January 2014 effective unemployment rates:9.8 % – Headline 15.9 % – All-In (U-6)
by Alan Harvey, Institute for Dynamic Economic Analysis.
The apparent confusion over labor market conditions among policy-makers, particularly at the Federal Reserve Board, is entirely unnecessary. The unemployment rate is up and the labor force participation rate is down, and we hear much discussion about how to disentangle the effects. But in fact, under reasonable assumptions, they measure the same phenomenon — unemployment.
A simple and methodologically robust calculation can produce a single metric, which we call the “IDEA effective unemployment rate.” This is the headline rate provided by the Bureau of Labor Statistics plus or minus the difference between the stable pre-crisis participation rate and the contemporaneous rate. Beginning with the March 2014 BLS release, IDEA will produce this metric in real time for your Twitter feed. You will see the release, and less than a minute later, you will see the effective rate of unemployment. (A good reason to follow us on Twitter, @IDEAeconomics.)
Discussion of Method
Prior to the crisis the participation rate (Civilian Labor Force Participation Rate – LNS11300000) was stable between 65.8 and 67.3 for more than a dozen years (see chart far below). The IDEA effective rate begins with a 66.2 participation rate benchmark. This is not the arithmetic mean, but is a combination of weighted average and judgement. Any number close would serve as well. We simply calculate the between this historically stable rate and the recession rate, then add to it the headline unemployment rate (Unemployment rate – LNS14000000) or the all-in rate (Alternative measure of labor underutilization U-6 – LNS13327709). This simple calculation produces the effective rate. Our single assumption is that the entire change in participation is due to would-be workers dropping out, making arrangements, by necessity, to support themselves because they cannot find suitable work. They choose these alternatives when they would rather be working.
A great deal of thought and discussion has been generated recently which attributes weakness in participation to alternative explanations. While there is no doubt the population is aging, to say that the enormous drop in participation was due to this demographic factor that just happens to correspond to the crisis and recession beggars credibility. The oldest baby boomers are not yet 70.
Other research was recently cited by Philadelphia Reserve Bank president Charles Plosser recently on Bloomberg. Plosser, which we heard on Bloomberg, where he said there is evidence that much of the change in participation is due to people going on disability or taking retirement. As if this were independent of the job market.
You have to apply the smell test. When I go outside into the relatively clean air of relatively prosperous Seattle, I do hear people who are retiring in the normal course of their lives. But I also hear “I can’t afford to retire because my house isn’t the asset I thought it was going to be,” or I can’t afford to sell it.” Elsewhere I hear, “I have to retire because I can do better on social security than I can working at minimum wage.” The stress at the point of retirement is palpable.
Disability likewise is also sometimes a choice taken as a route out of joblessness. Easing of disability standards actually started affecting the participation rate going back to the early Clinton era. It is not as if people are suddenly becoming disabled at a faster rate for some non-economic reason.
In our metric we avoid this entire elaborate literature of alternative assumptions and note that countervailing literature could be generated as well, say, around the fact that an historically low minimum wage actually depresses the participation rate, as people literally cannot afford to work at poverty wages. Or that the observed rise in participation in times of low unemployment is confirmation (see graph far below).
Again, in the IDEA effective rate all of that literature is replaced by the single simple assumption: The change in the participation rate from its previously very stable rate to its post-crisis recession rate comes from would-be workers not finding a place in the labor market.
Having derived the difference between the historically stable rate and the recession rate, the IDEA effective unemployment rate simply adds to it the headline unemployment rate (Unemployment rate – LNS14000000) or the all-in rate (Alternative measure of labor underutilization U-6 – LNS13327709). This simple calculation produces the effective rates.
We call this the “IDEA effective unemployment rate.” The advantage of using the IDEA label to those who cite the statistic is little more than that they can then link to this page and description, and thus deflect the need to explain or defend it themselves. The effective rates will be available, as we said, in real time via Twitter and on the website. In addition, beginning with the March 2014 BLS release, we will provide updated graphic displays — see below for a sample — which you can use freely. (Our attribution is embedded in gray within the graphs.)
There is no doubt that officials and policy-makers are aware that both the unemployment and the participation rates directly describe the strength or weakness of labor markets. There is also no doubt that they can make this simple calculation, or a variation of it. It is conceivable that they do, but choose not to cite it for public consumption because it creates a larger, scarier number. This they would like to avoid, either for reasons of promoting “confidence” or to protect their policy decisions from criticism.
Side-by-side, the IDEA effective unemployment rate when compared to the official BLS rate is at least three full points more accurate in describing real labor market conditions.
[iframe src=”/files/ad_openx.htm” width=”600″ height=”300″ frameborder=”0″ scrolling=”no”]
[iframe src=”http://econintersect.com/authors/author.htm?author=/home/aleta/public_html/authors/harvey_alan.htm” width=”600″ height=”400″ frameborder=”0″ scrolling=”no”]