With a key Fed meeting coming up, it’s worth thinking about the advice we’d give the Fed if for some bizarre reason they actually asked you or I for advice. Suppose we just had one shot, and we put zero weight on our own ego and 100% on the well-being of the public. Then what? Here’s what I would not do:
1. I would not recommend NGDP targeting.
2. I would not recommend level targeting.
Oh I would mention as an aside that I still think those are the two keys, but I see both as being politically impossible right now. The Fed just made a big deal about their commitment to an official 2% inflation goal, and they aren’t about to tear that up because of some outside advice. Even I can see that they’d look silly. That’s an issue to work on for the next recession. If smaller countries adopt it, maybe it will catch on internationally as inflation targeting did after New Zealand first adopted it.
It’s kind of an interesting challenge to think of the most aggressive policy stance that actually has a prayer of being adopted. Something that would not be viewed as being inconsistent with the past actions and comments of Fed officials, especially Bernanke. I’ll give you my thoughts, but I’d also like to know what commenters think. I see the need for a three-pronged attack:
1, Weekly QE as the policy instrument/communication device.
2. Target the forecast using internal Fed forecasts of core PCE inflation and unemployment
3. Emphasis on equal weights in the dual mandate, with inflation averaging 2% over the cycle.
First let me explain why this is politically feasible. The Fed has already done several QEs, so nothing radical there. They’ve lost their normal communication device (fed funds targets) and could use another. Bernanke has admitted that they’d cut rates if they could, which suggests a need for a new instrument. My commenter Benjamin Cole used to always recommend a given level of monthly QE until the targets were hit. Tim Duy recently mentioned it as well.
Target the forecast is also a respected procedure. The Fed’s never formally adopted it, but it has distinguished academic support (including Bernanke’s colleague Lars Svensson) and the Fed has informally nodded in this direction. When they cut rates after the 1987 stock crash, or 9/11, they were basically doing a forward-looking policy. The recent pattern of publicizing both the expected instrument paths (for short term rates) as well as expected paths for inflation and unemployment, is obviously extremely close to targeting the forecast. So why not just go all the way? Remember, I’m not asking the Fed to change its 2% inflation goal. But its short term inflation forecast may have to rise to fulfill the dual mandate.
And finally the dual mandate, which is widely misunderstood. Bernanke has often said that the Fed takes both sides equally seriously. Nonetheless many people seem to wrongly assume that the Fed is targeting inflation at 2% each and every moment in time. IF THEY DID THAT THEY’D HAVE A SINGLE MANDATE, AND THEY’D BE VIOLATING THE LAW. Sorry for all the caps but the Fed needs to make it really clear that if the dual mandate means anything, they can’t aim for 2% inflation at each moment in time. Rather inflation should average 2% over the cycle. It also means that any deviations from 2% inflation must help achieve its unemployment objective. Now even if we take the most conservative assumption, which is a vertical long run Phillips Curve, the Fed can hit the dual mandate with an inflation rate that average 2% over the cycle, but is allowed to rise above 2% when unemployment is above average, and run below 2% when unemployment is below average. (And recall that Krugman recently argued the Phillips Curve is not vertical at low inflation, with some evidence to back up his claim–so I’m actually making a highly conservative assumption in this post.) BTW, my approach would reverse the slope of the Phillips Curve—the dual mandate requires the Fed to aim for a positively sloped PC.
Let’s summarize. QE is not at all radical. It’s been used already. Why not use it systematically for the communication device that even Bernanke admits the Fed has lost and wishes he still had? “Target the forecast” is no longer radical. The Fed’s been clearly moving in that direction, step by step. And Bernanke has repeatedly emphasized that the dual mandate is taken seriously, and that the needs of the unemployed require the Fed to occasionally miss on inflation for part of a business cycle. Put it all together as follows:
The New York Fed will be instructed to buy $30 billion a week in T-securities of various maturities, until the Fed’s forecasting department is able to forecast a path of unemployment and inflation over the next 5 years that minimizes the sum of the deviation of inflation from 2% and unemployment from its long run average (or estimated natural rate.) At that point it will stop. If forecasts of inflation and unemployment change in such a way as to under or overshoot the previous expectation, the New York Fed will either buy or sell T-securities, as appropriate.
Any better ideas, which aren’t politically infeasible?