US in Recession: A new definition

This post on WSJ blogs caught my attention.

August 18, 2008, 5:13 pm
UCLA Professor Says U.S. Is Still Far From Recession

Though U.S. economic activity remains subpar, it is still well above recession thresholds thanks to resilient industrial production that has offset a recession-like rise in unemployment, according to a UCLA forecaster in a National Bureau of Economic Research paper.

In his paper, Edward Leamer created an algorithm of three economic indicators — payroll employment, the unemployment rate and industrial production — that he said “nearly perfectly reproduces the NBER official peak and trough dates.”

“Bottom line: things have to get much worse to pass the recession threshold,” Leamer wrote.
The NBER, an academic association, is considered the official arbiter of whether the U.S. is in recession, a determination it usually makes many months after the fact. There have been nine NBER recessions since 1950 — the last one being from March 2001 to November 2001.

Leamer isn’t a member of the NBER business cycle committee.

NBER bases its determination on monthly indicators including personal income less transfer payments, employment, industrial production and manufacturing sales volume. According to NBER’s Web site, “there is no fixed rule about which other measures contribute information to the process.”

Leamer said the purpose of his model “is to take the guesswork out of the recession definition.” The thresholds include falling industrial production for six months at a rate of at least 6% per year; declining payroll employment for six months at a minimum 1% rate per year; and a six-month rise in the unemployment rate of at least 0.8 percentage point.

“Every recession since World War II has included months that satisfied all three of these limits. And there has never been a time in the expansions during which all three of these limits were satisfied,” Leamer wrote.
The one instance when Leamer’s algorithm didn’t line up with NBER was in the early 1970s. NBER had a recession starting November 1973. Leamer’s model had it beginning later in 1974.

As for 2008, the unemployment rate satisfies the recession threshold, but the decline in payrolls hasn’t hit the recession cutoff yet, while “the problems in industrial production are nowhere close to the limit,” Leamer wrote.
His paper examined data through June. In an email, Leamer said those trends hold even including July data released since his paper was written. –Brian Blackstone