Once you start seeing rents at the high end of the US compensation scale, it’s hard to stop (“rents” in this context means being paid well above your actual contribution to your firm’s value-added).
There’s all the “front-running” stuff by high-frequency traders that’s finally getting a lot of press, which is a close cousin of the early-info traders I’ve written about a few times. Then there’s the collusion among Silicon Valley CEOs to suppress competitive wage bargaining in the labor market (h/t: KA).
And then there’s the extreme cyclicality of top-tier wage trends.
Start with EPIs revealing wage series, developed from Social Security administrative wage data, a highly reliable source. The figure below, from their State of Working America, shows the extremes of growing earnings inequality since the late 1970s. It plots real earnings by income group, indexed to 1979, to enable comparison of very different scales—by 2012, the average annual earnings of the top 1% was about $650,000, while that of the bottom 90% was $32,000 (and that of the top 0.1%, not shown, was $2.5 million).
Now, OTEers know I’m always going on about tight labor markets and wage trends, but the point I’ve stressed is that full employment boosts the pay of low-wage workers the most, middle-wage a good bit, and not at all at the top of the scale (see, for example, figure 5 here).
One can see the full employment impact of the latter 1990s in the figure’s bottom line for the 0-90%, as those are about the only years in the series where the bottom 90% gets a clear boost. But look at those very large cyclical bips and bops amidst the top 1%–and they’re even more pronounced for the top 0.1%. What the heck’s up with that?
Did our supermanagers, as Thomas Piketty calls them, suddenly all become super-responsive to the business cycle? Did they all bang their heads at the same time and suddenly become terribly unproductive, only to recover from their amnesia shortly thereafter? Under what version of marginal product does this pattern prevail? In fact, that pattern looks a lot like the movements in the stock market! J’accuse!
The Real Earnings of the Top 0.1% (left axis) and the Dow Jones Index (right axis)
Now, as EPI’s Larry Mishel points out—and this post comes out of discussions with Larry about these data—the salary of these top earners is increasingly a function of stock options (though the truly “skilled” ones know how to time, if not backdate, the options to avoid such capital losses). So the pattern in the above figure isn’t so surprising. But unless you want to bend yourself into a pretzel to do so, explaining why a bubble-induced implosion in equity values should sharply and temporarily reduce to the marginal product of a narrow group of execs seems like a fool’s errand.
A slightly more formal analysis reinforces the point about whose pay is more sensitive to tautness and slack in the job market–and whose isn’t. The table below shows a simple Phillip’s wage curve regression of the traditional format (see data note if interested). An extra percentage point of unemployment (technically, an added point to the unemployment gap) lowers the average wage growth of the bottom 90% of workers by a highly significant one percent and that simple model explains almost a third of the variance in this series. For the top 0.1%, however, the coefficient is insignificant and the model explains nothing.
It’s neither a coincidence nor a surprise that the earnings of these tippy-top earners moves with the market these days. But it sure makes it hard to tell a story that doesn’t involve a hefty dose of rent seeking—and not just seeking, but finding!
Data note: The dependent variable in the regression is the log change in nominal salary minus the rate of inflation lagged one year (using the CPI-RS inflation index). The sole independent variable is actual unemployment rate minus the CBO’s NAIRU, or their unemployment rate associated with full employment.