Sunday, March 27, 2011

Andy Harless pulled back from comments

quotes of my old post in quotes. His comments in plain text. My replies in italics.

"QE2...had no measurable effect on anything"

Yes, because it was just coincidence that every single economic indicator turned around and started giving positive surprises instead of negative ones almost the very day QE2 was announced.

what very day was QE2 announced ? You will note that Janet Yellen said that there was almost no sign of any effect on the day of the official announcement because it was anticipated. I believe that QE2 can only affect the economy via real interest rates (either lower nominal rates or higher expected inflation). So, since real interest rates went up when it was officially announced in November and are now at the same level as they were in August when Bernanke first publicly mentioned QE2, I believe it has had no measurable effect on anythying

My prior is that the effect isn't exactly zero, but after looking at the data, I can't believe it is large. Your claim about indicators should be specific. In particular tell me who was surprised. I certainly haven't been plesantly surprised in the past 14 months

"a 'more expansionary' monetary not would require the ability to precommit to making inflation high when unemployment is low and there is no way anyone can convince anyone that he will do that."

Nonsense. Bernanke has already convinced many people (admittedly almost all lunatics) that he will do that. On the margin, the easier his policy now, the more near-lunatics will be convinced that he will let the inflation rate rise. There is a wide range of opinions about the likely future course of monetary policy, and there are plenty of things the Fed can do to influence the mean of that distribution. In this respect, recent Fedspeak constitutes a drastic tightening of monetary policy at a time when Yglesias and Drum rightly suggest that a drastic loosening would be in order.

hacks not lunatics. Republicans claim they believe that, but they don't put their money where there mouths' are. Forecast inflation can be read off by comparing TIPS and regular treasuries. Bernanke hasn't convinced anyone that there is going to be higher than declared target (2%) inflation. The green line in the FRED graph is the implied average expected inflation over the next 7 years from Treasury notes and TIPS. People say all sorts of things, but no one with serious money is betting on inflation

"current Fed policy is...extremely radically expansionary..."

Until one agrees on a measure of policy ease, this question is completely subjective.

Long long ago in 2007 people agreed on measures of policy ease. They include the Federal funds rate (can't be negative and is now zero or pedal to the metal) and open market operations. The Fed's liabilities are triple what they were then. The fact that some people say that monetary policy is not radically expansionary just shows that they have decided to redefine terms with well established meanings. Imagine going back to distant distant 2006 say. If you went back in time and told an economist, any economist that $500,000,000,000 in open market operations aren't radically expansionary, you would convince that economist that you were insane.

Certainly if you fit a regression of the federal funds rate to the core inflation rate and the unemployment rate, you will find that recent residuals have been strongly positive. Perhaps this was unavoidable because of the zero constraint, but given your contention that QE is ineffective, the amount of QE necessary to be consistent with past policy (i.e. to simulate the impact of the negative federal funds rate predicted by the regression) must be much larger than what we have actually seen. By this criterion, current Fed policy is extremely radically contractionary.

You just said that the unemployment rate is a measure of Fed policy. You declare your faith that monetary policy is always effective *and* that it is the only thing that affects the economy. Since the entire debate is over how effective monetary policy is right now, your assuming that you are right and I am wrong in your redefinition of a standard term is not responsible debate.

Or if you prefer to use quantities rather than interest rates as your basic criterion, we can look at money. To be properly interpreted, measures of the money stock have to be adjusted for changes in velocity. Therefore, ultimately, the only proper quantitative measures of monetary ease are measures of nominal spending. Of course all such measures plummeted in early 2009 and have not since recovered (indeed, they have diverged further from the earlier trend). So in terms of velocity-adjusted monetary aggregates, monetary policy has again been extremely radically contractionary.

What ? Why is velocity part of a measure of monetary policy. The money supply expanded enormously. That's what the Fed can do. When you say that a change in velocity is a change in Fed policy, you assume again that the Fed must always have complete control over the economy. You are attempting to prove your claim by redefinining terms such that it is tautological.

If "the standards of all central bankers post Rudolf Havenstein" indicate otherwise, then there is a problem with those standards, and perhaps we need to appoint central bankers who are better at measuring the ease of monetary policy.

5:57 PM

Andy Harless said...
"Economists tend to agree that monetary policy can't affect the unemployment rate averaged over a long period of time"

Umm...greasing the wheels effects? What are Tobin, Akerlof, etc., chopped liver? Not to mention the recent PLOG evidence from the IMF, which indicates that, as an empirical matter, large output gaps can persist for long periods of time (and person-years of unemployment thus accumulate) when inflation rates are low.

Akerloff is a founder of New Keynesian economics in which the effects of monetary policy are short lived. The statement about what economists generally think is absolutely conventional. Note for example Krugman

OK Akerlof. I googled "Akerlof" "long run neutrality of money" I got a pdf

The claim is the absolutely standard way of declaring that one is a new Keynesian. For example see Brad DeLong quoted by Markos Moulitsas

You will note that I don't agree with Akerlof, Krugman, DeLong and Mankiw on this one.
Moreover, even if we pretend that the actual long run Phillips curve is the vertical one about which they teach in Ec 10, most evidence (as well as the fact that the unemployment rate is bounded below at zero) suggests that the short run Phillips curve is convex. Over time, a monetary policy that misses the NAIRU will result in a net employment sacrifice even if the inflation rate at the end of the time period is the same as at the beginning.

Note again that I am not a New Keynesian and I don't agree with Akerlof, Krugman, DeLong and Mankiw on this one. For convenience models are used in which the short run Phillips curve is linear or linear approximations are used. I agree with you that this is a silly mistake. Non quadratic loss functions imply that stabilization policy can affect long run growth. However, most economists consider this negligible. This is simply a fact. Like you, I don't agree with most economists. Actually the non linearities aren't what convinces me. I believe in Hysteresis. I also am sometimes convinced that endogenous growth models have something to do with reality and the long run short run dichotomy breaks down (there is no convergence to what might have been in such models). I just said that almost all economists divide macro into long run and short run and claim that monetary policy affects short run (and so can affect welfare but not growth). Now they don't literally believe this -- it is a standard assumed because everyone assumes it thing. But it's just the way the profession is. I think I made it clear in my most that I don't agree.

Anyhow, it isn't clear that 30 years is really "a long period of time" for this purpose. The inflation rate has been trending downward that whole time -- which is to say, we have been mostly on right side of the long-run Phillips curve. I met a lot of people in 1981, and we are not all dead yet.

Well nothing is clear if you want to contest everything. It is very clear that 30 years is treated as a long period of time for that purpose. This widespread belief is used to identify time series models. The conventional meaning of long run and short run (as used by economists) are clear enough that no one can honestly claim that 30 years is not the long run to a conventional salt water New Keynesian macroeconomist. The models are simulated numerically. The time unit is defined when justifying parameters. New Keynesian models give an impulse response which fades to negligible in much much much less than 30 years. You must know this.

"Monetary currently perceived to be about as loose as it can possibly be"

And yet there are trillions upon trillions of dollars worth of debt instruments that the Fed is legally allowed to hold and that are not currently in the Fed's portfolio. How is it that those instruments are somehow invisible?

Here I concede that what I wrote was, at least, unclear. I really honestly don't consider anything that the Fed is allowed to do to be monetary policy. Bailing out big banks was Fed policy, but not monetary policy. There is more to central banking that monetary policy. To me we are at the limit of monetary policy, because the Fed funds rate is essentially zero. The Fed can still intervene in financial markets, but it would affect the eocnomy by affecting say Treasury bonds outstanding or AIG debt in the hands of private agents or something. It's liabilities are not at the moment anything special -- they are treated as perfect substitutes for T-bills.

No comments: