I have said this before. But I seem to need to say it again…

The very intelligent and thoughtful David Beckworth, Simon Wren-Lewis, and Nick Rowe are agreeing on New Keynesian-Market Monetarist monetary-fiscal convergence. Underpinning all of their analyses there seems to me to be the assumption that all aggregate demand shortfalls spring from the same deep market failures. And I think that that is wrong.

Simon Wren-Lewis writes:

I really like David Beckworth’s Insurance proposal against ‘incompetent’ monetary policy. Here it is: 1) Target the level of nominal GDP (NGDP). 2) “The Fed and Treasury… agree… should a liquidity trap emerge anyhow… quickly work together to implement a helicopter drop….” Market Monetarists and New Keynesians [do not] suddenly agree about everything… for David this is an insurance against incompetence by the central bank, whereas Keynesians… view hitting the ZLB as unavoidable if the shock is big enough. However this difference is not critical…

And Nick Rowe concludes:

How much money should the central bank print and buy things with? As much as is necessary to hit the NGDP target. And if it runs out of… government bonds [to buy]… it should buy newly-produced things…. What particular things should be bought and held on the asset side of the consolidated balance sheet of the government plus central bank? That is a micro public finance question. What particular things should be held on the unconsolidated central bank’s balance sheet rather than on the government’s balance sheet? That is a public choice question…. I don’t think there’s anything left to argue about. Except a lot of micro public finance and public choice stuff.

But I’m sure we will think of something.

I, however, find myself unsatisfied with this convergence. Finally summarize it, it seems to go like this: In a depression in which the problem is that not enough stuff is being bought, (i) the central bank should buy bonds for cash in order to put more cash in people’s pockets so that they will buy more stuff, and (ii) if that doesn’t work–if for some reason people still don’t buy enough stuff even though the Federal Reserve has driven interest rates to zero–the central bank should print money and give it away so that people will buy more stuff, and (iii) if that doesn’t work, the central bank and government should jointly print money and buy more stuff.

This seems to me to be largely unexceptionable. But I think it misses a couple of subtleties. And I think these subtleties are important because they undermine the lexicographic preference for (1) open-market operations, (2) helicopter drops, and (3) fiscal expansion in that order that is at the base of this Oxford-Bowling Green-Ottawa compact. I suspect that the reason that Simon, David, and Nick are able to reach agreement is that they do not foreground the issues that I learned when Olivier Blanchard forced me to study that key paper of the greatest of University of Chicago macroeconomists, [“Wealth, Saving, and the Rate of Interest” by Lloyd Metzler][4].

Whether you want to use fiscal or monetary (or banking!) policy to fix the depression depends critically on why not enough stuff is being bought–on where the market failure is.

  1. The problem could be Monetarist. If the problem is a scarcity of liquid cash money (with associated sticky-price and nominal-debt-contract market failures, so that dropping the price level to instantaneously raise real money balances to their first-best level just isn’t a good option), then, yes, conventional open-market operations can do the job–but whether you want to use monetary or fiscal policy hinges on what you want the post-depression real interest rate to be.

  2. The problem could be Keynesian. If the problem is a scarcity of savings vehicles for transferring purchasing power from the present into the future (with the same associated sticky-price and nominal-debt-contract market failures)–a shortage of money-plus-bonds–then conventional open-market operations will not do the job: they get you to the zero lower bound, and once there your swapping of cash for bonds has no effect on anything. What you need to do is either (a) increase the supply of bonds by summoning the Confidence Fairy and getting private entrepreneurs to issue more bonds as they borrow-and-spend, (b) decrease demand for bonds by summoning the Inflation Expectations Imp and thus imposing a tax on nominal assets, or (c) increase the supply of bonds directly by having the government buy stuff. And there are a host of issues as to which is the best to do.

  3. The problem could be Minskyite. If the underlying market failure is that the credit channel has broken down, so that savers do not trust the securities that financial intermediaries originate to be of the safety that they seek, then there are definite drawbacks to summoning the Inflation Expectations Imp or indeed to keeping interest rates low: the economy then winds up with an incentive to produce too many long-duration assets and to invest too little in risky projects. A better solution is to have the government bear risk. And the best solution is to restore the credit channel.

So let us start with Metzler’s basic model: three commodities and two prices, one of them sticky;

  • Commodity: liquid cash money M–in order to create the trust that you need to transact.
  • Commodity: bonds B–savings vehicles for transferring purchasing power from the present into the future.
  • Commodity: currently-produced goods and services Y.
  • Price: the sticky price of currently-produced goods and services in terms of money P.
  • Price: the interest rate i, the inverse of the price of nominal bonds in terms of money.

A Monetarist depression is then when there is not enough money M in the system. The interest rate i is then high–the price of bonds is low because the marginal utility of holding the scarce money is high–and is high enough to equalize supply and demand for bonds at that i. But because P is sticky, there is an excess demand for money and so, by Walras’s Law, an excess supply of currently-produced goods and services.

You can fix a Monetarist depression by having the central bank buy bonds for cash until the stock of money M is high enough to balance supply and demand. But is that what you want to do? You could also (a) summon the Confidence Fairy and so get private businesses to issue more bonds, (b) summon the Inflation Expectations Imp and so reduce demand for both money M and bonds B relative to currently-produced goods and services, or (c) have the government issue more bonds. Such expansions of B would drive the interest rate i up further, and with enough bonds the interest rate would be high enough that the sticky value of P would also be the Walrasian equilibrium value of P, and there would be neither excess demand nor excess supply for money M or for currently-produced goods and services Y.

In order to decide whether you want to fix a depression characterized by a high interest rate i on bonds via monetary expansion or fiscal expansion (or by summoning the Confidence Fairy and getting a private investment boom going), you need a way to rank full-employment equilibria. What share of production on currently-produced goods and services Y should take the form of government purchases G? What cost should we put on businesses undertaking private capital formation I via the value of the interest rate i? Robert Solow’s chapter for the 1962 Economic Report of the President argued, I think wisely, for easy money and tight fiscal policy–that there was no obvious reason for an anti-depression policy to have the side effect of reducing the post-depression capital stock below what it would otherwise be. But it is possible to imagine situations in which the Wicksellian natural rate of interest consistent with full employment when the government’s budget is balanced is below the proper social rate of time preference, and induces private investments that do not cover their properly-measured social costs. It’s not a slam dunk.

A Keynesian depression is then when there is not enough money plus bonds M+B in the system, when businesses are not confident enough to undertake enough investment projects to back enough bonds to meet private-saver demand for savings vehicles. Because bonds are scarce people bid their price up to par, and bid the interest rate i down to zero. And there the price of bonds becomes sticky too–it cannot go any higher, and money and bonds become perfect substitutes, and open market operations do nothing because they simply swap a zero-interest asset for another. Then there is an excess demand for the aggregate that is money-plus-bonds M+B, and an excess supply for currently-produced goods and services Y.

Since open-market operations no longer do anything, policies are down to:

  • (a) summon the Confidence Fairy and so get private businesses to issue more bonds.
  • (b) summon the Inflation Expectations Imp.
  • (c) have the government issue more bonds.

Summoning the Confidence Fairy to get private businesses to print up more bonds and so increase the supply of M+B will work, as long as you can actually summon spirits from the vasty deep. Raise M+B enough, and there will be no excess demand for M+B at the sticky price P, and so no excess supply of currently-produced goods and services. Raise B even more and i will start to rise and the infinite elasticity of demand between B and M will break.

Summoning the Inflation Expectations Imp to reduce the demand for the money-and-bonds aggregate M+B will also work–once again, as long as you can actually summon spirits from the vasty deep.

Having the government issue more bonds, however, works directly on the supply of the aggregate M+B.

Which is preferable? Well, how good are your spirit-summoning skills? What are the costs of accumulating government debt? What are the costs of raising the inflation target if you can summon the Inflation Expectations Imp?

We have by now a number of balls in the air: a (sticky) price of currently-produced goods and services in terms of money P, an expected rate of inflation π, a degree of business confidence, an interest rate on bonds, a supply of money M, a supply of bonds B, and a division of the bond supply between bonds that fund private investment and government debt. And there is a shadowy concept of the “first best”–of what the full-employment economy ought to look like in terms of the proper incentive to invest, the proper supply of liquidity services, and the proper inflation target.

But we have still more: we have a Minskyite depression, in which on top of our other market failures–our sticky-wages and our nominal-debt contracts that make deflation impossible or unwise, and our zero lower bound that makes the price of bonds become sticky too–we also have a collapse of the credit channel, an inability of financial intermediaries to do the risk transformation and so turn claims on the risky investment projects that make up the private capital stock into bonds of a high-enough quality for private savers to want to hold. We now have a fourth commodity: in addition to money M, (safe) bonds B, and currently-produced goods and services Y, we now have junk bonds J. Now our economy gets wedged with an interest rate i=0 at the zero lower bound and a price level P such that there is an excess demand for the perfect-substitutability aggregate M+B, with a high interest rate j on the junk J, and by Walras’s Law with an excess supply of currently-produced goods and services Y.

Now our admissible policies are:

  • (b) summon the Inflation Expectations Imp.
  • (c) have the government issue more (safe) bonds.
  • (d) summon a different Confidence Fairy and restore saver confidence in the ability of financial intermediaries to properly do the risk transformation and originate not junk J but safe bonds B.
  • (e) reform and recapitalize the financial intermediaries so that they can in fact do the risk transformation and originate not junk J but safe bonds B.
  • (f) use the government as a loan guarantor and use its own (or, rather, the taxpayers’) risk-bearing capacity to do the risk transformation.

And what about (a), summoning the (business) Confidence Fairy? It is not a solution: getting businesses to issue more debt does not help matters as long as financial intermediaries are unable to sell it as safe bonds B rather than junk J.

Now note that our first-best has required an additional dimension. We had been thinking about the proper incentive to invest, the proper supply of liquidity services, and the proper inflation target. Now we are thinking about the proper inflation target the proper supply of liquidity services, the proper real societal rate of time preference (the interest rate on safe bonds), and the proper risk spread between safe bonds and junk.

And if the root market failure is the collapse of the credit channel–a greatly enhanced spread between B and J, or, indeed, credit rationing–that (b)-(d) no longer look as attractive. When the problem is an inability of financial intermediaries to do the risk transformation, restoring full employment by reducing the real rate of time preference on safe bonds (which is what (b) does) provides the economy with too-great an incentive to produce long-duration assets. Having the government issue more (safe) debt (c) produces an economy with too much current government spending on social welfare or on infrastructure and too little on risky private investment projects. And (d) restoring saver confidence in financial intermediaries is not a good thing unless savers should in fact have confidence in financial intermediaries.

There is a very large number of different positions you can take on what set of macroeconomic policies will get the economy closest to its first-best when it is subject to what kinds of shocks–Monetarist shocks to liquidity demand and supply, Keynesian shocks to business investment committee animal spirits, or Minkyite shocks to the ability of financial intermediaries to do the risk transformation. And to sideline them–as Nick Rowe does–as “micro public finance and public choice stuff” seems to me to overlook several large elephants that are not just in the room but on the couch and pressing the remote…


[4]: “http://delong.typepad.com/1825743.pdf” title=”1825743.pdf (Lloyd Metzler (1951): “Wealth, Saving, and the Rate of Interest”, Journal of Political Economy 59:2 (April), pp. 93-116. Cf. also James Tobin, passim.)