Scott Sumner writes:
I don’t think Keynesians should be arguing that lower real interest rates are the key to recovery. A bold and credible monetary stimulus that was expected to produce much faster NGDP growth might well raise long term risk-free interest rates.
This point doesn't get explained well - and I probably won't do any better, but I will give it a try anyway. A simple way to think about this is the basic IS-LM story (without wanting to get into a big debate about the efficacy of IS-LM):
In this version, the IS curve has shifted so far to the left that it intersects with the LM curve at the horizontal section - the zero bound problem. If I set i equal to the nominal interest rate and assume positive inflation, this translates to a negative real rate. Full employment, however, is only consistent with a nominal interest rate below zero, which implies a lower real interest rate as well. Given the zero bound on nominal rates, Keynesians (using the term loosely; labels can get sloppy), turn their attention to reducing the real interest rate.
Given the zero bound, we talk about ways to shift the IS curve to the right. Usually, these discussions take on two forms. The first is fiscal policy via deficit spending, which I am very confident will do the trick, but I am also very confident it really doesn't "fix" the economy. The instant you back off the fiscal accelerator, the economy falters. In my mind, fiscal policy is undoubtedly necessary in the near-term as a stop-gap measure, but in the long-term is leading the US down the Japanese path of endless defict spending.
The second policy response is monetary, typically raising inflation expectations. This in turn lowers real interest rates - and this is the important part - at all nominal interest rates. This, like fiscal policy, induces a rightward shift in the IS curve:
At the zero bound, higher inflation expectations lowers the real interest rate, hence the Keynesian preoccupation. But I think the key here is the rightward shift of the IS curve past the zero bound "kink" at which point nominal and real rates begin to rise and we lift off the zero bound. We can talk about different mechanisms to accomplish this, but moving sustainably beyond that kink should be the ultimate policy goal.
Thus, ultimately I think you can have a focus on negative real interest rates as a stop on the path to Sumner's desired outcome. And I completely agree with Sumner (and I think I am paraphrasing him correctly here) in that the failure of interest rates both real and nominal to rise represents an absolute, unmitigated, unacceptable, and quite frankly irresponsible failure on the part of the Federal Reserve:
One would think the Fed would sit up and take notice that the US government sold 10 year debt at a record low interest rate today as a sign that they need to do more, not less. Notice also the failure of either real or nominal rates to get a boost after Operation Twist. This is evidence of the pointlessness of that effort. For all the grief I have given St. Louis Federal Reserve President James Bullard, he certainly had it right last year when he said:
A strategy aimed at lowering longer-term borrowing costs, sometimes referred to as a twist operation, would help drive down longer-term borrowing costs for businesses, economists say.
But James Bullard, president of the St. Louis Fed, said the effectiveness of such a strategy is questionable.
"A twist operation would not have very much effect," Bullard told Reuters Insider in an interview. "It's been analyzed many times, and the general tenor of that analysis is that it did not have very much effect."
Finally, notice that I also put the variable "confidence" into the specification for the IS curve. Here I am offering another mechanism by which we can think that QE has an impact by signalling that policymakers have an intention and a desire to maintain the pre-recession path of nominal spending (here I am paraphrasing Brad DeLong). The failure to maintain that path has undermined confidence in that agents now have less certainty in the future path of income. If policymakers let the path of nominal spending shift downward once, why should we not expect them to do it again?
Bottom Line: I don't think what Sumner describes as a Keynesian focus on negative real interest rates is inconsistent with his views on what should happen in the presence of a credible monetary policy committed to actually lifting us off the zero bound.