What Do Inequality Data Mean?

Steven Horwitz has an interesting post on inequality data. Here is the way he understands what the data means:

Carroll uses a nice analogy from Schumpeter that I'd never heard before: the distribution of income is like a hotel with some really fancy rooms on the top floors and some very basic ones on the bottom. All the rooms are always full, but who occupies which rooms changes from year to year.

If one wants to stretch the analogy a bit more, it's also the case that each year brings a new upgrade to every room. What constitutes a "basic" room gets slightly more luxurious each year as standards of living rise, and the same is true on other floors. It might be the case that the upgrades to the top floor rooms are proportionally greater than those to the basic and middle floor rooms, but given that the occupants of the rooms switch around from year to year, those greater improvements at the top are still consistent with improvements in the absolute standard of living for many.

And to take the analogy even further: if we account for immigration and other new entrants to the labor force, it's as if the hotel keeps adding rooms/floors on each year at the lower/basic level, enabling everyone else to potentially keep moving up (assuming that some occupants die or leave the country!).

I've often noted in my courses that aggregate data hide a large amount of important information. In the case of the static aggregate income distribution data, we cannot see what happens over time and within each income quintile.

It seems that many people assume a person who is in the lowest income quintile in any year will also be in the lowest income quintile 5 years and 10 years later. It also seems many assume that a person in the top income quintile in one year will still be in that top income quintile 5 years and 10 years later. If you check out Mr. Horwitz's post you can learn that such conclusions are not well-founded. For example:

Of those taxpayer households in the lowest quintile of income in 1999, 57.5% had moved up at least one quintile by 2007 and over 30% jumped two quintiles or more.

Of those in the top 1% in 1999, only 44.6% were still there in 2007.