Monday, December 6, 2010

Why We Need an Explicit Nominal Target for Monetary Policy

So we learn from the 60 minutes interview with Ben Bernanke that if necessary there will be a QE3:
Scott Pelley: Do you anticipate a scenario in which you would commit to more than 600 billion?

Ben Bernanke: Oh, it’s certainly possible. And again, it depends on the efficacy of the program. It depends, on inflation. And finally it depends on how the economy looks.
This ad-hoc approach is exactly why we need a rules-based approach to QE.  If the Fed had adopted an explicit nominal target--preferably a nominal GDP level target--and forcefully committed to maintain it no matter the cost, it is unlikely the Fed would need to keep announcing new rounds of QE. All the market would need to know is that the Fed is serious about hitting its nominal target.  The rest would take of itself.  The market would do the heavy lifting by automatically increasing nominal expectations to a level consistent with the target. 

Instead, we get a piecemeal approach where explicit, large dollar security purchases are announced presumably to impress the market.  We  now know there are two problems with this approach.  First, it invites criticism that may hinder the Fed's ability to carry out QE.  Some observers, for example, are fearful that the large expansions of the monetary base under QE will become inflationary.  If these observers are influential they can create political pressure for the Fed.  If the Fed were instead aiming for an explicit nominal target where the expansion of the monetary base would be endogenously determined, this concern would be muted.  For there would be no need to announce large dollar  security purchases. And, if the market itself did most of the heavy lifting as noted above, the increase in the monetary base probably would be less than under the current QE program.  Second, this approach has not convinced the market that the Fed is serious about raising nominal expectations and thus nominal spending. This can be seen in the implied inflation rate from TIPS.

It is time to adopt an explicit nominal target. I nominate a nominal GDP level target.


  1. I always say "growth path." Is that terminology too indiosyncratic? To me, "level" targeting suggests a zero growth rate. Growth path, to me, suggests a series of levels growing at a certain rate. When you say, "growth path" you define the rate and the initial level (or the level at one point) and then you know what ngdp (or final sales) is going to be.

    Perhaps it is just that these days when economists say "price level" targeting, then mean a target for a growth path of the price level?

    Or is it that I am too specific. Level targeting really is thought about as a rule for growth rates. The growth rate is, say, 5%. If it is 6 percent, then it is lower, say, 4 percent for enough time before going back to 5% that it averages at 5 percent.

    I am just puzzled. Remember the guy who insisted that since GDP in europe is now higher than it was in 2007, it is doing just fine by "level" targeting? And frankly, if someone is focused enough on growth rates, this faster and slower growth business sounds difficult to predict. Always going up 5% sounds easier than sometimes 5%, then sometimes 4%, then sometimes 6%. But, if you define a growth path, these notion that the future is hard to predict because the growth rates might vary is absurd.

  2. If the target variable is P, rather than NGDP, the clearest terminology is "Price level path targeting". So, by extension, it should be "Nominal GDP level path targeting". Bit of a mouthful though.

  3. I don't think Bernanke bought himself any capital with conservatives when he waded into the income redistribution argument as well. Wow, I thought that was bizarre.

  4. Interesting blog. Kepp reading Scott Sumner.