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

Time Paths of Weather-Induced Mood Effects on Stock Returns 54 Table 1 Descriptive Statistics Panel 1.1 Index Return, Variance, and Raw Weather Variables Statistics Return 1 Variance 1 Raw Weather Variables 2 Air Pressure (Hectopascal) Cloud Cover (Decile) Mean 1.20E-04 0.0090 97.0589 5.4800 Standard Deviation 0.0154 0.0070 29.7507 1.4028 Skewness 0.0147 2.9282 0.3972 -0.5737 Excess Kurtosis 7.5761 19.2257 0.0484 -0.2306 Minimum -0.1606 0.0000 0.0000 0.0909 Maximum 0.1135 0.1109 250.5455 8.0000 Jarque-Bera Statistic 1.52E+04*** 1.07E+05*** 254.7168*** 545.8528*** Augmented Dicker-Fuller Statistic (Lag) -72.5261*** (0) -11.5031*** (11) -29.0519*** (2) -21.6138*** (6) Observations 6,335 6,335 9,651 9,566 Note: *** = significant at the 99% confidence level. 1 and 2 = statistics computed from the observed data on trading days from February 17, 1992, to December 29, 2017, and calendar days from January 1, 1991, to December 31, 2017, respectively. Panel 1.2 Correlations 1 of Deseasonalized Weather Variables Weather Variables Air Pressure Cloud Cover Ground Visibility Air Pressure 1.0000 Cloud Cover -0.1110*** 1.0000 Ground Visibility 0.0067 -0.1107*** 1.0000 Rainfall -0.0051 0.1849*** -0.1694*** Relative Humidity -0.1101*** 0.5549*** -0.2391*** Temperature -0.3454*** -0.3178*** 0.1233*** Wind Speed -0.0673*** -0.0516*** 0.2112*** Note: *** = significant at the 99% confidence level. 1 = statistics computed from the deseasonalized observed data on calendar days from February 17, 1992, to December 29, 2017 (6,335 observations). If all the weather variables are included in X t , the model in equation (1) will be complicated, and the interpretation of the results will be difficult. Because the weather variables are correlated, my strategy is to use one of their principal components (PCs) to summarize the common factors and serve as W t . In most weather (Dehaan, Madsen, and Piotroski, 2017) and sentiment (e.g., Baker and Wurgler, 2007) studies, the first PC is chosen because it best explains the common