Target Ratcheting in Common and Unique Performance Measures

While most performance measurement systems use multiple measures, research on target revision in multi-measure systems is rare. Examining the target-setting practice in such a system, we hypothesize that the performance variances in common and unique measures are differentially informative of an agent's effort level and this, in turn, affects a principal's decision on the degree of commitment to incomplete incorporation of past performance information into new targets. We analyze proprietary data on the performance evaluation of strategic unit managers in a Korean conglomerate to report (1) intertemporal serial correlations of target achievement in both types of measures and (2) partial (full) target difficulty revision in common (unique) measures. We also confirm the results, while classifying measures into financial and nonfinancial ones. With the findings from outside an experimental setting, we contribute to the accounting literature on a common measures bias as well as on target ratcheting