The Household Survey: A Note on the December 2, 2011 Report

We do not typically see 40 basis-point declines in the unemployment rate in a single month.

The decline that we would like to see would be this one: a 60 basis-point decline generated by an increase in employment, partially offset by a 20 basis point-rise generated by an increase in labor force participation. That would show very strong demand for labor by firms. And that would show workers confident that demand for ther labor would be strong and hence reentering the labor force.

That is not what we saw in November with our 40 basis-point decline in the household-survey unemployment rate to 8.6%. What we saw in November, instead, was a 15 basis-points decline in the unemployment rate generated by more people at work. That is welcome. But most of the decline was a 25 basis-point decline generated by a fall in labor force participation. Workers as a group did not become more optimistic about their long-run employment opportunities, but rather less. That is not welcome. It is harder to pull people into employment if they are out of the labor force than if they are in the labor force and unemployed. Hence the fall in the labor force participation rate leads us to mark down the long-term potential output growth path of the American economy.

All in all, therefore, this decline in the unemployment rate is bad news: a little bit of good news about the current state of the labor market, coupled with rather more bad news about the sustainable level of potential output in the United States.

The inconsistency between the household survey increase in employment and the much smaller payroll survey increase in unemployment reinforces the bad news. From the payroll survey we would have expected the household survey employment-to-population ratio to be flat. The employment gain is the household survey is thus likely to be partially due to sampling error, and thus offset in future months.

I would rather have had a decline in the unemployment rate from 9% to 8.9% coupled with flat labor force participation rate than this household employment report. Would I rather have had a completely flat report for both the unemployment rate and the labor force participation rate? I am not sure. That is a tough question.

How large should we mark down our estimate of the long-run economic growth path of the United States from today's household employment report? How much should we mark down our estimate of the long-run economic growth path of the United States from the household employment reports of the past two years? Two years ago, after all, the recession was over. The employment-to-population ratio was where it is. The labor force participation rate was a full 100 basis points higher.

If you believe that the unemployment rate follows a stationary stochastic process without much memory--that it by itself encodes the state of the business cycle and thus the gap between current and potential GDP--then the past two years' employment news would lead us to mark down the future growth path of U.S. potential output by 1.54%. Given a trend level of real potential GDP right now of $15T/year, a projected real growth rate of 2.7%/year, and the assumption that real GDP is a cash flow of equity riskiness with an equity required real rate of return of 6%/year, that would mean that the bad news about labor force participation over the past two years has reduced our estimate of the present value of future American real GDP by:

($15T x 1.54%)/(6% - 2.7%) = $7T

$7 trillion of bad news from the fact that over the past two years, as the employment-to-population ratio has flatlined, labor force participation has fallen.

Now surely this is an overestimate. Surely the unemployment rate is not a sufficient statistic for the state of the business cycle and the gap between actual and potential output.

But by how much is this an overestimate?

How would we estimate the bias?

And if the bias is small, then the U.S. government's delaying the actions to jumpstart the recovery that would turn it from a jobless one into a real recovery costs Americans an extra $300 billion/month in foregone future economic growth, over and above the current $90 billion/month in idle workers and factories that could perfectly well be producing useful goods and services without triggering inflation.

For if the bias is small, than that is the right estimate of the cost of having a jobless rather than a real recovery: $400 billion/month--$5,000/month for each family of 4--with 1/4 coming from the current output gap and the delay in returning to the normal of potential output, and 3/4 coming from the damage that the awful state of the labor market is doing to the skills and attachment of the American labor force.