The January Jobs Numbers

Last Friday, the Bureau of Labor Statistics released its report on the employment situation for the month of January. According to the payroll survery, 257,000 private-sector jobs were created in the first month of the year. Moreover, revisions added the number of people working in previous months. According to the household survey, the unemployment rate dropped from 8.5% to 8.3%. The news is positive but the magnitude of the change is still too small to affect the overall jobs outlook, which remains pretty miserable.

First, let’s put the new data in perspective:

The actual number of people employed by the private sector since 2006 is in blue. The recent acceleration in job creation is barely visible to the naked eye. The red line roughly represents how many people would have jobs under conditions resembling full employment. The green line projects what employment growth will look like if jobs are created at the same average rate as they have been since the trough in February, 2010. Finally, the purple line shows what will happen if the economy adds jobs at the rate seen in the past two months, which has been markedly faster than any time since the beginning of 2011.

As you can see from the chart, the recent acceleration in job creation is real. Before the new data that came out in the past few months, the gap between the green line and the red line would not close until 2031. Now it will close in 2026. However, even if jobs continue to be created at the same rate as they have been, we will not return to conditions resembling full employment until 2019. That is not a recovery. That is a catastrophe.

For comparison, this is what an actual recovery looks like:

What about the drop in the unemployment rate? Since the start of this downturn, the official unemployment rate has been deeply misleading because millions of people have been disappearing from the official statistics. The result is that the official unemployment rate is far too low. Even the Congressional Budget Office has commented on the discrepancy. Between October and December, 170,000 people disappeared from the dubious number known as the labor force, which makes the decline in the official unemployment rate during that time (from 8.9% to 8.5%) less impressive.

Remember, the labor force generally should grow at the same rate as the population over short periods, although over longer periods demography makes a difference. There is no reason to think that demography can explain what happened in the past five years, however:

One piece of good news from the recent data is that the job creation that occurred in January was accompanied by an increase in the official labor force, which is now at its highest level in absolute terms since April, 2010. This means that the drop in the official unemployment rate was not fake, even if the level is still far too low. In fact, if we believe the employment numbers and use a more realistic number for the labor force (the red line above), the real unemployment rate has been collapsing since the summer:

On the downside, the recent decline in unemployment mirrors what occurred between December, 2009 and April, 2010. After declining from 12.3% to 11.6% in five months, the real unemployment rate rose steadily until it reached 12.3% again in November, 2010, after which the real unemployment rate stayed above 12.0% until this month, when it dropped to 11.6%. Is the American job market actually improving or is this another false dawn? At this point it is simply too early to tell.

One wrinkle you may have read about elsewhere was the sudden increase of the number of people officially counted as “not in the labor force.” In the span of one month, 1.2 million people were added to this statistic. It turns out that this was caused by the discovery of about 1.7 million people previously ignored by the BLS and that most of those people were elderly and not working:

The “civilian noninstitutional population” is defined as every American over 16 who is not in the military overseas, in prison, or in a mental hospital. I generally refrain from using this statistic when considering the health of the American labor market because crime and war should not be used to inflate the private employment/population ratio. I prefer to use the total population, which is produced by the Census Bureau rather than the BLS.

The point of this chart is that you can clearly see that the January increase in the number of people not in the labor force matches the unusual increase in the noninstitutional population. Moreover, as discussed above, the official labor force actually increased in January. If it had declined by the same amount as the number of people not in the labor force grew, then the recent drop in the unemployment rate would be bogus. This does not excuse the BLS for letting six million people disappear from the official statistics over the past few years, but they should not be accused of any special act of malfeasance this month.