Are We at Full Employment Yet?

With 4.4% unemployment, one AB reader asked me if we were in danger of demand-pull inflation. My answer was that this was not likely with the employment-population ratio still at only 63.2% – and I had put forth the notion that the Natural Employment-Population Ratio was near 64%.

But why not the natural rate of unemployment you ask. Dean Baker provides one answer:

In the 90s, the unemployment rate first fell below 5% in May of 1997 and didn’t cross 5% again until October of 2001. During this four and a half year period, there was only a modest uptick in the inflation rate, much of which was attributable to a sharp rise in oil prices in 1999-2000. Back in the 90s, the Post would tell the public that economists believed that the unemployment rate could not get below 6.0 percent without generating inflation. This was the consensus within the profession. The low unemployment of the 90s proved that the mainstream consensus within the economics profession was wrong. But, rather than expand its array of sources to include economists who were not wrong about this issue, the Post simply goes back to the same folks who now use the same disproven economic theory with 5% as the magic number instead of 6%.

In other words, the natural rate of unemployment tends to move over time – and we are not really sure what the current level is. As long-time AB readers know, I have pointed out that the unemployment rate is one minus the ratio of the employment-population ratio relative to the labor force participation rate. I have also argued that the labor force participation rate will tend to rise back towards that circa 67% rate we saw before the 2001 recession. Other economists have argued that there has been a voluntary decrease in labor force participation rate so not all of the decline represents the discourage worker effect (so much for the supply-side rational for those tax cuts).

Well it turns out, that Mark Thoma has spotted at least some evidence that my proposition is qualitatively correct:

Participation has been increasing since the trough in January 2005, not decreasing (this graph ends at December 2005, but more recent participation figures do not alter the picture – see below). Furthermore, the scale of the graph is deceptive and the overall variation is small. Participation is 66.8% in January 1995 at the beginning of the graph, and 66.2% for the latest available observation for September 2006, only a .6% difference. The peak is 67.3% in January 2000, and the trough is 65.8% in January 2005.

Mark’s post was reacting to Grep Ip noting that the Federal Reserve believes that the labor force participation rate will naturally trend downwards with demographic charges to lowering the amount of job creation needed to keep the economy at full employment.

Menzie Chinn also has some interesting thoughts on this debate.

There has been much hullaballoo about how tight the labor market is given the upward revisions in the August and September figures, on top of the preliminary benchmark revision reported last month … The current expansion’s employment growth (still) looks rather lackluster. However, as noted above, various observers have argued that the lower particpation rate of recent years implies a lower rate of employment growth necessary to accomplish labor market equilibrium. In order to investigate this assertion, I ignore the fact that the participation rate is an endogenous variable, and plot in Figure 2 the official payroll series (blue), the adjusted-for-preliminary-benchmark revision payroll employment series (red), and the BLS’s household employment series adjusted to mimic the establishment series, all divided by the labor force (variable CLF16OV, in FRED II mnemonics). Obviously, since the household and establishment series are collected in different ways, dividing the payroll series by the labor force is not a perfect measure, but hopefully it will pick up the relevant trends … What is clear is that only in the case of the BLS’s “adjusted” household series is the labor market nearly as tight as it was at the peak of the 1991-01 expansion. Obviously, using the official payroll employment series, the labor market is not so tight, roughly comparable to the conditions prevailing in 1996. Even incorporating in an ad hoc manner the preliminary benchmark revision, one finds the current market, as measured by the establishment survey, is about at the same level as it was in 1997. Indeed, the strong performance in the last three months now looks very much like a slackening market. Finally, my skepticism regarding the tight market scenario is buttressed by an inspection of the employment cost (wages plus benefits) index, deflated by either the CPI or the Personal Consumption Expenditure deflator (chain-type index).

This last point is quite important. One would think that real compensation would be higher if we were at full employment.