Discoverer of Wallace Neutrality Ricardian Neutrality Revivalist
Scott Sumner posted a speedy reply to the question I posed for him in “Should the Fed Promise to Do the Wrong Thing in the Future to Have the Right Effect Now? His answer, like mine, is “No!”
On the whole, it turns out that the two of us are in relatively close sympathy. On my side, I certainly think that the Fed should be clear about its medium-term objective, in the spirit of nominal GDP targeting, though there are some details to come in my upcoming post “Optimal Monetary Policy: A Primer” that will diverge from overly simple versions of nominal GDP targeting. As for shorter-run targets to get to a medium-run target in the spirit of nominal GDP, near Wallace neutrality means that the changes in quantities needed to be on target for a given effect on nominal GDP can change dramatically as key assets reach the zero lower bound. (See my post “Trillions and Trillions: Getting Used to Balance Sheet Monetary Policy.”) So especially when the zero lower bound is an issue, I think stating things in interest rate terms (despite the complexities that entails) has some advantage over stating things in terms of asset quantities.
On Scott’s side, he clearly states he does not believe that Wallace neutrality is an accurate description of the real world:
In theory Wallace is right and Kimball is wrong, but in practice Kimball is right, because the conditions for Wallace neutrality to hold would probably never occur in the real world. This will take some explanation, so bear with me.
I wouldn’t quite say I am wrong in theory, but it is true that simple theoretical models definitely “want” to have Wallace neutrality. It takes some doing—on the research frontier—to theoretically model departures from Wallace neutrality like those that I believe exist in the world. This is a great research area for graduate students looking for a research topic. It is worth many dissertations because it is not good enough to have a departure from Wallace neutrality in your model, it needs to be plausible as a description of why the real world departs from Wallace neutrality. Since different economists will have different judgments about how realistic various alternative mechanisms are, it will be good to have many different models of departures from Wallace neutrality, using different mechanisms.
In terms of how economic theory works, let me draw a history-of-economic-thought analogy between Wallace neutrality and Ricardian neutrality. Ricardian neutrality is the proposition that for a given path of government purchases, tax rebates won’t have any important effect on the economy because the tax rebates lead to higher government debt, which unavoidably leads to higher future taxes to pay for that debt, which in turn leads, according to the theory, to enough extra private saving to pay for those extra taxes, which is exactly enough extra saving to satisfy the government’s extra borrowing needs. Despite the complexity of this story, simple macroeconomic models very much “want” to have Ricardian neutrality. But in the history of 20th century economic thought, most economists immediately disbelieved Ricardian neutrality as a description of the real world when Robert Barro revived interest in Ricardian neutrality. These disbelievers were very successful when they set out to find cogent reasons why Ricardian neutrality might not hold in more realistic models. (However, the tendency of simple models to exhibit Ricardian neutrality has ensured that an important minority of economists still believe to this day that it is an accurate description of the real world.)
The difference in the theoretical status of Wallace neutrality as compared to Ricardian neutrality is that we are earlier in the process of putting together good models of why the real world departs from Wallace neutrality. Studying theoretical reasons why the world might not obey Ricardian neutrality was frontier research 25 years ago. Showing theoretical reasons why the world might not obey Wallace neutrality is frontier research now.
Let me make one last point. Scott mentions various ways to stimulate the economy when the fed funds rate is already at zero:
Consider all the “foolproof” escapes from the liquidity trap. My NGDP targeting idea, Svensson’s currency depreciation, Friedman/Kimball’s massive QE, etc.
It turns out that every single one of these require a departure from Wallace neutrality. I will talk about the purchases of foreign assets that make Svensson’s proposal work in a future post about international finance. Here I want to point out that NGDP targeting to work also needs either (a) implicit promises to overstimulate the economy in the future to get stimulus now OR (b) departures from Wallace neutrality. In particular, it seems to me that if nominal GDP targeting were combined with a rule that the Fed would only purchase short-term Treasury bills that it would indeed work through (a) implicit promises to overstimulate the economy in the future. (I will explain how I am defining “overstimulating” in my upcoming post “Optimal Monetary Policy: A Primer.”) But if the Fed is willing to contemplate purchases of other assets that still have a positive interest rate, then nominal GDP targeting would work by (b) implicitly committing to buy enough of those other assets to head toward the nominal GDP target, even if it take trillions and trillions of dollars worth of asset purchases. That is very close in spirit to the kind of thing I am recommending.
Post in the Monetary Policy Thread