Consequences of Data Errors for the Validity of the E-index as a Proxy for Corporate Governance

The E-index is one of the most widely used proxy variables for corporate governance in empirical research. We show that ISS data contain a significant number of measurement errors and misreported values. We use alternative databases to manually check deviations across databases and to construct a corrected E-index. We then revisit analyses conducted in three well-known and widely-cited studies using the E-index and compare results from using either ISS or corrected data to construct the E-Index. The results show that data errors in ISS result in a substantially overestimated negative association between the E-index and firm value. Moreover, data errors in ISS lead to incorrect conclusions regarding the relationship between the E-index and different measures of operating performance and analysts’ earnings forecasts. In summary, our results show that data quality is an important issue when measuring corporate governance.