Policy Options at the Zero Lower Bound When Foresight is Limited

Abstract

We reconsider several monetary and fiscal policies that have been proposed as tools of stabilization policy when conventional interest-rate policy is constrained by the zero lower bound on interest rates, assuming that households and firms are capable of explicit forward planning over only a limited horizon. The predicted effects of all of the policies are somewhat different than under rational expectations, but credible announcements about future policy can still influence behavior, and there is, if anything, an even stronger case for pursuing systematic policies outside crisis periods in order to shape expectations during a crisis.

Fiscal and Monetary Stabilization Policy at the Zero Lower Bound: Consequences of Limited Foresight

“…These include consideration of the eventual effects of commitment to an interest-rate peg for a long period of time, as in Woodford (2019); empirical modeling of US economic data over a period of decades that included significant shifts in both output and inflation trends, as in Gust et al (2019); analysis of the conditions under which joint monetary-fiscal policy regimes imply sustainable longrun dynamics, as in Xie (2020); and consideration of the difference between commitment to a systematic price-level targeting rule and adoption of an ad hoc "temporary price-level target" when the ZLB binds, as in Woodford and Xie (2019).…”

“…This condition holds if and only if µ < μ, (2.14) where μ is a bound between zero and 1 that depends on the values of κσ and of β; this is the case considered by Eggertsson and Woodford (2003). 34 In this case, the Markovian RE solution under which the crisis state persists is given by…”

“…emphasized in the RE literature, following Eggertsson and Woodford (2003); we show below that restriction of attention to this RE solution can be justified as the limit of the unique solution associated with a model with finite planning horizons, when the length of the planning horizons is made arbitrarily long. 34 A Markovian rational-expectations solution can also be defined when µ exceeds the bound (2.14), but in this case it does not correspond to the limit of an equilibrium with finite-horizon planning, as planning horizons are made arbitrarily long. 35 This is a common assumption in New Keynesian models used for monetary policy analysis, though in models where Ricardian Equivalence would hold, it is without loss of generality.…”

Expectations Formation, Sticky Prices, and the ZLB

“…Farhi and Werning (2017) study the combination of level-k thinking, incomplete markets, and occasionally binding borrowing constraints and find that all features together provide a solution to the forward-guidance puzzle, but they abstract their analysis from the ZLB. 7 Recent work by Woodford (2018) and Woodford and Xie (2019) studies a different aspect of limitations to fully rational expectations formation -namely, finite planning horizons. A related, but conceptually different, literature addresses higher-order beliefs.…”

Does rationality matter to the central bank?

“…At the same time, Woodford and Xie (2019) theoretically argue that the rational expectation is an idealisation and show that under finite forward expectation, Ricardian Equivalence is violated and aggregated demand can be stimulated by the government transfers. Moreover, based on the simulation result that in absence of fiscal policy changes, inflation targeting causes the price level to fall below the target path when zero lower bound is binding so that temporary price level is more effective to achieve the target path, they show that different finite forward-expectation of economic agents can generate different predicted dynamics of price level and conclude that the finite expectation is shorter, the rule based temporary price level targeting is more efficient.…”