Wednesday, July 13, 2011

John Cochrane and the Krugmaniacs

John Cochrane has apparently created a stir among the usual blogosphere types, not so much for what he writes about, but for who he eats dinner with and how much they pay for the wine. This has caused Krugman to come unhinged.

To sort out the players here, remember that Cochrane wrote this in response to Krugman's 2009 New York Times Magazine piece. It's likely that, in the absence of his Krugman-critique, Cochrane would never have appeared on Krugman's radar screen, much like the rest of the the economics profession or, indeed, economics in general. However, Cochrane now has a place of honor on Krugman's list of bad guys, and serves as a convenient foil, particularly given his University of Chicago affiliation.

Krugman's blog piece gives us the usual message. In Krugman's mind there is a Chicago school of arrogant freshwater economists who can't bear the humiliation of facing up to the fact that Old Keynesians are oh-so-right. Since this is more of the same, I'll refer you to my earlier post, and this one.

Now, to get to something more constructive, this reminded me that I had not looked at this paper by Cochrane, which someone recommended to me. The paper was written almost a year ago, and it's Cochrane's attempt to make sense of the financial crisis and policy in the context of the crisis, looking through a fiscal-theory-of-the-price-level lens.

Cochrane claims that he can make a model-free argument based on two equations. The first is a "valuation equation for government debt," which states that the current real value of the debt (money plus bonds) is equal to the expected discounted value of future primary government surpluses. Essentially, the total consolidated-government debt is valued according to "fundamentals." The second equation is an old-fashioned money-demand function, of the sort that would fall out of a cash-in-advance environment, for example.

There are two ideas in the paper that I like. First, Cochrane points out that the "flight to quality" during the financial crisis represented an increase in the demand for all consolidated-government liabilities, not just currency, as occurred for example during the banking crisis in the Great Depression and in National Banking era banking panics. You'll find that idea in these slides as well, and in the associated paper.

Second, I liked the emphasis on the relationship between fiscal policy and monetary policy. As Cochrane points out, those ideas go back to the 1980s, and can be found in Sargent and Wallace's "Unpleasant Monetarist Arithmetic" paper, and other work by Tom Sargent.

What I do not buy into are the ideas that come from the fiscal-theory-of-the-price-level literature, which Cochrane contributed to, along with Sims and Woodford. This is where Cochrane gets his conclusions about the relationship between anticipated government deficits and inflation. Cochrane is incorrect in stating that he can get results from two equations, without fleshing out a complete model. Indeed, I can think of models where his equation (1) on page 2 of the paper does not hold, and this actually has an important bearing on how we think about the financial crisis.

Here's a counterexample. Take any pure-currency monetary model where we derive a role for money from first principles. There is just one asset, fiat money, in fixed supply, and the environment is stationary. The deficit is zero forever, so that the right-hand side of equation (1) is zero. But there exists an equilibrium where the price level is constant, money is valued, and the left-hand side of equation (1) is greater than zero, so the equation does not hold. That's if the money is an endowment for economic agents in the first period. But what if money is injected initially by way of lump sum transfers? Then the right-hand side of the equation is negative, so it still does not hold.

What is going on? Money is a bubble. It deviates from its fundamental value, which is zero. In general, any asset can exhibit such bubbly properties if it is used in exchange and is sufficiently scarce. For example, there can be a liquidity premium on Treasury bills due to their value in exchange. During the financial crisis, it seems clear that all government liabilities were scarce, real rates of return on safe assets were low, and the marginal liquidity value of government debt in financial exchange was high.

In principle, one could construct a model of the runup in house prices after 2000 and the subsequent bust, based on incentive problems leading to lending to the wrong people, and the use of mortgage-backed securities (MBS) in financial exchange, which not only creates a liquidity premium for MBS, but a liquidity premium on the underlying assets - houses. The key question is whether the theory could produce a liquidity premium large enough to match the data. Note that this has nothing to do with the way that Robert Shiller, for example, thinks about bubbles.

On another Cochrane matter, I ran across this piece from last month, which I thought was pretty good. Cochrane looks at the bond yield data and argues that QE2 was a failure, and goes on with this:
Both sides ignore an inescapable conclusion: With near-zero short-term interest rates, and bank reserves paying interest, money is exactly the same thing as short-term government debt. A bank doesn't care whether it owns reserves or three-month Treasury bills that currently pay less than 0.1 percent.

This is what drove the Fed to QE2 in the first place. Conventional easing -- buying short-term Treasuries in exchange for reserves -– obviously has no effect now. Taking away your green M&Ms and giving you red M&Ms instead won't help your diet.

But if exchanging money for short-term debt has no effect, it follows inescapably that giving banks more money is exactly the same as giving them short-term debt. All QE2 does is to slightly restructure the maturity of U.S. government debt in private hands.

Now, of all the stories we've heard to explain our sluggish recovery, how plausible is this one: “Our big problem is the maturity structure of Treasury debt. If only those goofballs at Treasury had issued $600 billion more three-month bills instead of all these five-year notes, unemployment wouldn’t be so high. It’s a good thing the Fed can undo this tragic mistake.” That makes no sense.
I like that, and it's consistent with what I wrote here.

Cochrane finishes with this comment on QE2:
Mostly, it is dangerous for the Fed to claim immense power, and for us to trust that power, when it is basically helpless. If Bernanke had admitted to Congress, “there’s nothing the Fed can do. You’d better clean this mess up fast,” he might have had a much more salutary effect.

Now, I'm reading Cochrane's stuff, and I'm finding it useful. I don't agree with everything, but he is using the economics he knows to effectively make some very useful points. I don't see any connection between what Krugman and friends are writing about the guy and what the guy actually writes.


  1. I love this. Yglesias writes, "I obviously don’t have the technical chops to wrangle with Cochrane’s model in detail." Of course, he then goes on to attack the model. Fantastic.

  2. The next time somebody sees Krugman eating dinner, demand to know how much he paid for his bottle of wine.

  3. Scott Sumner on Krugman vs Cochrane:

  4. "Scott Sumner on Krugman vs Cochrane:"

    Wow is that a festering sore of a website. Your old buddy Greg Ransom shows up there, and Sumner is a crank.

  5. I do not profess to understand Cochrane's paper but surely his equation number (1) should be an inequality bound. I woould have thought that the value of the debt is not more than the net present value of future primary surpluses. There is no reason why future surpluses should increase the amount of current debt.