8342937138

SHAMSHER MOHAMAD AND ANN U AR MD. NASSIR

where S² is the residua! variance for security i from ihe market model regression, L is the number oł observationsduring the estimaiion period (i.e. 106 days). R , is the return on the Ali Share Market Index for the k^lh dayof the esumation period. R_mn is lite return on the Ali Share Market Index for day n in the analysis period, and R_m is the average return of the Ali share Market lndex (i. e. market portfolio) for the estimaiion period. Assumingthat

summary of the results is presented in Table | be Iow.

Second, the differences in the abnormal re-tums for the eleven days surrounding the aiv nouncement for bidders offering the differeni means of payment were calculated. The dctailed resultsare presented in Tables2(A), (B) and (C) i_r, the Appendix_t and a summary of the results is pre sented in l abie 2.

individual abnormal returns are approximately normally distributed and independent across time and across securities. the statistic t and l_{| nt}, which follow unit-normal distribution (Docld and Warner. 1983) are used to test the hypotheses that the average standardised abnormal returns (AR ^ and average c umulalive standardised abnormal rei (CARj.jp,) equal zero.

where

t = VN x (AR*) and

Vn

TIT2 “

VT2 - Tl + 1

CAR_m2 (6)

TABLE 1

Summary of the abnormal returns (AR) and the two-<la\ announcement cumulative abnormal returns (CAR) f_or bidders in the cash, shares and a combination categorict The i-siatistics are in parentheses.

Means of payment

ft i 1 % *	Eventday	(^ash N=30	Shares N W	Combination N=30 ^1
	-1	0.1129	-0.3662	-0.0296
(5)		(0.618)	(-2.006)	(-0.162)
	0	-0.3699	-0.8898	-0.8871
(6)		(-2.026)	(-4.874)**	(-4.858)**

To compare the abnormal returns between the different methodsof payments, the daily mean differences in abnormal returns from these different methods were derived. For example, for day zero, the daily mean difference between bidders offering cash and lhose offering shares was calculated, and the stalisticał signiflcance in the dif ference using l-statistic was compuled as follows:

CAR (-1.0) -0.257	-1.256	-0.917
(-0.995)	(-4.860)**	(-3.550)**

** Significantly different from /.ero at 1 % lcvel

ARj) ARjjjJj    AR jharc*» *

.AR

D

S

(7)

Nc Ns

where N_t and represent the number of sampled firms in the cash and share offers respec-tively. S - and S ~ are the variances of the abnormal returns in the respective cash and shares cate-gories.

FINDINGS

Information Hypothesis

A testofthe Information hypothesis’ wasconducted in two steps. First, the abnormal returns were observed for the eleven days surrounding the takemer announcement for bidders offering cash and shares offer. I hecłetailed results are presented in Tables 1 (A), (B) and (C) in the Appendix and a

Share Bidders

The results in Table 1 indicate that, for share bidders, the abnormal return on the day prior u the First announcement(day-l) of the offer is 0.3662 which is not significantly different from zerr (t = -2.006). The abnormal return on tlu announcement day (day 0) is -0.8898 which i-signigicant at 1 % level (t =-4.874). The two-da) (day-1 and 0) announcement CAR is—1.256 whidi is also signiflcant at 1 % level (i = -4.860). The* results indicate that, on average, the shareholden of bidding Firms offering shares to the targei shareholders in takeovers experience significanth negative abnormal returns at the announcementol the offer. These results are consistent with the prediction of the information hypothesis' that the announcement of a share offer corneys a ncgativr signal to the market about the true value of the bidder s shares.

Cash Bidders

For bidders offering cash as lite means of payment

I able 1 shows that the abnormal returns on dav-

/

(AR=0.1129,1 = 0.618) and the announcement da

96

PERTAN1KA VOL. 14 NO. 1, 1991