A BRIEF RESPONSE TO JACOB T. LEVY

Professor Levy of McGill University responded to my thought experiment with this comment:

"This proves too much, doesn't it? The thought experiment as stated doesn't even depend on increasing income inequality. What it seems to show is "*any* income inequality must lead to increasing wealth inequality," since even at static income inequality, the rich will save more than the poor, year after year.

But if things were that simple, presumably we wouldn't have needed Piketty at all.

Of course, there's a lot more to it-- losses as well as gains in the value of stocks or real estate are concentrated at the top; taxes and spending matter; new fortunes aren't typically the result of ordinary savings behavior; etc. But all of that means that, *regardless* of whether income inequality rises or not, it can't be axiomatic that wealth concentration increases.

Krugman's "that can't be right" is, and had better be, based on a sense of what various data have looked like for a generation, not on pre-empirical certainty about what *must* be true."

Thus far, Professor Levy.

In point of fact, my thought experiment was not intended to prove anything, and of course it does not.  [I leave to one side serious thought experiments, like that of Einstein.]  It was intended to explicate Paul Krugman's gut reaction to Giles: "That can't be right."

What Krugman clearly meant was that on the basis of his long experience analyzing macro-economic data, Giles' claim was prima facie implausible, and hence called for some very close examination.  Since I was not sure that everyone would see the manifest implausibility, I decided to spell it out with an elementary arithmetic example, my reason for this being my frequent observation that people can read general remarks and nod agreement without really understanding the quick process of reasoning that underlies them.

Professor Levy is quite right, by the way, that the reasoning in the thought experiment can apply also to situations in  which income inequality is not rising.  That is, put in a polite, politically safe fashion, the central point Marx was making about the way in which the workers, by their labor, create the accumulations of capital that then stand over against them and tyrannize over them.

Krugman's point, I take it is this:  In a stable situation with constant, not rising, income inequality, one would have a constant distribution of accumulated wealth only if the savings proclivities of the rich were the same as those of the poor [adjusting, as Professor Levy suggests, for the vagaries of stock tumbles, profligacy among the rich, and so forth.]  But if income inequality is soaring, as Giles agrees it has been, then a constant distribution of accumulated wealth would require that the ever richer rich actually save a smaller proportion of their income than the poor, and that, as Krugman says, "can't be right."

Let me be clear now, since apparently I was not the first time round.  "Can't" does not mean, in this context, that the case is closed and no further investigation is required.  It means something like "Whoa!  That need some careful looking into.  Let us see what data we have that can throw light on this, and while we are at it, let us look closely at how Giles arrived at his conclusions."

Let me close with some words about the practical problems Piketty, Saez, and others have faced in carrying out their investigations, problems that exist as well for those who wish to check or contradict their findings.  A large national economy is an extraordinarily complex social phenomenon, not to speak of the world economy.  Piketty spends a good deal of time in his book explaining the nature of his data sources and indicating their fragmentary and nationally specific origins and character.  Even statistics as constrained and specific as the monthly employment numbers for the American economy are the product of a complex sampling procedure, carried out by the Bureau of Labor Statistics, that poses all manner of theoretical and practical problems.  [Incidentally, these problems are well understood by the people who generate the numbers, and in their publications they talk about their processes and the limitations of those processes with admirable objectivity, openness, and professionalism.  I have spent some time reading those publications, and I recommend them to you.  They are real eye-openers.]

Piketty has been compelled to make a series of judgment calls in order to aggregate data from diverse sources along the way to producing his long-run secular time series and graphs.  He knows that these choices are not inevitable or open-and-shut, and that serious scholars can perfectly well argue for somewhat different choices that in turn generate different results.  One of the ways that macro-economists weigh the defensibility of their choices is by their "robustness," which is to say, the extent to which their overall conclusions remain even when different plausible choices are made regarding which data to use, how to make varied data sets compatible, and so forth.  If a small, reasonable difference in one of those choices generates very large differences in the results, then the results cannot be said to be robust.

Krugman and Piketty [and Giles?  It is not clear] understand this, and recognize that the conclusions of Capital in the Twenty-First Century must be capable of withstanding this sort of critical examination by scholars well-versed in the data sources and techniques of analysis.  I completely lack the knowledge and experience to carry out such a critique, but I think I do understand what is required and what is at stake.

One final word:  There appears to be no disagreement whatsoever about the extreme inequality in the distribution of both income and wealth, however one measures them, although there may be legitimate disagreement [I am not sure] over the trends now manifesting themselves.  From my point of view, that agreed-upon inequality is all that is required to justify and underpin a radical critique of the capitalist system that produces and reproduces it.