臺大管理論叢 NTU Management Review VOL.30 NO.2

155 NTU Management Review Vol. 30 No. 2 Aug. 2020 (i.e., -0.091 and -0.144, respectively) and highly significant at the 1% level (i.e., t -statistics = -21.58 and -31.02, respectively). They make OCI adjustments 0.09% or 0.111% smaller, respectively, when information inputs reveal deterioration rather than improvement in business conditions. Note that is -0.002 and loses its significance ( t -statistic = -0.8). Thus, forward market-based inputs contain little information about current OCI adjustments. In contrast, is 0.295 ( t -statistic = 70.54), suggesting that forward accounting-based information has explanatory power for current OCI adjustments. Table 3 Time Span and Asymmetric OCI Adjustments Expected (1) (2) (3) (4) Variable sign p = 2, q = 2 p =3, q =3 p =4, q =4 Forward β 0 -0.056*** -0.098*** -0.143*** 0.01*** (-25.6) (-29.96) (-30.72) (6.27) β M 1 + 0.104*** 0.134*** 0.153*** -0.002 (33.07) (35.04) (34.03) (-0.8) β M 2 – -0.042*** -0.051*** -0.029*** -0.091*** (-6.56) (-6.01) (-2.62) (-21.58) β S 1 + 0.396*** 0.439*** 0.466*** 0.295*** (79.36) (72.32) (64.99) (70.54) β S 2 – -0.032*** -0.05*** -0.107*** -0.144*** (-3.5) (-3.36) (-5.24) (-31.02) Adj. R 2 0.281 0.312 0.318 0.119 Hypothesis test of regression β M 1 + β M 2 = 0 0.062*** 0.083*** 0.124*** -0.093*** F-value (165.40) (156.94) (196.44) (1,512.23) β S 1 + β S 2 = 0 0.364*** 0.389*** 0.359*** 0.151*** F-value (1,813.49) (843.45) (382.65) (1,441.57) N 28,100 22,629 18,384 49,408 Note: Models (1) to (3) present results from a pooled cross sectional OLS regression of the following model: