Table 4

A multiple regression of the total premium paid to target stockholders (Premium) on the pre-bid runup (Runup) and dummy variables for various charactersitics of successful and unsuccessful mergers and tender offers for exchange-listed target firms, 1975-91:

Premium = a0 + b0 Runup + ak Dki + bk (Dki Runup) + ui, k = 1, . . ., 9

where Premium = Runup + Markup. The sample excludes deals that took longer than one year to consummate and target firms whose equity value is small (below $10 million) or whose pre-runup stock price is low (below $2 per share). The dummy variables appear separately to represent changes in the intercept ak, and they interact with Runup to represent differences in the effect of pre-bid runups on the total premium paid, bk. Runup is the cumulative abnormal return to the target's stock from day -42 to day -1 relative to the first bid. Markup is the cumulative abnormal return to the target's stock from the day of the first bid through delisting or 126 trading days after the first bid, whichever comes first. The characteristics of deals that are used in the regression include: the target firm is taken over (Success), there is news implying that a bid might be forthcoming (News), the target firm has a poison pill in place (Pill), there are multiple bidders (Auctions), the deal is a tender offer (Tender Offer), the deal is a management buyout (MBO), there is an all-cash payment to target shareholders (Cash), there is an all-equity payment to target shareholders (Equity), and the S.E.C. later accused someone of engaging in insider trading prior to the takeover (Insider). The substitution hypothesis implies b < 1, while the markup pricing hypothesis implies b = 1. Rsquared is the adjusted coefficient of determination. White's (1980) heteroskedasticity-consistent standard errors are used. The t-statistic for the runup coefficient tests whether it is equal to one; the other t-statistics test whether the coefficients equal zero. The tests for whether all of the coefficients representing intercept (ak) and slope changes (bk) equal zero, which have large-sample chi-square(9) distributions, and their p-values, are also shown.

SampleRunup Coefficient, bkT-statistic, bk = 1 or 0Intercept, akT-statistic, ak = 0
Poison Pill0.0780.480.0792.54
Tender Offer0.1051.270.0783.72
Degrees of freedom1,503
Standard error of the regression0.268
Test for joint significance, Chi-square(9)21.2259.9

© Copyright 1996, G. William Schwert
Return to Markup Pricing Page