Why Productivity Slowdown is Likely Largely Illusory

Some bearish commentators, such as Nouriel Roubini, Paul Kasriel and Michael Shedlock argues that employment growth is overestimated due to the flaws of the so-called birth/death methodology the Bureau of Labor Statistics use. They're probably right. However, unless output growth is lower, a lower employment growth only implies higher productivity growth. Output growth in itself may be overestimated due to the likely underestimation of inflation in official price indexes, but considering the more positive tone of private sector indicators, it seems reasonable to assume that growth is still positive, if only slightly.

Anyway, there are increasing signs that employment growth was probably underestimated back in 2002-2005, while it is likely overestimated now. Something which in turn implies that productivity growth was overestimated in 2002-2005 and underestimated now. The reason for this, aside from the birth/death model, is the factor of illegal immigration. In 2002-2005, employment growth was much higher in the household survey than in the payroll survey. Given the fact that most of the employment growth was concentrated to hispanics, it seems reasonable to assume that employed illegal immigrants not reported by employers made up the difference. But while employment growth was underestimated, output growth was not underestimated, meaning that the alleged productivity boom in 2002-2005 was largely an illusion.

Now however, we see how construction spending is falling-while construction employment have increased somewhat. Most of the productivity decline is a result of this surprising statistic. With illegal immigrants being overrepresented in construction, this probably reflects that they're the ones losing their jobs now-after having received all the jobs during the housing bubble.

This is also confirmed in the fact that the hispanic employment rate have fallen a lot more than the overall employment rate. Between January and May, it fell from 65.5% to 64.6%, compared to a overall decline from 63.3% to 63.0%.

Between January 2002 and January 2007, the hispanic employment rate rose more than average (despite a lot faster population growth), from 64% to 65.5%, compared to the overall increase from 62.7% to 63.3%.

All of this indicates that the productivity slowdown is largely a statistical illusion, with productivity growth being overestimated in the past and underestimated now.