This dataset contains the EQSCORE, a fundamental-analysis based firm-year measure of earnings quality, for US-listed firms from 2010 to 2020. The methodology of the EQSCORE is described and validated in Khan, Peddireddy, and Rajgopal (2024).
The sample includes firms listed on the three main exchanges (NYSE, NASDAQ, AMEX) with share codes 10 and 11, and at least five years of existence on COMPUSTAT. Additionally, we exclude industries with SIC codes 6000 - 6999, 9000 - 9999, 4950 - 4961, 4970 - 4971, and 4990 - 4991.
Our methodology for computing the EQSCORE involves assigning weights to each constituent variable based on its ability to predict AAERs within each industry (Fama-French 48 classification) over the past five years. The 'SUBSAMPLE' sheet provides the EQSCOREs for firms in industries with a history of AAERs. To address the issue of sample attrition, we compute the EQSCOREs for firms in industries without a history of AAERs by applying average weights across the entire sample over the past five years. Consequently, the 'FULLSAMPLE' sheet provides the EQSCOREs for firms in all industries. Our findings regarding the validation of the EQSCORE are consistent when using both the partial and full sample.
This dataset is freely available for non-commercial uses. We kindly request that if you use this data, you reference our paper and acknowledge the data source.
Reference:
Khan, U., Peddireddy,V., & Rajgopal,S. (2024). “Earnings Quality on the Street”.