Must-Read: Mariana Garcia Schmidt and Michael Woodford: Are Low Interest Rates Deflationary? [No:] A Paradox of Perfect-Foresight Analysis
Must-Read: Are Low Interest Rates Deflationary? [No:] A Paradox of Perfect-Foresight Analysis: “￼￼￼￼￼￼￼￼￼￼Some… suggest that prolonged low nominal interest rates…:
…and central-bank promises to maintain such policy–may bring about lower inflation (rather than higher)…. ￼The “Neo-Fisherian” View (Cochrane, 2015b)…. Is it really true that “modern theory” — deriving aggregate demand and supply relations from intertemporal optimization — implies the neo-Fisherian view? Can one maintain the orthodox view — that maintaining a lower nominal rate for longer should cause higher inflation and capacity utilization — while having a view of expectations that implies that central-bank commitments regarding future policy should have any effect?…. The assumption of perfect foresight is… very strong…. PFE predictions are relevant only to the extent that
the PFE is the limit of an iterative process of belief revision… [that] converges fast enough for the limit to well approximate the outcome from a finite degree of reflection…. Temporary equilibrium (Hicks, Grandmont, etc.): endogenous variables determined by optimizing behavior of economic agents, under subjective expectations that are specified as part of the model (and need not be correct)…. Dixit-Stiglitz monopolistic competitors. Calvo-Yun… staggered price adjustment….
Cochrane (2015a) proposes a particular criterion for selecting one among this continuum of PFE solutions as the prediction of the model: the “backward stable” equilibrium…. But is this a reasonable view of what should follow from intertemporal optimization, and an ability to reason about the implications of the central bank’s policy commitments for future economic outcomes?… Consideration of the set of PF equilibria is especially misleading in the (somewhat artificial) thought experiment of a permanent interest-rate peg–a case in which none of the PFE paths approximate TE paths for finite n, even for very large n…. So we should not necessarily expect… that a long-enough lasting commitment to remain at the ZLB must eventually make inflation lower, rather than higher…. Especially for larger values of T, the approach recommended here leads only to a set of possible predictions for a given policy, but this still allows qualitative conclusions that remain very useful for practical policy analysis, and insisting on PFE analysis simply because it makes more sharply-defined predictions may lead to large errors.