March 02, 2015
Jared Bernstein has a great blog post at the WaPo on the unemployment rate associated with full employment. Bernstein refers to this as the FEUR, or “Full Employment Unemployment Rate”.
When economists talk about the FEUR, they’re referring to the lowest rate of unemployment that can be achieved before inflation takes off. The concept behind the FEUR is that high levels of employment will be associated with levels of demand that are high enough to induce wage-price spirals.
Bernstein notes that, according to the Congressional Budget Office (CBO) and the Federal Reserve (“Fed”), the FEUR is about 5.4 percent. He argues that this estimation is far too high because the official unemployment rate understates the level of slack in the labor market. Bernstein writes:
“There are times when the unemployment rate is not up to the task of capturing the true amount of slack in the labor market. It counts only as unemployed people actively seeking work, so if discouraged job-seekers give up looking, it paradoxically falls.”
I agree with Bernstein’s assessment, and I thought I would provide some extra evidence illustrating his point.
In 1994, the Bureau of Labor Statistics (BLS) began publishing six different monthly unemployment rates. They are referred to as the U-1 through U-6 unemployment rates, respectively. The U-1 rate provides the most restrictive definition of unemployment (it was 2.7 percent in January) while the U-6 rate provides the most inclusive definition (11.3 percent in January). Historical data on the six unemployment rates can be found here.
These measures can prove useful when assessing the claim that 5.4 percent unemployment is full employment. The official unemployment rate, which is used in determining the FEUR, is the BLS’s U-3 unemployment rate. In order to be counted as unemployed in the U-3 measure, a prospective worker must “have actively looked for work in the prior four weeks“. The U-4, U-5, and U-6 measures of unemployment expand upon this definition in the following ways:
- The U-4 unemployment rate takes the U-3 rate and adds in “discouraged workers”. These are prospective workers who have searched for work in the last 12 months, but haven’t searched within the past four weeks specifically for the reason that they didn’t think any jobs were available;
- The U-5 unemployment rate takes the U-3 rate and adds in “marginally attached workers”. Like discouraged workers, marginally attached workers have searched for jobs within the past 12 months but not the past four weeks. However, they can give any reason for why they stopped looking for work; and
- The U-6 unemployment rate takes the U-5 unemployment rate and adds in workers who are employed part-time (less than 35 hours per week) but would like to work full-time. (These people are sometimes referred to as “involuntary part-time workers”.)
The U-3, U-4, U-5, and U-6 unemployment rates present us with different methods of estimating the degree of slack in the labor market. And when we compare the U-3 measure today with the more inclusive measures, we see a discrepancy.
With the CBO and the Fed claiming that a U-3 unemployment rate of 5.4 percent represents “full employment”, an independent observer would think that there’s very little slack in today’s labor market: in January, the U-3 unemployment rate stood at 5.7 percent, just 0.3 percentage points above the CBO and the Fed’s estimates of full employment. And this doesn’t even represent the lowest rate of unemployment we’ve seen since the end of the recession. That was achieved the prior month when the U-3 unemployment rate hit a trough of 5.6 percent, the lowest rate since June of 2008.
The problem here is that today’s labor market contains plenty of slack that is picked up by the more comprehensive measures but not by the official unemployment rate. To illustrate this point, I took all the months since 1994 in which we’ve had 5.6 percent unemployment, according to the U-3 measure* and then looked at the U-4, U-5, and U-6 unemployment rates from those months. And as the reader can see, the U-3 unemployment rate is capturing far less of today’s labor market slack than it has at any point in the past (at least any point that we have data for):
This problem also shows up when we look at the employment-to-population (EPOP) ratio for working-age (25 to 54 years old) Americans, which I’ve previously argued is the best measure of employment:
In every instance, December 2014 is a significant outlier. Between 1994 and 2008, all months with U-3 unemployment rates of 5.6 percent had U-4 unemployment rates of either 5.8 or 5.9 percent; in December 2014, it was slightly higher at 6.0 percent. Excluding December 2014, the average U-5 unemployment rate associated with a U-3 rate of 5.6 percent is 6.6 percent; in December 2014, it was 6.9 percent. The average for the U-6 rate is 9.9 percent, yet in December 2014, it was much higher, at 11.2 percent. And finally, the average for the working-age EPOP ratio is 79.5 percent, 2.5 percentage points higher than December’s working-age EPOP ratio of 77.0 percent. This tells us that in December there was far more slack in the labor market than we would normally associate with a U-3 unemployment rate of 5.6 percent.
To be clear, I’m not arguing that I have a better estimate of what the U-3 rate will be when, or if, we achieve full employment. Instead, I’m arguing against using the U-3 measure in the first place. When groups like the CBO and the Fed refer to a FEUR of 5.4 percent, what they’re effectively referring to is a “FEU3R”, a “Full Employment U-3 Rate”. But this obsession with the U-3 rate is clearly misplaced. In the case of the Fed, this preoccupation is especially worrisome: if the Fed prematurely comes to the conclusion that we’ve reached full employment and raises interest rates in 2015, millions of Americans could be thrown out of work. So when we look at the state of the labor market today, it’s vitally important that we disregard the U-3 unemployment rate – or, at the very least, we should recognize that right now it is significantly understating the level of slack in the labor market. The U-3 measure isn’t the holy grail of unemployment rates, and it shouldn’t be treated as such.
*Sixteen months had a U-3 unemployment rate of 5.6 percent. These were November of 1994; January, May, June, September, November, and December of 1995; January, April, and May of 1996; February, April, May, and June of 2004; June of 2008; and of course, December of 2014. I also included an average rate of unemployment from November and December 2001. The unemployment rate was on the upswing in late 2001, and unemployment jumped from 5.5 percent in November to 5.7 percent in December. Since the unemployment rate averaged 5.6 percent during those two months without ever actually hitting 5.6 percent, I thought it would be useful to include some measure of unemployment from that time period.
