Page 213 - 34-1
P. 213

NTU Management Review Vol. 34 No. 1 Apr. 2024




                   NG_COMP  = α+ β  POST  + β  CONTROL  + ε  ,
                                     i
                                               n
                                           it
                                                          it
                                                              it
                              it
               where i is the firm index, t is time, and α and β  are the fixed effects for the year and firm
                                                         i
               i that control for market-wide changes in the information environment and unobserved
               heterogeneity between firms. The variable of interest is POST, an indicator variable
               representing the number of years since a comment letter is received. For the 4 years prior
               to receiving the first comment letter, POST is set as 0; the post period is defined as the 4

               years after a comment letter is received.


                                           4. Research Results


                   The empirical results reveal that a significant positive association existed between

               NG_COMP and POST. Therefore, the comparability of non-GAAP earnings improves
               after the firm received a comment letter. To enhance our confidence in inferring causality,
               we perform two robustness tests. First, we repeat the analysis with periods of 6、8 and 10
               years. Second, considering research and development expenditures could affect the use of

               non-GAAP earnings, we control the effect of R&D expenditures in the regression model.
               Further, we conduct four additional tests. First, the issuance of non-GAAP Compliance
               and Disclosure interpretations (C&DIs) could have effects on the non-GAAP comparabil-
               ity because C&DIs provide guidance of disclosing non-GAAP earnings. The main results

               still hold after we control for the non-GAAP C&DIs. Second, we use analyst non-GAAP
               forecast error and dispersion to test the effect of non-GAAP comment letters. We find both
               the non-GAAP forecast error and dispersion reduce if companies receive non-GAAP com-
               ment letters. Third, we control the inconsistency type of exclusions in our regression mod-

               el. Our results still remain the same. Fourth, we control firms matched on industry and size
               that receive no comment letter in the test firm’s event year. Our empirical results remain
               the same.



                                    5. Originality and Contribution


                   Our study examines the monitoring effect of SEC comment letters on the
               comparability of non-GAAP earnings. We measure the comparability of non-GAAP



                                                     205
   208   209   210   211   212   213   214   215   216   217   218