** DATA DESCRIPTION: MA-SCORE DATA 2020 **

** Note: This is an update of the dataset used in Demerjian et al. (2012); please cite the original 
paper if you choose to use this data. **

Citation: Demerjian, P., B. Lev, and S. McVay. 2012. Quantifying managerial ability: A new measure and 
validity tests. Management Science 58(7): 1229-1248.

** Note: We have changed our procedure for calculating MA-Score. In prior iterations, we estimated both
the first and second stage by Fama-French 48 industry. In this updated version, we estimate not by 
industry but rather by year for both stages. Others aspects of the estimation (inputs/outputs, control
variables) remain the same. Please contact Peter Demerjian (pdemerj@uic.edu) with any questions on the
calculation of MA-Score.

** Note: Two updates for 2022: First, the change in lease accounting has implications for the first 
stage calculation, affecting both the calculation of capitalized operating leases and net property, 
plant, and equipment as reported in Compustat. Having examined the impact of this change on 
calculation of MA_SCORE, the influence is minimal and does not require modification of the original 
formula. Please see details in the research note. Second, following Banker and Park (2021), the second 
stage regression is now estimated using OLS instead of Tobit regression; Banker and Park determine this 
to be the most robust method.

Citation: Banker, R., and H. Park. 2021. Three-stage approach to analyze managerial ability. Data Envelopment 
Journal 5(1): 27-54.

The dataset ma_score_2020 includes observations for fiscal years 1980 through 2022. This dataset was 
generated on February 17, 2024. There are 238,518 observations of nine variables; PROC MEANS output 
from SAS for the numerical variables is presented below. GVKEY is the Compustat firm descriptor. 
FISCAL_YEAR_END is the reported fiscal year end from Compustat (datadate). FIRM_EFFICIENCY_2022 is the 
first stage, DEA-based measure of total firm efficiency, with values ranging from zero to one. 
MA_SCORE_2022 is the residual from the second stage estimation, controlling for various firm-
specific factors. Decile ranks (by industry and year) are included for both firm efficiency and 
MA_SCORE_2022. INDCODE is the Fama-French 48 industry code. The dataset also includes a character 
variable, the COMPANY NAME, and the CIK code.

Variable                          N            Mean         Std Dev         Minimum         Maximum
---------------------------------------------------------------------------------------------------
FIRM_EFFICIENCY_2022         238518       0.2860597       0.1642672               0       1.0000000
FISCAL_YEAR_END              238518        15075.57         4242.17         7335.00        23010.00
GVKEY			     238518        46071.14        59222.33         1001.00       353945.00
INDCODE                      238518      27.3738083      12.1270699       1.0000000      44.0000000
MA_SCORE_2022                238518    -0.000882217       0.1197273      -0.2706511       0.6616877
FIRM_EFFICIENCY_2022_RANK    238518       0.5506796       0.2855771       0.1000000       1.0000000
MA_SCORE_2022_RANK           238518       0.5506805       0.2855875       0.1000000       1.0000000
---------------------------------------------------------------------------------------------------