He accuses Piketty of confusing gross returns to capital with returns net of depreciation. But in the book, Piketty specifically says that his figures are net of depreciation. If you want to quibble with his specific data or how he accounts for depreciation in it, then you can do that, but you can’t just say “I think he misreads the literature by conflating gross and net returns to capital.” He doesn’t conflate them. He’s careful to explain the importance of depreciation and tries to account for it.
Pointer from Mark Thoma. Unlike Bruenig, Summers is very clear to distinguish Piketty’s data from Piketty’s analysis. The data may include depreciation, but the analysis, in which Piketty argues for an elasticity of substitution between capital and labor greater than one, evidently does not include depreciation. I give this point to Summers, not to Bruenig.
Bruenig goes on,
Additionally, when economists start going into the substitution elasticity stuff (on which Summers himself admits there is not good data), they appear to me to be pushing Piketty into a physicalist capital framework that is totally different from what he is talking about. As I explained in a prior post, Piketty has a social constructivist account of capital. The “capital” he is discussing in his book refers to all tradeable assets that deliver passive returns, not just physical buildings and machines and the like. Models that try to show adding more machines will cause the rate of return to fall proportionally such that the owners of the machines won’t grab increasing shares of the national income do not actually address Piketty’s “capital.” Summers falls into that trap here.
Right. Who cares about all this neoclassical production function, Solow growth model stuff, anyway? Instead, give a “social constructivist account of capital,” and just make statements about how this capital accumulates, earns returns, and becomes concentrated in the hands of a few.
I can be sympathetic to that approach. I would be the last person to defend the neoclassical aggregate production function. But a “social constructivist account of capital” leaves a lot of room for reasonable people to disagree about whether Piketty has discerned the true nature of capitalism or is instead being highly speculative.
Orginazational capital is (to use the Garret Jones term) is real. But I suspect it also depreciates, in the sense that it will lose its value if not constantly renewed by new investment. The Rognlie critique does not depend on the capital being physical.
The funny thing to notice is that lower / diminishing returns to capital will make Piketty’s measure look even more impressive even while he denies it is happening.
A social constructivist account of capital is a useful metric to use when, since 1980, there has been a secular increase in P/E ratios and the corresponding secular decline in interest rates.
In other words, the actual productive income that the wealthy are able to achieve with the same amount of capital has not changed very much, but the price of that capital has risen because of the diminishing returns to capital, the savings glut, and the fact that there are not enough alternative sources of high real yield.
So, let’s say I’m a landlord and the economy shifts from an era of high entrepreneurial growth to one of stagnation. In the past, my building is only worth 10 times my net rent income, because a potential buyer could invest their money in some new business and earn 10% interest. But when those opportunities dry up, my cap rate will collapse, and even if rent holds constant, my building could rise in value by five fold if prevailing interest rates fall to 2%.
So now I will seem super wealthy and wealth will seem more concentrated. But the actual ‘output’ which I can use for my own income with which to consumer has not changed at all. I have not even increased my wealth by accumulated savings or retaining earning, it is entirely a result of a the lower P/E rations.
It would ironic if most of Piketty’s result has nothing much to do with inevitable social processes that produce inequality but is merely an the automatic function of the most recent phase in the economic cycle.
“Not just”, perhaps. But doesn’t the reply seem like “not at all”?
And in either case, wasn’t that “just” what he was measuring? That is the complaint of Summers, et al: “You used the wrong numbers. They measured less of the picture than you thought.”
The reply can’t be, “You are perfectly right, and more than that, these numbers don’t measure anything at all related to my claim.”
“I used numbers unrelated to my point to prove my point” can’t be the argument. Just can’t be…