Cash deals are more attractive as the correlation between TCAR and the cash dummy is positive and generally significant at the 1% level.
More importantly, TCAR is positively related to the G index and, as such, negatively related to governance.
We find that while the market condition factor is positively related to TCAR as expected, the G index remains significantly positively related, in fact slightly more so.
Consistent with Jensen (1993), TCAR are significantly positive at 13.3%, 13.9%, and 17.5% over the event windows [-2, +2], [-5, +5], and [-10, +10], respectively, with significance calculated using the Wilcoxon p values.
To add further granularity, Panel A of Table V graphs the mean TCAR and ACAR versus G Index for each value of the G index from 2 to 19, the respective minimum and maximum in our cross-border dataset.
In Table V, Panel B, we provide both the linear and quadratic fitted results where the dependent variable is first TCAR as follows:
TCAR = Intercept + [[beta].sub.1] x G Index + [[beta].sub.2] x [(G Index).sup.2] + error term.
As a result, the highest TCAR may lie in the middle of the G index for cross-border deals.
Specifically, it tests whether the dictatorship premium, that is the difference between TCAR for dictators versus democracies, is significant.
Note that the G Index itself is not consistently related to TCAR, although it appears to be positively related when at all significantly related, primarily in the longest horizon tests.
More importantly, in Model 3, when all of the controls are included, the dictator premium is evident as the G Index is significantly positively related to TCAR in all but the shortest horizon.