Monday, January 3, 2011

An Inequality Post

. Monday, January 3, 2011

In the middle of a pretty good post on inequality, Ezra Klein writes:

In the 1970s, median household income begins stagnating. But it's not until the mid-1980s -- and really beginning in 1987 -- that the income share of the top 1 percent begins skyrocketing. ...

What happened in 1987? From about 1952 to about 1986, the top 1 percent's share of income fluctuates between 7 percent and 10 percent. But between 1987 and 1988, it jumps sharply -- rising from 10 percent to 13 percent in a single year -- and never comes back down. So what happened in 1987? There's a massive stock market crash that year, but it's not a crash that's considered to have had profound or lasting impacts on the real economy. Most explanations peg it as a market-driven, rather than economy-driven, event. And yet something that year does seem to have profoundly changed income equality in this country, and in a lasting way. But what?

As it happens I wrote my senior undergrad thesis largely on this question. It's almost a very good question. I say "almost" because Klein (I assume) is using pre-tax/pre-transfer tax return data. If so, 1987 returns reflect 1986 income, so the question should really be "What happened in 1986?" And the answer to that is, quite simply, one of the largest restructurings of the tax code since WWII, the Tax Reform Act. As Showdown at Gucci Gulch describes in detail, TRA86 was all about the distribution of benefits via the tax code, but a few changes would have an especially large effect on measured income inequality after passage.

First, it eliminated many deductions and loopholes, especially for real estate holdings. That meant that a lot of income that was going essentially unreported in 1986, because it was being sheltered, was now reported in 1987. Almost all of that income belonged to the upper tiers of the tax code that could take advantage of those loopholes, so TRA86 shone the light on a lot of income that was previously held in the dark. Second, capital gains income was taxed at the same rate as labor income. Previously, capital gains were taxed at much lower levels than ordinary income -- 20% versus 50% for the top earners. After TRA86, capital gains taxes would rise from 20% to 28%. This heavily incentivized workers that were able to do this to take income in the form of capital rather than cash. These gains would be taxed when realized... but not until then. So quite a lot of income was given to richer workers in the form of stock options and the like, and these were all exercised in 1986 to take advantage of the low rate before it went up. All of this is explained in the very long "summary" report issued by the Joint Committee on Taxation (very large pdf). Third, once corporate and individual tax rates were brought into balance by TRA86, there was a lot of "income shifting" from corporate to individual tax returns.

The net result of these for our purposes is that, among the rich but not the poor and middle classes, a lot of income was reported in 1987 that was not reported previously. So the jump in 1987 was largely a statistical artifact. The rich in 1987 were essentially as rich as they were in 1986, but the vagaries of the tax code meant that the situation looked quite a bit different when looked at in a time series. Alan Reynolds of Cato has written about this quite a lot, see e.g. here. He goes so far as to say that income inequality has basically not changed at all since 1988 (and thus since 1979 or so), and for this he has been beaten down quite severely (e.g. here and here and here and here etc.). But I think he's got this one point -- about 1987 -- basically right.

So do the academics. In their landmark inequality 2003 QJE study, Thomas Piketty and Emmanuel Saez write:

One additional motivation for constructing long series is to be able to separate the trends in inequality that are the consequence of real economic change from those that are due to fiscal manipulation. The issue of fiscal manipulation has recently received much attention. Studies analyzing the effects of the Tax Reform Act of 1986 (TRA86) have emphasized that a large part of the response observable in tax returns was due to income shifting between the corporate sector and the individual sector [Slemrod 1996; Gordon and Slemrod 2000]. We do not deny that fiscal manipulation can have substantial short-run effects, but we argue that most long-run inequality trends are the consequence of real economic change, and that a short-run perspective might lead to attribute improperly some of these trends to fiscal manipulation.

So the shift in 1987 is probably just a statistical mirage. The longer-run shift, encapsulated somewhat by this graph reproduced by Klein, is not.

Derek Thompson takes a stab at it here, and comes away perplexed. Scott Sumner characterized this as a shift in compensation from "producers" (the 1945-1973 economy) to "discoverers" (1973-2010 economy):

Today the most productive members of society are not those who produce things, they are those who discover the things that need to be produced. Once you have the blueprint, it is easy to produce many types of software and pharmaceuticals. The big money goes to those who figure out the blueprint, but also to those who allocate capital to the guy who has the idea for a Google, or Facebook, or Twitter. In contrast, the technicians who actually implement the vision often earn modest salaries. Thus companies are “discovered” in much the same way as an iron deposit is discovered by a skilled geologist.

I think that's part of it. Viewed in that light, the "breaking" of productivity and median compensation comes from the fact that productivity has not gone up because the skills of the median worker have improved, but because the skills of the "discoverer" and those who give him capital have improved. They have improved because of political and technological developments over the past quarter-century or so have pushed the production possibilities frontier way out for those with good ideas and access to capital to develop them. The nouveau riche are not Andrew Carnegie and John D. Rockefeller, but Bill Gates and Roc-A-Fella, and those who finance them.

"Discovery" has become more important because the world's labor supply and consumption markets are now essentially globalized. There isn't anything an American worker can do that a worker someplace else can't do. That wasn't necessarily true in 1945. And even if it was, corporations didn't have the same access to foreign workers that they have now. At the same time, we now have many machines that allow us to produce much more with much less labor. So the supply of labor accessible to markets has increased tremendously over the previous three or four decades at the same time the demand for labor has slowed. If the labor supply curve shifts right, the price of labor goes down. If the labor demand curve shifts left, the price of labor goes down. Viewed in that light it's no surprise that median incomes have stagnated in richer countries.

So, basically, I'd boil it down to two factors:

1. Labor supply (demand) has gotten larger (smaller), depressing the price of labor.

2. Consumption markets have gotten much larger, heightening the price of invention.

Put together, this means rising inequality and stagnating median wages.

This view is not incompatible with Cowen's "short on volatility" story, nor with the economics of superstars. But it doesn't require either of those either, and ultimately I find it more satisfying because it incorporates elements outside the domestic economy. It is mostly incompatible with stories from Pierson/Hacker and Krugman and others that inequality is about rent-seeking and manipulation of the state for private ends, although I think a lot of that goes on too (maybe another post on that soon). In other words, it's mostly an economic story rather than a political story, which is why we see some common trends across countries and not just within them.

The plus side of this, for workers, is that nonmonetary standards of living are going up even if monetary rewards are not. I find the argument that consumption inequality has decreased even as income inequality has increased to be pretty persuasive, and I don't think all of that was pre-crisis credit-based. In 1975 the richest person in the world couldn't own an iPhone; now everyone I know has one. I don't, but I have a Macbook and a flat screen television and easy access to more information and entertainment than almost anyone in history. To paraphrase Eddie Izzard, my standard of living would make King Solomon blush. It just doesn't show up in the statistics.


Alan Reynolds said...

Asin 1986-88, we also get a similar break in the Piketty-Saez data after the 2003 tax cut, which was followed by huge surge in reported dividends, capital gains and business income among the top 1percent. The figures are in my recent Wall Street Journal article available under my bio at

An Inequality Post




Add to Technorati Favorites