臺大管理論叢 NTU Management Review VOL.29 NO.1

239 NTU Management Review Vol. 29 No. 1 Apr. 2019 5.5 Recognized versus Non-Recognized Assets Impairment Examination The decision to recognize asset impairment includes two facets: 1. Do managers recognize asset impairment (a binary decision)? 2. What magnitude of impairment is to be recognized (the magnitude of impairment loss decision)? This study thus uses a binary variable to replace the initial magnitude of asset impairment variable and reexamines the earnings informativeness of managerial recognition versus non-recognition decisions. In this case, the magnitude of asset impairment variable (IM t ) in Regs. (1) and (2) is replaced by the dummy variable of recognized impairment (D_IM t ). The D_IM t variable in this analysis is denoted as one for firms recognizing asset impairment (winsorized if IM t less than 0.1%), otherwise, D_IM = 0. Empirical results are reported in Table 10. The IM model in Table 10 shows that the coefficients of IM t *X t and IM t *X t3 are respectively -0.208 ( t = -1.78) and 0.050 ( t = 1.75), both statistically significant at the 10%. The coefficients of IM t *X t and IM t *X t3 are respectively -0.304 ( t = -2.08) and 0.070 ( t = 2.03) in the REV model, both statistically significant at the 5% level. We also find that the coefficient of REV t *IM t *X t3 is -0.287 ( t = -1.85), which is negative and statistically significant at the 10% level. The combined coefficient of IM t *X t3 and REV t *IM t *X t3 ( β 8 + β 13 ) is -0.217 ( t = -1.43), negative and statistically insignificant. The additional diagnoses do not qualitatively change the primary results. Thus, both managerial decisions to recognize asset impairment and the magnitude of assets impairment to be recognized are related to earnings informativeness.