Sunday, August 28, 2011

Conversations with Brad

Tyler Cowen linked to one of my posts and seems to find something in there to like. Brad DeLong does not like it.

Cowen wanted to think of what I do as "neoclassical economics." Some people like to call it "micro-founded macreconomics," and Barro recently called it "regular economics." The people I have learned from, work with, and talk to do economics - they take theory developed by other people, build on it, match the theory to data, and attempt to evaluate macroeconomic policies in a sensible way. Some of those people are sympathetic to Keynesian ideas; some of those people are critical of those ideas.

Here's what Brad has to say:
From my perspective this "neoclassical macroeconomics" is merely Hicks (1937) (or perhaps Wicksell (1898)) "plus", as RĂ¼di Dornbusch liked to say, "original errors".
This will be a theme. The pristine ideas are apparently in the classics - Hicks, Wicksell - and we have just thrown sand into the gears.

I said something in my piece about how the inefficiency I wanted to isolate was not a Keynesian inefficiency. The Keynesian inefficiency is a too-high safe real rate of interest; my inefficiency was a too-low safe real rate of interest. Here's Brad again:
In the Keynesian-or perhaps it would be better to say Wicksellian--framework, when you say that real rates of return are "too high" you are saying that the market rate of interest is above the interest rate consistent with full employment, and with savings equal to investment at full employment. Wicksell called that interest rate the "natural rate of interest" and it is relative to that natural rate of interest that Wicksellian (and Keynesians) speak of interest rates being "too high" and "too low". Thus Williamson is wrong when he say that what we have now--when the natural rate of interest on relatively safe securities is negative and the market rate of interest is not--is "not a Keynesian [or Wicksellian] inefficiency". It is precisely such an inefficiency. To claim that it is not misinterprets Keynes (and Hicks, and Wicksell), and misleads readers trying to understand what they did and did not say.
I'm taking my cue here from Mike Woodford, for example "Interest and Prices." Woodford certainly thinks that he is channeling Wicksell, though maybe Brad thinks he's not. In a Woodford New Keynesian model, monetary policy is about moving the market nominal interest rate around, and the transmission mechanism for monetary policy works through the real rate. Once you hit the zero lower bound on the nominal rate, the real rate can't go lower. That's the way New Keynesian economists inside the Fed system frame the monetary policy problem in the current circumstances. The real rate is too high, you want it to be lower, but monetary policy can't do that in the conventional manner.

Here's the interesting part. Brad is characterizing the current state of affairs and says "the natural rate of interest on relatively safe securities is negative and the market rate of interest is not." What I think he is saying is that the safe real rate of interest is low, but the relevant "market rate of interest" is in fact high. I don't think you can find that feature in any "Keynesian" framework where you would be able to correct the problem through some kind of policy Brad might want to prescribe in the current circumstances.

But whether you can find it in the General Theory or not, Brad has brought up something very useful. The fact that the safe real rate is low is intimately related to the fact that the "market rate of interest" is high. In fact, you can find this in this paper. The effect is more pronounced in the financial crisis, but I think it persists. The idea is that greater uncertainty and higher costs associated with evaluating collateral and unwinding debt acted to increase interest rate spreads - the spreads between the safe rate of interest and "market rates of interest" - reducing the quantity of privately-produced liquid assets, creating the asset scarcity that lowered the safe real rate of interest. To understand that idea, you don't need to go digging in Wicksell, Hicks, Fisher, or the General Theory. It's elucidated much more precisely in the work of Rob Townsend, Doug Diamond, and other people who worked on modern information theory, contract theory, and the theory of financial intermediation.

Brad goes on:
A second thing I think is wrong is Williamson's claim that while things could be (or perhaps should be) improved by shifting the IS curve out and to the right...
Don't go there Brad. As I said, it's not in Hicks.

The relevant question he asks is this one:
Why does pulling spending forward into the present fail while pushing taxes back into the future works?
In my post, I was arguing that the problem was a shortage of safe assets, monetary policy could not solve the problem, and we could think of an otherwise Ricardian fiscal experiment that would do the trick. I wanted to avoid talking about government expenditures on goods and services, as that would open up a whole set of issues I did not want to get into - redistribution, alternative types of spending and the government's advantages relative to the private sector, political economy, etc.

Then, Brad says:
And yet a third thing I think is wrong is Williamson's claim that "the Fed is powerless" because "swap[s of] reserves for T-bills or reserves for long-maturity Treasuries… essentially amounts to intermediation activities the private sector can accomplish as well, this will have no effect". But such swaps take various forms of duration and default risk onto (or off of) the Fed's and thus taxpayers' balance sheets and off of the private market and thus investors' balance sheets. These are different (but overlapping) groups who perceive risks differently, have different resources, and react to risks differently. The fact that the private market could undo any particular Federal Reserve policy intervention does not mean that it will.
This is a key part of my argument - i.e. under the current circumstances, quantitative easing is irrelevant. In order for asset swaps to have any effect the Fed has to have some advantage in the intermediation activity it is engaging in relative to what the private sector can accomplish. Currently the Fed has no advantage in turning long-maturity Treasury debt into overnight assets, so QE is irrelevant.

I would say go back to Hicks, Keynes, Fisher, and Wicksell, and think about them carefully: they were smart. A "neoclassical macro" that does not start from them has little chance of getting much of anywhere.
Yes, Hicks, Keynes, Fisher, and Wicksell were smart, but Keynes did not spend all of his time with his nose in Adam Smith, otherwise he would not have got any work done. Maybe someone got something wrong, and there are nuggets of insight in early work that were not passed on through the economic writings that succeeded it. However, most of what is useful from Hicks, Keynes, Fisher, and Wicksell is baked in the cake. Modern students of macroeconomics are much better off reading Lucas, Woodford, Prescott, Gertler, Gali, Farmer, Rogerson, Kehoe, Wright, Christiao, Kiyotaki, etc. from the older generations, or any number of excellent young macroeconomists. However, here is what I read the other day in Hicks's 1937 paper:
The General Theory of Employment is a useful book; but it is neither the beginning nor the end of Dynamic Economics.
Hicks was smart.


  1. If everything we need to know was in Hicks, Keynes, Fisher, and Wicksell - why did Brad even become an economist? Since we know everything we need to know from them already, I can only conclude academic economists shouldn't exist.

  2. @Ted: maybe Brad's definition of an economist is someone who repeats what Hicks, Keynes, Fisher, and Wicksell said over and over...

    Of course, History of Economic Thought is what economists engage in once they are too old to have any fresh ideas.

  3. I think this is a bit unfair to Brad.

    My interpretation of the "original" errors comment was that trying to build micro-foundations is very hard. For purposes of tractability, you compromise on realistic models of risk, heterogeneity, market structure, etc.

    And the claim that Brad is making is that for purposes of being able to give useful/practical policy advice in response to financial crises, the regular economics program has not really delivered anything substantially new that improves on the classics, and has delivered some prescriptions that have been harmful.

    That view is completely consistent with the view that there is more to learn, and that the Lucas-Sargent program might one day be able to provide more effective policy advice than was provided by Hicks/Keynes/Wicksell.

  4. 1. On the "original errors" comment, we would have to get Brad to clarify, but I interpreted it in the context of other things he has said. If I understand you correctly, you think that we don't have "realistic" models of "risk, heterogeneity, market structure, etc., and that's not true. Start with the paper I link to above. Now, of course you are not going to find much that's actually explicit about risk, heterogeneity, and market structure in Hicks.

    "for purposes of being able to give useful/practical policy advice in response to financial crises, the regular economics program has not really delivered anything substantially new that improves on the classics..."

    Again, not correct. Read the paper I link to, or better than that go to this site:

    and look at the papers on the program.

    "has delivered some prescriptions that have been harmful."

    You might say that Woodford's adaptation of Prescott's RBC model - the New Keynesian model with sticky prices and wages - was not successful in guiding central banks as to what to do in the crisis. But that was the best Keynesians could do in channeling Keynes and Hicks into modern economics. If Woodford+Keynes+Hicks doesn't work then how does Keynes+Hicks work?

    "the Lucas-Sargent program might one day be able to provide more effective policy advice than was provided by Hicks/Keynes/Wicksell."

    That day is here. It already does, if you use it properly.

  5. And what is the "effective policy advice" regarding our current predicament that is flowing from "the Lucas-Sargent program"?

  6. Well, almost everything post-1972 in macroeconomics can be thought of as part of the "Lucas-Sargent program" unless you want to be more specific. There's stuff right under our noses here, in fact. This paper:

    which was part of what informed the blog post we are talking about

    could indeed be considered part of the "Lucas-Sargent program," as I think it conforms to the rules of that game. Indeed, Lucas once read the paper and seemed to like the thing, though you would have to ask him about that. I think he thought the "New Monetarist" program needed some more quantitative work. In any case this is my very serious attempt to provide "effective policy advice." Why don't we discuss that? Do you have a problem with it?

  7. Greg:

    If you are looking for more immediate things, part of what you are calling the "Lucas-Sargent" program drove ideas about commitment and forward-looking policy. If you know that stuff, you would be able to make sense out of Plosser's and Kocerlakota's dissents at the last FOMC meeting. Both of those people were well-schooled in the "Lucas-Sargent" program, and their dissents made good sense. Effective policy advice at work.

  8. Well if Rudi Dornbusch would have said they are original errors, they must be original errors.

    I think that I have more of an appreciation of the history of thought than most commenters here. I think that there are insights that can be gleaned from previous work and all economists would do well to familiarize themselves with the likes of Hicks, Wicksell, etc. (No doubt Steve heard this from David Laidler, no?) HOWEVER, to claim that we have accomplished nothing since Hicks (1937) is not only an insult to the profession, but it is an insult to Hicks himself. Ugh.

  9. Yes, David is very good on that stuff, and can tell you exactly where it matters. Once David brought Hicks himself over from England to give a talk at Western Ontario.

  10. Stephen, you aren't answering Greg's question.

  11. No, I did answer it. I could go on at great length on that subject, but you have to give me some prompts. It seems you think the "Lucas-Sargent" program was not good for much, unless I'm misunderstanding you. I have to figure out where you are coming from. What is (are) your complaint(s), exactly?

  12. Stephen,

    Thanks for the references. I will look at them.

    When you say "almost everything post-1972 in macroeconomics can be thought of as part of the 'Lucas-Sargent program,'" you're ruling out a lot of Macro that had illuminating things to say about the crisis, some of them even before the crisis unfolded. Was there any Macro coming out of the "Lucas-Sargent program" that "predicted," or warned, of the crisis, perhaps something comparable to the steady stream of papers by Wynne Godley and others at the Levy Institute or Robert Shiller's work?

  13. Greg:

    It is easy to predict a crisis when you predict one all the time. The Levy Institute is worthless.

  14. It's pretty complicated to determine the lineage of some of these ideas, schools of thought, and research programs. I'll give this a try though. In the 1980s, some academics starting using the information theory developed in the 1970s by people like Akerlof, Stiglitz, Rothschild, Townsend, Holmstrom, among others, to think more deeply about banks and other financial intermediaries and what they do. One key development was Diamond and Dybvig's (JPE 1983) banking model. That's a basic model of a crisis - i.e. a banking panic, which a lot of people found attractive as a way of thinking about panics and runs. Diamond and Dybvig were two graduate students in Yale's PhD program. The basic model was something that Neil Wallace, at that time at Minnesota, found attractive. What he liked about it was that it fit what we might think of as the rules of the program. It was a model based on preferences, endowments, technology, and an equilibrium concept, and it stripped things down to the essentials you could use to address the problem. Neil extended that framework in late 1980s, and so did Ed Green, using some of the mechanism design machinery developed by Hurwicz and others in the 1970s. Gary Gorton, now at the Yale Business School, has been an expert on securitization since the 1980s; in fact I first heard about the phenomenon from him circa 1988. Gary also worked on banking theory, along with banking history, and he finds it useful (and he has a point) to think about the collapse of repo markets and "shadow banking" during the financial crisis in Diamond-Dybvig terms. Ed Prescott also thought about financial intermediation in the 1980s. One interesting, though esoteric, paper on intermedation was one that he wrote with John Boyd in the 1980s. Thus, the ideas, though not confined to what I would think of as "Minnesota economics" have strong ties to the general program and way of looking at the world, and involved people like Hurwicz, Wallace, Prescott, and Green, who indeed worked at the Minneapolis Fed and the University of Minnesota. That's just an example. Here's another one. John Kareken and Neil Wallace published this paper in 1978:

    It's the classic paper on deposit insurance and moral hazard. Moral hazard is of course a key aspect of the financial crisis - excess risk-taking, too-big-to-fail, etc. In 2004, Gary Stern, then President of the Minneapolis Fed, and Ron Feldman (also Minn. Fed) publish this book:

    That's essentially a warning about risk taking in the financial industry and moral hazard, pre-financial crisis.

    Of course, none of this sounds like the narrow view that some people have of what went on (or still goes on) in Minneapolis, for example. That was not all about permutations of the neoclassical growth model.

  15. "This is a key part of my argument - i.e. under the current circumstances, quantitative easing is irrelevant. In order for asset swaps to have any effect the Fed has to have some advantage in the intermediation activity it is engaging in relative to what the private sector can accomplish. Currently the Fed has no advantage in turning long-maturity Treasury debt into overnight assets, so QE is irrelevant."

    You just would be a lot more persuasive and teach better if you would specifically say why something that's wrong is wrong, this leads to this, leads to this,…,leads to this not happening, instead of just saying here is the vague general principle for why the right thing is right.

    For example, if the Fed buys huge amounts of five year and ten year securities it will push down their rates. That will make more long term corporate projects positive NPV, so more will be done, employing more people. It will make mortgage payments lower, making more houses purchased and bulit than otherwise, etc. Why Exactly Is This Wrong?

  16. "Why Exactly Is This Wrong?"

  17. Add Professor Perry Mehrling to the list of people unimpressed with DSGE. Talking about interbank liquidity problems, he claims DSGE models give you no language to discuss the issue.
    I imagine you feel differently, Professor.....

  18. Here's the critical sentence in Mehrling's piece:

    "The line of analysis that I just explored would be impossible if I had constrained myself to talk the language of the Dynamic Stochastic General Equilibrium model, the lingua franca of academic macroeconomics before the crisis, and today as well."

    Question 1: What does Mehrling think a DSGE model is? It would help if he supplied a brief description of the thing he thinks is DSGE so we can actually match that up with what some macroeconomists do, and evaluate whether he knows what he is talking about.

    Question 2: Later in the post, he refers to Woodford, and he seems to think that Woodford does DSGE. So, suppose by DSGE, he means New Keynesian models - i.e. Prescott RBC with sticky wages and prices, plus some other bells and whistles. That's certainly a popular framework that you see a lot in some central banks, and some academics work with it. But modern macroeconomics is much richer than that. We work on a wide array of models, some of which can be used to discuss the things Mehrling wants to discuss in his piece - and actually shed more light on those issues than he does.

  19. When Hicks says "Dynamic Economics" he's mostly talking about Hayek's dynamic economics -- and his own Hayek-inspired dynamic economics.

    But I don't expect any contemporary economists to know this, or much if anything at all about economics published 4 or 5 years prior to their first year of graduate education.