David Glasner asks some very good questions about real rates of interest. Here is one bit:
…what we see is a steady decline in real interest rates from over 2% at the start of the initial QE program till real rates bottomed out in early 2012 at just over -1%. So, over a period of three years, there was a steady 3% decline in real interest rates. This was no temporary phenomenon; it was a sustained trend. I have yet to hear anyone explain how the Fed could have single-handedly produced a steady downward trend in real interest rates by way of monetary expansion over a period of three years. To claim that decline in real interest rates was caused by monetary expansion on the part of the Fed flatly contradicts everything that we think we know about the determination of real interest rates.
These days, Fed policy is more contractionary (if only verbally) and real and nominal rates are rising. The most plausible account of this process is that flows not stocks drive the market, and thus the Fed has enormous power over real rates, although like Krugman I find this tough to swallow. “Most plausible” does not equal “plausible.”
(By the way, the older literature does find liquidity effects on short-term real interest rates, but it is not unusual for them to disappear well within the space of a year; see for instance this John Cochrane piece (pdf). That said, Cochrane also finds a liquidity effect on the twenty-year note, which combined with the temporary nature of the short rate effect seems to suggest an unexploited profit opportunity. Note by the way that Cochrane did his dissertation on this topic. Or try this Grier and Perry paper (pdf). Evans and Marshall (pdf) find no significant liquidity effect on long rates and that is the same Charles Evans who is now president of the Chicago Fed.)
Another possibility is that we are seeing a recovery and real rates of return are rising. I don’t find that crazy per se, but it doesn’t explain why the rise in real rates marched in lockstep with Fedspeak. I don’t think “the Fed has private information about the recovery, so their announced contraction caused everyone to break out the champagne” is going to work, especially once we look at asset price movements.
Krugman yesterday argued that his earlier prediction was not wrong about interest rates, but he doesn’t commit to a mechanism for what is driving real interest rates up as of late and the plausible mechanisms do seem to contradict his (and my) earlier views. So what if the Fed backs off on QE over an extended period of time? That shouldn’t be significantly driving real rates ten years out unless of course one believes in the “primacy of flows” view, which Krugman does not. He’s already told us he doesn’t see a major recovery, or even much hope from the obsolescence of capital, so he can’t invoke rising returns from the real side of the economy either, even if we set aside the problems mentioned in the paragraph above.
Another “way out” is to claim we are mismeasuring inflation and also real interest rates, but I don’t see much promise in that. The most plausible CPI mismeasurements span very long time periods (say now vs. 1900), not now vs. ten years out. Alternatively, one might argue “inflation will be higher than we think in years 3-10 looking forward, so the ten year real rate isn’t actually up very much.” That will fail on numerous grounds, for one thing, money just got tighter.
So we have one implausible theory — the primacy of flows view — and a lot of worse contenders. The only people with a good predictive record on this particular issue are the Old Old Keynesians (e.g., money pushes around real interest rates more or less forever) and perhaps some of the cruder, pre-adaptive expectations Austrian views, although neither of those approaches has a good predictive record more generally.
Arnold Kling makes numerous good points here. He suggests that the profession’s understanding of real and nominal interest rates is extremely poor. He is right.
In the meantime, we should put more effort into thinking about how a version of the flows view could be made more plausible.