Over the last couple of years Scott Sumner has published posts from time to time on the MoneyIllusion that discuss the consequences of attempting to raise interest rates with tighter policy and I think can be briefly summarized as predicting the impact of temporarily raising short rates only to have them become lower later. I’d bet that this is what we are seeing here in the recent flattening of the yield curve, a sort of ultra slow-motion train wreck in progress that became more pronounced as OMO’s to reduce the balance sheet began in October. Eventually, short rates will have to be “normalized” or we take the consequences.


When I ponder public discussions of monetary policy by policymakers having to do with plan of “removing accommodation,” it stirs up the illusion of what some may judge as easy policy with nominal rates being lower than the Wicksellian rate. But five 25-basis point rate hikes later, the flattening yield curve makes it appear as if policy really wasn’t wildly accommodative at all, and characterization of those discussions as exaggerations of the circumstances is reasonable.

In a very recent post, however, Sumner makes the following claim:

It’s difficult to evaluate Fed policy in isolation; one needs to consider the regime over an entire business cycle.  Thus the current sub-2% inflation rate is entirely consistent with the dual mandate, assuming that inflation is appropriately countercyclical (which is what the mandate implies.)  But in the past inflation has tended to be procyclical (in violation of the dual mandate), in which case current policy is inappropriately tight.

So how do we know if the Fed policy today is appropriate? We don’t know, and won’t know until the next recession. If inflation rises during the next recession, then current policy will have been appropriate—even though inflation is now under 2%. If inflation falls during the next recession then current policy will have been inappropriately tight. Based on past experience, the latter assumption is more plausible, but again, current policy is exactly right if the Fed were taking its dual mandate seriously. That mandate calls for slightly above 2% inflation during periods of high unemployment, and slightly below 2% inflation during periods of low unemployment. And right now unemployment is well below average. The dual mandate calls for sub-2% inflation at this point in time.

If we are to take the Fed seriously about its mandate, it’s useful to consider the rate of inflation required in order for it to average 2% inflation over the medium term should a recession occur, and imagine the equation one would need to compose in order for the Fed to be consistent in its own interpretation of its mandates. To do otherwise would be to hand the Fed a pass on current policy omissions and the consequences of credibility loss on the nominal anchor in a tight-biased regime. I am unwilling to give the Fed a pass because the omissions aren’t cost-free, and my preference is to prevent further damage rather than wait until it adds up unnecessarily and becomes noticeable to just about everyone.

Movement in the headline unemployment figure is useful in detecting nominal influence like when policy has been much too tight, but whether its present level is lower than the peak of the previous business cycle after a trend of decline spanning a handful of years is not of much consequence. Likewise, it doesn’t seem appropriate to me to say that because that figure continued to decline and is at a low point now, tight policy is appropriate or even exactly right when there is another side to the mandate that I assume is the reason the Fed has an inflation target: achieving the preamble of managing policy commensurate with the economy’s long run ability to increase production. At what unemployment level would tight policy not be fine, 5%, 7%, 10% – when?  Without a nominal target, one simply cannot tell whether any given level is appropriate.

My suggestion is that the NGDP indicator of policy that currently has an ~2% southern gap from the trend that the Bernanke Fed ended with would be more useful in characterizing the recent performance of monetary policy, assuming we might agree that is where the trend line should be drawn. I can understand that hashing out where to set the NGDP trend line in order to normalize discussions about the Yellen Fed’s performance might expose points that would be inconsistent with the claim that Yellen did a spectacular job. But setting the NGDP trend line in 1Q 2011 would provide a reasonable comparison that is likely to be most charitable to Yellen due to allowing the bygones that drawing the line earlier as I have seen other do does not.