Trillions and Trillions: Getting Used to Balance Sheet Monetary Policy tumblr_m51wyl95O61r57lmx.jpg

There is something wonderful about having earned one’s first big critic in the blogosphere, so I do want to answer Stephen Williamson’s post “Quantitative Easing, the Conventional View."   But before the end of this post I promise to also make two important points that go beyond this small dustup with Stephen.  I will talk about:

  1. the role of economic models in informing economic policy and 
  2. why effective use of balance sheet monetary policy can involve open market asset purchases that are in the trillions of dollars.

Stephen’s objection 1  is to my statement from "Balance Sheet Monetary Policy: A Primer”

Above the natural level of output, core inflation rises.  Below the natural level of output, core inflation falls. 

He then criticizes this statement by saying two closely related things I heartily agree with:  (a) the current situation is hard to interpret and (b) in general, at any given time it is hard to know what the natural level of output is at that moment.  He also characterizes the theoretical provenance of this statement as New Keynesian.  That is also true.  I take the idea that inflation adjusts gradually from my main graduate school advisor Greg Mankiw, one of the most eminent New Keynesians: both from his textbook where he gives his view of the facts and from his theoretical 2002 paper with Ricardo Reis trying to explain those facts: “Sticky Information Versus Sticky Prices: A Proposal to Replace the New Keynesian Phillips Curve.”Michael Kiley anticipated Mankiw and Reis in his 1995 job market paper.  He used the nice phrase “sluggish inflation” to describe what he was explaining with his model.      

In criticizing this view, Stephen is presumably alluding to the fact that in other models the price level jumps up instantly above the natural level of output, and in others, the inflation rate jumps up instantly above the natural level of output and then gradually returns to normal.  But I don’t need sluggish inflation for the thread of my argument in “Balance Sheet Monetary Policy: A Primer”.  The key points in that section were only 

  • It is possible in some circumstances for monetary policy to be too expansionary.
  • It is possible in some circumstances for monetary policy to be too contractionary.  
  • Monetary policy does not have the power to permanently raise the level of output. 

(The history of thought for macroeconomics makes the phrase “the natural level of output” a reference to the third bullet.)   I would be shocked if Stephen disagrees with any of these statements.  

I actually began addressing Stephen’s objection 2, even before he made it, when I wrote:

It is logically possible that sellers might sell all of a particular asset to the Fed before its interest rate gets down to zero.  Then it has to find some other asset to buy that it doesn’t already have all of.  Notice that this only happens when people don’t feel they really need that particular asset very badly, otherwise they wouldn’t sell it to the Fed so cheaply.  So nothing bad happens as a result of the Fed buying all of an asset that people are willing to let go of that easily.  (Remember that an interest rate above zero has to be associated with a lower price than what the asset would have at an interest rate equal to zero.)  

Stephen’s objection 2 boils down to saying that, because of the Modigliani-Miller theorem, this logical possibility is what would actually happen: no matter how much the Fed buys of an asset, the price and interest rate of the asset will not change.  I have several responses:  

A. Scientifically, the best way to find out would be: try balance sheet monetary policy on a massive scale and see what happens.  If you really believed in this strong version of Modigliani-Miller as applicable to the real world, there would be no policy downside to such an experiment, since what the Fed does in terms of balance sheet monetary policy would have no effect on anything that matters.  

B. As Noah Smith points out in a recent post, the Modigliani-Miller theorem was originally applied to corporations.   But Stephen needs a version of the Modigliani-Miller theorem that applies to the government (the Wallace theorem), which is tougher to justify than the version that applies to corporations.   I want to clarify one thing about Noah’s post.  Stephen clearly does think that the money supply can have an affect on the economy as long as the Fed funds rate is above zero.  But Noah is right that Stephen is implicitly claiming that “quantitative easing" cannot cause inflation.

("Quantitative easing" is the very confusing phrase that the press has decided to use to refer to what I am more accurately calling balance sheet monetary policy.   Explaining why I think "quantitative easing” is such a misleading term requires explaining the difference between the history of Japanese monetary policy and what the Bernanke Fed has been doing, something I will save for another post someday.)

C. The original corporate version of the Modigliani-Miller theorem serves as the “frictionless case” of Corporate Finance.  As in Physics, the frictionless case is of great value as a starting point for teaching Corporate Finance.   But I don’t know of any real-world application of the Modigliani-Miller theorem where it applies as well as the the frictionless case in Physics applies to, say, planetary orbits.  It would be giving the frictionless case of the Modigliani-Miller theorem its due (and maybe more than its due) as a useful approximation if we think of the real world economy as being like the case of a golf ball sailing through the air.  Thinking about how the golf ball would behave in a vacuum is a good start for understanding its trajectory, but a golfer who ignores the wind is unlikely to win the U.S. Open.   

Stephen’s objection 3 has several elements.  He seems to assert even more forcefully that the assumptions needed for a Modigliani-Miller result–or Wallace result–to apply to the real world.   He indicates using technical language that he believes that what the Fed does matters as long as the Fed funds rate is above zero.  At least that is my interpretation of this passage:

A central bank is a financial intermediary. Its power to alter the allocation of resources and economic welfare derives from its monopoly over the issue of some special kinds of liabilities (currency and reserves) which are used in retail transactions and large-value financial transactions. 

But most important is what Stephen is suggesting in his objection 3 about the proper role of economic models in informing economic policy.  He is saying that (at least in advance of solid empirical evidence) we should use either the frictionless model to guess what will happen in the real world if we use balance sheet monetary policy or, a formal model with the relevant friction spelled out.   It is easy to see the attraction of this view.  In the next few years, young economists will make their mark–or older economists will burnish their reputations–by laying out many different formal models of frictions that could make balance sheet monetary policy work in those theoretical models.  (Perhaps someone has  the killer model in this category already, but if so, they clearly haven’t convinced Stephen.)  It would be better if we had all of those models in hand now, instead of a few years from now.  But the European debt crisis could send the U.S. economy into another full-scale recession before economists have all of those models worked out.  

In the meantime, the key question for economic policy making is “Do we believe that the real world is like the frictionless case or not?”  To make an analogy to math, the judgment involved is less like the judgment of whether something is proven than it is like the question of whether a mathematician believes strongly enough that something is true–and that she understands well enough what is going on–that it is worth betting a substantial chunk of time on trying to prove it.  Andrew Wiles made a judgment of exactly this sort when he set out to prove Fermat’s last theorem.  According to the current text of the wikipedia article on Andrew Wiles

Starting in the summer of 1986, based on successive progress of the previous few years of Gerhard Frey, Jean-Pierre Serre and Ken Ribet, Wiles realised that a proof of a limited form of the modularity theorem might then be in reach.

Based on his intuition that he understood what was going on in relation to Fermat’s last theorem, Andrew Wiles proceeded to spend the next 7 years working on the theorem.  The corresponding kind of judgment in Economics involves trying to read tentative bits of early empirical evidence (as well as thinking about possible theorems and approaches) in a way that Mathematics does not, but the kind of judgment involved is similar in spirit.  All of the theoretical models and all of the data analysis we have in hand goes into making that judgment, as well as our ability to extrapolate from what we know to guesses about what we don’t know.  In particular, Ben Bernanke and his fellow decision-makers about monetary policy have to make decisions based on available theory and evidence and their best judgments if proven results are unavailable in time of crisis.  

So for now, in the absence of a formal model of frictions that I am willing to point to as decisive for understanding balance sheet monetary policy, let me state as my carefully considered judgment (revisable in the light of further evidence and theory, but a bet on the truth with reputation at stake nevertheless) that there is enough friction modifying the Modigliani-Miller logic in relation to balance sheet monetary policy to be like the case of wind resistance.  In other words, there is some friction, but not much.  Balance sheet monetary policy is like moving the economy with a giant fan.  It can be done, but it takes huge open market purchases of assets to move the economy much once the Fed funds rate is more or less at zero.  

The key insight is that it is perfectly possible for the Fed to buy trillions and trillions of dollars of assets other than Treasury bills–or for other central banks to take corresponding actions–if that is what it takes.  The key issues are about the side effects and dangers of doing so.  Balance sheet monetary policy can powerfully stimulate the economy if the Fed does enough.  But we might have to get used to open market purchases in the trillions and trillions.