Central-Bank Independence, Economic Behavior, and Optimal Term Lengths: Comment
评论Waller和Walsh（1996）关于央行行长最优任期的研究，通过内生化保守程度，扩展了其模型，并澄清了相关论点。
In an earlier article in this Review, Christopher J. Waller and Carl E. Walsh (1996) (hereafter, WW the greater this influence, the weaker is the independence of the central bank. Institutional arrangements designed to overcome the time-inconsistency problem and political uncertainty constitute another dimension of central-bank independence, since these legal arrangements are ex ante choices which should not be dismissed ex post. W&W examine two legal arrangements, the length of the central banker’s term in office and the degree of conservativeness expressed as the weight attached to the inflation target. While partisanship and the degree of conservativeness are assumed to be exogenous, W&W only allow the length of the term in office to be optimally determined. They successfully obtain a finite optimal term length by introducing persistent shifts in long-run social preferences of inflation. Furthermore, they conclude that the appointment of a conservative central banker increases the optimal term length and results in lower average inflation, but need not necessarily increase output volatility. The obvious question is then: Why should a conservative central banker be appointed? As shown in many studies (e.g., Kenneth Rogoff, 1985), the highest degree of conservativeness is actually not optimal from the point of view of social welfare maximization. Moreover, since several types of economic distortions exist in W&W’s economy, that is heterogeneity of the preferred inflation rate, political uncertainty over the short-run median voter’s preferred inflation rate, persistent shifts in social preferences of inflation, and discretionary ways of implementing monetary policy, social welfare would further increase if more “instruments” were available to society. In this Comment, I extend W&W’s study by endogenizing the determination of the optimal degree of conservativeness. A simplified W&W model is set up in Section I, and some of the arguments in W&W are then clarified in Section II. Concluding remarks are given in Section III.