INTRODUCTION

It is apparent from the results so far that there is a gap between the qualitative and econometric evidence. The former predicts not only that central banks can be fairly easily classified according to the degree of statutory autonomy enjoyed vis-à-vis government but that there is also a clear empirical connection between their independence vis-à-vis government and average inflation performance. By contrast, the econometric evidence would lead one to conclude that central banks in the industrial world are not as different as the qualitative evidence implies. This chapter attempts to provide explanations for the conflicting evidence. It is argued that certain elements in the measurement of central bank behavior are difficult to quantify, are imperfectly measured, or have evolved over time in a manner that is not easily reconciled by the two approaches.

This possibility was already discussed in Chapter 2. Moreover, while the exchange rate regime clearly matters, it appears to matter less than would be suggested by the attempts to refine existing classifications in relation to textbook descriptions. Of course, as we have seen in Chapter 2, dating a change in policy regimes is a tricky matter. Conflict between official dates and ones estimated via econometric methods rest in part on how fast individuals' expectations respond to actual changes in the variables of interest. In addition, until recently, exchange rate regime considerations rarely figured as an explicit feature of the statutory relationship between the central bank and governments.