Monday, September 09, 2013

No warning label.

Perceived inequalities are on people's minds.  We spend a lot of time worrying over our status and devising signals that might boost our place in the pecking order.  The measured indicators, including what we earn, get the most attention.  No wonder this makes it into politics; class warriors feast on it. "Fairness" and "equitable" are almost impossible to define in practical terms but serve as indispensable rhetorical devices that are used interchangeably with "justice" in normal political discourse.

Economists are involved because they are skilled at interpreting the data that we have. Data on income distributions are widely available but all of the shortcomings associated with these measures as indices of well-offness are less discussed. This is why one would think that a symposium on "The Top 1 Percent" in the current Journal of Economic Perspectives (open access) is especially useful.

It's easy to be misled. Count dollars per person or dollars per household?  For inter-temporal comparisons, household size changes.  Count in-kind transfers?  How? These are bigger than ever?  What about consumer surplus?  Much of what I get via the internet is practically free but quite valuable.  What about person-to-person swaps?  The list goes on.

But the biggest data challenge of them all involves the "comparing snapshots" problem. Inter-temporal comparisons are complicated by the fact that many people move in the pecking order. Young people, for example, are most likely to move up as they leave school, go to work and eventually achieve seniority. But most of our data on income distributions do not account for this phenomenon. Not acknowledging the problem prompts many people to make the mistake of asserting that the passage of time accounts for increasing inequality.

Comparing the shares of national income that accrue to any group (such as the top 1%) over many years misleads. The normative claim that we do want to be able to test is one that recognizes income mobility. How much is there? How does that compare to other times and other places?

Yet, of the six papers included in the JEP symposium, only one (by Miles Corak) addresses the mobility question. The other six should carry a warning label. I say this in all seriousness because the eagerness to misinterpret the data -- and to suggest ever more redistribution -- practically defines the politics we have.