Sticky Prices as Coordination Failure

Abstract

This paper links the and menu cost approaches to the microeconomic foundations of Keynesian macroeconomics. If a firm's desired price is increasing in others' prices, then the gain from price adjustment after a nominal shock is greater if others adjust. This leads to multiple equilibria in the degree of rigidity. Welfare may be much higher in the equilibria with less rigidity. Thus, nominal rigidity arises from a failure to coordinate price changes. (JEL E12, E30) Keynesian macroeconomics waned in the 1970's because economists grew disenchanted with its weak microeconomic foundations. The central difficulty was that Keynesian models were based on ad hoc rigidities in nominal wages and prices. The 1980's produced two approaches to reviving Keynesian theory. The menu cost literature (N. Gregory Mankiw, 1985; George Akerlof and Janet Yellen, 1985) seeks to provide rigorous explanations for nominal rigidities. These papers argue that small frictions in price setting are enough to produce large nominal rigidities. In contrast, the literature (Russell Cooper and Andrew John, 1988) abandons nominal rigidities and seeks alternative foundations for Keynesian models. The central idea is that many economic activities, such as production (e.g., John Bryant, 1983), trade (e.g., Peter Diamond, 1982), and investment (e.g., Nobuhiro Kiyotaki, 1988), exhibit synergism or complementarity: one agent's optimal level of activity depends positively on others' activity. Strategic complementarity can lead to multiple equilibria, with high-activity equilibria superior to low-activity equilibria. Thus, an economy may be stuck in an underemployment even though a superior equilibrium exists. Models with nominal rigidities and models with coordination failures are often presented as competing paradigms.1 This paper shows that this view is incorrect. We take a step toward unifying the foundations of Keynesian economics by showing that the two sets of ideas are highly complementary. Nominal rigidity arises from a failure to coordinate price changes. This failure has the essential features of coordination failures in previous models. Flexibility in one firm's price increases the incentives for other firms to make their prices flexible. This strategic complementarity leads to multiple equilibria in the degree of nominal rigidity. Equilibria with less rigidity (more active price adjustment) are often Pareto superior to equilibria with more rigidity. These results contribute to our understanding both of coordination failure and of nominal rigidity. The range of Keynesian phenomena explained by coordination failures is greatly expanded. Previous coordination-failure models contain only real variables and thus ascribe no role to monetary policy or other determinants of nominal spending. Our results suggest that coordination failure is at the root of inefficient nonneutralities of money. Theories of nominal rigidities gain new empirical and policy im