Tyler Cowan complains that much of the debate about Thomas Piketty’s Capital in the 21st Century is ignoring the existing science on wealth inequality, which Cowan believes “has already offered a wide range of insights on these topics, as well as having rendered some of the more extreme claims unlikely.” And Cowan offers this paper by Ana Castañeda, Javier Díaz-Giménez and José-Víctor Ríos-Rull as an instance of the kind of science that is being ignored. Cowan says the paper he cites contains “a reasonably good and accurately calibrated model” and as a result “we already have a theory which does quite well in explaining U.S. wealth inequality, and it isn’t based on the total centrality of a comparison of r and g, as you find in Piketty.” And he complains that “no one in the current debates is citing this piece, Piketty included.”
But it doesn’t surprise me that serious researchers on wealth and income inequality would not cite the paper by Castañeda et al. Note this passage from footnote 10 in the paper:
Note that throughout this article our deﬁnition of earnings both for the U.S. and for the model economies includes only before-tax labor income. Consequently, it does not include either capital income or government transfers.
In other words, the authors’ model assumes there is no capital income or government transfer income, only labor income. That is a strange restriction to say the least.
And how did the authors obtain their estimates of these all-important earnings from labor? Well, for various reasons, they don’t like the estimates that are provided by some of the available sources of direct estimates of labor income data such as the Panel Study of Income Dynamics, the Current Population Survey, or even the Consumption Expenditure Survey. So what they do instead is use estimates of their own that generate the results they are seeking to predict within the oddly limited resources provided by their own model:
To get around these problems, instead of using direct estimates from earnings data, we use our own model economy to obtain a process on the endowment of efficiency labor units that delivers the U.S. distributions of earnings and wealth as measured by the SCF.
Explaining this procedure further, the authors say:
We calibrate our model economy to the Lorenz curves of U.S. earnings and wealth as reported by the 1992 Survey of Consumer Finances (SCF). We do this instead of measuring the process on earnings directly, as is standard in the literature. This feature allows us to obtain a process on earnings that is consistent with both the aggregate and the distributional data on earnings and wealth. It also enables the earnings-rich households in our model economy to accumulate sufficiently large amounts of wealth sufficiently fast.
In other words, in order to generate the observable results, results which include rapid accumulation of wealth by earnings-rich households, and do so in a way that takes no cognizance of the fact that these households have capital income, the authors have to ignore the available data on labor income and tweak their own self-generated estimates of labor income until they have derived the observed wealth and income distribution results.
Now you might call these results “science”. But if I understand these parts of the paper correctly, then what the authors are doing is similar to the results one could get if one attempted to model the muscle mass distribution of American adults by assuming a model in which there is no protein ingestion via meat consumption, but an adjustable parameter for protein ingestion via legume consumption. You might discover that if you completely ignore the manifest empirical fact that Americans consume a lot of meat, but then assume the right value for bean consumption, you can generate the observed muscle mass distribution. This might be an interesting exercise, but one that would quickly be dismissed as an explanation of the actual phenomena under investigation, since its assumptions are clearly at odds with the empirical facts.
In a world in which it is obvious and well known that people earn substantial income from their existing wealth, and not just from labor, there is no reason to pay much attention to other-worldly models that seek to save the observable income and wealth distribution phenomena by positing a non-capitalist world in which all income is labor income.
4 thoughts on “Cowan on Piketty and Science”
doesn’t what they’ve done strengthen their result? I should say at that outset that I agree they are in “interesting result” territory not “here is the correct explanation of reality” territory, but what I mean is that if they added capital income etc. to their model it would make it easier to explain income distributions? I haven’t properly read the paper, but I presume they end up with a distribution of ability across individuals necessary to explain the observed earnings and wealth data in SCF, and we are quite a liberty to think that distribution is not plausible and the true explanation of earnings and wealth involves things they ignore? In a way I agree with Cowan, in that it might be worth taking the time to say why alternative explanations (i.e what we see reflects the distribution of ability combined with the presence of uninsurable shocks) don’t add up.
If the point is to come up with an accurate and scientifically well-confirmed understanding of what causes income inequality, then there is no value in any explanation that is inconsistent with known empirical facts about the actual sources of large portions of personal income.
Cowan wants to put into consideration the hypothesis that the source of income inequality is the differences among people in terms of their endowments of labor power, not differences that are due only to differences among people in the amount of capital they own, and the tendency of capitalist economies to increase the return to labor. So along those lines he has cited a paper that uses a model that simply assumes at the outset that there is no capital income, and in which the authors then simply make up out of whole cloth whatever labor endowments they need to assume in order to generate the income inequalities that are actually observed. It’s a ridiculous and pointless exercise, and the fact that it employs mathematical deduction doesn’t make it science. It’s similar to the process one might follow if one assumed Neptune, Uranus and Saturn did not exist, contrary to the directly observable evidence, but then held that there were 15 invisible planets in the solar system, and one then used whatever made up masses of these planets one needed to use to account for the orbits of the inner planets.
Pingback: "Consistent with" | Homines Economici
Pingback: My Piketty Series Resurfaces | Samma Vaca