As I wrote about last month, economists keep expecting NFP (nonfarm payroll) numbers to come in positive though this has yet to happen (except the paltry 4,000 increase in November but that was a revision and doesn't count). For the NFP January numbers that came out last Friday, analysts at BoA, DB, JPMorgan, Barcap, and MS were predicting in excess of 25,000 new jobs were added during January. Just like in December, NFP numbers have disappointed at -20,000 jobs.
Its surprising that even though non-farm payrolls showed the economy lost 20,000 jobs, the unemployment rate still fell. This decline is especially strange considering the economy needs to add about 125,000 jobs a month just to keep up with demographic trends. This seeming contradiction is a result of seasonal adjustments (from CreditWritedowns):
Notice on the chart above that seasonally adjusted unemployed decreased from 15,267 in December 2009 to 14,837 in January 2010. The unadjusted numbers, however, tell a completely different story, showing that the number of unemployed actually increased from 14,740 in December 2009 to 16,147 in January 2010.
Seasonal adjustments are based on historical regressions. The problem is that economic trends are significantly different today than they were in the past. The recession has severely impacted hiring and firing trends. As best I understand it, the Bureau of Labor Statistics' model expects high levels of firing in January as the holiday season comes to an end. Since hiring was weak to begin with, there isn't much fat left to trim. The BLS' model overcompensates with its adjustments, assuming traditional firing levels, and computes the unemployment rate as lower than it is.
So the "bad" NFP numbers are actually worse than they seem. Furthermore, the -85,000 January result was revised to -150,000! Initial unemployment claims are still as high as they were last summer on an absolute basis:
The unemployment picture still looks pretty terrible.
