Reactions to Solow's Choice

Below, my reactions to comments on my post Solow’s Choice:

1. Brad DeLong notes that work on the large simulation models continued outside of academia, so when I wrote that work on these models collapsed, I should have been sure to qualify this by saying that academic work on these models collapsed.

I bet that even the private forecasters have moved away from large structural models along the lines of the 1978 Fair model, with its 97 equations, but no matter. (Note added Jan 2016: Since I wrote the previous sentence, I have been in touch with Ray Fair, who says that contrary to my conjecture, work on large models has continued. Moreover, I’ve become more familiar with work on deep learning, which uses models with many parameters, so I’m no longer as confident that good science rules out the use of such models.) Brad’s point is surely right that some private forecasters continued to work with models that evolved from the large structural simulation models that academics simply abandoned. This makes sense because as I wrote, it is quite possible that the combination of a statistical model plus expert judgment can produce better forecasts (both conditional and unconditional) than either alone. Nevertheless, expert judgment is not the same as science. Science must involve codification and communication, and small SAGE (simple/silly applied general equilibrium) models are better suited for this.

So it makes sense to have a division of labor in which forecasters sell their services as experts and academic economists work toward scientifically valid answers to such questions as why the nominal stock of money matters for the real economy.

Being explicit about this division of labor makes it easier to summarize my general point. For the scientific side of this division of labor to function properly, the academic economists working on the scientific questions should tolerate lines of inquiry that do not hew closely to the prevailing expert consensus. However, the intellectual freedom that this grants has to be matched by restraint. The scientists should avoid premature claims that an exploratory line of inquiry yields reliable policy advice.

This means that policy makers should rely on the advice of the experts, not the academic scientists. The academics have influence by persuading the experts. To use a medical analogy, a sick patient relies on an MD with clinical experience. Lab scientists with new insights about disease change treatment patterns by convincing the MDs. Lab scientists are free to ask such provocative questions as whether preservatives used in vaccines might be harmful (Answer: No) or whether ulcers can be caused by infection (Answer: Yes.) The lab scientists should not try to bypass the clinicians by offering patient advice on the basis of an unverified conjecture. A conjecture is verified not when the person who makes it convinces him/herself that it is true. It is verified when there is a professional consensus among researchers and clinicians that it is correct.

2. Robert Waldman and Joshua Gans both noted that Solow worked on some SAGE models that tried to introduce various types of wage rigidity. In particular, Waldman cites a paper that Solow published soon after the 1978 conference, in which he proposes an early version of an efficiency wage model.

In the same way that I gave credit to Lucas for going on to explore a menu cost model and to Sargent for going on to explore the effects of nonstationarity and learning, I should have given Solow credit for this contribution to the general line of work that I recommended, figuring out what was the right imperfection to add to the existing SAGE models.

That said, in his 1978 response to Lucas and Sargent, Solow chose not to give any indication that he agrees that the way forward in academic macroeconomic theory is to develop extensions of SAGE models, or that he intended to do work along these lines.

3. Waldman says that Solow must have been making a joke when he said “… identifying restrictions … whatever those words mean.” I do not see anything funny in this remark.

Instead, it seems clear that he is using exactly the same rhetorical strategy that George W. Bush signaled in his debate with Al Gore with the phrase “fuzzy math.” In each case, the speaker signals a bond with the audience. “You and me, we are sensible people. We can ignore the irrelevant concerns of those other people, the ivory tower academics, who are not like us.”

My conjecture is that it was this stance that encouraged people at Chicago to disengage. The issue was not hurt feelings. It was the apparent futility of continuing a scientific discussion when the other side simply denies that there is anything to discuss.

4. A number of people have contacted me to defend Joan Robinson. They argue that there was lots of substance to the capital controversies and that I am being unfair by attributing political motives to her.

I am not persuaded. I do not see any reason for backing off my claims:

Here, for example, is what one of her supporters wrote to me:

But of course the difference of views between the two Cambridges had political underpinnings: Cambridge England sought to explain the distribution of income between capital and labour in terms of theories inherited from Ricardo and Marx; while the neoclassical economists explained it in terms of relative factor scarcities and technical conditions of production as captured by the aggregate production function. In this respect, there is little question that neoclassical economists had a political agenda as much as the Cambridge school.

This is yet another version of “ok, we did it. But so did everybody else”. In such cases, it seems safe to trust the “ok we did it” part of the statement.

For more, see for example this retrospective by Roger E. Backhouse.

Backhouse quoting Harcourt, another of Robinson’s supporters:

In criticizing marginal productivity theory, Robinson, Sraffa, and their followers saw themselves as pointing out that income distribution was not simply a by-product of the pricing system but was the result of class conflict rooted in the structures of capitalism, including the institution of private property and the existence of entrepreneurial and wage-earning classes (Harcourt 1972, 2).

Blackhouse summing up:

The controversy between the two Cambridges eventually came to be seen by MIT economists (and most of the economics profession) as a waste of time.

5. The persistent support for Robinson is yet another indication of how unusual the social dynamics of science is compared to the social dynamics of politics and how easy it is to fall back into the norms of politics.

Politics maps directly onto our innate moral machinery. Faced with any disagreement, our moral systems respond by classifying people into our in-group and the out-group. They encourage us to be loyal to members of the in-group and hostile to members of the out-group. The leaders of an in-group demand deference and respect. In selecting leaders, we prize unwavering conviction.

Science can’t function with the personalization of disagreement that these reactions encourage. The question of whether Joan Robinson is someone who is admired and respected as a scientist has to be separated from the question about whether she was right that economists could reason about rates of return in a model that does not have an explicit time dimension.

The only in-group versus out-group distinction that matters in science is the one that distinguishes people who can live by the norms of science from those who cannot. Feynman integrity is the marker of an insider.

In this group, it is flexibility that commands respect, not unwavering conviction. Clearly articulated disagreement is encouraged. Anyone’s claim is subject to challenge. Someone who is right about A can be wrong about B.

Scientists do not demonize dissenters. Nor do they worship heroes.