There is no question that we've been experiencing an episode of economic stagnation in which key indicators of output and employment are not growing at their usual historical rates. Some analysts take the position that this is all that really matters, and whether it meets the formal definition of a "recession" is a pointless question of semantics. In support of such a position, there are a number of economic models quite popular among academics today that would warrant just such a perspective, in that they presuppose linear dynamic systems in which a "recession" is indeed just an arbitrary definition you would make up to characterize a string of bad luck.

I espoused the opposing view in a recent academic paper. I argued that recessions represent distinct and objectively identifiable episodes in which the usual dynamic factors that drive economic growth-- technological progress, population growth, and capital accumulation-- are replaced by a distinctly different dynamic in which lost income in some sectors feeds back into declines in output for others. One of the defining characteristics of this phenomenon is the rapid rise in the unemployment rate that we've seen in every historical recession. It's very difficult to generate that kind of pattern from a system governed solely by linear dynamics, or interpret it from a perspective in which there's nothing special going on during a true economic recession.

BLS monthly civilian unemployment rate (from FRED), with NBER-dated recessions as shaded regions.

I was therefore quite alarmed by the 5.5% unemployment rate reported yesterday by BLS. Although the rate itself remains moderate by historical standards, the increase last month is the biggest monthly change that we've observed in over 20 years. King Banian, Paul Krugman, and Howard Gold argue the numbers aren't that bad, reflecting a surge in young job seekers rather than a big decline in employment per se. I personally share the more pessimistic interpretations of Jared Bernstein, Calculated Risk, Michael Mandel, and Mark Thoma; (you can find other interesting discussion from Phil Izzo, Barry Ritholtz, Andrew Samwick, and Angry Bear). It is true that the estimated May employment decline from the BLS payroll series was only 49,000, more moderate than the 285,000 drop implied by the household survey (from which the 5.5% unemployment rate is derived). But we've now seen 5 consecutive months in declining nonfarm employment as estimated from the BLS establishment survey, putting the overall level essentially back to where it was a year ago. Year-on-year employment stagnation is another thing we just don't see outside of an economic recession.

Year-over-year percent change in seasonally unadjusted number of U.S. workers on nonfarm payrolls, from BLS.

And while I'm collecting graphs that make me feel bearish, I can't resist updating this one, despite Menzie's cautions against reading too much into consumer sentiment numbers.

Reuters/Michigan index of consumer sentiment. Data source: FRED and Market Watch.

And how would one reconcile this view that the economy may already be in a recession with the very strong gains in the stock market on Thursday? Oops... never mind.

Source: Yahoo.

Technorati Tags: macroeconomics, employment, economics, unemployment, recession