Today, I’d like to explain the cause of the current high inflation. Not the part of inflation caused by supply problems; rather the portion caused by excessive nominal spending, sometimes called excessive aggregate demand.

The proximate cause of this excessive spending is obviously an overly expansionary monetary policy. But I’m interested in the deeper cause; why did the Fed allow spending to grow so rapidly?  I’ll show that the fundamental problem is that the Fed has the wrong policy framework. 

In 2020, the Fed adopted flexible average inflation targeting (FAIT).  That’s actually a great idea!  So why didn’t it work?  The problem is that the Fed interpreted FAIT in a very strange way.  They decided to correct undershoots of their 2% inflation target with higher than 2% inflation in the future, but decided against correcting overshoots of 2% inflation with below 2% inflation in the future.  That’s why we are where we are—the Fed didn’t make the policy symmetric.

But why did the Fed make this mistake? The deeper problem is that the Fed doesn’t understand its role in the economy.  It sees itself as a sort of fireman, fixing mysterious “shocks” that occasionally hit the economy.  In fact, the Fed is more like an arsonist, occasionally over or under-stimulating nominal spending. This creates business cycles and unstable inflation.  

But why should confusion over the Fed’s role have led them to the asymmetric form of FAIT, correcting inflation undershoots but not overshoots?  The answer is not at all obvious, and indeed didn’t occur to me until today.  So here’s my theory:

Suppose you adopted the fireman view of the Fed.  Why might that make asymmetric FAIT seem more appealing?  The answer is pretty simple; recessions are much more painful than booms.  If the economy is in a recession, output is low and hence living standards suffer.  You’d like to correct that problem as quickly as possible, and FAIT allows for a more rapid recovery than a simple “let bygones be bygones” inflation targeting approach.  In contrast, periods of economic overheating are much less costly, as employment and consumption are quite high.  Indeed the main problem with overheated booms is that they make future recessions more likely.  

[As an aside, supply side inflation is more painful than demand side inflation (and is politically very unpopular as we now see), but in this post I’m only addressing excess nominal spending, i.e. the inflation from excess demand.]

If I believed that the Fed was a sort of fireman, fixing problems caused by mysterious economic “shocks”, then I’d also support asymmetric FAIT.  In that case I would be less critical of the Fed’s mistake in 2021.  Even in that case they would have raised rates too slowly (in retrospect), but the error would have been forgivable.

But I believe the Fed is more like an arsonist, creating economic shocks through unstable monetary policy.  That’s unforgivable.  And as we will see, this is why the Fed needs to adopt a symmetric FAIT policy.

The real reason to adopt FAIT (or better yet NGDP level targeting) is not to get a faster economic recovery; it’s to prevent the recession from occurring in the first place.  I use italics, as this is the central point of this post, and indeed of much of my blogging over the past 13 years.  The mere expectation that the Fed will quickly push the economy back to the trend line will make the recession much milder than otherwise, as current nominal spending depends heavily on future expected nominal spending.  Investment projects cancelled in a deep prolonged recession will not get cancelled if the recession is expected to be followed by a quick recovery back to the trend line.

But this argument is entirely symmetric.  If the Fed were expected to offset a period of above normal NGDP growth with a future period of below normal NGDP growth, then the initial overheating will also be less dramatic.  Even if the Fed is slow to raise rates during an unexpected upswing in inflation, the markets will do their work for them.  Longer-term interest rates will rise sharply in anticipation of the future Fed policy moves required to hit the 2% average inflation target.  You correct overshoots with future undershoots not because overshoots are painful, but to prevent overshoots from occurring (or at least make them milder).

In 2021, I was too complacent about inflation because I thought the Fed actually had a symmetric policy in place.  (Indeed some people at the Federal Reserve also seemed to view the policy as symmetric.)  I assumed that the Fed would do whatever it takes to assure that inflation averaged roughly 2% in the longer run.  I thought average inflation targeting meant average inflation targeting. 

In retrospect, the Fed did an exceedingly poor job in communicating its policy change.  There was very little discussion of the fact that the new policy would mean that future inflation would average more than 2%.  Most discussion implied that this would assure an average inflation rate of 2% over time, instead of the below 2% inflation experienced during the 2010s.  Instead, they went from one extreme (too low inflation in the 2010s) to the other extreme (too high inflation in the 2020s.)  They should have gone from below 2% inflation to 2% inflation, on average.

Jim Bullard once suggested that the “flexible” part of FAIT should be interpreted as something like NGDP level targeting.  I wish that were what the Fed intended.  Under NGDPLT, you do allow average inflation to occasionally vary from 2% during supply shocks, as long as NGDP growth stays on track. This would have allowed for somewhat higher inflation during the current supply shocks, but without the current overheating.  If the Fed wants to fix what went wrong in 2021, then they should announce that henceforth FAIT will be symmetrical, and also that the “flexible” part of FAIT implies that some variation in inflation will be allowed as necessary to stabilize NGDP growth. 

Of course the simplest solution would be to move directly to NGDPLT.  But the politically easier path would be to do a FAIT that mimicked NGDPLT for a decade or two, and then formally switch over if the policy seems effective.  Given that the Fed screwed up its inflation targeting policy over the past 12 months, now’s not the time to formally abandon inflation targeting.