(a) Descriptive Statistics and Univariate Analysis
In successful takeovers, a target firm director is considered replaced if he or she does not sit on the acquiring firm board two years after the date of takeover completion. For unsuccessful takeovers, turnover is assessed by examining the target firm board two years after offer termination. A two-year period is chosen as it provides sufficient time for any adjustment to directorships to take place. This length of time is comparable to the two-year post-takeover period used to determine management turnover in previous studies (e.g., Kini et al., 1995; and Kini et al., 2004; in the United States and Franks and Mayer, 1996; and Kennedy and Limmack, 1996; in the United Kingdom). Furthermore, it is a more extensive period of time than that used in Harford's study (2003), which establishes director turnover using the first proxy statement following either takeover completion for successful bids or takeover termination in unsuccessful offers.
Details on target firm director turnover are provided in Panel A of Table 2. Information is provided on an overall basis and then partitioned by takeover outcome. The sample size is reduced to 574 directors, as it is not possible to assess the impact of the takeover on directors whose firm is successfully acquired by a foreign company not listed on the ASX or an Australian non-listed entity.
Turnover of Target Board Members
|Percentage turnover of directors|| || || || z stat|
|All directors (n= 574)||79.26||86.76||66.18||5.84***|
|CEOs (n= 80)||71.25||84.31||48.28||3.42***|
|Executive directors (n= 147)||74.83||85.71||53.06||4.30***|
|Non-executive directors (n= 427)||80.61||87.13||69.68||4.40***|
|Blockholders (n= 53)||72.60||86.11||59.46||2.55**|
|Cross-sit directors (n= 35)||42.86||38.46||62.50||−1.20|
|Accept recommendation (n= 321)||85.40||85.37||88.89||−0.50|
|Reject recommendation (n= 148)||74.32||100.00||60.53||5.35***|
|Percentage of board replaced (mean)|| || || || t stat|
|Percentage of boards where|| || || || z-stat|
|Entire board is removed||56.07||63.08||46.34||1.69*|
|Entire board is retained||3.77||0.00||9.30||−2.59***|
|Panel B: Director Turnover in Non-target Firms||2000||2001||2002|| |
|Percentage turnover of directors|
|All directors||32.69||34.23||31.80|| |
|Executive directors||31.66||31.71||30.87|| |
|Non-executive directors||33.18||35.37||32.21|| |
As in Harford (2003), target firm directors are significantly more likely to lose their board seat following a successful takeover than an unsuccessful bid. In completed takeovers, 87% of target directors lose their position, compared to 66% in unsuccessful offers.8 The significant difference in director turnover across takeover outcome holds for executive and non-executive directors, CEOs, and directors who are target blockholders. The impact of takeover outcome on the target board is also highlighted in the last two rows of Panel A of Table 2. In 63% of successful takeovers, the entire board is removed, compared to 46% in unsuccessful bids. Where the target firm board recommends offer acceptance, 85% of directors are replaced within two years of the bid, with the rate of turnover statistically similar for unsuccessful and successful bids. In contrast, in hostile takeovers, director retention varies by takeover outcome. In unsuccessful takeovers, 61% of directors no longer retain their board position. In comparison, there is a real cost to the board of hard bargaining if the bid succeeds as 100% of directors lose their board seat.
Our results indicate that a failed bid does not ensure retention of a board seat, as only 34% of the target board remains in position two years after bid termination. Although, retention rates are higher for executive directors than non-executive directors after an unsuccessful takeover, still approximately half of executive directors no longer hold a board seat. This finding stands in contrast to Harford (2003), who finds that 85% of US boards are retained following a terminated merger. A partial explanation for this different finding is the longer period of time used to determine director turnover in this study. When turnover is recalculated using one year after the termination of a takeover, director turnover is reduced to 35%.
As a means of comparison with the rate of director turnover following a takeover offer, we calculate the number of other ASX directors (i.e., non-target) on the Connect 4 Boardroom database that were replaced in the subsequent two years for 2000, 2001 and 2002. These statistics are presented in Panel B of Table 2. The rate of turnover for directors not involved in takeover offers is similar for each of the three years, with approximately 30–35% of directors being replaced in the subsequent two-year period. Highlighting the impact of a takeover, the frequency of replacement of directors not involved in a takeover is significantly lower to the turnover of directors following both successful and unsuccessful takeovers.
As our findings indicate a higher rate of director turnover in unsuccessful bids than documented in the United Kingdom and the United States, we determine the reason target firm directors are replaced by searching through announcements made to the ASX and disclosures in target firm annual reports. These explanations are summarised in Table 3. The majority of directors leave the target board following their resignation. Unfortunately, in the vast majority of cases no reason is provided for the director's resignation and where a reason is given, the explanation is essentially ‘the pursuit of other interests’ by the director. Similar to the findings of Denis and Serrano (1996), a substantial proportion of directors cease their board position after the purchase of a substantial stake in the target by a new shareholder. In only three cases was the director voted out by the company's shareholders.
Reasons for Director Turnover Following Unsuccessful Takeovers
|Resigned|| 76|| 56.7|
|Removed following purchase of major interest by a new stakeholder|| 34|| 25.4|
|Dismissed on appointment of liquidator|| 14|| 10.4|
|Retired|| 6|| 4.5|
|Voted out by shareholders|| 3|| 2.2|
|Replaced as shareholder nominee|| 1|| 0.8|
(b) Multivariate Analysis
To gain insight into factors that influence the removal of directors from their target board seat post-takeover, we estimate a logit regression model, equation (1a), that includes explanatory variables for target firm performance, director characteristics, and takeover characteristics. The dependent variable (Turnover) is coded as 1 if the director does not retain his or her board seat post-takeover. To allow the influence of the independent variables to vary across takeover outcome, the model is estimated separately for successful and unsuccessful takeovers.
We include a number of variables to test for the ex post settling-up hypothesis. These include target firm performance (Perf), measured as the target firm buy-hold abnormal return over the 24 months to 6 months prior to the takeover announcement. Abnormal returns are calculated by subtracting the return on the All Ordinaries Accumulation Index from each firm's return. Share prices are extracted from the Core Research Database maintained by the Securities Industry Research Centre of Asia-Pacific. It is expected that the probability of turnover will be negatively related to target firm performance.
Directors' hostility is included in the model of equation (1a) using a dummy variable indicating takeovers where the initial recommendation of the target board is to reject the offer (Hostile). The recommendation of the target board is obtained from the target response to the takeover lodged with the ASX. Consistent with Harford (2003), we expect that director turnover will be higher in successful hostile takeovers. In unsuccessful takeovers, if shareholders benefit from the defeat of a hostile takeover, the rejection of the takeover may increase director retention. Alternatively, if hostility is not considered to be in the interest of shareholders, director turnover may increase. As hostility that follows a period of poor prior performance is unlikely to be in the interest of shareholders, we also include an interaction variable in equation (1a) between target firm performance and hostility (PerfHost).
We also include dummy variables to denote takeovers where there are competing bidders (Multiple) or where the bidding firm increases its offer price (Revision). The level of effort exerted by target firm directors is proxied using the relative size (Relsize) of the target to the bidding firm and a dummy variable denoting foreign acquirers (Foreign). The relative size of the target to the bidder is calculated using the market capitalisation of the target and bidder three months before the takeover announcement. The ex post settling-up hypothesis predicts that in unsuccessful takeovers, these four variables will have a negative coefficient. In successful takeovers, encouraging a competitive takeover contest or extracting a higher price may lead to antagonism towards the target firm board from the successful acquirer, resulting in an increased rate of director turnover. Directors of relatively small targets are expected to have a higher rate of turnover in successful bids owing to their reduced bargaining power and the lower demand by the acquiring firm for the expertise and experience of directors of a smaller firm. In a successful bid by a foreign firm, whilst target directors may be more likely to retain their board seat for their local expertise, it is also possible for them to be removed from the board because of cultural and language differences.
To test the reputation and experience hypothesis, we include the number of other directorships held by the director at the time of the takeover announcement (Otherdir). We expect this variable will be negatively related to the probability of turnover from the target firm board.
The remaining variables included in equation (1a) are primarily control variables. We use dummy variables to control for four director characteristics: These variables denote whether the director is a blockholder (Block), an executive director (Exec), a managing director/CEO (CEO), or a director who cross-sits on the board of the bidder (Cross). It is predicted that the turnover of a target director's board seat will be lower for executive directors and CEOs, as their expertise is likely to be of greater value post-takeover. Directors who already hold a seat on the board of the acquiring firm are predicted to be removed less frequently. Harford (2003) finds that director blockholders in successful takeovers have a lower rate of turnover, consistent with these directors using the influence of their shareholding to negotiate a position on the acquiring firm board. Following the results in Kini et al. (1995), we include a dummy variable highlighting target firm boards with a majority of non-executive directors (Boardind). If boards dominated by insiders are less likely to discipline management, turnover will be negatively related to an independent target firm board.
We control for the takeover premium, as a higher premium may indicate greater possible efficiency gains from removing poorly performing managers (Kennedy and Limmack, 1996). A positive association would also be consistent with Hartzell, Ofek and Yermack (2004), who report that CEOs may be willing to sell their firms at a lower premium in exchange for a board seat post-takeover. The takeover premium is measured as the offer price less the target share price 20 days prior to the takeover announcement, divided by the target share price 20 days before the announcement (Premium). The final control variable included in the model of equation (1a) is an indicator variable signifying bidder and target pairs that are in the same Global Industry Classification Standard (GICS) industry (Industry). It is expected that turnover will be higher where the bidder and target firm operate in the same industry, as the acquiring firm is less likely to require the target director's expertise.
The model of directors' turnover can be summarised as follows:
As equation (1a) is estimated at the director level, each target firm is included multiple times. We also estimate an OLS variation of the model at firm level (after excluding director level variables), with the dependent variable being respecified as the proportion of the target board that is replaced (Propturn). This alternative model is:
The results of estimating equation (1a) are provided in Table 4. The model is estimated separately for successful and unsuccessful takeovers and then within takeover outcome for executive and non-executive directors.9 We first discuss the results for successful takeovers, which provide significant coefficients only for non-executive directors. Non-executive directors are significantly more likely to be replaced following a period of poor performance, consistent with ex post settling-up. This result is in contrast to Harford (2003), who reports better performance increases retention for insider directors. Given that non-executive directors are not involved in the day-to-day operations of the firm, it is surprising that they be penalised for poor prior firm performance. Perhaps in line with the argument of Jensen (1986) and Florou (2005), outside directors are punished for not disciplining poorly performing executives. Alternatively, the turnover of non-executive directors suggests that their expertise is no longer required. Consistent with the reputation hypothesis, an increasing number of other directorships reduce the likelihood that non-executive directors will be removed from the board.
Model of Turnover of Target Directorship at the Director Level
|(−0.68)|| || ||(−2.23)**|| || |
|(−0.16)||(−0.32)|| ||(−0.19)||(0.75)|| |
|% Classified Correctly||89.19||87.76||91.18||75.00||77.61||74.84|
As expected, the control variables indicate that turnover is decreased for non-executive directors, when the target and bidder are in dissimilar industries and when the target director has a seat on the board of the bidder. Unexpectedly, the results indicate that turnover is higher for directors when the board is independent. This finding is likely explained by the bidder not requiring the services of a greater number of non-executive directors who lack management expertise with the target firm.10 Similar to Martin and McConnell (1991), we find that neither takeover hostility nor bid premium influences director turnover. The insignificant coefficients on Multiple, Premium, and Revision indicate that engaging in hard bargaining does not influence the probability that a director will be retained by the acquiring firm.11
The results for unsuccessful takeovers indicate that executive directors have a lower frequency of separating from the target firm than non-executive directors, thereby suggesting that non-executive directors believe they can no longer make a worthwhile contribution to the firm. Executive directors are also more likely to retain their board seat if the takeover was hostile, indicating that the failure to complete a friendly takeover results in pressure being applied for their resignation. Alternatively, the reputation of these directors allows them to retain their board seat. Although the insignificant coefficient on Perf and Revision is inconsistent with ex post settling-up, non-executive directors are more likely to retain their board seat when they exert greater effort during the takeover (negotiating with a relatively large or foreign bidder) or conduct an auction for the firm.12 The significant negative coefficient on PerfHost also supports the ex post settling-up hypothesis, as director turnover is higher when a poorly performing target defeats a takeover.13
The results for the firm-level variation of the model of equation (1) are presented in Table 5 and provide results broadly consistent with the director-level model. In successful takeovers, the proportion of the board replaced is significantly higher in hostile bids and following a period of poor target performance. For unsuccessful takeovers, the proportion of directors who leave the target firm board is reduced following a hostile takeover offer, a takeover contest that involved multiple bidders, or the negotiation of an increased offer price. The significant results on Multiple, Revision, Relsize and PerfHost support the ex post setting up hypothesis.
Model of Turnover of Target Directorship at the Firm Level
(c) Sensitivity Analysis
A major limitation with our analysis of target director turnover is that target directors are classified as being replaced following a successful takeover when they are not appointed to the acquiring firm board. This approach overclassifies director turnover, as some board members will obtain a position with the acquiring firm without being appointed to the firm's board. To ascertain whether this problem impacts our findings, we searched through the acquiring firm's annual report and Jobson's Yearbook of Public Companies to determine if each target director is employed in a senior position with the acquiring firm. This search determined that 16 additional directors originally classified as ‘replaced’ still hold an office with the acquiring firm. Re-estimating our regression analysis after reclassifying these observations does not change the results.