Due to Falling Unemployment Claims in January, Recession Odds Have Fallen From 35% to 15%

Last week there was a CD post featuring the graph above of initial claims for unemployment benefits, which was mentioned on Greg Mankiw's blog.

Now a recently released study by labor economist Tim Kane for the Joint Economic Committe of Congress "Employment Numbers As Recession Indicators" reports (p. 12) that:

The most surprising finding, contrary to conventional wisdom, is that both payroll and household employment measures are of no value as recession indicators. The 1-month change in the unemployment rate has value, but is perhaps prone to false signals due to its moderately high variance. Finally, we have learned that unemployment insurance claims are very valuable and reliable pre-recession indicators. The fact that the UI data are released weekly makes it even more timely, and so it merits close attention.

From the Executive Summary: The best pre-recession employment indicator is actually weekly claims for unemployment insurance (UI).

From the WSJ:

Kane based his model on the three-month change in the unemployment rate and initial jobless claims. Both rose in December, which pushed up Kane’s model to signal 35% recession odds, which was still below what many on Wall Street and academia have thought.

Yet the surprising decline in weekly jobless claims this month to around 300,000 — which is usually consistent with a very healthy labor market — has brought those chances down to around 15%-16%, Kane said.