What causes recessions?

More specifically, what causes non-Covid recessions? Here’s Nick Rowe’s answer:

At some level, Nick and I basically agree.  But I am going to answer the question slightly differently.  It’s not necessarily wrong to say recessions are caused by money hoarding, or by interest rates being set too high, or by a drop in spending on C+I+G+NX, but I don’t find any of these explanations to be the most useful way of framing the issue.

In my forthcoming book (July 16, 2021), I argue that if monetary policy X is the most reasonable way to prevent recessions, then recessions are caused by not doing monetary policy X.  More specifically, I believe the most effective way to prevent recessions (except the Covid recession) is to stabilize 1-year forward NGDP growth expectations at around 4%/year.  So in that sense, recessions are caused by sharp declines in NGDP growth.

As an accounting matter, NGDP is the monetary base times base velocity.  So you could say that Nick is describing the empirical fact that declines in M*V are mostly due to declines in V.  And Nick would probably agree with me that it’s the central bank’s job to offset decreases in velocity by increasing the supply of money during periods where money demand is increasing.  (Lower velocity is equivalent to higher money demand.)

The basic monetary model of recessions is symmetrical, and thus falling NGDP could be produced by either an increased demand for money or a decreased supply of money.  That raises an interesting question.  Are recessions merely “errors of omission”, periods where the central bank fails to accommodate an increase in money demand, or are some recessions caused by declines in the growth rate of the (base) money supply.

Here’s the US monetary base growth rate from 1918 to the mid-1960s:

The sharp drop in base growth seems to have lagged in the 1920-21 recession, but that’s a bit misleading.  That recession was quite mild during the first 8 months of 1920, and thus the (plunging) supply of base money actually correlates pretty well with falling NGDP during the short but severe 1920-21 slump.  Base growth also turned negative in 1929-30, although velocity also declined sharply during that year.  Base growth fell especially sharply before and during the 1937-38 slump.  It also turned negative right before the 1949 recession, and the growth rate fell to roughly zero during the three Eisenhower recessions.

Of course this is all very unscientific.  I’m pretty sure that in an accounting sense the money demand shocks that Nick refers to played the larger role.  I’m not arguing for a simplistic monetarist explanation of recessions.  But I’d make two additional points:

1. The various slowdowns in monetary base growth were errors of commission, not errors of omission.

2.  To some extent the accompanying slowdowns in velocity were endogenous, caused by the slowing economy, which itself was caused by the slowdown in the growth in the monetary base.

Distinctions between errors of omission and errors of commission are quite fuzzy, and in my view not particularly useful.

Here’s the more recent data:

Now the correlation seems weaker, although in a few cases you can still see money growth slowing slightly during recessions.  Perhaps the most interesting case is 2007-08, where base growth slowed to near zero as we fell into recession, before soaring much higher in late 2008.  I’d make two points about the latter period:

1. After the Fed began paying interest on bank reserves (IOR) in late 2008, the demand for base money soared much higher.  The Fed accommodated that demand with various QE programs, which sharply boosted the growth rate of the monetary base.

2.  Even without IOR, the decline in market interest rates to near zero would have led to a big rise in money demand, i.e. a big fall in velocity.

Of course the decision to pay IOR is also an “error of commission”.  Thus the Fed triggered the Great Recession in early 2008 with the error of commission of sharply reducing the growth in the monetary base, and worsened the recession in late 2008 with another error of commission—paying IOR.  That’s not to say those two errors completely explain the Great Recession; errors of omission also played a big role.  Rather it’s important to recognize that the problem was not solely errors of omission.

Of course this is all just accounting.  One could also say the recession was caused by the Fed allowing NGDP growth to fall sharply.  Or that the recession was caused by the Fed not adopting a regime of 4% NGDP growth, level targeting.  Or that the recession was caused by the Fed not pegging the price of one year forward NGDP futures contracts at a level 4% above its current level (in 2008).

PS.  Nick lives in Canada, which has a more stable banking system than the US and thus more stable demand for base money.  During the Great Depression, the Canadian base fell much more sharply than the US monetary base, as they did not have banking panics.  So at least in that episode, the crude monetarist model works better for Canada than the US.  Was that an error of commission?  Or should we excuse the BOC because they had to reduce their monetary base under the international gold standard?  On second thought, I guess I won’t blame the BOC for the fall in the Canadian base during the early 1930s, as the Bank of Canada did not yet exist.