Scott Sumner  

Always double check your claims with the AS/AD model

No, Bill Poole, the Fed Does N... Giving Thanks...

Benny Lava sent me an article discussing the German opposition to Mario Draghi's proposed monetary stimulus:

But on Monday, Lautenschlaeger broke with normal etiquette to publicly criticize this stance, saying that ever looser monetary policy had its limits and that money printing had yet to stabilize sinking price inflation, its formal goal.

"I don't see any reason for further monetary policy measures, especially not for an extension of the (asset) purchase program," she said, ten days ahead of a meeting at which policymakers will set the ECB's course.

"It buys time but does not heal the structural causes of a slack economic recovery... We should give the numerous and massive monetary policy efforts time to take full effect."

Suppose you tried to illustrate Lautenschlaeger's concerns using as AS/AD model. The most straightforward approach would be to shift the AS curve to the left. But this would cause higher inflation, whereas the eurozone is facing excessively low inflation, and is trying to bring inflation up to its 1.9% target.

One could argue that structural problems might also interact with monetary policy, indirectly leading to tighter money and a drop in AD. But the Germans insist that money is easy. And if money were tight, then you'd want monetary policy to offset that tightness, and prevent AD from declining in response to a fall in AS. If Lautenschlaeger believes structural problems are causing low inflation, then the solution is more monetary stimulus. To summarize, a casual look at the AS/AD model suggests that the Germans have no plausible explanation for why the eurozone faces sub-par inflation rates.

The Fed is currently wrestling with the issue of when to raise rates. Janet Yellen and Stanley Fischer are both fans of the Phillips Curve, which predicts that inflation will rise as the economy reaches full employment. I'm much less confident. Consider the following:

Fed policymakers say confidence that inflation will rebound should be enough to pull the trigger on rates for the first time, most likely at their Dec. 15-16 meeting. The second move may need more proof that prices are in fact rising, so its timing will offer a better clue to how the tightening cycle will unfold.

It may also show to what extent the Fed still relies on the theory of an unemployment-inflation trade-off commonly known as the Phillips curve.

"You hear arguments inside and outside the (Fed's policy) committee that in this circumstance you should just look at inflation by itself," Bullard, who will vote on rates next year, said. "Maybe it is the appropriate thing to do in this environment,(but)if we are really going to give up on the Phillips Curve at the heart of central banking that would be a major change," he told Reuters from his St. Louis office.

There are two ways of thinking about Bullard's comments. But first let's think about how central banks should view the Phillips Curve, which is the idea that inflation is a procyclical variable, rising during booms and falling during recessions. If we refer back to the AS/AD model, it's easy to see that the Phillips Curve is only correct when the economy is being buffeted by AD shocks. When there are AS shocks, then the inflation rate is countercyclical. But here's the irony---the Fed is supposed to try to prevent AD shocks, and can't do anything about AS shocks. So if the Fed is doing its job, then the Phillips Curve model will be wrong. And not just slightly wrong, but the opposite of the truth. If they are using the Phillips Curve to predict the economy, then they are using a model that can only predict accurately when monetary policy is inept.

So does that mean that Bullard is wrong? Not necessarily, and that's because there are actually several issues the Fed needs to address---the business cycle and also the trend rate of inflation. Perhaps Bullard is implying that if the Fed were not to raise interest rates, then AD would rise, causing both output and inflation to accelerate. That would be a procyclical inflation rate, which violates the Fed's dual mandate. Fed policy is supposed to make inflation countercyclical.

So then maybe the Fed should raise rates to keep inflation below 2% during the coming boom, so that the Fed can run inflation over 2% during the next recession. That would be consistent with their mandate. But there's just one problem; the Fed doesn't have a system in place to produce higher inflation during recessions. During the previous recession, inflation fell to zero, and there's no reason to suppose inflation won't fall again during the next recession. In that case, if the Fed tolerates sub-2% inflation during the coming boom, and inflation falls during the next recession, then their long run 2% inflation target will lose credibility.

I see this as the biggest dilemma facing the Fed. The policy that will smooth the cycle right now, and would be appropriate if the Fed were doing NGDP targeting at 3%, is tighter money. But this would also cause them to miss their inflation target in the long run, if inflation goes back to being procyclical in the next recession.

That's why I keep saying the real issue is not whether the Fed does or doesn't raise rates in December, but rather whether the Fed is able to finally come up with a way of meeting their inflation/employment dual mandate, which requires a countercyclical inflation rate.

PS. Some people are confused as to why the dual mandate requires a countercyclical inflation rate. To see why, first imagine a single mandate, with inflation stabilized at 2% year after year, regardless of unemployment. That's a sort of benchmark. Then ask how things would be different if the Fed also cared about stabilizing unemployment. In that case, policy would be a bit more expansionary during periods of high unemployment, and vice versa. This monetary policy would be a bit more countercyclical than a straight up inflation targeting single mandate, and hence inflation will also be a bit more countercyclical than under simple inflation targeting. Since inflation is 100% acyclical under pure inflation targeting, making it somewhat more countercyclical in a relative sense, also makes it countercyclical in an absolute sense.

Comments and Sharing

COMMENTS (5 to date)
Don Geddis writes:

The "P.S." was a very helpful clarifying comment. Thanks!

ThomasH writes:

"the Fed doesn't have a system in place to produce higher inflation during recessions."

I guess that could be true given some definitions of "in place." The Fed certainly does have instruments to produce inflation during AD recessions. The fact that there was too much heat from "macro media" for the Fed to just keep buying stuff until the market twigged is another matter. According to the quote, Draghi has the same problem.

Scott Sumner writes:

Thanks Don.

Thomas, It's much worse than that. The key mistakes made by the Fed and ECB occurred before they hit the zero bound, and before monetary stimulus became controversial.

dbeach writes:

Whereas I agree that the dual mandate implicitly demands countercyclical inflation, it's not at all clear that the Fed believes that. I don't recall anyone from the Fed back in 2009 saying they needed to get inflation above 2%. The most expansionary members of the FOMC supported the Evans rule, which permitted 2.5% inflation, but did not prescribe it.

James Alexander writes:

Lautenschlager wasn't all bad. Her opposition to more QE seems a bit half-hearted, possibly for internal German consumption only, and the (obviously backward-looking) data isn't looking so bad.

Comments for this entry have been closed
Return to top