Friday, August 12, 2011

Kocherlakota Elaborates

Here is Narayana Kocherlakota's explanation for his dissent with respect to the most recent FOMC decision. As I mentioned in the comment section of my previous post, Kocherlakota actually has some New Keynesian (NK) leanings, and his dissent is entirely consistent with an NK view of the world. In this view, a Taylor rule dictates monetary policy. In Taylor-rule land, only the current state of the world, defined by the current inflation rate and the current unemployment rate, matters (though some Taylor rules have anticipated inflation on the right-hand side). On those terms, the current state actually looks better than the state in November 2010. Inflation is higher and unemployment is lower, which should dictate a less accommodative policy, not a more accommodative one.

Thus, if we buy NK, Kocherlakota's dissent makes sense. However, this guy, for example, seems to think that Kocherlakota is ridiculously hawkish and callous toward the unemployed. Seems he just wants to enforce New Keynesian consistency though.

I live in non-NK land (or willing-to-be-convinced-but-still-unconvinced-NK-land), so if Kocherlakota's dissent makes sense to me, and it can make sense in NK-land, it must be good, right?

25 comments:

  1. Kocherlakota's dissent seems rather mild to me, whereas critics seem to portray him as being that fringe nut in the back shouting about Zimbabwe.

    What concerns me about the language is that it risks the Fed's hard-earned credibility. Nobody has a good enough model to project inflation and employment out to mid-2013 - especially in the odd times we are in now with an extended period of near-zero IROR and a huge quantity of excess reserves. To tie your hands the way the Fed has with pretty clear ignorance about the path of inflation going forward seems pretty foolish. Let's say in late 2012 inflation is accelerating, are they going to renege on their promise and raise the policy rate - reducing their credibility. Or, are they going to just let inflation keep rising until mid-2013.

    If the FOMC wanted to make a long-term policy commitment to a certain rate path, they needed to introduce qualifiers. They might have said that we will keep the policy rate near-zero through mid-2013, unless inflation exceeds some unacceptable value. That would have been more sensible and realistic about the limitations of their models of inflation, and wouldn't pose potential harm to their credibility in the future.

    ReplyDelete
  2. Steve,

    "This guy" is Mike Konczal. Why don't you use his name?

    PS

    ReplyDelete
  3. I've done that before. I'll use "this guy" when I have a dim view of "this guy's" economics writing. For example, I call Yglesias "this guy," to the extent that I ever mention him.

    ReplyDelete
  4. Since we are on the topic, a question for you about the last Kocherlakota flare up (last August or so?).

    I remember it involving a lot of discussion of Wicksell vs Fisher.

    Do you think this post captures the essence of Kocherlakota's position?

    http://coffeehouse-economics.blogspot.com/2011/08/kocherlakota-vs-selgin-amv.html

    Specifically, the author points out that Wicksell's concept of the spread between the market interest rate and the natural rate breaks down if you substitute rational for adaptive expectations. A spread just cannot exist. Does this do an accurate job of explaining the models you and Kocherlakota work with?

    ReplyDelete
  5. Steve,

    I know you've done that before, and it's always sounded dismissive to me. Guess I was right, though I hadn't assumed it was deliberate.

    FWIW, I think this is a mistake. Having a dim view of someone's work isn't an excuse to be disrespectful. I think it reinforces the impression by some (not me, yet), that you're being arrogant and pulling rank.

    PS

    ReplyDelete
  6. Pete,

    I figured that was where you are going. Do you have a different opinion of Konczal's stuff? If so, point me to something good. I'm always willing to be convinced.

    ReplyDelete
  7. JP,

    "Do you think this post captures the essence of Kocherlakota's position?"

    No, not really. In the thing from a year ago, what K is saying is quite conventional. Suppose the fed funds rate is 5% and the Fed then pegs the rate to zero. What happens? In the short run, we think there is a short run nonneutrality of money. For whatever reason, the real interest rate falls. However, if the Fed pegs the nominal rate at zero forever, then, given the neutrality of money (or, roughly, super-neutrality of money) in the long run, we must get deflation in the long run. That's all he's saying.

    "Does this do an accurate job of explaining the models you and Kocherlakota work with?"

    Narayana and I work on some models that are the same and some that are different. Recently, he has revealed himself as something of a New Keynesian, and I would not describe myself that way. The argument he is giving for his policy decision is essentially a New Keynesian one.

    ReplyDelete
  8. "For whatever reason, the real interest rate falls."

    These are the liquidity effects that are outlined in the papers mentioned here?

    http://artsci.wustl.edu/~swilliam/papers/syl10.pdf

    ReplyDelete
  9. Yes, there are basically two effects on the nominal interest rate, a liquidity effect (or real-rate effect) and a Fisher effect. The Fisher effect comes from anticipated inflation, and there is not much in the way of contention about that. The liquidity effect is more contentious. In the papers on my reading list, that comes from a segmented markets friction. In New Keynesian models you can get it from sticky prices. In the short run, the liquidity effect dominates, and in the long run, the Fisher effect dominates.

    ReplyDelete
  10. "In the short run, we think there is a short run nonneutrality of money. For whatever reason, the real interest rate falls. However, if the Fed pegs the nominal rate at zero forever, then, given the neutrality of money (or, roughly, super-neutrality of money) in the long run, we must get deflation in the long run."

    But isn't the critique that you won't necessarily make it to the long run, that with the rate pegged at zero you can instead get a short run where inflation keeps increasing to infinity, and the currency becomes unused?

    "In Taylor-rule land, only the current state of the world, defined by the current inflation rate and the current unemployment rate, matters (though some Taylor rules have anticipated inflation on the right-hand side). On those terms, the current state actually looks better than the state in November 2010. Inflation is higher and unemployment is lower, which should dictate a less accommodative policy, not a more accommodative one."

    The current state may be better, but in NK models is it the direction or the absolute level? Because the absolute level of unemployment still looks very high and the absolute level of inflation, even headline, still looks pretty low.

    And again, why do you think the markets have so little concern about inflation, with 5 year T-bonds at about 1% and 10 year at about 2%? People who favor structural explanations for unemployment tend to also think that the markets are very efficient – and they don't mean just rational investors by that; they mean that prices are controlled by marginal investors who are not just rational, but also have tons of public knowledge in their heads, years and years of education and training in investments, lots of time and willingness to analyze that information, and lots of cash to push prices before their portfolios become too undiversified for it to be worth it. But how can the markets be so efficient, and our policies be so inflationary at the same time?

    ReplyDelete
  11. So if a model finds that liquidity effects are low to non-existent in the short and long run, and this model does not allow for a Wicksellian cumulative process to break out because a persistent spread between the real interest rate and the market level of the real rate is impossible, then you would always hold that raising interest rates leads to inflation (and has no real effects)?

    ReplyDelete
  12. 1. You'll have to explain to me what a "Wicksellian cumulative process" is.

    2. Suppose you have a model where there is no liquidity effect. There are lots of models like this. Standard cash-in-advance, or money-in-the-utility function models give you that, and you can be more explicit about it - e.g. Lagos-Wright. Now, consider two alternatives, a low money growth rate forever, or a high money growth rate forever. The higher money growth rate will give you a higher rate of inflation and a higher nominal interest rate (the Fisher effect). What happens to the real interest rate? It may be unchanged, but maybe not - depends on the model. One possibility is a so-called Tobin effect. With a higher inflation rate, people substitute from money to other assets, including capital, which lowers the marginal product of capital and the real interest rate. I think the upshot of the empirical evidence is that the Tobin effect is small, so it's reasonable to think that higher inflation will have neglibible effects on the real interest rate in the long run. What about other real effects? Inflation introduces an intertemporal distortion in essentially any monetary model and that is costly in terms of aggregate welfare. Anticipated inflation matters, and so does unanticipated inflation, and there are important empirical arguments over how large these costs are.

    ReplyDelete
  13. Steve,

    First, sorry for the delay in getting back to you. I've been thinking about how best to respond, and have decided that my first inclination was the best.

    Which is...you're missing the point. Frankly, that seems to be something else you do fairly often (in addition to the "this guy" stuff).

    I don't really care about your opinion of Konczal's economic reasoning, or anyone else's. I'm also not interested in trying to change that opinion. My point was to argue that having a negative opinion of his writing isn't an excuse for being rude, arrogant & dismissive. Said another way, referring to him by name doesn't mean that you agree with everything -- or anything! -- he says.

    How often have you seen Krugman/DeLong/Thoma refer to you as "this guy," no matter how strongly they disagree with you? (Have a grad student search their blogs if it's too much for you to bear.)

    PS

    ReplyDelete
  14. "Now, consider two alternatives, a low money growth rate forever, or a high money growth rate forever. The higher money growth rate will give you a higher rate of inflation and a higher nominal interest rate (the Fisher effect)..."

    You never explained the low money growth rate forever scenario. I guess its just symmetrical to the high money growth rate, and also has no effect on the real rate (absent liquidity effects)?

    If one's model has no liquidity effects, then the Fed is irrelevant, it would seem.

    ReplyDelete
  15. Re: Wicksell's cumulative process. I don't think I can do it justice.

    But Thomas Sargent, who if memory recalls has been mentioned on this blog, models the Wicksell effect in "Commodity Price Expectations and the Interest Rate". See: http://www.jstor.org/pss/1883997

    What do you think?

    ReplyDelete
  16. Pete,

    Thank you kind sir for your extremely helpful advice. I could not be more thankful, and have taken it to heart. I will try ever so hard to live up to the standards of decorum of the honorable Krugman/DeLong/Thoma, particularly Mr. DeLong.

    ReplyDelete
  17. "If one's model has no liquidity effects, then the Fed is irrelevant, it would seem."

    No, they still control inflation, which matters. This is an issue that is at the heart of debates about monetary policy and how to conduct it. Everyone agrees that the Fed can and should control inflation, and that a high rate of inflation is a bad thing. People disagree about what causes short-run nonneutralities of money and their quantitative significance.

    ReplyDelete
  18. Are there any New Keynesian models that take Wallace's dictum seriously? Have any tried to incorporate Kiyotaki/Wright search theory for instance? Is the divide between NK and New Monetarist models really that set in stone? Do you guys talk together at lunch break? (Yes, I know, off topic but since you are generously taking questions...)

    ReplyDelete
  19. Randy Wright and I explain some of this in our two "New Monetarist Economics" pieces. In the Handbook of Monetary Economics chapter, there is an example of sticky prices in a New Monetarist model. It's pretty simple and doesn't have everything in it that you might want, but it gets at the general idea. Sticky prices and monetary frictions are not incompatible. Woodford has never taken to deep models of money, though. I asked him about it, and he said something like "there are so many frictions, and it's so complicated that I would rather not deal with it." There is plenty of interaction in the profession among people with different ideas. For example, I have known Woodford for a long time, and I think he is an excellent economist who deserves to be taken seriously.

    ReplyDelete
  20. Interesting that Woodford would say that he doesn't disagree with deep models, only that he thinks they're too complicated. If that's the sole line of contention, what prevents the two (NK and NM) from being folded into each other? Surely there must be deeper and more contentious issues at the heart? Are the policy prescriptions similiar? Is it politics?

    ReplyDelete
  21. Yes, it's no problem "folding" these things together, and there's nothing preventing anyone from doing it. You'll still get the NK policy implications of course, but then optimal policy will give you some tradeoff between mitigating the sticky price/sticky wage frictions and mitigating the other frictions (monetary exchange and credit frictions) in the model. One thrust of New Keynesian economics was a desire to simply get the job done. A modern model was needed that captured the key innovations in modern macro but also had a non-neutrality of money and could be useful in guiding policy. This is what the NK people delivered, and it wasn't like the deep money people had a ready-made policy tool. Some of that has changed, but NM people have not as yet done much quantitative work.

    ReplyDelete
  22. I can understand why putting money into the utility function is done. It's a quick and dirty way to get over the hassle of introducing money into a model that has no reason to contain money.

    Why do some models include assets like government bonds in utility functions or production functions?

    ReplyDelete
  23. It's a short cut. Long ago, Neil Wallace convinced me that this was a bad idea. Read the introduction to "Models of Monetary Economies," and Wallace's paper in that volume (available on the Minneapolis Fed web site). Also look at "A Dictum for Monetary Theory," by Wallace.

    ReplyDelete
  24. Ok, I see. Putting money into utility functions leads to a problem: what constitutes the specific item money? Putting all paper assets in the utility function heads the question off. But then you lose the idea the fiat money, equities, bonds etc are all intrinsically useless.

    ReplyDelete
  25. "The Fisher effect comes from anticipated inflation, and there is not much in the way of contention about that. The liquidity effect is more contentious. In the papers on my reading list, that comes from a segmented markets friction. In New Keynesian models you can get it from sticky prices. In the short run, the liquidity effect dominates, and in the long run, the Fisher effect dominates."

    I was googling "negative Fisher effect" yesterday and stumbled on a bunch of papers by yourself. I couldn't quite understand what you meant by that term. How does it fit into the above story?

    ReplyDelete