臺大管理論叢 NTU Management Review VOL.30 NO.3

63 NTU Management Review Vol. 30 No. 3 Dec. 2020 Impulse Response (Day) Cumulative Response (250 Days) 5 6 7 8 9 -8.06E-06 5.44E-07 1.28E-06 8.01E-07 -0.0156 -0.0298 3.9671 2.4400 2.1040 1.3776 1.3019 19.08721** -2.61E-03 -2.32E-03 -1.63E-03 -1.30E-03 -1.11E-03 -0.0650 4.9403 4.0537 3.5287 3.1656 2.6567 20.2683** -0.0016 -0.0010 -2.65E-04 -4.26E-04 -3.50E-04 -0.0552 7.9630 5.4851 4.6967 4.4587 3.5432 35.83848* Based on this common factor, I find no price reversal and concluded that the weather effects are permanent. It is possible that temporary effects exist for some individual weather variables. But when all the variables are represented by the common factor, the temporary effects for those weather variables cancel. I examine how each of the seven individual weather variables contributes to temporary and permanent effects on stock returns. I substitute the fourth principal component in equation (1) by the individual weather variables. I report the estimation results in Table 5. Air pressure and wind speed contribute significantly to the effects on stock returns. For the air pressure variable, the impulse response is significant for day 1; for the wind speed variable, the impulse responses are significant for days 1 and 2. The cumulative impulses are significant for both variables. I observe price reversal for neither the air pressure nor the wind speed. The temporary parts of the effects do not exist. They do not cancel out among the different individual weather variables. The weather effects have only the permanent parts. 9 and 26 for the triennial setting and the annual sample with control variables, respectively. a and b = the last and the first subsamples are two years. *and** = significant at the 90% and 95% confidence levels, respectively. The full tables are available from the author upon request.