[Noah Smith] Inequality is up a lot. The question is: how much?

Inequality in the US has risen a lot during the past few decades. This has sparked outrage among segments of the public, raised concern among economists and other social scientists, and revitalized America’s socialist movement.

Much of the debate has been driven by the work of three French economists -- Thomas Piketty, Emmanuel Saez and Gabriel Zucman. The trio has put out a huge amount of impressive work, digging into historical archives and patching together a diverse array of modern data sources to graphically depict how income and wealth have become more concentrated. They have also helped design some of the bolder Democratic tax plans.

But what if the three French economists have overstated their case? A number of other economists are beginning to question whether inequality has risen quite as much as Piketty, Saez and Zucman claim. Even the prettiest graph is only as good as the assumptions that underlie it and calculating inequality numbers requires a great many assumptions.

For wealth distribution, the big problem is that good data is hard to come by. Because Americans aren’t required to report their total wealth for tax purposes, this must be estimated. One way to estimate it is to look at reports such as the Survey of Consumer Finances, then try to guess how representative those surveys are. Another way is to look at the capital income of top earners (which does have to be reported for tax purposes) and to try to estimate the value of the underlying assets that generate that income.

Both of these methods involve a lot of guesswork. For example, a 2016 paper by Saez and Zucman found that the top 1 percent of the wealth distribution held 28.1 percent of the national wealth in 1990 and that this increased to a whopping 41.8 percent by 2012. But some economists from the Federal Reserve found less concentration and a more moderate rate of increase.

Meanwhile, a team of economists from Brookings found slightly higher wealth concentration than the Fed estimated, but still much lower than the French economists found. And economists Matthew Smith, Owen Zidar and Eric Zwick, using a variant of the Saez-Zucman approach, found numbers broadly similar to the Fed’s.

In other words, most economists think that the top 1 percent hold about 30 percent of the wealth in the country, but the French trio believes that it’s more than 40 percent. Everyone agrees that wealth is very concentrated at the top, and that this concentration has risen, but the two numbers are different enough that they tell different stories about what happened during the past few decades. The difference might change some people’s minds about whether a drastic and immediate overhaul of the US economic system is necessary or whether inequality can be corrected without revolutionary change.

Income inequality is also hard to measure for different reasons. Income is reported on taxes, though economists have to guess how much evasion is going on. A lot of income is earned on paper by corporations, and there’s the question of how much of that actually belongs to the owners of those corporations. This becomes more important when calculating the income share of the very highest earners, who get most of their income from asset holdings (at least on paper). Another problem is how income is changed by taxes and transfers. Should a dollar of government health care spending count as a dollar of income for the recipients?

Different answers to questions like these can produce very different pictures of income concentration. Piketty and Saez estimated that the top 0.1 percent earned 12 percent of national income in 2012; the Federal Reserve and the Brookings team both came up with a number closer to 8 percent, only slightly more than in the late 1980s. Meanwhile, a 2018 paper by Piketty, Saez and Zucman finds that even after accounting for taxes and transfers, the top 1 percent went from taking about 10 percent of national income in the late 1980s to more than 15 percent by the mid-2010s. But a paper by Gerald Auten and David Splinter finds numbers closer to 7 percent and 9 percent, respectively. A Congressional Budget Office analysis gets numbers somewhere between the two, though a little closer to the French trio’s.

How should we react to these differing numbers? In this age of data, policy is going to be influenced by whose numbers you believe. But even most economists will find it difficult to dive into the various sets of assumptions and determine which are more plausible. And most observers aren’t equipped to do that much.

Ideology also makes it harder to evaluate the competing claims. If you believe the story that leading American economists are bought and paid for by plutocrats, then you’ll be inclined to trust the French trio; if you think that most economists are generally trying their best to get at the truth, or that the French team is as biased as others, then you’ll probably assume that the truth is somewhere between the two extremes.

The most prudent course is probably to believe in the overall shape of the trends. Almost everyone agrees that wealth and income have become more concentrated at the top in recent decades. Therefore, whatever policy changes are made, they should be in the progressive direction. But simple prudence dictates that the US should start with moderate correctives rather than wholesale change based on the most extreme and alarming set of numbers. 


Noah Smith
Noah Smith is a Bloomberg Opinion columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion. -- Ed.

(Bloomberg)