As in the previous Part I include controls for Delaware incorporation (DEINC), whether the
controller is a founder or family group, and the consideration used (STOCK, set to 1 for stock
deals). I also control for the log of the transaction value (LNVAL), on the view that tender offers
entail higher fixed costs than mergers, and so a tender offer becomes more likely as these fixed
costs can be amortized over a larger deal value. In certain specifications I also include PCTINST
and NUMINST to control for a potential influence of institutional investors on the controller’s
choice of transactional form.
The model is run as a probit regression, though results are virtually identical if I run the
model as a logit. Because decisions on transactional form may not be independent within law
firm, standard errors are adjusted for clustering by the controller’s outside counsel. The results
are reported in Tables 5A and 5B.
[insert Tables 5A & 5B here]
coefficient is positive and highly significant in all models, in the predicted direction. In Table
5B, the CONTROL coefficient continues to be positive but is statistically significant at 95%
confidence in only one out of four specifications.
The results also provide some support for Hypothesis H4 and the interaction between H3 and
H4. The coefficients for the law firm experience variables are positive in all models. In Table
5A, the EXPERIENCED? dummy variable and the interaction EXPERIENCED?*CONTROL
are statistically significant at 95% confidence. In Table 5B, the law firm experience variables
are statistically significant in all specifications. One possible interpretation of these findings is
that law firm experience plays an increasingly important role in the “sweet spot” of a large
controller in a Delaware target. Outside of this sweet spot, the influence of the controller’s