Thoughts on Methodology and Macroeconomics from Scott Sumner

|Peter Boettke|

I have voiced my doubts about Scott Sumner's ideas about quantitative easing several times on this blog.  But those criticisms were always voiced from the perspective of political economy, not from the perspective of monetary economics.  Bill Woolsey, Steve Horwitz, Larry White and George Selgin have persuaded me that Sumner's position is technically less problematic than my Sennholzian training would imply.  So, in fact, I've argued all along that my endorsement of older ideas in monetary economics falls out of my pessimism concerning the end game in politics.  In short, I don't think the Fed can credibly commit, and thus quantitative easing policy translates into another round of inflationary policy in the future. The Fed cannot accomplish the dance of providing the liquidity and sopping the excess simultaneously.  Much of the evidence Sumner points to is the failure of the Fed to attempt the dance, rather than examples of where the dance was performed beautifully.  So I am still not persuaded that we can do what it is that he argues in theory would be ideal to do to avert disaster.

But I always benefit from reading Scott, and his new essay at The Money Illusion is no different.  It is actually a very thoughtful discussion of blogging, methodology, and the 'doing' of economics.

HT: Tyler Cowen.

Comments

[Pietro M.]

I have theoretical doubts about NGDP targeting.

Starting from stationary conditions, excessive credit creation by the banking system would almost immediately start to rise employment and nominal wages, which would increase consumption and put pressure on final prices. Under a NGDP rule the central bank will have to raise interest rates and relieve the inflationary pressure, thus effectively preventing, as far as it is possible, capital disequilibrium. The same applies in case of growth, with P forced to fall by a - hopefully sustainable - increase in Q.

So far, all is ok. However, I have to problems with applying the rule once a malinvestment boom has been under way for a decade or more. (In other words, NGDP targeting may be ok unconditionally, but conditionally on being at the peak of the boom it may be dangerous).

FIRST DOUBT (apparently unrelated)

Is it necessary to look for explanations in terms of monetary disequilibrium to explain this crisis? What would an unaugmented Austrian theory say about a recession occurring after a binge of a decade? What would a credit-channel financial accelerator model of massive deleveraging say about the pain of the restructuring?

It is often said that ABCT is not a theory of the recession, but this statement is exaggerated. It would be more correct to say that it is not a theory of an unending and deep depressions, because it does not consider the actions of Hoover and Roosevelt (Mises's theory of interventionism, which he applied in 1931 to explain the Great Depression, is an unsophisticated but insightful version of the Cole & Ohanian model, and both are necessary to understand that specific event because the problem was NOT malinvestment).

That's not the present case: this recession can be explained either by Austrian, credit-channel, MET, or regime uncertainty theories, and there is an evident identification problem in finding the empirically more relevant framework. I still haven't seen the smoking gun for the empirical relevance of MET (or any other theory) as opposed to real explanations, such as financial deleveraging and malinvestment liquidation.

SECOND DOUBT:

I tried to imagine how to keep NGDP=MV=PQ from falling: the only solution that came to my mind has been to do what Bernanke is already doing.

The banking system is broke or close to be broke; the money multiplying process, which has been overexploited during the boom, is now no longer functioning; banks are deleveraging, and a return to normal levels of leverage, multipliers, maturity mismatch etc is under way.

How is it possible to avoid a major contraction in NGDP? Q is falling because malinvestments are becoming evident, as a consequence P should be forced to rise. However, banks have structural (real) problems: P, without some policy step, will fall.

There are a few ways to make P rise: increase consumer spending through fiscal policy, printing helicopter money, or forcing banks to channel funds toward the end users of money (unlikely).

In order to avoid the depletion of investable funds, credit needs, moreover, to be kept from falling and, even better, further expanded: this can only be obtained by recapitalizing the banks and removing their toxic assets from their balance sheets.

If these policies are not put in practise, NGDP falls. It looks like being unavoidable.

[Bill Woolsey]

Pietro:

Credit expansion immediately impacts expenditure of capital goods. Capital goods are final goods. While there may also be an almost immediate impact on wages and consumption exenditure, the first step is an increase in the demand for final output. Interest rates must reverse the excess credit expansion, right off.

Rather than stablizing the CPI, so that one must go through the steps of more consumption spending, higher demand for consumer goods, firms perhaps using inventories and even expanding production, before, or simultaneous to price increases, and then some period before the statisticians measure it, the depletion of inventories and pumped of production of capital goods is what is being "controlled." Of course, there is the statisticians measuring it problem.

Deleveraging is irrelevant. It is the fallacy of composition. What do the lenders do with the money they don't lend or that is repaid? Malinvestment is relevant and the live issue is how much of the drop in output (it is about 10% below trend) is due to a drop in capacity vs how much is do to an excess demand for money?

While it is true that nominal expenditure targeting does imply that prices will rise if productive capacity falls, (which is a bit of a least bad response rather than a good thing,) you focus too much on "trying" to get prices up.

Anyway, it is not necessary to buy up toxic assets or reacapitalize the banks in order to expand either the quantity of money or nominal expenditure. Bernanke "needs" to do all that so that past relationships between the Fed's traditional target--the Fed funds rate--and everything else, will return to "normal."

[Jerry O'Driscoll]

The Fed's problem is twofold.

First, there is the knowledge problem emphasized by Milton Friedman. The Fed does not have the information in real time needed to engage in effective counter-cyclical policy. That is the technical argument for a monetary rule.

Second, there is the Public Choice problem raised by Pete Boettke. The Fed will bend to political pressure. In practice, that means it will lexicographically order the goals of its dual mandate and give priority to employment (as argued by Allan Meltzer in his 3-volume history of the Fed).

When the Fed does act aginst inflation, it does so belatedly, half-heartedly and abandons the fight too soon. That leads to what the Brits call a stop/go cycle, and ended in the 1970s in stagflation.

Only when popular opinion turned decisively against inflation late in the Carter administration, and Carter appointed Volcker as Fed chairman, did the alter the ordering. Even so, there was no permanent break in the uptrend in the price level. Click on this link to see the dismal record.

http://research.stlouisfed.org/fred2/series/CPIAUCNS?cid=9

[Mikeikon]

Doesn't quantitative easing just create more false signals?

[Pietro M.]

Mr Woolsey: thanks for the answers.

"you focus too much on "trying" to get prices up."

That's how I interpret a MV-stabilization policy: Q falls, so P must rise. White talked about it at the Roberts podcast, and I found this policy defended in books by White, Selgin and Horwitz. Shall I interpret it differently?

"Deleveraging is irrelevant. It is the fallacy of composition."

I'm not convinced. Malinvestment is created and sustained by the capacity of the financial sector to create credit: deleveraging has real effects exactly like the creation of credit. There are also other channels which may play a role: the reduction of excessive debt burdens, the reduction in financial leverage (if Modigliani-Miller fails), the reduction in the value of collateral (if collateral is used to reduce moral hazard between lenders and borrowers), the existence of nominal contracts (irrelevant today because prices are not falling)...

"the live issue is how much of the drop in output (it is about 10% below trend) is due to a drop in capacity vs how much is do to an excess demand for money?"

I agree, but it's not a dicotomy: I would use a... ehm... tricotomy, hoping it's not hairsplitting (the following reasoning can't be more sophisticated than this pun). In Austrian economics as I read it, there are two components of the recession. One that would exist even if the economy became suddenly neoclassical and another one of additional and neoclassically avoidable pain. The latter component is not only "monetary", however: there are costs in the recessions that are neither a pure adjustment to malinvestment nor an aggregate problem.

The first effect is that the money is destroyed first in credit markets: there is an undershoot in the purchasing power of borrowers. THis is the real impact of the deleveraging process I described before. It is a relative price problem and has no aggregate solution.

The second effect is that there is a bout of uncertainty and relative price movements that makes for an increase in risk, thus postponing some investments: this I would call the "recalculation" problem, although it is also due to real restructuring and reallocation. It is not an aggregate problem, too.

The third effect is that prices need not only to change in relative terms, but also in absolute terms because money is being destroyed and prices have to fall: this problem is aggregate and may have a monetary solution.

A fourth effect (unlikely relevant now) is that the fall in final prices will cause a ceiling in the real interest rate, which will not be able to fall below the deflation rate, thus resulting in a problem identical to price controls. This problem would be solved by aggregate stimuli.

Given the tricotomy "unavoidable costs" / "relative frictions" / "absolute frictions", it appears to me that monetary policy, if it could be divorced from credit policy, would only solve the third type of problem.