Page 201 - IJOCTA-15-1

Key drivers of volatility in BIST100 firms using machine learning segmentation
Table 5. Descriptive statistics of variables

Variables Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Obs.
vol 0.056 0.052 0.432 0.020 0.022 6.684 104.813 828
co 5.396 1.525 402.990 0.238 25.072 9.916 119.203 828
bo 0.487 0.511 1.039 0.003 0.239 −0.275 −0.790 828
adh 1.090 0.069 290.941 0.000 13.774 17.366 327.384 828
roa 0.080 0.065 0.635 −0.323 0.095 1.190 4.859 828
roe 0.142 0.147 21.819 −39.741 1.658 −15.151 470.799 828
pddd 3.173 1.462 249.481 0.000 12.760 15.853 281.071 828
beta 0.830 0.817 1.656 0.234 0.229 0.220 0.063 828
Sources: Authors’ Finding.

Table 6. Correlation of “vol” variable with independent variables
vol
Variables Model 1 Model 2 Model 3
(46 Firms) (37 Low Volatility Firms) (9 High Volatility Firms)
co −0.014 −0.056 −0.078
bo −0.012 0.061 −0.028
adh 0.016 0.030 0.061
roa 0.011 0.073 −0.006
roe 0.004 0.008 0.043
pddd 0.037 −0.014 0.000
beta 0.293 0.387 0.180
Sources: Authors’ Finding.

Table 7. Cross section dependency test results

B-P LM Test Pesaran CD Test
Variables
Statistics Prob. Statistics Prob
vol 6705.27 0.001 76.89 0.001
co 3674.58 0.001 2.06 0.038
bo 4375.35 0.001 12.53 0.001
adh 3468.27 0.001 29.31 0.001
roa 2237.23 0.001 10.29 0.001
roe 2322.83 0.001 14.75 0.001
pddd 2812.91 0.001 24.37 0.001
beta 2316.85 0.001 30.65 0.001
Sources: Authors’ Finding.

Table 8. Cross section dependency test results

Harris-Tzavalis Test
Variables
Statistic Prob.
vol 0,274 0,0001
co 0,570 0,0001
bo 0,743 0,0001
adh 0,513 0,0001
roa 0,382 0,0001
roe −0, 429 0,0001
pddd 0,530 0,0001
beta 0,209 0,0001
Sources: Authors’ Finding.

Panel regression analysis was performed with the classical (pooled) model for Model 3. The
fixed effects for Model 1 and Model 2, and with results of the panel regression analysis for each

195