Identifying firms' tax loss carry-forward status: The accuracy of database-ariven methods

摘要

Due to data restrictions, empirical tax research commonly relies on database-driven methods as a means of identifying firms' tax loss carry-forward (TLCF) status. Employing a panel of listed Italian parent companies, I am the first to empirically examine the accuracy of database-driven methods in predicting the availability and the amount of TLCF at single-firm level. In order to assess the accuracy of database-driven identification methods, I compare firms' true TLCF status, as determined based on IFRS statement information, to the TLCF status predictions of the methods examined. I find that database-driven methods do not perform well in predicting the availability of TLCF. They perform poorly in predicting the amount of TLCF available to firms. Empirical studies that rely on database-driven identification methods might thus not be able to derive reliable results regarding the impact of TLCF. My findings thus indicate that there is a strong need for firm-specific TLCF information provided by local authorities in empirical tax research.