Value of Corporate Takeovers

by Mark L. Mitchell The Value of Corporate Takeovers

Thisarticle summarizes the resultsof threestudies of the valueof corporatetakeovers. The first study suggests that takeoversdiscipline some managerswho makevalue-reducing decisions.Specifically, firms that have made acquisitions that reduced their stock values tend to becometakeover targets, while firms that have madeacquisitions that increasedtheir valuesdo not. Furthermore,acquisitions associated with abnormalstock price declines tend to be divested,either in subsequentbust-up takeovers or duringand followingsubsequent takeoverattempts. Thesecond study suggeststhat thequality of acquisitiondecisions tends to increasewith the degreeof the acquiringfirm's leverage.The more levered the acquiringfirm, the more likelyit is that its stockprice will increaseat theannouncement of an acquisition.However, it is difficultto determinewhether managers of highly leveredfirms makebetter decisions (presumablyas a resultof the oversightprovided by debtholders)or whethermanagers who makegood decisionsare simplyable to obtainmore debt. Thelast studyargues that Congressional action to limittakeovers contributed to the10.44 per cent marketdecline on October14-16, 1987, which in turn may have triggeredthe marketcrash of October19. The stock marketreactions following Congressionalaction suggest that marketparticipants view takeoversfavorably.

D URINGTHE 1980s, mergers and acqui- dence in support of takeovers and leverage- sitions, takeovers, restructurings and restructuringactivities. other corporate-controlactivities considerably altered corporateAmerica. With junk- Do Bad Bidders Become Good Targets? bond financing allowing a firm to acquire an- Commentators have long recognized that the other firm several times its size, no firm interests of managers and shareholders can di- appeared safe from takeover. While some com- verge.2 Many mechanisms exist to solve the mentators argue that this recent wave of take- "agency" problems resulting from the separa- overs and restructuringshas increased the na- tion of ownership and control. These mecha- tion's wealth, others assert that these activities nisms include competitive labor markets, man- only serve to redistributewealth from labor and agerial compensation plans linked to stock price other stakeholders to target shareholders. performance, outside board directors, stock Since 1983, the Office of Economic Analysis ownership by managers and corporate take- (formerly, Office of the Chief Economist) at the overs. We focus on the extent to which the last U.S. Securities and Exchange Commission has mechanism, corporate takeovers, disciplines released numerous empiricalresearch papers on target management. takeovers and other corporate-controltransac- Robin Marris and Henry Manne argue that tions. This article summarizes two of these the stock prices of firms whose managers devi- papers and a third study currentlyin progress.1 ate from profit maximization are less than they While these studies do not settle the ongoing would otherwise be.3 They contend that this debate, they do provide some empirical evi- difference between actual and potential stock prices creates incentives for outside parties to acquire the firms and operate them in profit- 1. Footnotes appear at end of article. maximizing ways. Michael Jensen argues that

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O 21

The CFA Institute is collaborating with JSTOR to digitize, preserve, and extend access to Financial Analysts Journal ® www.jstor.org takeovers mitigate manager-stockholder con- flicts that arise when firms generate free cash Glossary Nontargets: Firms that do not receive friendly or flow (cash flow in excess of that necessary to hostile bids, pay greenmail, file for bankruptcy, finance positive-return investment projects).4 significantly restructure or become subject to Jensen asserts that the managers of such firms large open-marketpurchases. often use free cash flow to finance unprofitable Hostile Targets:Firms that are targets of success- ventures, ratherthan pay it out to shareholders; ful and unsuccessful hostile tender offers, proxy takeovers discipline such firms. contests (in which the dissenting shareholder Anecdotal evidence suggests that some take- attempts control) and large unsolicited open- overs are motivated by the poor acquisition marketpurchases in which the purchaser seeks records of target firms. When Sir James Gold- control. smith attempted a hostile takeover of Goodyear Friendly Targets:Firms that are targets of successful and unsuccessful friendly tender offers, Tire & Rubber Company in October 1986, for mergers and leveraged buyouts. example, he stated that he intended to sell Greenmail: The repurchase by a target firm of a Goodyear's petroleum division and concentrate large stake held by an individual shareholderor on tire operations. Goodyear had diversified small group of shareholders at a price higher into the petroleum industry in 1983, with its than the market price. The purpose of the repurchase of Celeron Oil for approximately$800 purchase generally is to end the threat of a million. On the day the Celeron acquisitionwas hostile takeover; often, the shareholder agrees announced (February8, 1983), Goodyear'sstock not to buy any more stock in the target firm for price declined 10 per cent, never to rebound. a specified period of time. The stock market had correctlyanticipated that Event Study: Empiricalexamination of an occur- Goodyear would have difficulty with petro- rence that causes investors to change their ex- pectations regardingthe discounted future cash leum. flows of a stock. Event studies rely extensively The premium Goldsmith offered presumably on the EfficientMarket Hypothesis, which holds reflected the value of Goldsmith's divestment that the price of a stock incorporates all cur- plan, which was designed to recoup share- rently availableinformation and quickly adjusts holder losses sustained as a result of Goodyear's to the release of new information. diversificationefforts. Goodyear defeated Gold- Antitakeover Provisions: Measures taken to re- smith's bid by instituting a major restructuring duce the probability of a corporate takeover, program similar to the one that Goldsmith had especially a hostile takeover. Firms, especially promised. The restructuringprogram included potential takeover targets, as well as govern- the sale of much of its petroleum assets. ments (noted in this article) implement these Many commentatorsview Goldsmith'shostile provisions. takeover attempt as a social waste, because he failed to acquire Goodyear. The empirical evidence suggests otherwise. The bid forced Good- year to institute a restructuringplan similar to sisted of 228 firms (19.7 per cent of the sample) Goldsmith's, and Goodyear's stock price re- that were targets of successful and unsuccessful mained well above the level it had been at prior hostile takeover attempts. The friendly target to Goldsmith's takeoverattempt. Can the Good- group contained 240 firms (20.7 per cent) that year case be generalized to a large sample of were targets of friendly takeovers. The miscel- takeovers? Do firms that become takeover tar- laneous category contained 90 firms (7.8 per gets have bad acquisition histories, and do cent) that paid greenmail in the absence of a bust-up takeovers correct bad prior acquisi- tender offer, filed for bankruptcy, were subject tions? to large open-market purchases where the purchaser expressed no interest in securing control, The Sample or significantly restructured (absent a takeover We studied 1,158 corporations covered by attempt). The remaining 600 firms (51.8 per Value Line. Based on its history between Janu- cent) were classified as nontargets. ary 1980 and July 1988, each firm was classified Next, we examined the Dow Jones Broadtape into one of four groups-(1) nontargets, (2) for announcements of acquisitions by the 1,158 hostile targets, (3) friendly targets and (4) mis- firms during 1982-86, including acquisitions of cellaneous firms. The hostile target group con- public and private companies and purchases of

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O 22 Table I Abnormal Stock Market Performance Associated with Firms Announcing Acquisitions During 1982-86 (t-statistics in parentheses)

Event Windowa Category 0 -1,1 -5,1 -5,40 -20,40 EntireSample -0.21** -0.08 0.14 0.70 0.57 (N = 401) (-2.18) (-0.45) (0.53) (1.05) (0.75) Nontargets 0.09 0.49** 0.82** 3.32*** 3.48*** (N = 232) (0.66) (2.19) (2.42) (3.80) (3.46) All Targets - 0.78*** -0.93*** - 1.27*** -3.38*** -3.46*** (N = 113) (-4.59) (-3.16) (-2.82) (-2.93) (-2.60) Hostile Targets -0.95*** - 1.50*** - 1.34** -3.37** -3.19** (N = 70) (-4.64) (-4.22) (-2.46) (-2.42) (-2.00) Friendly Targets -0.50* -0.01 -1.17 -3.39* -3.91* (N = 43) (-1.68) (0.02) (1.47) (-1.67) (-1.67) Miscellaneous -0.31 -0.69 0.14 -1.93 -3.33 (N = 56) (-1.01) (-1.32) (0.17) (-0.94) (-1.42) a. The event window is the length of the period for measuringabnormal returns around acquisitionannouncements. 0 is the acquisition- announcementday, -1 is the day priorto the acquisition-announcementday, 1 is the day after the acquisition-announcementday, and so forth. * Significantat the 10 per cent level. ** Significantat the 5 per cent level. ***Significant at the 1 per cent level. the assets, divisions and stock of other compa- average stock price for both hostile and friendly nies. We limited the sample to acquisitions in targets declined over 3 per cent. Almost all these which the purchase price was at least 5 per cent stock price declines are statistically significant, of the market value of the acquiring firm's especially for the hostile targets.6 In contrast, value. During this period, 280 firms (24 per cent the average stock price for nontarget firms in- of the sample) made 401 large acquisitions. This creased by 0.09 per cent on the announcement sample of 401 can be broken down into (1) 232 day and by over 3 per cent for longer trading acquisitionsby nontargets, (2) 70 acquisitionsby periods around the acquisition announcement. 48 hostile targets, (3) 43 acquisitions by 29 In most cases, these increases are statistically friendly targets and (4) 56 acquisitions by 38 greater than zero. miscellaneous firms. The results presented so far show that targets, We employed conventional event-study especially hostile targets, tend to have made methodology to measure the stock price effects acquisitions that diminished their stock prices, associated with announcements of the 401 ac- whereas nontargets tend to have made acquisi- quisitions.5 Table I provides a display of the tions that increased their stock prices. One stock price movements for several time periods plausible explanation is that the takeovers aim around the acquisition-announcementdate for to undo the inefficient acquisitions made by the the entire sample and the various subgroups. targets. If this is correct, two things should be Overall, the data reveal that target firms, espe- true. First, the divestiture rate should be higher cially hostile targets, had systematically made for targets than for nontargets. Second, the acquisitions that reduced their equity values, divested assets should be those that had the whereas nontarget firms had made acquisitions most adverse effect on stock price around the that increased their equity values. These results acquisition announcement. support the argument that the market for corporate control disciplines inefficient manage- Target-FirmDivestitures ment. To test the first proposition, we compared the The target firms on average realized a stock rate at which acquisitions made by target com- price decline of 0.78 per cent on the day of the panies were subsequently divested with the acquisition announcement. The average stock corresponding rate for nontargets. The target prices for hostile targets and friendly targets divestiture sample consisted of acquisitions by declined 0.95 and 0.50 per cent, respectively, on targets that were divested during a period rang- the acquisition-announcementday. Over longer ing from three months prior to the target's intervals around the announcement day, the reception of a takeover bid through the end of

FINANCIALANALYSTS JOURNAL / JANUARY-FEBRUARY1991 O 23 Table II AbnormalStock MarketPerformance Associated with Announcements of Acquisitions that were Subsequently Divested Versus Acquisitionsthat were Not SubsequentlyDivested (t-statisticsin parentheses)

EventWindouw Category 0 -1,1 -5,1 -5,40 -20,40 A. Acquisitions that were Subsequently Divested EntireSample - 1.26*** - 1.75*** - 1.53*** -4.01*** -5.59*** (N = 81) (-6.15) (-4.93) (-2.81) (-2.88) (-3.48) Nontargets - 1.16*** - 1.66** -0.57 2.55 2.48 (N = 21) (-2.86) (-2.30) (-0.53) (0.92) (0.78) All Targets - 1.45*** - 1.56*** -2.07*** -7.04*** -8.91*** (N = 46) (-5.58) (-3.46) (-3.01) (-3.99) (-4.38) Hostile Targets -2.01*** -2.59*** - 1.84** -4.96*** -6.35*** (N = 28) (-7.13) (-5.30) (-2.46) (-2.59) (-2.88) FriendlyTargets -0.58 0.04 -2.44* -10.27*** -12.90*** (N = 18) (-1. 19) (0.05) (-1.89) (-3.09) (-3.37) Miscellaneous -0.75 -2.38** -0.45 -3.21 -5.91 (N = 12) (-1.16) (-2.11) (-0.26) (-0.73) (-1.16)

B. Acquisitions that were Not Subsequently Divested EntireSample 0.05 0.35* 0.56** 1.89*** 2.13** (N = 320) (0.50) (1.90) (1.99) (2.63) (2.57) Nontargets 0.21 0.70*** 0.96*** 3.40*** 3.58*** (N = 211) (1.59) (3.07) (2.78) (3.80) (3.47) All Targets -0.32 -0.50 -0.72 -0.87 0.28 (N = 67) (-1.47) (-1.33) (-1.26) (-0.59) (0.17) Hostile Targets -0.25 -0.77* -1.00 -2.31 -1.08 (N = 42) (-0.94) (-1.70) (-1.45) (-1.31) (-0.53) FriendlyTargets -0.45 -0.05 -0.25 1.56 2.56 (N = 25) (-1.24) (-0.08) (-0.27) (0.64) (0.91) Miscellaneous -0.18 -0.23 0.30 -1.58 -2.63 (N = 44) (-0.56) (-0.40) (0.34) (-0.71) (-1.02) a. The event window is the length of the period for measuringabnormal returns around acquisitionannouncements. 0 is the acquisition- announcementday, -1 is the day prior to the acquisition-announcementday, 1 is the day after the acquisition-announcementday, and so forth. * Significantat the 10 per cent level. ** Significantat the 5 per cent level. Significantat the 1 per cent level. the sample period. These include divestitures Table II displays for target and nontarget by targets to defend against takeovers, divesti- firms abnormal returns for two sets of acquisi- tures made as part of restructuring programs tions-acquisitions that were subsequently di- after defeat of takeover attempts, and divesti- vested during the sample period and those that tures by acquiring firms following successful were not. For the subsample of 81 acquisitions takeovers of targets. The nontarget divestiture that were divested, average stock price declined sample consisted of acquisitions that had been 1.26 per cent on the acquisition-announcement divested by the end of the sample period. day; over a multiday trading interval around the For the entire sample of 401 acquisitions dur- acquisition announcement, average stock price ing 1982-86, 81 (20.2 per cent of the sample) had declined as much as 5 per cent. All declines are been subsequently divested by July 1988. The statisticallysignificant. For the 320 acquisitions divestiture rates differ significantly between that were not divested, average stock price nontargets and targets. Only 9.1 per cent (21/ increased 0.05 per cent on the acquisition- 232) of the acquisitions made by nontargets announcement day; over a longer window were subsequently divested, whereas 40.7 per around the announcement, average stock price cent (46/113)of the acquisitions made by targets increased up to 2 per cent. In most cases, these were subsequently divested, either in response increases are statisticallysignificant. to or following successful or unsuccessful take- The finding of significant negative abnormal over attempts. This evidence suggests that the returns associated with acquisitions that were takeover market during the 1980s functioned in subsequently divested and significant positive part to reverse the bad acquisitions of poorly abnormal returns associated with acquisitions managed firms. that were not subsequently divested suggests

FINANCIALANALYSTS JOURNAL / JANUARY-FEBRUARY1991 O 24 Table III Shareholder Equity and Long-Term Debt for Goodyear Tire & Rubber

Shares Shareholder Long-Term Stock Outstanding Equity Debt Debt/Equity Year Price (in 000's) (in 000's) (in 000's) Ratio 1982 35.000 73,917 2,587,100 1,037,100 0.40 1983 30.375 105,212 3,195,810 961,700 0.30 1984 26.000 106,113 2,758,940 950,700 0.34 1985 31.250 107,209 3,350,280 997,500 0.30 1986 41.875 109,435 4,582,590 2,487,500 0.54 1987 60.000 56,882 3,412,920 3,282,400 0.96 1988 51.125 57,246 2,926,700 3,044,800 1.04 1989 43.500 57,779 2,513,390 2,963,000 1.18

Source:Standard & Poor's CompustatServices and Standard& Poor's DailyStock Price Record.

market efficiency. On average, the market is the target firms divested their bad acquisitions, able to identify acquisition failures before any they may have avoided a takeover attempt. cash flows from the resulting business combina- The empirical evidence supports the argu- tion are known. ment that many corporate takeovers effectively The results from the sample of 46 target-firm discipline managers who do not maximize acquisitions that were subsequently divested shareholder wealth. The results also provide support the argument that the motive behind evidence that many hostile bust-up takeovers many takeovers is to undo inefficient acquisi- promote economic efficiency by moving assets tions previously made by targets. For these 46 to higher-valued uses. acquisitions, average stock price declined 1.45 per cent on the announcement day and up to Managerial Decision-Making and Capital almost 9 per cent over multiday windows. All Structure these negative movements are statistically sig- Goodyear successfully thwarted SirJames Gold- nificant. In contrast, for the 67 acquisitions that smith's bid, but only by instituting a restructur- were not divested, there were no significant ing program similar to the one proposed by negative stock price movements. Goldsmith.7 First, Goodyear repurchased half The significant differences between stock its common stock outstanding, using bank price reactions to acquisitions that were and loans. The bank loans and stock buy-back in- were not subsequently divested hold for both creased Goodyear's debt-to-equity ratio sub- friendly and hostile targets. In short, the data stantially. Second, Goodyear sold its non-tire reveal that the average negative stock price assets, including its petroleum division. Even effect associated with acquisitions by targets is after the sell-off of its assets, Goodyear's debt/ driven almost exclusively by the subset of acqui- equity ratio remained high relative to pre- sitions that were subsequently divested, either restructuringlevels. in bust-up takeovers or during or following an Table IIIdisplays Goodyear's capital structure unsuccessful takeover attempt. for the period 1982-89. During 1982-85, Good- The divestiture rate for nontargets (9.1 per year's debt/equity ratio remained relatively flat, cent) is significantly lower than the rate of ranging from 0.30 to 0.40; it increased substan- divestments by targets surroundingtheir receipt tially in 1986 and 1987 as a result of the restruc- of takeover bids (40.7 per cent). What about turing program.The increase in the debt equity/ target-firm divestitures well in advance of the ratio to 0.54 in 1986 reflects the new bank loans, receipt of takeover bids? In only two cases did which increased the firm's debt burden from $1 the target firms divest acquisitions prior to re- to $2.5 billion. Goodyeardid not repurchaseany ceiving takeover bids for themselves. The vol- of its stock until 1987;as a result of the substan- untary divestiture rate for targets (1.8 per cent) tial equity retirementin that year, its debt/equity is considerably lower than the divestiture rate ratio increased again, to 0.96. Goodyear also for nontargets. Given the results in Tables I and increased its debt burden to $3.3 billion in 1987. II, this finding suggests that those nontarget Even after majorasset sales in 1987 and 1988, firms that divested acquisitions may have Goodyear's debt burden did not fall substan- avoided takeover attempts on themselves; had tially. In 1989, its long-term debt was $3 billion,

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 G 25 only 10 per cent less than in 1987. Its debt/equity project selection. Debtholders require periodic ratio actually increased to 1.04 in 1988 and to payments. If Goodyear doesn't meet these pay- 1.18 in 1989. Given that Goodyear doesn't plan ments, debtholders can take the firm to court. many more asset sales, it appears that its debt/ Also, given the limited amount of cash on hand, equity ratio will remain relatively high for quite expansion projects must meet the test of the some time. marketplace; lenders will scrutinize potential What impact will this heavy debt burden have projects to ensure the return of their money. on Goodyear management? Newly appointed Management must be able to convince existing Chairman Tom Barrett does not feel that the debtholdersand new debtholders that proposed high debt burden has harmed Goodyear in its projects will yield profitable returns. In cases tire operations. He states, "Our debt is higher where managers have positive inside informa- and our interest payments are higher . . . but tion about a project's feasibility, but are unable we've been able to invest and continue to invest to convey that good informationto lenders, they and do the things we've needed to be competi- will have to bypass positive-net-present-value tive. "8 Barrett'sclaim is not without empirical projects. support. While Goodyear hasn't made any di- versifying investments since Goldsmith's take- An Empirical Test over attempt, it has expended more dollars on The foregoing discussion suggests that man- its core business. On March 14, 1988, Goodyear agers of highly levered firms will tend to make disclosed a plan to build a radial-tires plant, acquisitions that are favorably viewed by the costing up to $500 million. This new plant market. Furthermore, debt-burdened firms would be Goodyear's first major expansion may not be able to take on some positive-net- since Goldsmith's takeover attempt. The stock present-value projects, hence may engage in marketapproved of this decision, as Goodyear's fewer and smaller projects than firms with low stock price increased over 4 per cent during the leverage. We tested these arguments using the two-day announcement period (the day of dis- sample of 401 acquisitions described above. closure and following day). The empirical results indicate that the more What accounts for the difference between levered the acquiring firm, the more likely its Goodyear's 1983 strategy and its 1988 strategy? stock price will increase at the announcement of Perhaps the threat of takeover was greater in an acquisition. Thus managers of highly levered 1988 than in 1983; certainly Goldsmith's at- firms make better decisions, ceteris paribus. tempted raid demonstrated that a takeover was However, the highly levered firms actually ac- not out of the question. Increased threat of quired more and largerfirms than the firms with takeover could, arguably, lead to better deci- low leverage. The imposition of debt does not sion-making on the part of Goodyear manage- appear to restrict positive-net-present-value ment. Given that Goodyear had already di- projects. vested most of its earlier mistakes by 1988, While the empiricalresults support the argu- however, there would seem to be little threat of ment that a heavy debt burden induces manag- a bust-up takeover. And while a change in ers to make better decisions, they don't provide management had occurred following Gold- any direct evidence on this hypothesis. It may smith's takeover attempt, it was due not to the simply be that good managers can obtain more takeover attempt, but rather to Goodyear's debt than bad managers, and that the positive mandatory retirement policy. relation between the debt/equity ratio of the We argue that Goodyear'sincreased debt bur- acquiring firm and its return at the acquisition den is the reason behind its change in strategy announcement is picking up this good-manager and the market's assessment of its strategy.9 effect. This is not to say that the debt-monitor- Encouraged by Goldsmith's restructuringplan, ing effect is not also present; it may, however, the market did not allow Goodyear simply to be entangled with a manager-qualityeffect. sell assets and then distribute the proceeds to One way to determine the extent of the man- shareholders. Instead, the market forced Good- ager-quality effect is to examine acquisitions year to incur the heavy debt burden, only part made by firms that have significantly increased of which Goodyear could repay with the pro- their leverage. Differences between the quality ceeds from asset sales. The remaining debt of acquisitions made before and after an in- burden restricts Goodyear's discretion over crease in a firm's debt/equity ratio can reason-

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 D 26 ably be attributed to the increase in firm lever- would arguablyhave lowered the value of most age, rather than a change in manager quality. firms, and not necessarily only those firms ac- We are currently examining acquisitions made tually "in play." before and after leverage-restructuringevents to test whether the debt-monitoringeffect system- Chronology of the Bill atically operates independently of a manager- We examined the chronology of the House quality effect. Ways and Means Committee tax bill to determine the dates on which informationabout the Did Antitakeover Provisions Trigger the takeover restrictions reached the stock market. '87 Crash? Our review of the financial press revealed no Over the three trading days preceding the stock mention of the proposals before October 14. At market crash on October 19, 1987, the market 5:33 p.m. on October 13, the Dow Jones Broad- declined 10 per cent.10 Many observers claim tapereported that the Democratson the commit- that the market declines on Wednesday, Thurs- tee were near an agreement on a tax package, day and Friday (October 14-16) precipitatedthe but made no mention of changes in the tax marketcrash on the 19th. We set out to test this treatment of corporate-controltransactions. argument. We found that proposed tax changes One hour after the October 13 Broadtapestory, restrictingtakeovers contributed to the October Democratic members of the committee, in a 14-16 decline. closed caucus, agreed to tax increases, including On the evening of Tuesday, October 13, 1987, the takeover tax proposals. On the evening of the House Ways and Means Committee intro- Thursday, October 15, the full committee ap- duced a tax bill that would have affected the proved the antitakeoverprovisions. market for corporate acquisitions and financial Immediately following the October 19 crash, restructurings,especially hostile takeovers. The investment banking firms, citing the potential bill eliminated deductions for interest expenses effects of the proposed antitakeover tax provi- exceeding $5 million a year on debt incurred to sions as a cause of the crash, began lobbying to acquirethe majorityof another firm's stock or to eliminate the provisions from the House tax bill. repurchasea majorityof a firm's own stock over On Wednesday, October 28, Representative a three-year period. The bill also prohibited Dan Rostenkowski, chairman of the House interest deductions on any debt used to finance Ways and Means Committee, testifying before a hostile bid for over 20 per cent of a target's the House Rules Committee, indicated that the stock. Finally, the bill contained provisions that antitakeover tax provisions could be changed. would tax asset sales after takeovers, especially On the evening of October 29, Chairman Ros- hostile takeovers. tenkowski made a formal statement agreeing to These tax changes would have had far- modify, though not drop, the takeover tax pro- reaching effects. First, eliminating the deduct- visions. During the next month and a half, ibility of interest expenses for debt incurred in ChairmanRostenkowski maintained his willing- hostile takeovers would have reduced the num- ness to modify the tax rules on takeovers but ber of such takeovers. Second, the proposed refused to drop all the provisions. On the morn- interest restrictionswould have limited not only ing of December 16, during negotiations with takeovers, but also leveraged buyouts and re- the Senate, the House abandoned almost all the capitalizations,including stock repurchasesand takeover tax provisions. debt-for-equity swaps. Third, the debt restric- We identified five event dates where new tions would have increased the agency costs of information about the proposed takeover re- free cash flow; the more debt a company has, strictions reached the market. The top half of the more cash flow it has to pay out to claim- Table IV displays the chronology, source of holders and the less cash flow it has to invest in news and corresponding event dates for the projects with negative net present value. proposed takeover restrictions. Under the hy- Fourth, the taxation of post-acquisition asset pothesis that the proposed takeover restrictions sales would have reduced bust-up takeovers. would reduce shareholder wealth, the market Finally, inasmuch as takeovers and the threat of should have declined on October 14 and 16 and takeovers reduce agency costs arising from the increased on October 29 and 30 and December separation of ownership and control of public 16. Much of the price change on the first four corporations, the proposed changes, if enacted, event dates should have occurred during early

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O127 Table IV Chronology, Source of News, CorrespondingEvent Date and Daily and Intraday Stock MarketMovements Associated with the U.S. House Ways and Means Committee Proposed Changes in the Tax Treatmentof Takeoversand other FinancialRestructurings (t-statistics in parentheses)

Tuesday evening, October 13: Democratson the House Ways and Means Committeeagreed to takeoverrestrictions. Reported in the WallStreet Journal on October 14. Correspondingevent date: October 14 Thursdayevening, October15: The full committee approved the takeoverrestrictions. Reported on the Broadtapeand in the WallStreet Journal on October16. Correspondingevent date: October16 Wednesday afternoon, October28: ChairmanRostenkowski in House testimony indicated that the takeover restrictions could be changed. Reported on Broadtapeat 2:08 p.m. on October28 (marketclosed at 2:00) and in the WallStreet Journalon October29. Correspondingevent date: October29 Thursdayevening, October29: ChairmanRostenkowski strengthened his remarksfrom the day earlier, releasing an officialstatement that he would agree to a 'reasonablecompromise.' Reported in the WallStreet Journal on October30. Correspondingevent date: October30 Wednesday morning, December 16: RepresentativeDowney of the committeeannounced that almost all the takeover restrictionshad been dropped. Reportedon Broadtapeat 11:58a.m. on December 16 and in the WallStreet Journal on December 17. Correspondingevent date: December 16

Oct. 14 Oct. 16 Oct. 29 Oct. 30 Dec. 16 Daily S&P 500 -2.95% -5.16% 4.93% 2.87% 2.17% Return (-2.86)*** (-5.00)*** (4.77)*** (2.78)*** (2.11)** IntradayS&P 500 -1.39% -1.18% 2.23% 2.99% 0.80% Return (-2.21)** (-1.88)* (3.56)*** (4.77)*** (2.80)*** Daily Abnormal -1.43% -5.25% 5.00% 4.39% 1.79% Return for Takeover (-2.03)** (-6.92)*** (6.13)*** (5.62)*** (2.42)** Portfolio IntradayAbnormal -0.31% -2.51% 3.65% 4.02% Return for Takeover (-1.60) (-6.15)*** (4.03)*** (4.21)*** Portfolio

Significantat the 10 per cent level. * Significantat the 5 per cent level. Significantat the 1 per cent level. trading, because the first opportunity to trade conferees announced their decision to abandon on the antitakeover news was at the open. On the antitakeover provisions, the S&P 500 in- December 16, some of the market reaction creased 2.17 per cent. All these movements are should have occurred immediately after noon, significantly different from zero, with the pre- because the news came across the Broadtapeat dicted sign. 11:58 a.m. Because the provisions would have Table IV also reports the S&P 500 return from had the greatest impact on companies that were the close on the day of each of the first four actually in play during this period, those com- announcements (October 13, 15, 28 and 29) panies should have experienced even greater through 11:00 a.m. on the (following) event price changes than the overall market over the date. During early trading on each of these four entire day and during the period immediately days, the market moved as predicted-down after the announcement. 1.39 and 1.18 per cent on October 14 and 16 and The lower half of Table IV displays the move- up 2.23 and 2.99 per cent on October29 and 30. ments of the overall market (representedby the During the hour after the December 16 an- S&P 500 index) on the five event dates. These nouncement that the antitakeover provisions movements are consistent with the hypothesis had been dropped from the House tax bill, the that the antitakeover provisions of the House S&P 500 rose 0.80 per cent, providing further proposal had a negative impact on the market. support for the hypothesis that the overall mar- The S&P 500 declined 2.95 per cent on October ket reacted to the bill. All these returns are 14, following the introduction of the provisions, significantly different from zero, with the pre- and 5.16 per cent on October 16, following dicted sign. committee approval. On October 29 and 30, To analyze the effects on takeover targets of after reports that Chairman Rostenkowski the antitakeover provisions of the tax bill, we might be flexible on the antitakeoverprovisions, constructed a portfolio of 19 firms that were the S&P 500 increased 4.93 and 2.87 per cent, targets of an outstanding offer on October 13. respectively. On December 16, when House We excluded in-play firms for which the take-

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O 28 over was substantiallycompleted by October 13 deficit announcements from April 1987 (Febru- and thus exempt from the provisions of the bill ary 1987 trade deficit) through December 1988 (which applied to distributions made after Oc- (October 1988 trade deficit). The data suggest tober 13). We estimated the beta-adjustedreturn that the announcement on October 14 caused to the takeover portfolio for each of the five little of the stock market decline. The difference event dates, using CRSPdaily stock returndata. between the predicted and actual trade deficit To determine immediate market response, we announced on October 14 is the fourth smallest calculatedintraday beta-adjusted returns for the of the 21 announcements. In contrast, the S&P first four event dates, using data on individual 500 daily return on October 14 is the second- transactions from the Securities Industry Auto- largest absolute return of the 21 announcement- mation Corporation. (Intradaytransaction data date returns, and the intraday return on Octo- were unavailable for December 16.) ber 14 (from the previous day's close to 11:00 Table IV displays the takeover portfolio's be- a.m.) is the fourth-largestin absolute terms. ta-adjusted returns. The data support the hy- The only surprise macroeconomicnews other pothesis that takeover targets were more sensi- than the trade-deficitannouncement during Oc- tive than the overall market to the antitakeover tober 14-16 was an increase in interest rates on provisions of the tax bill. The beta-adjusted October 14. But commentators have suggested takeover portfolio declined 1.43 per cent on that this increase was not independent of the October 14 and 5.25 per cent on October 16 and trade-deficit announcement, as traders feared increased 5.00 per cent on October 29, 4.39 per that government actions to lower the deficit cent on October 30 and 1.79 per cent on Decem- could increase interest rates. Other fundamental ber 16. These beta-adjusted returns are all sig- factors frequently cited in connection with the nificantlydifferent from zero, with the predicted October 14-16 market decline (such as the bud- sign. get deficit and Persian Gulf tensions) were not The intraday risk-adjusted returns indicate the subjects of unexpected news. that the takeover portfolio responded signifi- RichardRoll has argued that the crash did not cantly to the takeover tax news in early trading. begin in the United States, as many other world The intraday takeover portfolio declined 0.31 markets experienced a severe decline before the per cent on October 14 and 2.51 per cent on U.S. markets opened on October 19.11 He rec- October 16 and increased 3.65 per cent on ognizes that the U.S. decline during October October 29 and 4.02 per cent on October 30. 14-16 may have precipitated later international With the exception of October14, these intraday declines, but notes that some other world mar- returns are all significantly different from zero, kets also declined during October 14-16. Roll with the predicted sign. concludes that "the overall pattern of intertem- poral price movements in the various markets Other Possible Factors suggests the presence of some underlying fun- Factors other than the takeover tax bill may damental factor . . . but . . . seems inconsistent have contributed to the stock price drop during with a U.S.-specific macroeconomicevent." October 14-16. These include a higher-than- A decline in the rest of the world's markets expected trade deficit, rising interest rates, and during October 14-16 that is insignificantly dif- increased worries about the government deficit ferent from the contemporaneous U.S. decline and possible recession. would be inconsistent with our hypothesis that At 8:30 a.m. on October 14, the Commerce the proposed takeover restrictions caused the Departmentreleased the trade-deficitfigures for decline in the U.S. market, because the pro- August 1987. The $15.68 billion deficit for Au- posed restrictions did not affect foreign firms gust was smaller than the July deficit of $16.47 directly. We compared the performance of the billion, but it declined by a smaller amount than S&P 500 with the FT-Actuaries World Index expected. Several sources attribute the market consisting of the value-weighted stock move- decline on October 14 to this higher-than- ments of 22 countries. expected deficit. To test this explanation, we TableV displays U.S. versus non-U.S. market investigated whether prior trade-deficit an- movements during October 14-16. On October nouncements were associated with comparable 14, the S&P500 declined 2.95 per cent, while the stock market movements. World Index actually increased 0.84 per cent. We examined the market impact of 21 trade- The difference is statistically significant. On

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 n 29 Table V S&P 500 vs. World Index During October 14-16, Securities and Exchange Commission. Kenneth 1987 (t-statistics in parentheses) Lehn coauthored the first essay; Michael T. Ma- loney and Robert E. McCormickcoauthored the Oct. 14 Oct. 16 Oct. 14-16 second essay; Jeffry M. Netter coauthored the S&P 500 Return -2.95% -5.16% -10.44% third essay. (-2.86)*** (-5.00)*** (-5.84)*** 2. This section is an abbreviatedversion of an article World Index Return 0.84% -0.67% -0.60 (0.87) (-0.70) (-0.35) I coauthored with K. Lehn (Securities and Ex- S&P 500 Return- -3.79% -4.49% -9.84% change Commission), "Do Bad Bidders Become World Index (-2.68)*** (-3.16)*** (-4.01)*** Good Targets?"Journal of PoliticalEconomy, April Return 1990; it is used here with the permission of the publisher. * Significantat the 10 per cent level. 3. R. Marris, "A Model of the Managerial Enter- ** Significantat the 5 per cent level. ***Significant at the 1 per cent level. prise," QuarterlyJournal of Economics,May 1963, and H. Manne, "Mergers and the Market for October 16, the S&P 500 declined 5.16 per cent, CorporateControl," Journalof PoliticalEconomy, April 1965. while the World Index declined only an insig- 4. M. Jensen, "Agency Costs of Free Cash Flow, nificant 0.67 per cent; the S&P 500 decline is CorporateFinance, and Takeovers," A.E.R. Pa- significantly greater than the world decline. persand Proceedings,May 1986. Only Mexico had a larger decline than the U.S. 5. Using CRSPdaily stock returns, we estimated the on October 14, and only Mexico, Ireland, France abnormal return, arj, for each acquiring firm and Belgium had a larger decline than the during the period from 20 days preceding the United States on October 16. announcement date (the first date on which the Overall, during October 14-16, the U.S. mar- Dow Jones Broadtapereported a story about the ket declined 10.44 per cent, while the world acquisition) through 40 days following the an- market fell only 0.60 per cent. These data dem- nouncement. The abnormalreturn is: onstrate that the U.S. decline significantly ex- arit = rit - i - P3irmt, ceeded the non-U.S. decline. If the October 14-16 decline did indeed trigger the October 19 where ritis the returnto firm i at time t, rmtis the crash, the evidence suggests a U.S.-based event return to the CRSP value-weighted index of as the trigger. NYSE and AMEX stocks, and ai and Pi3are There was no significant news over the Octo- market-modelparameter estimates from the pe- ber 17-18 weekend that could have caused eq- riod 170 through 21 trading days preceding the uity values to fall over 20 per cent on the 19th. announcement date. But the decline on the 19th began beforethe We averaged the daily abnormalreturns across market opened, or, as Grossman and Miller firms in each group to obtain the portfolio abnormal return, AR, and cumulated over various state, "some precipitating trigger before the windows to obtain the cumulative return, CAR, . 19th caused a massive liquidity event .. at the for the portfolio. Tables I and II report the AR for opening of the markets on the 19th.",12While the acquisition-announcementdate (0) and the the mechanism of the crash itself is beyond the CAR for the corresponding four windows-(a) scope of this article, we suggest that the 10.44 one day before the announcement through one per cent market decline during October 14-16 day after the announcement, -1,1; (b) -5,1; (c) may have triggered the down open and subse- -5,40; and (d) -20,40. quent drop on October 19. Before October 14- 6. See B. Hagin, "WhatPractitioners Need to Know 16, 1987, the market had not experienced a About t-Tests," FinancialAnalysts Journal,May/ decline of over 10 per cent during a one, two or June 1990, for a discussion of gauging statistical three-day period since May 13-14, 1940, when significance. 7. This section describes the results of an unpub- German tanks broke through the French lines. lished paper, "ManagerialDecision Making and Given the rarityof the event, the October 14-16 Capital Structure,"I coauthored with M. T. Ma- decline may arguably be related to the October loney (Clemson University and the SEC) and 19 crash, especially since no trading days inter- R. E. McCormick(Clemson University). vened between the two events.13 m 8. "TalkingBusiness: The Importanceof Being Big- gest," New YorkTimes, June 20, 1989. Footnotes 9. This argument follows the work of Jensen, 1. The article summarizes three essays on corporate "Agency Costs," op. cit. takeover I coauthored while employed at the 10. This section is an abbreviatedversion of an article

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O 30 I coauthored with J. M. Netter (University of Quantitative Research in Finance Seminar. The Georgia), "Triggering the 1987 Stock Market presentation won the Institute for Quantitative Crash: Antitakeover Provisions in the Proposed Research in Finance's 1989 Roger F. Murray House Ways and Means Tax Bill," Journalof Prize. FinancialEconomics, September 1989. It is used The views expressed in this articleare those of here with the permission of the publisher. the author and do not necessarily reflect the 11. R. Roll, "The International Crash of October views of the Securities and Exchange Commis- 1987," Financial Analysts Journal, September/ sion or the author's formercolleagues on the staff October 1988. of the SEC. The Securities and Exchange Com- 12. S. Grossman and M. Miller, "Liquidityand Mar- mission, as a matter of policy, disclaims respon- ket Structure,"Journal of Finance,July 1988. sibility for any private publication or statement 13. I presented this paper at the Fall 1989Institute for by any of its employees.

FINANCIAL ANALYSTS JOURNAL / JANUARY-FEBRUARY 1991 O 31 “The Value of Corporate Takeovers” Copyright 1991, CFA Institute. Reproduced and republished from Financial Analysts Journal with permission from CFA Institute. All rights reserved.