James Galbraith weighs in:
Brad DeLong's Semi-Daily Journal: Dean Baker on Judges Behaving Badly: I keep harping on a theme, but I can't help it.
The one thing missing from this discussion, at every level from Brad DeLong and Paul Krugman through the comments, is actual data.
D-A-T-A. Numbers. Measurement.
Everyone is talking about inequality, attributing 2/5ths to this cause and the rest to some other.
But what are you talking about, actually?
Income inequality? Earnings inequality? Wage inequality? They are not the same, and in the U.S. they sometimes don't even move in the same direction.
Income inequality soared in the late 1990s. Why? A decomposition by region and sector can tell you pretty much exactly: it was the tech bubble and the stock boom. Capital gains and stock options realizations. Much of it in just five places in the whole country: Manhattan, King County WA, and Santa Clara, San Francisco and San Mateo Counties, CA. Take out those five, as Travis Hale and I showed in a paper, and the between-counties component of income inequality (which isn't all of it, but it isn't chopped liver, either) doesn't go up at all.
Meanwhile, earnings inequality went down in the same time. Why? Full employment. This component of inequality is closely tied to utilization rates and unemployment. It varies with hours worked, and overtime earned, more than anything else. It is, in short, a macroeconomic phenomenon.
Most of the discussion here and elsewhere, sadly, is about a third concept -- inequality in hourly wage rates. That is the subject of Dean Baker's anecdote about Northwest airlines: the flight attendants do not wish to work for the wage rate the company is offering.
In a decade of working on this topic, I have yet to see any measure of inequality that is based exclusively on movements in relative hourly wage rates. Not one. No doubt, changes in the structure of hourly rates can affect inequality, in principle -- changes in the minimum wage do have a measurable effect in past data.
It is quite certain, however, that the effect of changes in relative hourly wage rates on income inequality is very, very small, compared to that of changes in non-wage income (such as capital gains), and changes in the macro and demand factors that cause variations in relative earnings.
Understanding the hierarchy of components of income is the first, small step toward a realistic grip on the inequality issue. Any takers?
I would say there are five things going on: (a) income and wealth at the top end, driven by changes in finance; (b) the effects of unemployment on the wage share; (c) skill-biased technological change; (d) declining union power; and (e) shifts in policy that erode equality-supporting measures like the minimum wage. I can't untangle them.