INTRODUCTION

The quantitative study of central bank behavior has a long history. Introduced in the heyday of trust in fine tuning, functions describing how central banks react to economic conditions, called reaction functions, were intended to convey the belief that a central bank or government could achieve a set of economic goals by solving an optimal control problem. Political or institutional considerations did not matter initially since policy makers were assumed to have the requisite instruments at hand to optimally achieve desired objectives.

Until recently, and other than general dissatisfaction with the concept of fine tuning born out of the stagflation of the 1970s, two other rather technical issues led economists to shy away from estimating reaction functions for a time. These were the temporal instability of estimates and the inability of standard functions to reveal policy makers' preferences. By contrast, political scientists never lost their enthusiasm for the approach as their concerns primarily dealt with political influences on the macroeconomy in general. The ability to separately identify the preferences of the central bank or government from those of the public was considered secondary, perhaps because the State and the central bank were not viewed as separate institutions as such.

Recently, reaction functions have been interpreted, as we shall see, as a device to reflect rules like behavior apparently adopted by several central banks. Some of the earlier technical problems remained but economists overcame, to some extent, their displeasure with the reaction function approach thanks in part to several important developments the econometric analysis of time series.