Page 161 - 35-2

NTU Management Review Vol. 35 No. 2 Oct. 2025

Table 5 (continued) Additional Analysis: The Effect of Fair Value Model Versus
Historical Cost Model on Analyst Forecast Error

Panel B: Significance Tests of Sum of Coefficients
Coef. t-stat p-value Coef. t-stat p-value
UK + UK×POST 0.710 2.13 0.034
UK + UK×YEAR_0506 1.715 2.09 0.037
UK + UK×YEAR_0708 0.766 1.66 0.099
UK + UK×YEAR_0910 0.667 1.28 0.202
UK + UK×YEAR_1112 0.523 1.15 0.249
UK + UK×YEAR_1314 0.558 1.43 0.155
Notes: This table presents the results of the multivariate analysis of the determinants of EPS
forecast error using UK and US firm-year observations from 2002 to 2014. Panel A presents
the regression results, and Panel B details whether the sums of the coefficients of interest
are significant. The dependent variable is ABS_ERROR, measured as the absolute value
of the difference between I/B/E/S actual earnings and the mean EPS forecast, scaled by
the absolute value of the mean EPS forecast. POST is an indicator variable that equals one
after 2005 and zero otherwise. In Column (2), we replace POST with a vector of time-period
indicator variables: YEAR_0506, YEAR_0708, YEAR_0910, YEAR_1112, and YEAR_1314.
YEAR_0506 equals one for the period 2005-2006 and zero otherwise; all other time-period
indicator variables are defined accordingly. The t-statistics are calculated using robust
standard errors clustered at the firm level.

153