. But he also probably didn’t choose the title. However, he did choose this statement:

But [N. Gregory] Mankiw’s book, like every introductory econ textbook I know of, has a big problem. Most of what’s in it is probably wrong.

Here’s what’s striking: In an article that purports to show that Mankiw is wrong on many issues, he doesn’t point out how he’s wrong on any issues.

Moreover, he doesn’t even try. At no point in his piece, does Smith ever relate anything he says to specific things that Mankiw claimed.

Of course, it’s possible that Smith doesn’t think he needs to do so because he takes as given that his audience knows what’s in Mankiw’s text.

So let’s look at that. On the minimum wage, Smith writes:

For example, Econ 101 theory tells us that minimum wage policies should have a harmful impact on employment. Basic supply and demand analysis says that in a free market, wages adjust so that everyone who wants a job has a job — supply matches demand. Less productive workers earn less, but they are still employed.

If you set a price floor — a lower limit on what employers are allowed to pay — then it will suddenly become un-economical for companies to retain all the workers whose productivity is lower than that price floor. In other words, minimum wage hikes should quickly put a bunch of low-wage workers out of a job.

And Smith gives his criticism in the next paragraph:

That’s theory. Reality, it turns out, is very different. In the last two decades, empirical economists have looked at a large number of minimum wage hikes, and concluded that in most cases, the immediate effect on employment is very small. It’s only in the long run that minimum wages might start to make a big difference.

In other words, in most cases there is a small, presumably negative, effect on employment. And presumably in the other cases there is a large effect. How, exactly, does this contradict the claims that Mankiw makes and that many of us teach in our equivalents of Econ 101? It doesn’t.

Now it is true that in the 5th edition (2009) of his text, Mankiw writes:

Although there is some debate about how much the minimum wage affects unemployment, the typical study finds that a 10 percent increase in the minimum wage depresses teenage employment [by] between 1 and 3 percent.

In light of the more recent studies that Smith is referring to, Mankiw might need to soften that statement. But he need not change his conclusion that the minimum wage puts some teenagers out of work. So Smith, in an article purporting to disagree with Mankiw on this, finds himself agreeing.

The other issue on which Smith takes issue with how Econ 101 is taught — or is it Mankiw’s text? — is on welfare. Smith writes:

Another example is welfare. Econ 101 theory tells us that welfare gives people an incentive not to work. If you subsidize leisure, simple theory says you will get more of it.

What’s Smith’s objection? He writes:

But recent empirical studies have shown that such effects are usually very small. Occasionally, welfare programs even make people work more. For example, a study in Uganda found that grants for poor people looking to improve their skills resulted in people working much more than before.

But here he’s attacking a straw man. Economists who have claimed that welfare discourages work have generally had in mind welfare programs that impose a very high implicit marginal tax rate because the people on welfare lose a lot of their welfare payments when they work more.

Go to the link he cites and you find that he’s talking about the kind of welfare is typically in the form of unrestricted cash grants that, presumably, they don’t lose if they work more. That means that such welfare programs do not — repeat do not — subsidize leisure.

That certainly doesn’t contradict the standard exposition in Econ 101 or the exposition in Mankiw’s text. Mankiw discusses a hypothetical welfare program in which the government guarantees an annual income of $15,000 and then takes away one dollar of welfare for every dollar earned. He writes:

The incentive effects of this policy are obvious: Any person who would make under $15,000 by working has little incentive to find and keep a job.

Why? Mankiw explains:

In effect the government taxes 100 percent of additional earnings.

Do the studies the linked article that Smith cites contradict this? No. In fact, here’s what the linked article states:

There’s no doubt that poorly designed social programs can deter work. Aid to Families With Dependent Children, the pre-welfare reform welfare program, was found to decrease hours worked by 10 to 50 percent among recipients; that likely has something to do with the fact that AFDC benefits were taken away at a rate of 100 percent, so every dollar earned on the job was a dollar not received from AFDC. Who would work under that condition?

Exactly."