A couple of days ago I was invited to give comments on Joe Stiglitz’s presentation of a new paper on theoretical models of evolution of wealth and income inequality. The other two commentators were Duncan Foley and Paul Krugman. Stiglitz’s paper is not, as far as I know, on the Internet yet, so I cannot give the link. (I also had my own slides; they can be seen here.)
Stiglitz’s is a long paper (some 60 pages) and is in reality composed of two independent papers. The first one, on which I mostly commented, is a continuation of the discussion started by Piketty’s “Le Capital…”. Stiglitz points out to several very important puzzles that cannot be easily accommodated in the current neoclassical framework: broadly constant rate of return despite massive capital deepening, rising share of capital incomes even if the production function studies tend to find elasticity of substitution between capital and labor of less than 1, and stagnant wages despite the increase in K/Y ratio. The second paper is the extension od Stiglitz’s 1969 paper on the theory of wealth and income inequality whose objective is to model the long-run distributions among households that differ in terms of labor and capital incomes and savings behavior.
I will mostly describe and comment on Stiglitz's first paper. The puzzles that I mentioned can however be solved if one distinguished capital and wealth which Piketty treats interchangeably. Piketty’s treatment of wealth as capital has already attracted attention, and I also wrote about it here. What I called the Pikettian “transformation problem” is the use of the results that come from the production function world, the K world, into the results that come from the world of personal income distribution, the W world. Combining the two, and linking the theory of growth to the theory of distribution was one of key Piketty’s contributions. But that conflation leaves several points (“puzzles”) unanswered, and can be solved either by (i) arguing that the K world behaves really very differently from what we believe to know (e.g. marginal product of capital does not decline even if K/L or K/Y increase), or (ii) by arguing that K and W are two different things.
Stiglitz takes this second route. K is productive capital, the number we put in a production function (I leave aside here all the issues linked with the Cambridge controversy), and W is private household wealth which includes not only productive capital but housing, paintings etc. Now, Stiglitz argues, W can increase, as shown in Piketty’s data, while K remains constant or even declines. Then the puzzles disappear: the return on capital need not decrease if K is constant; wages too can now be stagnant.
What created the wedge between W and K is, according to Stiglitz, land. Suppose that land (valuable real estate) is owned mostly by the rich. Suppose also that most of it does not enter the production function. Let then this land be suddenly demanded by other rich, say the rich living outside the country with whose income distribution we are concerned. Clearly, as the real estate becomes more valuable, the average wealth of the country increases, and also becomes more unequally distributed. This happens because the asset, predominantly held by the rich, has gone up in price and made the distribution of wealth more skewed.
Moreover, more unequal wealth distribution spills into a more unequal income distribution because income includes (as per Piketty but also as per common sense used in all empirical studies of income distribution) higher imputed rent for the real estate owners whose housing has gone up in price and who have not sold it. (Just to be clear: if I own an apartment in New York whose price has doubled because Chinese oligarchs have bought scores of similar apartments in New York, both my income and wealth are up.)
Stiglitz provides a number of caveats and nuances to that story, but its essential contours should have been clear by now: (1) the puzzles disappear because capital is not wealth, (2) greater demand for an asset owned by the rich increases wealth and income inequality, and (3) long-term production potential remains the same. The increase in wealth thus does not imply that there would be an increased output forthcoming in the future. (This, of course, is very different from the situation where rich people invest in productive assets.)
I think that Stiglitz’s story represents a very reasonable attempt to both save the neoclassical production function and explain the increase in the empirical wealth/income ratio evident in Piketty’s numbers. Implicitly, the paper raises a number of important issues about the nature of economic progress. If the demand of the rich were, over the longer-term, directed mostly toward non-productive assets, that is, to the Riviera and Manhattan real estate, old paintings or similar luxury items with fixed supply, we would go back to a pre-capitalist system, justly criticized by Adam Smith, for having led the people (mostly because of insecurity of property) to bury huge sums into unproductive assets like gold. The prospect for further increases in output will thence be limited. Rich countries could end up with both greater wealth and income inequality, and stagnant real output of goods and services.
Now, the real important question is, why would the rich prefer to invest in non-productive assets? A simple answer is that they do so because the expected return on such assets is greater than the expected return on equities or direct investment. But the broader question, to which I have no answer, is whether such expectations simply partake of the usual bubble phenomenon, where eventually, the price of the asset will have to go down, or are we witnessing a shift in the preferences of the rich, away from production and into old cognacs, Picassos, Rothkos, and Riviera real estate? And what does it imply for real economic growth, let alone the various theories of the trickle-down economics? For if there is no growth, there is nothing to trickle down.