A Feasible and Objective Concept of Optimal Monetary Policy: The Quadratic Loss Function in the Postwar Period

Abstract

The 1960s and the 1970s saw an intense debate about the optimal monetary policy instrument, to which Poole, in 1970, and Sargent and Wallace, in 1975, made seminal contributions. According to the practitioners' narrative, these authors popularized the use of quadratic loss functions in monetary economics. This narrative argues that monetary economists adopted linear-quadratic methods mainly because they delivered easy solutions to complex stochastic models. The practitioners' account identifies also the centrality of Henri Theil and Herbert Simon, focusing on their proofs that models with quadratic objective functions have the certainty equivalence property: this particular feature made the solution of these models feasible for the computers available at that time. In this paper I argue that, in addition to offering solutions feasibility, the use of a quadratic loss function to characterize the behavior of central banks also inaugurated an objective or uniform way of talking about optimality. In this respect, the tool stabilized the discourse on optimal monetary policy. My account provides a richer understanding of the quadratic approach to the monetary policy debate by analyzing, among other things, how quadratic loss functions were used in operations research and management science problems in the 1950s and the early 1960s. This account thus sheds light on the reasons why monetary economists adopted quadratic loss functions at that particular historical moment.

Keywords: quadratic loss function, linear-quadratic methods, optimal monetary policy, central bank, William Poole, Thomas Sargent, Neil Wallace, Henri Theil, Herbert Simon

JEL Classification: E61, B22, B23, E50