Evan Soltas
May 29, 2012

The Missing Mini-Recession

So I've been thinking about this idea of the "mini-recession" -- low-duration, low-intensity periods of economic contraction -- and how the U.S. does not seem to have them, which Scott Sumner first brought to my attention in this excellent post in December.

Sumner's explanatory story was that "inertial central banks that target nominal rates and observe the macroeconomy with a lag might occasionally produce short contractions, typically 9 to 12 months." He looked at the change in unemployment rates, and found that increases beyond a small amount appear to trigger far larger recessions.

We see the same effect in the NBER's recession dating -- see the graph above, which is a histogram of the length in months of economic recessions (negative) and expansions (positive). Short recessions are very rare. Instead, recession lengths appear tightly clustered around 12 months, with a pronounced negative skew, i.e. there are more recessions of length > 12 m. than < 12 m. Similarly, expansion lengths are more loosely clustered around 24 months, with a positive skew.

This is the signature of the absence of mini-recessions.

My method to arrive at the graphed data above was pretty basic. I downloaded the monthly NBER business cycle dating data from fred, calculated using Excel the length of each recession or expansion, and then used Excel's "FREQUENCY" function to create the histogram. (The graph is in Google Docs so it can engage you interactively.)

It's more or less impossible to explain this without a lagged policy response.