More Inflation Yes, Inflation Targeting No!

Scott Sumner has two new posts up that speaks to something that has been bugging me lately: the increasing popularity of an explicit inflation target for the Fed.   Many bloggers, including myself, have been calling for the Fed to create more inflation or at least stabilize inflation expectations.   I have been particularly vocal on the latter point.  Others have been  more forceful in their call for an explicit inflation target as a means to increase the inflation rate.  All along, my reason for arguing for the Fed to stabilize inflation expectations is that doing so would indicate the Fed is stabilizing expectations of future  aggregate demand (given that productivity does not appear to be contributing to drop in inflation expectations).  Such actions, in turn, would also serve to stabilize current aggregate demand as  well.  

Now, I have never been enthusiastic about stabilizing inflation as an end in itself.  The reason being is straightforward: inflation is a symptom, not an underlying cause.  More generally,  the percentage change in the price level could be the result of shocks to  aggregate demand (AD), aggregate supply (AS), or both.  Currently, it seems clear that the drop in inflation expectations and the drop in core inflation are reflecting faltering aggregate demand.  Thus, it makes sense to talk about the need to arrest these drops.  However, it need not always be the case--low inflation could also be the symptom of a positive AS shock (e.g. productivity boom).   Imagine, for example, aliens land and give us new technology that makes our computers faster, gives us clean energy, and allows us to travel to  distant galaxies.  Such an alien encounter would create mother of all productivity booms.  Among other things, this productivity boom would imply a higher neutral interest rate, lower inflation rate, and robust AD growth (given  the increase in expected future income).  In such a case a rigid inflation target of say 4%, as some have proposed, would not make  sense here.  Most likely it would be too high an inflation rate to keep AD stable.   

Another way of saying all of this is that monetary policy should focus only on that over which it has meaningful control:  total current dollar spending or AD.  It should ignore AS shocks, both the good and the bad, because all it can do by responding to such shocks is to make matters worse  as alluded to above.   Focusing too narrowly on an inflation target--which assumes every shock is an AD one--can cause a central bank to make this very mistake.  I made this case before in  more detail in this post which got some play time in the economics blogosphere (e.g. Mark Thoma reposted it here).  My hope was that this post would help folks see that the stabilization of AD rather than inflation targeting should be the key  objective of monetary policy. 

So what kind of monetary policy would serve to consistently stabilize AD?  The answer is one that directly targets a stable growth path for AD. This could be a NGDP target or a final sales of domestic output target or any measure that directly aims to stabilize the growth of total current dollar spending.  Scott Sumner outlines its advantages in a recent post, but let me add a few thoughts.  First, an AD target is easy to implement.  All it requires is a measure of the current dollar value of the economy.  It does not require debates over the proper inflation measure, inflation target, output gap measure, and coefficient weights that plagued inflation targeting and the Taylor Rule.  Second, it can easily be made into a forward-looking rule by having the Fed targeting the market's forecast of AD.  This would require some innovations such as  NGDP futures market, but it is within the realm of possibilities as shown by Scott Sumner.  Finally, it could be easily communicated to and understood by the public by labeling it along the line of a "total cash spending target." 

Let me end by noting some of the famous observers who have called in the past for some form of AD targeting: Bennet McCallum, Greg Mankiw, Robert Hall, Menzie Chinn, Jeffrey Frankel, Martin Wolf, Samuel Brittan, and Frederick Hayek. I like this crowd, how about you?

Update: Karl Smith replies to Scott Sumner and me on this issue.  One point Karl brings up is the possibility that a NGDP target could increase macroeconomic instability.  Bennett McCallum addressed this issue a while back and showed that this need not be the case if there is forward looking behavior.  More recently, Kaushik Mitra showed that even with adaptive expectation-type behavior NGDP targeting can work well.