Incoherent and Incoherenter

I have comments on Brad DeLong’s question asked of Richard Koo.

Re: Brad Delong

which is

Re: Richard Koo
“I am not sure that I can ask this question coherently.”

I think the question is plenty coherent. The post-script does a better job of capturing not your personal confusion but just how confusing the issues are.

Brad’s first question

The conventional arguments of those whom Martin Wolf calls the Austerians runs more-or-less like this: someday QE will succeed in shifting beliefs from an expectation of permanent depression to an expectation of rapid normalization. Savers then look at their holdings of maturing government bonds and roll them over only if they are offered a normal and positive real interest rate. And then the price level will rise very rapidly to a value at which–as John Maynard Keynes said of France during its inflation of the 1920s–real future government primary surpluses discounted at the normal real interest rate are equal to the nominal debt divided by the price level.

In your framework, that would be a sudden very large shift in private-sector net savings behavior from surplus to deficit. And in your framework such a shift is almost inconceivable. But in their framework such a shift seems almost inevitable. Can you tell me why judgments of likelihood of a near-hyperinflationary collapse upon normalization are so different in the two frameworks?

I think even in your charitable inerpretation, the neo-Austrian theory is based on a false dichotomy, a falacy more common than any other and more common than any valid method of reasoning. In your story, there are two states of expectations, so a shift must be a jump so there is either liquidity trapping or high inflation.

In mainstream macroeconomics (including I think Koo) all functions are assumed to be continuous and all dynamics are assumed to be saddle path stable (exactly as many eigenvalues with positive real parts as forward looking variables). This means jumps can only occur when there is an exogenous surprise such as a shift in policy. It is an assumption (an independent core assumption — not an implication of rational expectations and all that) that economic fluctuations are stationary around a unique balanced growth path equilibrium.

So we have two inconsistent falacies and no reason to put any trust in either.

Brad asks other questions (or gets to his point)

I know that 25 years of history strongly suggest that they are wrong, but why? It was, after all, right for France in the 1920s. It was possibly right for peripheral European countries trapped in the eurozone. It was right for Argentina. Why is it not right–or not possible for any reasonable probability–for reserve currency-issuing credible sovereigns?

There is a key difference between debt denominated in the national currency and debt denominated in foreign currency. The case of France in the 20s is different from the other cases (Argentina was effectively using the dollar for finance as it guaranteed 1-1 convertability and allowed dollar denominated contracts). The only case relevant to the discussion is France in the 20s.

OK now Brad’s really confusing post script

UPDATE: perhaps the real issue is that we have three underlying models of macroeconomics. The first is the quantity theory of money MV = PY: the stock of money times its velocity equals the price level times production. The second is the Wicksellian savings investment equation S = I + (G-T): savings either finances investment or is absorbed by the government deficit. The third is the fiscal theory of the price level D/P = PV(-dp,r): the real stock of debt–the nominal debt divided by the price level–is equal to the present value of future primary surpluses discounted at the real interest rate. All three of these must be true at the same time, which means that at any time two of them are likely to be nearly redundant. For those two, shifts in what are supposed to be their driving variables are neutralized by countervailing forces. Right now, for example, increases in the money stock are offset one-for-one by reductions in velocity, and increases in the nominal debt are offset one-for-one by higher future primary surpluses and reductions in future real interest rates.

From this perspective, the key question of macroeconomics is always: when do each of these three models have primary traction, and why?


OK on the ps. You are making the problem harder by pretending to take MV=PY seriously. In fact you more nearly believe in an LM curve MV(i) = PY so the first two equations both hold where the IS curve meets the LM curve.

The fiscal theory of the price level clearly only holds with extra assumptions. A country can be France in the 20s or Russia in the 90s. Treasuries can default and have defaulted. In the equation (which is really just an accounting identity) default is a 100% tax on bonds, and part of the primary surplus. This is just one way in which future primary surpluses are endogenous.

I think that fiscal theory of the price level is the theory of the immaculate debt inflation. It must hold, but firms setting prices don’t consider the public debt and expected future primary surpluses. It is an identity not a behavioral equation (except to the extent that primary surpluses are the behavior of public officials — but primary surpluses and only primary surpluses seem to be considered exogenous in the theory).

I am quite confident that there was a more proximate cause of inflation in France in the 20s (could it be the inflation across the German border or, slightly further across the Rhine in more genuinely sovereign Germany ? You’re the economic historian, so I charge you to find the proximate cause. I’m pretty sure French WWI debt was inflated away, because the central bank decided to inflate it away — that the episode can be understood by looking at the money supply with only a footnote that monetary policy was inflationary because of fiscal dominance.