Policy Rate Decisions and Unbiased Parameter Estimation in Conventionally Estimated Monetary Policy Rules

The use of estimated policy rules has been on the rise over the past few decades as central banks have increasingly relied on them as policy benchmarks. While simple, conventionally estimated rules have proven insightful, their value is generally seen to depend, among other things, on the ability of the benchmark to accurately reflect the policy environment and on the relevance of the econometric assumptions behind the estimation method. This paper addresses a potential source of econometric bias that might naturally arise and adversely affect the accuracy of conventionally estimated policy rules as benchmarks. In particular, the discrete nature of the policy rate setting process at central banks leaves open the possibility that observed policy rate changes may include significant rounding errors. If so, parameter estimates using conventional econometric methods could be seriously biased; technically, this is an example of a censoring bias. To address this concern, the paper offers a new method for estimating monetary policy rules and demonstrates the ability of the resulting bias-adjusted policy rules to outperform conventionally estimated ones in characterizing the policy environments in the cases of the Czech Republic and the United States.