Last Friday, the government report that 169,000 new jobs were created in August, slightly below the expectations of economists and the average figures over the past year. Markets interpreted such weakness to mean that the Federal Reserve would be unlikely begin its tapering of bond purchases in September as rumored.
Now for the truth. According to the Bureau of Labor Statistics, the economy added 378,000 new jobs in August, but this number was "seasonally adjusted" downward to the reported figure of 169,000. "Well then, that's good, you say. The economy is stronger than we thought". No, not so fast. Last month when the government reported 104,000 new jobs (seasonally adjusted, of course), the economy really lost 1,186,000.
Confused? Allow me to clarify. The reason the government looks to adjust the monthly numbers is precisely so we don't all become schizophrenic as we see jobs swing from one million lost one month, to nearly 400,000 created the next. Precisely because there is a seasonality to hiring (for example retailers hire many temporary employees in the months before Christmas, only to let them go in January and college students flood the job market early in the summer, but then leave to go back to school in the fall), the government tries to make one month's figures more comparable to previous months by creating the adjustment. And it is this adjusted number to which the world so frenetically reacts. But the question remains, should we ascribe such certainty to a figure which is really just an (educated) guess?
The size of the monthly adjustments are huge. In fact, the median monthly adjustment (using data from January 1940 to the present), is more than 75% of the actual figure - and in the last decade, the adjustment has been bigger than the actual number in nearly 50% of the months.
I would also note, since the 1960's, the size of the adjustment (on a percentage basis) has grown as have the number of months where the adjustment is bigger than the actual figure. Usually when you've had practice doing something over and over, you get better. The government seems to be getting worse.
70's 80's 90's 00's
Pct of Months Where Adjustment > Actual 23% 29% 36% 48%
Median Size of Monthly Adjustment Relative to Actual 63% 69% 80% 97%
With such a margin for error, why do investors focus so intently on these monthly numbers? I argue, that they shouldn't.
If one really wants to understand the state of the labor market, a better way to use these figures is to look at a 12 month moving average of the unadjusted jobs created. By using a full year, seasonality is fairly scrubbed away and one does not need to rely on the accuracy (or lack thereof) of government adjusters.
When we do this we find:
- job creation remains anemic compared to historical norms (even moreso when thought of relative to the size of the economy and population)
- job creation is disappointing given the historic magnitude of the job losses that preceded the recovery
- job creation is disappointing given the amount of government spending that was employed to accelerate the recovery
- the monthly rate of job creation has not changed in the past two years
So, there really is no reason for the Fed to change their view of the economy as a result of Friday's numbers. The economy remains weak and is getting little better on the jobs front. Of course, any single monthly number can be used to justify any course the Fed decides to take, but from a fundamental perspective, the jobs picture has simply not changed.
