While you were grilling burgers over the long holiday weekend, a cadre of academics were locked away in their studies, poring over data sets by lamplight and banging out blog posts about the French economist Thomas Piketty. Did somebody say "party?"
As you may know, Piketty's dryly titled but exhaustively compiled "Capital in the 21st Century," an analysis of trends in global inequality, is the surprising best-seller of the early beach season. Since late last week, it's been the topic of considerable and highly technical debate as well. At issue is whether Piketty's data holds up to scrutiny or whether he's just pushing an agenda.
For those of us who didn't get beyond a couple of semesters of economics, there's not much hope of understanding the mathematical minutiae of this dispute, let alone settling it, and I wouldn't try. But by endeavoring to make sense of the broader issues here, we can learn something not just about inequality but also about the limits of data generally.
Let's first just crudely recap Piketty's main points (though really you should read the book, or at least skim it carefully). The essential thesis of Piketty's treatise is that the most affluent among us, both in the United States and globally, have been hoarding a sharply increasing share of wealth in their societies. He posits that we may have entered a prolonged period, similar to the age of the robber barons in the 19th century, when the value of capital will continue to outstrip economic and income growth generally, so that rich people will compound their riches while everyone else falls further behind.
He also offers a series of pretty radical solutions that would have seemed unmentionable just a few years ago, including a spike in tax rates for the wealthiest citizens (as high as 80 percent on income over $500,000, and 50 percent on income over $200,000) and a 2 percent global tax on wealth (which Piketty himself admits is more utopian than realistic).
Piketty and his most frequent collaborator, the Berkeley economist Emmanuel Saez, are recognized authorities in this field, so the populist left immediately embraced Piketty's book as a brief against the rich. Inevitably, of course, other experts immediately began looking for holes in it.
Last week, reporters at the Financial Times dealt what seemed like the most serious blow yet to Piketty's scholarship. They claimed to have discovered an array of mistaken or just plain inexplicable figures in Piketty's data, which he put online for anyone who wanted to see it. Taken together, the FT claimed, these mistakes had the effect of exaggerating trend lines, and this discovery seriously undermined Piketty's case.
The first question for us to consider, then, as informed but not expert readers, is whether we should trust Piketty on the basics of his data. As near as I can tell, the best way to think about this is to envision the data as being like the plumbing in your house. A sweeping historical study like the one Piketty has undertaken requires linking different data from long stretches of time, sort of like connecting pipes.
The problem any academic faces is that the kind of data governments keep isn't uniform from one country or one generation to the next. So the various segments of pipe aren't made from the same material — some may be lead, say, while newer ones are copper or plastic — and the researcher has to come up with ways to adjust or weight various data sets in order to make them fit together. Sometimes, in the course of doing this, the researcher miscalculates or makes a dubious decision that affects the findings.
The joints between Piketty's lengths of data may not be watertight, but the consensus, even among skeptics of the book, seems to be that nothing catastrophic to his argument has leaked from his data stream. If the FT is right, then Piketty's picture of rising inequality may be slightly alarmist, but its essential shape remains unchanged.
One of the more thoughtful Piketty critics, for example, is Scott Winship, an economist at the conservative Manhattan Institute. Winship started his career on the center-left at such think tanks as the Brookings Institution before rejecting the Democrats' populist bent and finding a home on the center-right. (In other words, he is the kind of nuanced thinker about whom Harry Truman despaired when he famously said he wanted a one-armed economist, so he wouldn't have to keep hearing about what was true on the one hand and what was true on the other.)
"I feel like I spent more time looking at the data this weekend than Piketty did," Winship told me when we spoke Tuesday. "Most of the adjustments that he makes, I think, are pretty easily defended. I have a hard time getting worked up about any of this."
But to focus on the granular data here, which only a relative handful of statistical experts can knowledgeably dispute, is to miss the larger issue surrounding Piketty's book. The deeper questions aren't so much about what he might have missed or miscalculated in his findings but about what he's chosen to leave out altogether.
For instance, Piketty's look at income disparities in the past few decades doesn't take into account the limitations in our reporting of capital gains taxes. This has the effect of making isolated windfalls among the wealthy at specific moments look like sudden spikes in income, as opposed to the cashing in of assets that actually gained their value years earlier. More to the point, it obscures the income that less affluent Americans get when they sell a house, say, because they're exempted from capital gains tax and don't have to report it.
Piketty seems to assume, too, that the recent trend he's identified will go on indefinitely ― a bit of not-so-scientific prognostication that apparently bothers the top economic advisor at the White House, Jason Furman. Speaking in Ireland earlier this month, Furman praised Piketty's grasp on historical data but questioned the idea that growth would continue to slow while the return on capital would continue to soar. That isn't in the data, because we haven't yet invented time travel.
Most consequential of all, though, Winship and other serious critics point out that Piketty only measures "market income" — that is, the amount of money you make before the government gets its hands on it. That means his findings ignore all of the transferring we do from the top few percent of earners or so to the rest of America, as well as what most of us get from our employers in health care and other tangible benefits that don't show up as income.
This is really important. Think about Social Security, which, like the rest of the social safety net, didn't exist in the Gilded Age we're supposedly on the cusp or reliving. Social Security is a huge and growing source of income for the middle fifth of American workers, an increasing number of whom are retirees with little or no market income to report.
Gary Burtless, an economist at Brookings, has noted that government transfer payments made up 17 percent of personal income in 2010 versus less than 1 percent in 1929, but none of that redistribution shows up in Piketty's assessment. "It should be plain," Burtless wrote last week, "that under a comprehensive income definition, inequality is far lower today than it was in the late 1920s."
All of which leads us to a couple of conclusions. One is that Piketty's book is probably a sound and damning measure of what's happened structurally in our economy over the past few decades, which is that a lot of wealthy, educated people are earning a whole lot more wealth, while everyone else is earning less. That's a serious contribution.
But while that data can tell us what people earn, we can't extrapolate from it what they actually end up with, nor can we evaluate what we've accomplished in a century's worth of trying to mitigate inequality through social programs ― a calculation that’s critical for policymaking. For that reason alone, it seems overly facile to use Piketty's book as a means of comparing our moment to the time of the robber barons or the Roaring '20s.
The second thing to think about is whether we should be careful about the almost mystical power we ascribe to data generally. There's a movement afoot ― it started in sports and politics and spread to most everywhere else ― to disparage theory and personal observation in favor of culling data. This kind of research, made possible by technological leaps, has tremendous value, because pure opinion is cheap and too often uninformed.
But data, like experience, has inherent limits; it can tell us some things but not others. And if that data gets to the heart of something important, as Piketty's does, it will almost always be distorted or overstated by partisans looking to discredit one agenda or rationalize another. Filtered through the usual lens of ideology, even the best data becomes oversimplified in the public square.
If Truman were alive today, he might be dismayed to find that not only do we have one-armed economists, but they pretty much dominate the debate.