BLS Jobs Report, March 2011: Is there really a there there?

In today's Bureau of Labor Statistics (BLS) report, the two main stories for the March numbers are that 216,000 jobs were created (from the establishment survey) and the unemployment rate declined to 8.8% (from the household data). This is not entirely unexpected. As I have said in the last couple of years February-April have been, per the BLS, the best months for job growth. We need to wait for the June report covering May to get a real idea of where the economy is headed, that is that the economy is doing better than our eyes are telling us it is. Despite these two good numbers in today's report, I was struck by how many times the report stressed that there had been little or no change in most areas.

Normally I begin with the household survey but since the big number today was the job number from the establishment survey I thought I would start there. The 216,000 is the seasonally adjusted number. The seasonally unadjusted number was only 25,000. So once again this comes down to how much you trust the BLS seasonal model. To be blunt, I do not trust it. My feeling is that it is growth weighted and will tend to assume growth even when there is none. In any case, our kleptocratic elites are going to spin this like crazy that it is a sign of a real, sustained recovery.

Even if you accept the BLS' seasonal modeling, 50.8 thousand of those 216,000 were temp jobs. And the model continues to predict only light job losses in the public sector with the number of state and federal positions remaining essentially unchanged and with local government losing 15,000.

In the household survey, March is one of the slow months for growth in the working population (149,000) so the number of jobs needed to keep up with population growth is especially low (87,000).

The household survey shows opposite trends between seasonally adjusted and unadjusted numbers from the establishment survey, that is the unadjusted numbers show much higher volatility. Unadjusted, employment rose 869,000 in March to 138.962 million. Adjusted, it rose 291,000 to 139.864 million. A sizable discrepancy (902,000) between these two numbers remains even with the large increase in the unadjusted number.

We see something similar in the unemployment numbers as well. Unadjusted, the unemployed dropped 482,000 to 14.06 million. Adjusted, the unemployed declined 131,000 to 13.542 million. The difference between the adjusted and unadjusted numbers is 464,000.

Taking employment and unemployment together, the discrepancy between the adjusted and unadjusted numbers is 1.366 million with the adjusted numbers both increasing employment and decreasing unemployment relative to the unadjusted figures. That's important because it represents 0.9% of the labor force.

With the big jumps in the unadjusted numbers, the unemployment rate declined 0.3% from 9.5% to 9.2% and the employment-population ratio had a similar increase from 57.8% to 58.1%.

The adjusted numbers reduced these swings with unemployment decreasing 0.1% to 8.8% and the employment-population ratio going up the same amount to 58.5%. This is just a restatement of what we said above: the adjusted rates, as with the numbers they are based on, start with lower unemployment and higher employment than their unadjusted counterparts. Importantly, the report notes that the percent of long term unemployed (those unemployed more than 27 weeks) increased from 43.9% to 45.5% of the total.

The broader U-6 measure of unemployment dropped unadjusted half a percentage point to 16.2% or 25.178 million Americans. Adjusted it decreased from 15.9% to 15.7% and represented 24.462 million.

The overall disemployment number based on the BLS figures unadjusted is 29.028 million or 18.2%. Adjusted disemployment is 28.312 million or 17.7%, down from 17.9% in February.

For my own determination of disemployment., I first calculate the BLS undercount by multiplying the civilian noninstitutionalized population over 16 (NIP), that is the potential labor force, by .67, the ratio of the actual labor force to the potential labor force characteristic of a solid recovery. .67(239 million) = 160.13. I then subtract the current labor force number from this: (160.13 million – 153.406 million (adjusted)) = 6.607 million. 2.4 million of this is already part of the U-6 so I subtract that out, leaving 4.207 million. I add this to the U-6 numbers above, getting a disemployment number of 29.385 million or 18.3% unadjusted and 28.669 million or 17.9% adjusted.

For comparison, the BLS estimates of its undercount: Not in labor force, want a job now are 6.25 million unadjusted and 6.509 million adjusted.

The other significant data in the BLS report are that overall neither wages nor hours increased. With the rise in commodity prices, this means in terms of wages Americans lost some ground relatively speaking in March.

While temp hires can be an indication of an improving job market, the change in those hires is a very small part of the overall employment situation. What we should really be seeing is an increase in hours, and that isn't happening. At the same time, the plight of the long term unemployed continues to worsen. And as I just said, even those with jobs are losing ground against inflation. Finally, the 216,000 jobs, are they real or a modeling error? I know it is confusing but this is why I'm giving both the unadjusted and adjusted numbers as often as I can now because it is important to see the discrepancies between them. I used to think the BLS numbers were indicative of what was happening in the economy. Now I don't think so. You see there should be some correspondence between them and what we see in our daily experience. We should be able to point to big companies hiring or new small businesses opening. As it is, it seems almost as if these are phantom jobs. The BLS says they're there but nobody can find them.