Scott Sumner  

The Fed: What's the minimum acceptable accountability?

Does Burning Your Money Make Y... A Partial Defense of David Fri...

Fed officials don't like the idea of a rigid policy instrument rule, such as the Taylor Rule. They don't even like less restrictive policy targets, such as 2% inflation; to be achieved in any way the Fed chooses. Currently the Fed operates under a dual mandate, but it has trouble spelling out what that means. For instance, they recently reduced their estimate of the natural rate of unemployment. A few years ago the estimate was 5.2% to 6.0%. Then it was reduced 5.2% to 5.5%. Now the estimate is 5.0% to 5.2%

These reductions point to the fact that there are other variables the Fed cares about, outside of inflation and unemployment. The Fed has looked at variables like part time employment (unusually high) and nominal wage growth (unusually low) and determined that the labor market has more slack than the official (U-3) unemployment rate indicates. That's fine. They want to keep rates low for a bit longer, and lowering their estimate of the natural rate of unemployment allows them to do that.

So let's say the Fed doesn't want to be told what to do. And they don't even want to be told to come up with their own rigid policy rule. It still remains true that the Fed must in some sense want to do SOMETHING. It must have some sort of policy objectives. And that means it ought to be possible to ascertain how effectively they have achieved those objectives, at least in a qualitative sense.

The Fed accepts the notion that it impacts inflation and unemployment by shifting the AD curve to the left and right. They do not control aggregate supply. In that case, it ought to be possible, once all the data comes in, to ascertain whether the outcome has been too much AD, or too little AD, relative to the Fed's hard to define policy objectives. That is what I'd call the minimum level of accountability.

My proposal is that after each meeting the Fed be instructed to provide a brief summary of the outcome of its previous monetary policy decisions, based on the latest available economic data. Again, the Fed is trying to achieve SOMETHING. Its only way of doing so is shifting AD to the left and right. So after each meeting they need to tell us whether, in retrospect, it would have been desirable for AD to have been higher or lower than what actually occurred.

This minimum level of accountability would not force the Fed to come up with any specific composite variable for its various inflation and employment goals. Merely to tell us whether, in retrospect, demand had been stronger or weaker than it would have liked.

I see this as a first step toward accountability. An absolute minimum level of accountability for a democratic society that delegates an important policymaking role to an unelected committee. But my hope is that it will also help the Fed to clarify its own thinking on monetary policy. Obviously it would be awkward to undershoot your AD target for 30 or 35 consecutive meetings (as they arguably did after the fall of 2008.) It would also help clarify decisions such as ending QE1 and QE2. Not that they couldn't end these programs with AD below target. They could do so based on forecasts of the future direction of aggregate demand. But then they would have to evaluate those forecasts at a later date. Was it wise to end QE1? How about QE2?

It would also help Congress. If the Fed said, "on balance, AD is about where we'd like it," then Congress would know that it's pointless to engage in fiscal stimulus. If they said AD is lower than they'd like it to be, then Congress could have a more intelligent conversation with the Fed. "Why is AD too low?" "Are you guys out of ammo?" "Do you want us to do more fiscal stimulus?" Note that this conversation might well lead to a fiscal policy outcome that I don't like (i.e. fiscal stimulus) so I'm not trying to stack the deck and set up a procedure that would lead to a market monetarist outcome. Rather I believe the first step in policy reform is clarity and accountability. And we don't even have that. The Fed does not currently evaluate whether its past decisions were wise, even in retrospect. That not science, it's medieval mysticism. It's "the secrets of the temple," and all the other accusations made by populists on both the left and the right.

Individual members of the Fed do occasionally admit to the Fed making mistakes in previous decisions. But we need an official vote of the entire FOMC, indeed preferably all 19 decision-makers, including nonvoting members. Was AD stronger or weaker than desirable? Even more accountability would be desirable, but that minimum level of accountability is a good first step.

Maybe in the more distant future they'd agree on single variable that best encapsulates the amorphous concept "aggregate demand." (I wonder what that single variable might be?) Then they'd set a target for that single variable. Later they might start targeting the forecast, i.e. setting their policy instrument such that they expect that AD proxy to grow right on target. Today, that all seems like a distant dream. But every long journey must begin with a single step.

PS. Bernanke has admitted that the Fed caused the Great Depression. I'd like those admissions to occur sooner than 70 years after the mistake occurred.

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (9 to date)
Brian Donohue writes:

Great post. This is an excellent drum to beat until your arms get tired.

Charlie writes:

We could start by just having a member of congress asking the Fed chairs during their regular testimony.

Michael Byrnes writes:

Isn't part of the problem that the FOMC is made up of individuals with widely differing views on what the goal should be?

sourcreamus writes:

Wouldn't this undermine the independence of the FED? Accepting responsibility for things going wrong would generate alot of publicity and attention from populist politicians. This would have to make the Fed more responsive to public sentiment. Would this be a good thing?

mark e writes:

How can there be a push for accountability at the Fed when there is almost none in any other area of the political spectrum?

Politicians constantly spout "tell-em-what-they-wanna-here" bromides to cheering crowds, and then rarely follow-thru, and are never held accountable to these campaign promises.

Scott Sumner writes:

Thanks Brian.

Michael, You said:

"Isn't part of the problem that the FOMC is made up of individuals with widely differing views on what the goal should be?"

Widely? They've all accepted the 2% inflation target, haven't they? There are differences in their estimate of the natural rate of unemployment, but those differences are small.

But even if you are right, that point has no bearing on the proposal in this post. The proposal is that the Fed be required to issue these reports. If they do it by a majority vote of the 19, that's fine.

sourcreamus, No institution should be so independent that it's not at all accountable. And this is the bare minimum of acceptable accountability.

Mark, Elections?

Jose Romeu Robazzi writes:

Great post, you stated your case without even mentioning the NGDP magical letters, actually i'm the first to mention it. But that is a diversion, the other thing this post shows unequivocally is that the FED should have a mandate to try to achieve AD stability, not all sorts of other things people expect from the economy and banking institutions ...

Scott Sumner writes:

Thanks Jose.

AS writes:

What if instead of a top-down banking system, we had a bottom-up banking system that didn't require central oversight? This entire problem of monitoring the Fed seems intractable, as most central planning problems are.

Comments for this entry have been closed
Return to top