Events

Prof. Matthew R. Lyle
Associate Professor
Goizueta Business School
Emory University

Compustat North America Fundamentals files are a cornerstone of empirical research in accounting and finance. While continuous adjustments to historical data can enhance comparability and correct errors, they also pose challenges to research validity. Using Compustat Snapshot, which preserves data as it existed at each point in time, we show that the same firm-fiscal-period observation is revised at least 5 times on average, with financial statement items changing by 9% of total assets. Adjustments can alter the information historically available to users, with implications for the reliability and reproducibility of research findings. Examining 35 accounting-based anomalies, we find that roughly 50% yield materially different inferences and economic magnitudes across data vintages. We examine firm, regulatory, and data provider events and characteristics associated with adjustments and offer practical guidance to researchers. Our findings highlight that data adjustments merit careful attention in empirical research.