Stock Trading Before the Announcement of Tender Offers: Insider Trading or Market Anticipation?

Transcription

1 Stock Trading Before the Announcement of Tender Offers: Insider Trading or Market Anticipation? GREGG A. JARRELL University of Rochester ANNETTE B. POULSEN University of Georgia 1. INTRODUCTION Recent insider trading actions by the government against Dennis Levine, Ivan Boesky, Martin Siegel, and others have influenced the public's perception of mergers and acquisitions activity.' These well-publicized cases generally involve illegal insider trading based on nonpublic information about impending bids for takeover targets. 2 Many regulators have interpreted public concern about illegal insider trading as political support for legislative proposals to restrict takeovers. These regulators argue that increased trading in stocks of tender offer targets before bid announcements indicates the pervasive nature of insider trading. Congressman Edward Markey (D- Mass.), chairman of the House Telecommunications and Finance Subcommittee, provides an example of how policymakers connect hostile takeovers We thank Robert Comment, Claudk) Loderer, Mark Mitchell, James Musumeci, Jeffry Netter, Michael Ryngaert, and the referees and editor of this journal for many helpful comments and suggestions. 1. In May 1986, the SEC charged Dennis Levine, an investment dealer at Drexel Burnham Lambert, with illegally trading in stocks and options of fifty-four companies. It alleged that Levine had made a profit of million by trading on inside information related to actual or proposed tender offers, mergers, and other business combinations. In November 1986, Ivan Boesky agreed to pay S100 million and to plead guilty to one felony count to settle charges of insider trading. According to the SEC complaint, Levine passed on inside information to Boesky, resulting in at least S50 million in profits to Boesky. Martin Siegel pleaded guilty to two felony count! for his role in Boc^y's iniider trading in February 1987 and implicated at least three additional Wall Street aroitragers. 2. See Netter, Poulsen, and Hersch for a further description of insider trading laws. Journal of Law, Economics, and Organization vol. 5, no. 2 Fall 1989 O 1989 by Yale University. All rights reserved. ISSN 8756-«

2 226 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 with insider trading: "In our current economic environment, corporate takeovers regularly provide a catalyst for insider trading.... The magnitude of the [stock-price] runups in many cases points to possible leaks or tips by insiders connected with the transaction." Markey also announced that he planned to soon introduce a comprehensive insider trading bill. 3 We study share price and trading volume data to determine how the market for information operates to anticipate tender offers and to analyze if it is appropriate to use prebid trading as evidence of insider trading. We examine market activity for 172 tender offers from 1981 to 1985, and, specifically, we study how several observable characteristics of the takeover bid affect prebid runup in stock prices and the "surprise" market reaction when the market formally learns the firm is "in play" (the unanticipated premium). These characteristics include reports of anticipated bids in the media, the foothold acquisition of the bidder, and the friendly or hostile nature of the bid. We also test whether unanticipated premiums and prebid runup are different in tender offers where there are later official allegations of insider trading. The presence of rumors in the news media about an impending bid is the strongest explanatory variable in accounting for unanticipated premiums and prebid runup. We also find evidence that unanticipated premiums are lower and prebid runup is greater when the bidder holds a relatively large position in the target at the time of the bid. We find no evidence suggesting that whether a bid is hostile or friendly affects unanticipated premiums or stockprice runup. Finally, we find no evidence that the twenty-six target firms in our sample that have been identified in government insider trading allegations show unusually low unanticipated premiums or high prebid runup, after accounting for the other variables discussed above. Our results suggest that increases in stock prices and trading volume before tender offer announcements are associated with several observable (and legal) factors. To argue that prebid runups necessarily reflect insider trading and that tender offers should be restricted to decrease the opportunities for insider trading is a misinterpretation of the data. 2. EXPLANATIONS OF RUNUP Stock prices and trading volumes of targets increase dramatically during the weeks immediately preceding public takeover bids. Several studies measure stock-price activity preceding mergers and takeover bids made between 1962 and 1980 and report that stock prices begin to move upwards in anticipation 3. Congressman Markey s bill was recently passed (Insider Trading and Securities Fraud Act of 1988, Public L. No Stat [1988]). Its main provisions include increased penalties, a reward program for turning in inside traders, and increased responsibility on investment banking firms to control illegal activities within their firms.

3 STOCK TRADING BEFORE TENDER OFFERS / 227 of bids as early as two weeks before formal announcement. 4 The studies also show that about one-half of the total appreciation is accounted for by the close of trading the day before the formal bid. Although most researchers do not ascribe these runups to widespread illegal activity, others consider this to be direct evidence of insider trading. 3 A relevant issue at this point is the definition of insider trading. Formally, illegal insider trading, as discussed here, refers to the restrictions imposed on insiders by Sections 10 (b) and 14 (e) of the Securities Exchange Act of 1934 and Rules 10b-5 and 14e-3 thereunder. Though the implications of these restrictions are complicated, the law requires basically that an insider must either disclose material inside information or refrain from trading. The courts have interpreted the rules to determine who is an insider (an individual with a duty to shareholders) and what is material information (information is material if there is a substantial likelihood that a reasonable investor would consider it important in his or her investment decision). Associating stock-price runups with insider trading has intuitive appeal. The government's success at identifying and prosecuting inside traders has reinforced the perception that such conduct is pervasive and that legitimate speculation is overwhelmed by illegitimate trading on nonpublic information. An active mergers and acquisitions market expands opportunities for profitable legal and illegal trading in anticipation of control bids. If illegal conduct is sufficiently widespread, then that conduct could be an important cause of prebid runup. Economists, on the other hand, generally view increases in stock prices before tender offers as supportive of the efficient markets hypothesis, which states that stock prices at any time fully reflect all public information." Researchers have documented stock-price reactions to many types of corporate announcements, including dividend changes, earnings reports, the sudden deaths of CEOs, proxy fights, charter amendments, regulatory changes, and macroeconomic events such as inflation, oil-price shocks, and interest-rate changes. In cases where it is possible for traders to discover information in advance of the news announcement, there is generally significant anticipation." Legitimate research and analysis of corporate information gives some trad- 4. See, for example, Bradley, Comment, and Keown and Pinkerton. Jensen and Ruback, and Jarrell, Brickley, and Netter review other articles in this area. 5. See for example, Stem and Jereski (33), Behr (Dl), and Stewart and Hertzberg (1). These articles include comparisons of stock prices and volumes of major takeover targets before and after the announcement of the offer and suggest that the runup figures reflect insider trading. Some academic researchers take this position also. Keown and Pinkerton state that most of the runup in their sample of mergers can be ascribed directly to insider trading, though they offer no evidence for this conclusion. 6. See Fama (chap. 5) for a complete exposition of the efficient markets hypothesis. 7. For example. Ball and Brown, in one of the earlier studies in this literature, show that 85 to 90 percent of the ultimate valuation effects of earnings reports are fully reflected in stock prices before their formal announcement.

4 228 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 ers informational advantages and their earnings serve as compensation for their efforts. 8 Their trading is beneficial to the extent it aligns stock prices with their theoretically correct values, promoting efficient allocation of capital. The prospect of large takeover premiums and the many kinds of clues legally available assures the existence of an active market for information on prospective takeover targets. 8 Therefore, much trading preceding important news can be attributed to a well-functioning market and not necessarily insider trading. 3. INFORMATION SOURCES FOR LEGITIMATE TRADING There are many sources of information on which trading in anticipation of a takeover may be based. Particularly important is the Schedule 13D requirement of the Williams Act, which mandates acquirers of more than 5 percent of a firm's equity to report the acquisition within ten days by filing a Schedule 13D. This filing requires the purchaser to disclose, among other things, his or her identity and the reasons for the purchase, specifically whether or not the purchaser intends to seek control. Once the original Schedule 13D has been filed, the acquirer must file amendments revealing any material changes such as a change in intent from "passive investor" to "investor seeking control." Several studies show that Schedule 13D filings cause significant increases in the target stock prices and surges in trading volume, particularly when the filer discloses takeover-type intentions or the filer is a known raider. l0 In addition to mandated disclosures, "street talk" can affect prebid market activity in takeover stocks. While some street talk might originate illegally from inside information, observers suggest numerous sources of legitimate information: [A]n army of traders, arbitragers and stock analysts is quick to spot unusual activity 8. When traders with better information (whether gained legally or not) profit from their investment positions, they presumably do so at the expense of less informed investors. The anonymous nature of most stock trading makes it difficult to determine who loses what the information specialist gains. Those selling into the market when the better-informed are buying probably would not have sold had they possessed the same valuable information. They still would have told, however, if the information specialist had refrained from buying, especially if the trading of the specialist did not affect significantly the stock price. This holds true whether the trading is based on insider information or on careful analysis and successful anticipation of the event. See Grossman and Hart for further discussion of the impact of asymmetric information. 9. Many studies have documented the large premiums paid to shareholders of takeover targets. Comment and Jarrell report that on average from 1981 through 1984, shareholders received a price 50 percent greater than the price at which the shares were trading prior to the offer. Bradley, Desai, and Kim report similar premiums. 10. See, for example, Mikkelson and Ruback, and Hoklerness and Sheehan. These studies report an average 3 to 7 percent appreciation in share price (net of market) at the Schedule 13D filing date. Macey and Netter summarize the theoretical effects of Schedule 13D filings.

5 STOCK TRADING BEFORE TENDER OFFERS / 229 of corporate executives, the calling and canceling of meetings and other seemingly insignificant corporate developments that might give a clue to an upcoming major development.... Experts closely monitor such things as which brokerage firm is doing the heaviest buying and selling of a particular stock. Does the firm have a relationship with a known corporate raider? Is the buyer bidding up the price of a stock, indicating that he is anxious to get the shares in a hurry? (Crudele) [I]t's not hard for all sorts of people to deduce that takeover action is about to occur in a stock. Computers are a big help here. Accumulation is instant news; so is unusual volume. Connected by a national network of direct phone lines, smart traders... talk to one another and to brokers constantly.... When investment bankers are hired, they frequently publish an internal memo that places their client corporation on a restricted list, to prevent conflicts of interest in trading and research recommendations. A company's name appearing on a restricted list often is all the information needed by shrewd investors. (Stern and Jereski) Arbitragers use the knowledge that takeovers often come in "waves" with targets clustered in certain industries as deregulation and other economic forces cause common adjustments in firms of common characteristics. Also, firm-specific knowledge, such as the impending demise of tight family control or internal disputes among managerial factions, often fuels takeover speculation. Finally, publicly available information disclosing large stock acquisitions and newspaper stories speculating on future takeovers can be important causes of prebid market activity, independent of illegal insider trading. 4. MEASUREMENT OF PREBID MARKET ACTIVITY We examine market activity (daily stock returns and trading volume) for 172 tender offers from 1981 to 1985 and present measures of runup in stock prices and volume surges. Additionally, we examine how several observable characteristics of the takeover bid are related to the surprise market reaction at the takeover announcement DATA AND EMPIRICAL METHODOLOGY Our sample is from the SEC's 1985 Office of the Chief Economist's study of tender offers from 1981 through 1984, supplemented with tender offers in 1985 identified from filings. The sample is limited to target firms that are listed on either the New York or American stock exchanges. All offers are cash offers and are successful in the sense that at least some shares were purchased by the bidder. The sample is listed in the appendix. We use stock return data obtained from the Daily Excess Stock Returns File provided by the Center for Research in Securities Prices (CRSP) to calculate the abnormal returns (ARs) to each firm around the announcement date. This file contains the daily returns for stocks in excess of the daily returns on a portfolio of securities with similar risk, estimated using the Scholes-Williams methodology (Scholes and Williams).

6 230 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 Cumulative abnormal returns (CARs) for each firm i are calculated summing the daily abnormal returns over the days of interest around the announcement date (from K to L) for the individual firms: CAR K,L) = 2 AH,.,-K (1) Portfolio CARs for N firms are formed and average cumulative abnormal returns are found: CAR(K,L) = (1/N) CAR, (K,L). '- (2) To test significance levels of the average CARs, t-statistics are calculated using the standard deviation of the daily average abnormal returns of the corresponding portfolio from 170 days before to twenty-one days before the announcement multiplied by the square root of (L - K + i), the number of days included in the CAR. U We measure the impact of runup in stock prices with two statistics. First, we measure the unanticipated premium from day t to day + 1, CAR,(t, +1). The unanticipated premium is the average additional premium received by shareholders between day t and the day following the bid and therefore indicates the value of the additional information contained in the period around the bid announcement. Since stock prices move only in response to new information, the unanticipated premium measures the market's surprise at the in-play announcement. To the extent that the market has already anticipated the announcement, runup has occurred. Therefore, the less the runup, the greater the unanticipated premium. Second, we calculate a portfolio runup index for each day t from twenty days before to five days after the bid announcement: {CAR( - 20,0 / CAR( - 20, + 1)} * 100. (3) The runup index equals the percent of the eventual premium that has been observed as of day t. Day +1 is used as the standard for measuring runup because by the close of the day after the announcement all investors will have had the opportunity to react to and trade on the information from the announcement. We also compute an analogous runup index for each firm. In addition to stock return data, we obtain data on daily trading volume from the Investment Statistical Listing (ISL) tapes provided by Interactive Data Services. The changes in the daily volume of shares traded before the 11. See Brown and Warner (1980, 1985) for full description of this procedure.

7 STOCK TRADING BEFORE TENDER OFFERS / 231 announcement of the tender offer is a measure of unusual activity before the public announcement of a takeover. In general, stock volume is positively skewed. 12 To offset overstatement of volume changes caused by this skewness, all volume calculations use the natural logarithm of the daily volume. ls We compare the volume around the takeover bid to a "clean" period (from 170 days to twenty-one days before the announcement) in which no takeover news should have affected trading. The average daily log volume for each firm j over the clean period is: -21 Meanlvol, = 1/150 X logfvolume,,). I--170 (4) Using the standard deviation of the mean log volume, S(Meanlvoli), we compute a z-statistic for log volume for each day t from twenty days before to five days after the bid announcement. The individual firm z-statistic equals: Z-Volume,, = [log(volume,,) Meanlvol,] / S(Meanlvol,). (5) The z-statistic for a portfolio of N firms is:.v "f l/nlnrvil. Z-Volume, = [l/n**.5)] * Z-Volume,, DETERMINING ANNOUNCEMENT DATE i-i (6) For each target firm, we collect news on corporate-control events during an extended period (at least one year) preceding the takeover bid, using the Dow Jones News Service, which covers all news releases on the Dow Jones News Service Ticker and in the Wall Street Journal (WSJ). The formal announcement of a tender offer often comes after widespread circulation of takeover-related news stories. In these cases, measuring runup in relation to the formal bid will bias the degree of anticipatory market activity due to illegal insider trades. Instead, we measure market activity in relation to the first day on which the target firm is in play. We call this the "news-adjusted" date. Specifically, the news-adjusted date is the earlier of: 12. See, for example, Harris for a measure of skewness of volume. 13. Zero trading volume on a given day may indicate decisions by potential traders to not buy or sell at the bid and ask prices or may reflect trading halts by the company or exchange. From 170 days before to twenty-one days before the bid, we assume that zero trading volume reflects no-trade decisions by potential traders and set log volume equal to zero. However, from twenty days before tlie bid to five days after, we assume zero volume results from a trading halt due to release of information and set zero volume equal to a missing value.

8 232 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, the day before the formal WSJ announcement date of a 14D-1 filing or tender offer proposal, or the day of the ticker announcement if before close of trading (the "formal" date), or 2. the public disclosure (usually over the Dow Jones ticker) of a Schedule 13D filing with a possible intention to seek a change of control, or 3. the public announcement of merger talks naming the target firm. We also require that to use dates (2) and (3) as the news-adjusted date, they must be followed closely by a formal bid or by sustained speculation of a takeover that materializes in a formal bid. For the 172 cases during 1981 through 1985 for which we have detailed information from the Dow Jones ticker and WSJ, seventy have news-adjusted dates earlier than the day before the formal WSJ announcement date. For all 172 firms, the average difference between the news-adjusted date and the formal date is 12.3 trading days. For the seventy firms where the news-adjusted date is not the formal date, the difference averages 30.3 trading days NEWS-ADJUSTED VERSUS FORMAL DATES Table 1 reports the daily CAR for 172 tender offers from twenty days before to five days after the news-adjusted in-play date. The first column indicates the trading day in relation to the event date. The second and third columns indicate the mean CAR and the (-statistic of the CAR. The fourth column reports the portfolio runup index, indicating the percent of the eventual premium observed on the indicated day. Our results verify the substantial amount of runup in stock prices before the announcement of a tender offer. The CAR is significantly different from zero and there is a steady increase in the runup index as early as fifteen days before the announcement. By one day prior to the announcement, the average stock price had appreciated to almost 39 percent of the total abnormal return one day after the announcement. The importance of selecting the correct announcement date to measure prebid trading is demonstrated when we compare runup from the sample of seventy firms in which the news-adjusted date differs from the formal date. Table 2 presents the CARs and runup index for this sample. For the seventy firms that were in play according to the Dow Jones ticker or the WSJ, the runup index one day before the formal announcement is 70 percent compared to 46 percent one day before the news-adjusted date. The formal announcement date would be the appropriate date to use if one were interested in finding premiums paid to shareholders once the offer is (nearly) certain. In many cases, however, market speculation about an impending bid may indicate a comparable level of certainty well before the formal announcement. Using the formal date for these already-in-play targets exaggerates the magnitude of runup. Thus, for the remainder of our work, we use the news-adjusted date as our announcement date.

9 STOCK TRADING BEFORE TENDER OFFERS / 233 Table 1. Cumulative Abnormal Returns (CARs) and Runup Index for 172 Tender Offers from 1081 to 1985, Using News-Adjusted Dates as Announcement Dates. Trading day(t) CAB(-20. t) f-ttatirtic Runup Index Notr. CAR) and the corresponding f-statistics use estimate! of a, and b, from 170 days to twenty-one days before the announcement date. The portfolio runup index equalj: {CAfl(-20, t) I CAR(-20, +1)} FACTORS INFLUENCING UNANTICIPATED PREMIUMS AND PREBID RUNUP We examine how four factors affect the magnitude of unanticipated premiums and prebid runup. The first three a^e proxies for the extent of legitimate prebid market activity, and the last looks directly at those tender offers involved in later government insider-trading-suits. The first factor is the presence of speculation in the media about control-related events. The second factor is the acquisition of a foothold block by the prospective bidder before a formal bid is made. The third factor is whether the tender offer is a hostile contest or a friendly, negotiated deal. If the dominant cause of prebid trading is illegal insider trading, these three factors should have only a negligible effect on bid anticipation. If these three factors, however, are able to explain differences in prebid trading across cases, it becomes difficult to necessarily attribute prebid runup to illegal insider trading unless illegal trading is correlated with the above factors. In our regression analysis, we specifically test for this correlation.

10 234 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 Table 2. Cumulative Abnormal Returns (CARs) and Runup Index for 70 Tender Offers from 1981 to 1985 with News-Adjusted Dates Not Equal to Formal Dates. Trade day ' Newt-adjusted date CAR t- statistic Formal date t- CAR statittic Newsadjusted runup Formal runup Note. CARj and the corresponding (-statistics use estimates of a, and b, from 170 days to twenty-one days before the announcment date. The portfolio runup index equals: (CAR(-20, () / CAH(-20, +1)) 100. Our fourth factor provides an independent test of the effect of illegal insider trading on runup by isolating cases in which insider trading allegedly occurred. Presumably, if insider trading is an important component of prebid trading, stock prices in cases where insider trading allegedly occurred should have above-normal runup levels and lower unanticipated premiums. The following sections discuss the expected effects of these variables on prebid runup and unanticipated premiums and explain how we construct the four determinants of prebid market activity. Table 3 provides summary information on these characteristics of our tender offer sample MEDIA SPECULATION CONCERNING A TENDER OFFER Runup in stock prices of targets of tender offers cannot be considered reliable evidence of insider trading if there has been public speculation or news reports that the firm may be the target of an offer. The substantial premiums paid to takeover targets encourage investor interest in the stock of firms

11 STOCK TRADING BEFORE TENDER OFFERS / 235 Table 3. Characteristics of 172 Tender Offers from 1981 to Category Number Percent Press Speculation No Press Speculation Average Percent Foothold of Bidder. All Firms Excluding Firms with no Bidder Foothold Hostile Friendly Insider Trading Allegations No Insider Trading Allegations identified as potential targets in thefinancialpress. Any credible speculation indicating an increase in the probability that a firm will be the target of a control bid is expected to result in increased trading and positive net-ofmarket stock returns for that firm. We expect to observe lower unanticipated premiums and more runup in those offers with press speculation. We search all news releases reported by the Dow Jones News Service during the year preceding the news-adjusted date to determine whether there is media speculation concerning a bid. There are sixty-nine cases of tender offers (40.1 percent) with press stories concerning an impending bid and 103 bids (59.9 percent) without press stories. The presence of any one of the following five kinds of news stories results in the case being considered a tender offer with press speculation: 1. rumored merger or takeover talks without confirmation by parties involved; 2. Schedule 13D filings or other substantial holdings reported without announcement of attempts at control changes; 3. formal merger or takeover negotiations that had been halted or terminated; 4. significant stock repurchases within the previous year with potential control impact; or 5. any story that would have qualified as denoting a news-adjusted date except for a six-month or longer delay between that date and the formal WSJ announcement. Note that in this process we rely on only two news sources the Dow Jones ticker and the WSJ. While these sources probably contain most important takeover-related news stories, relying on published press accounts will yield an imperfect measure of street talk. We expect that if we could somehow identify all tender offers that were the subject of takeover rumors, we could explain an even larger fraction of stock-price runup.

12 236 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, FOOTHOLD ACQUISITIONS OF BIDDERS Bidders often purchase large amounts of the stock of their targets before launching formal offers for control. Arbitragers attempt to track such bidder activity as a means of predicting whichfirmsmight be future targets. Unusual trading volume and an increased number of block trades are generally directly observable by arbitragers and can serve as important clues to improving returns to arbitrage activity. Therefore, we expect that assembling large footholds will "tip" the market to future tender offers and cause additional stock-price runup. Unanticipated premiums should be lower and runup should be higher when bidders accumulate significant shareholdings prior to the bid. We determine percent holdings of the bidder in the target firm at the time of the bid in each of the 172 offers included in our sample and relate those holdings to prebid trading activity. Holdings are obtained from disclosures in Schedule 14D-1 filings made by the bidder detailing the terms of the offer. The mean foothold position for the bidder at the time of the bid was 7.4 percent. In seventy-two of the offers, the bidder held no shares in the target firm at the time of the offer. The mean foothold across the other hundred target firms was 12.7 percent, ranging from a low of one-thousandth of 1 percent to a high of 50.5 percent. We use foothold data as an indicator of observable acquisitions by the bidder immediately preceding the bid. We emphasize, however, that our foothold data are only approximations of bidders' prebid trading. One potential difficulty with these data is that the footholds are measured at the time of the formal bid while we measure prebid trading activity as of the newsadjusted date. Thus, our foothold data may exceed actual holdings at the time of the news-adjusted date and overstate the amount of observable acquisitions by the bidder, at least for some of the seventy cases for which we use the news-adjusted date. Another potential problem is that the level of holdings at the time of the formal bid does not necessarily indicate when those holdings were accumulated. We would ideally focus on recently acquired positions; static, long-term positions are not as likely to provide any new information to the market. Nonetheless, the level of footholds at the filing date that we use here should yield reasonably good estimates of the relation between bidders' acquisitions and prebid runup. In our empirical work, we consider the effect of large footholds held for long periods on the expected negative relation between the foothold of the bidder and the unanticipated premium. We consider separately five tender offers in which bidders had large ownership positions (over 40 percent) that were held for at least six months. If these positions do not provide information about the probability of a bid occurring, we would expect the foothold information to have little impact on runup in the period close to the bid Thefivefirmsand ownership positions of the bidden are Guardian Industries (43

13 STOCK TRADING BEFORE TENDER OFFERS / HOSTILE VERSUS FRIENDLY TENDER OFFERS The differential effects of hostile versus friendly oflers are difficult to predict. A hostile bidder may attempt to surprise its target to lessen the target's opportunity to take defensive moves such as the adoption of a poison pill, selling significant assets, or restructuring its capital base to make the firm less attractive to the bidder. u Therefore hostile bidders may have more incentives to keep their bids secret than do those pursuing friendly transactions. Hostile bidders, however, may attempt to get the stock of their targets in the hands of arbitragers, who are expected to be more willing than other shareholders to tender into hostile offers." In these latter cases hostile bidders would be less concerned with keeping their bids secret and increased runup would result. In comparison to a hostile bid, a friendly, negotiated bid typically involves negotiations and considerable sharing of information between the advisors of the bidder and target parties before thefinalbid is made. This setting would seem to present greater opportunities for intentional or unintentional leakage of information concerning a proposed takeover, which should result in more prebid runup, all other factors held constant. Given these opposing considerations, we make no prediction concerning the direction of the relation between the hostile-friendly categorization and prebid runup. The tender offer is considered to be hostile if target management expresses opposition to the bid. Opposition may take several forms, including actions such as defensive stock repurchases, litigation, recommending that shareholders vote against the bid, or searching for alternative offers. While few successful takeover contests remain hostile to the end, we define cases here as hostile if the target, at any time, takes a defensive stance. We designate eighty-nine of the tender offers (51.7 percent) in our sample as hostile bids TENDER OFFERS WITH INSIDER TRADING PROSECUTION If prebid runup is the result of insider trading, then those cases in which insider trading did occur should have increased runup and lower unanticipated premiums. We identify twenty-six tender offers (15.1 percent) in our sample in which the Department of Justice or the SEC made formal allegations that insider trading occurred. These cases were identified by search percent), Braun Engineering (43 percent), Juniper Petroleum (49 percent), Gaylords (49 percent), and MGM (50 percent). 15. See Jarrell (1985) for a study of the effects of defensive litigation In defeating bids and in generating auctions for firms. Also, Dann and DeAngelo study the impact of defensive adjustments in asset and ownership structure. 16. Larcker and Lys provide evidence of substantial returns to rislc arbitrage activities in corporate control transactions. The importance of concentrating holdings in friendly hands is seen in recent insider trading allegations. Several recent cases have alleged "narking violations" in which third parties (including Drexel Buraham Lambert and Jefleries and Co.) held stock anonymously for hostile bidden.

14 238 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V: of the WSJ, the Dow Jones ticker and other publications and through the use of the Lexis data base. Unfortunately, those cases in which there were official allegations of insider trading do not necessarily form the complete population, let alone an accurate sample of all tender offers in which insider trading occurred. In some cases, officials may not choose to prosecute suspected inside traders because of difficulty in proving the case or, in other cases, officials may have incorrectly alleged insider trading. It is also certain that some cases of insider trading are never identified. 6. REGRESSION ANALYSIS OF UNANTICIPATED PREMIUMS AND PREBID RUNUP We use regression analysis to determine the effect of the explanatory variables discussed above on unanticipated premiums and prebid runup. Our results are reported in tables 4 and 5. The unanticipated premium for each firm is the dependent variable in our Table 4. Weighted Least Squares Regressions Explaining Unanticipated Premiums in 172 Tender Offers from 1981 to 1985 (t-statistics in parentheses). Dependent Variable = Unanticipated Premium between indicated day and day( + 1). Independent Variable* Intercept Preceding Runup Press Speculation Percent Foothold Hostile Insider Trading Large Foothold Dummy F-Value doy(-l) (9.86) (-1.85)' (-3.33) (-1.11) (0.47) (2.29)» 41.6 Panel A dny(-5) (11.89) (-1.68)' (-4.07) (-1.36) (0.93) (1.95)' 60.6 day(-10) (14.15)* (-1.88)' (-4.01) (-1.41) (1.27) (2Alf 91.9 day(-l) (9.94) (-1.68> (-3.46) (-1.88)' (0.85) (2.29) (1.70)' 36.4 Panel B day(-s) (13.60) (-2.44)* (-4.44)* (-2.47) (1.59) (2.04)* (1.68)' 67.7 day(-10) (14.16) (-1.78)' (-4.06) (-1.69)' (.145) (2.40) (0.93) Noter. a. Significantly different from zero at 99 percent probability level, b. Significantly different from zero at 93 percent probability level, c Significantly different from zero at 90 percent probability level. Preceding runup is the cumulative abnormal return Cor the firm from twenty days before to one day before the measurement dateforthe unanticipated premium. Press speculation equals one if there were news stories concerning the bid, zero otherwise. The percent foothold is the holding of the bidder at the time of the 14D-1 filing. Hostile equals one if the bid was hostile at any time, zero otherwise. Insider trading equals one if government agencies later alleged insider trading violations, zero otherwise. The large foothold dummy equals the foothold of the bidder in five cases with foothold over 40 percent heldforat least six months, zero otherwise. Weighted least squares estimation uses as the weighting factor the standard deviation of the dependent variable estimated from 170 trading days to twenty-one trading days before the event dates. 78.9

15 STOCK TRADING BEFORE TENDER OFFERS / 239 T a b l e 5. O r d i n a r y L e a s t S q u a r e s R e g r e s s i o n s E x p l a i n i n g I n d i v i d u a l F i r m R u n u p I n d e x i n T e n d e r O f f e r s from t o ( f - s t a t i s t i c s i n p a r e n t h e s e s ). D e p e n d e n t V a r i a b l e =» I n d i v i d u a l firm r u n u p i n d e x b e t w e e n d a y ( ) a n d i n d i c a t e d d a y. Independent Variables Intercept Press Speculation Percent Foothold Hostile Insider Trading Large Foothold Dummy R-squared F-Value Number of Firms day(-l) ( ) ( ) " ( ) ( ) ( ) ( ) Panel A Negative index d a y ( - 5 ) ( ) * ( ) U.97> (-1.70Y ( ) ( ) day(-10) ( ) " ( ) ( ) ( ) ( ) ( ) datf-l) ( ) ( ) ( ) ( ) ( ) ' Panel B Negative index = missing v a l u e d a y ( - 5 ) day(-10) ( ) ( ) ( ) * ( ) ( ) ( ) (3.78h ( ) > ( ) ( ) ( ) ( ) Notts: a. Significantly different from zero at 99 percent probability level. b. Significantly different from zero at 95 percent probability level. c. Significantly different from zero at 90 percent probability level Seven firms are excluded because both preceding runup and the full premium are negative. Press speculation equals one if there were news stories concerning the bid, zero otherwise. The percent foothold is the holding of the bidder at the time of the 14D-1 filing. Hostile equals one if the bid was hostile at any time, tero otherwise. Insider trading equals one If government agencies later alleged Insider trading violations, zero otherwise. The large foothold dummy equals the foothold of the bidder in five cases with foothold over 40 percent held for at least six months, zero otherwise. estimated regressions in table 4. The unanticipated premium measures the market value of additional information about a bid released during that time interval and thus the surprise component of the in-play announcement. To account for the possibility of heteroskedasticity, we use weighted least squares regressions, dividing the dependent and independent variables by the standard deviation of the dependent variable, CAR (t, +1). This procedure increases the efficiency of the parameter estimates. To consider the impact of runup prior to the day from which we measure the unanticipated premium, we include CAR(-20, t-1) as an additional explanatory variable in our regressions. We find a negative and significant relation between the preceding runup and the unanticipated premium, indicating that there is a trade-off in total premium between the earlier and later periods: that is, the higher the premium in the weeks prior to the bid, the lower the premium at the bid announcement. Table 5 reports regression results when the individual firm runup index is the dependent variable. The runup index might be a better measure of

16 240 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 prebid anticipation than the unanticipated premium since it represents the proportion of the total premium already received before the bid is announced formally. When the runup index for an individual firm is negative, however, it is difficult to interpret; for example, if CAR( 20, t) is negative and the total premium CAR( 20, + 1) is positive, the runup index would be negative and indicate that there is less than no anticipation. We find it difficult to make statements about runup based on negative anticipation. To adjust for this consideration, we estimate the regressions by setting the runup index equal to zero in all cases where CAR( 20, t) is negative and also by excluding these observations. The "press speculation" variable equals one if there was press speculation concerning the impending bid, zero otherwise. "Percent foothold" equals the percent holdings of the bidder at the time of the Schedule 14Dfiling.The "hostile" variable equals one if the bid was hostile, zero otherwise. The "insider trading" variable equals one if there were later allegations of insider trading, zero otherwise. In addition, we include an interactive dummy variable forfiveoffers in which the bidder held over 40 percent of the outstanding shares for more than six months. The variable equals the percent holding of the bidder in those five cases and zero otherwise. We discuss the results shown in tables 4 and 5 for each of these explanatory variables in turn PRESS SPECULATION The regression results in tables 4 and 5 verify the importance of media speculation in the prebid trading of tender offer stocks. Unanticipated premiums are reduced significantly and prebid runup is increased significantly when the Dow Jones News Service or the WS] had reported rumors of possible impending bids. The press-speculation variable is significantly different from zero at the 99 percent confidence level in all versions of the model. It can be argued that it is incorrect to treat the presence of press reports before bids as being independent of illegal insider-trading activity. Illegal trading could, perhaps by feeding Wall Street rumors, be directly responsible for many news stories. One obvious way to check for a relation between illegal trading and press reports is to see if the twenty-six cases involving insider-trading charges account for a disproportionate number of the cases with prebid media speculation. Fifteen of the twenty-six insider-trading cases (57.7 percent) had press speculation, compared with sixty-nine of 146 (47.3 percent) of the remaining tender offers. Although there is more media speculation among the insider-trading cases, this difference in incidence is not large and is insignificantly different from zero. We consider the role of insider trading and its interaction with the media variable in determining prebid market activity when we discuss the insider trading variable below.

17 STOCK TRADING BEFORE TENDER OFFERS / FOOTHOLD OF THE BIDDER Our results generally support the hypothesis that there is a negative relation between foothold acquisitions and unanticipated premiums, as reported in table 4. However, this relation is significantly different from zero only when the interactive dummy variable for large footholds is included (panel B), at the 90 percent confidence level for the day( 1) and day( -10) regressions and at the 99 percent level for the day( 5). The corresponding positive relation between foothold acquisitions and prebid runup is also observed in table 5, though it is significant only in the regression for runup as of day( 5). The inclusion of the interactive variable allows separate interpretation of the relation between bidders' acquisitions and market activity for the five large long-term footholds as compared to the remaining 167 tender offers. While the coefficient on the variable measuring bidders' footholds suggests that increasing footholds decrease unanticipated premiums, the interactive variable indicates that this is not true for the largest footholds all held for at least six months. For thefivecases of large long-term positions, the bidders' footholds have an insignificant impact on unanticipated premiums and prebid runup in the period closest to the bid. 17 Given that there had been no additional accumulation by these bidders in the period close to the bid, it is logical that the bids could have been unanticipated despite the large footholds HOSTILE VERSUS FRIENDLY OFFERS Our results reveal little statistical importance to the friendly-hostile distinction for explaining differences in unanticipated premiums or prebid runup among tender offers. Given the conflicting considerations in the impact of hostile versus friendly bids, the results are not surprising INSIDER TRADINC If insider trading results in significant prebid trading, we would expect to see lower unanticipated premiums and more prebid runup in those cases in which allegations of insider trading were made. The coefficient on the insider trading variable in our regressions, however, is always positive and significantly different from zero in the unanticipated premium regressions. These results perhaps reflect decisions by government officials to prosecute in those cases where insiders made substantial profits. The possibility of undetected inside traders in some tender offers also would weaken this test. At least to our ability to measure insider trading, however, these results contain no 17. The sum of the two foothold coefficients measure the full estimated impact of the foothold in these five cases on unanticipated premiums. The sum is insignificantly different from zero.

18 242 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 support for the view that insider trading causes greater stock-price runup. (The coefficient of the insider trading variable is generally insignificantly different from zero in the prebid runup regressions.) To further consider the relation between insider trading and the other explanatory variables, we test the impact of several additional interactive variables in our regression analysis, including the press speculation variable multiplied by the insider-trading variable and the hostile-friendly variable multiplied by the insider-trading variable. While most of the alternative specifications added little information to the preceding analysis, we found that the interactive variable representing hostile bids in which insider trading was also alleged was significantly different from zero as shown in the below estimated regression (f-statistics indicated below coefficients): CAR(-5, +1) = (Preceding Runup) (Press Speculation) (13.64) (-3.02) (-4.38) (Percent Foothold) (Hostile) (-1.96) (0.32) (Insider Trading) (Hostile 6: Insider Trading). (-0.10) (2.01) Thus, the seventeen cases where the bid was both hostile and there were later allegations of insider trading seem to drive the significance of the insider trading variable. The same coefficient, however, is insignificantly different from zero in the prebid runup regressions. 7. EVIDENCE ON TRADING VOLUME CHANGES FOR TAKEOVER TARGETS Table 6 reports the mean z-statistic for the difference between the natural logarithm of daily volume in the event period compared to the clean period for 164 tender offers from our sample. 18 The portfolio z-statistic is significantly different from zero for the full period reported, as seen in the second column. The third column of table 6 reports the number of firms with individual z-statistics where daily volume is significantly greater than its historical mean at the 95 percent confidence level. This measure shows above-normal volume was experienced by a large proportion of these targets in the days before the formal bids were announced. Ten days before the announcement of the bid, about 10 percent of the firms in our sample experience trading volume significantly different from mean volume in the clean period at the 95 percent confidence level. Five days before the announcement date, the percent of 18. If we were unable to obtain more than seventy-five days of volume data from 170 dayj to twenty-one dayi before the bid announcement, the firm wai excluded from the empirical work reported in thij section.

19 STOCK TRADING BEFORE TENDER OFFERS / 243 Table 6. Z-Statistic for Log Volume and Percent Significantly Different from Mean Volume for 172 Firms from 1981 through Portfolio Percent different Trade z- from mean volume at Day statistic N 95 percent confidence level Note- The ^-statistic for each firm equals daily log volume on the Indicated day minus the mean log volume for days ( 170) to ( 21) divided by the standard deviation of the mean log volume. The portfolio :-statisttc equals the summation of the firm z-statistics in the portfolio divided by the square root of the number of firms In the portfolio. firms with significantly different volume has increased to about 20 percent. Not surprisingly, about 80 percent of the firms have volume levels significantly greater than usual on the announcement date and the day after CONCLUSION This study presents data on stock prices and trading volume for targets in 172 successful New York and American stock exchange takeovers between 1981 and We examine how several factors affect market activity, including the presence of media speculation, the size of foothold acquisitions disclosed by the bidder, and whether the initial bid is friendly or contested. In addition, we consider whether unanticipated premiums were lower than 19. We attempted regression analysis to explain the individual firm z-statistici, using the same independent variables as in the analysis of unanticipated premiums. We were unsuccessful at explaining a significant portion of the variation in volume.

20 244 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 normal in cases where government agencies later prosecuted for insider trading. The 172 targets of successful bids show significant stock-price runup and surges in volume before the bids. Averaged over all cases, about 40 percent of the eventual takeover premium is anticipated in the form of prebid runup. The pattern is also consistent across time; the runup measures shown here are similar to earlier works based on samples covering 1962 to Our results indicate that the presence of rumors in the news media concerning an impending bid is the strongest variable in explaining unanticipated premiums and prebid runup for tender-offer targets. We also report evidence that unanticipated premiums are lower and prebid runup is higher in those cases in which the bidder held a relatively large position in the target at the time of the bid. Whether the bid is friendly or hostile has no statistical significance in explaining unanticipated premiums. We find a positive significant relation between insider trading allegations and unanticipated premiums, opposite of the prediction that insider trading causes lower unanticipated premiums and more stock-price runup. The significant effects of media speculation and of foothold acquisitions on unanticipated premiums and prebid runup are consistent with a legitimate market for information. They further imply that significant prebid market activity is consistent with little or no illegal insider trading. Our results thus indicate that aggregate runup statistics must be used cautiously as measures of illegal insider activity. APPENDIX Sample of Tender Offers Used in Study Target Firm A M F Inc. Aegis Corp. Adams Drug Inc. Amalgamated Sugar Co. American Sterilizer Amerace Corp. American Motor Inns Inc. American Nat. Res. ANTA Corp. Applied Data Reih. Inc. ARO Corp. AVCOCorp. Bacfae Croup Inc. Bangor Punta Belco Pete Co. Braun Engineering Breeze Corp. Cusip News- Adjuited Date Formal Date

21 STOCK TRADING BEFORE TENDER OFFERS / 245 Target Firm Cusip Newi- Adjusted Date Formal Date Brunswick Corp. Buffalo Forge Co. Bunker Raroo Corp. Burgess Inds. Inc. Burns IntL Sec. Svcs. Inc. Cannon Mills Co. Cardiff Equities Corp. Caressa Croup Inc. Carnation Co. Cenco Inc. Central Soya Inc. Cessna Aircraft Co. Chieftain Dev. Ltd. Chilton Corp. Cities Services Co. Clausing Corp. Cluett Peabody & Co. Inc. Coldwell Banker & Co. Compugraphic Corp. Connecticut Nat. Gas. Corp. Conoco Inc. Continental Airlines Corp. Core Labs Inc. Cox Communications Inc. Criton Corp. Crown Zellerbach Corp. Dan River Inc. Data Term. Sy$. Inc. Dean Witter Reynolds Inc. Delhi Intl. Oil Corp. Donaldson, Lufkin it Jenrette Inc. Dorchester Gas Corp. Easco Corp. El Paso Co. Enstar Corp. Esmark Corp. Faberge Inc. Fisher Scientific Co. Franks Nursery & Crafts Inc. Friona Inds. Inc. Frontier Hldgs Inc. CFCorp. Carfinckel Brooks Bros Miller Gas Svc. Co. Gaylords Natl. Corp- General Foods Corp. General Portland Inc. General Steel Inds. Inc. Cetty Oil Co. Ciddings & Lewis Inc. Ginos Inc. Grand Central Inc. Craniteville Co S

22 246 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 Target Firm Cusip New»- Adjusted Date Formal Date Gray Drug Stores Great Lakes Intl. Inc. Guardian Inds. Corp. Gulf Corp. H M W Industries Inc. Harsco Corp. Heublein Inc. Hobart Corp. Horizon Corp. Informatics General Corp. Interpace Corp. Itek Corp. James Fred S. & Co. Inc. Jewel Cos. Inc. Jonathan Logan Inc. Juniper Pete Corp. Kentron Intl. Inc. L F E Corp. Lane Bryant Inc. Lenox Inc. Levi Strauss & Co. Levitz Furniture Corp. Laghtolier Inc. Lowenstein M. Corp. Ludknv Corp. M G M Grand Hotels Inc. Malone & Hyde Inc. Marathon Oil Co. Marshall Field & Co. Masonite Corp. McGraw Ediion Means Services Inc. Mesa Royalty Trust Michigan Sugar Co. Midlands Energy Co. Miller Wohl Inc. Milton Bradley Co. Mission Ins. Group Inc. Mite Corp. Murphy G. C. & Co. N I Industries Inc. Nabisco Inc. Narco Scientific Inc. National Can Corp. National Mine Svc. Co. Northwest Energy Co. Northwest Inds. Inc. Northwestern Mutl Life Mtg. & Rlty. Norton Simon Inc. Opelika Mfg. Corp. Pacific Lumber Co. Parsons Corp. Pay Leis Drug Stores NW Peoples Drug Stores Inc. Petrolane Inc

23 STOCK TRADING BEFORE TENDER OFFERS / 247 Target Firm Ciuip News- Adjusted Date Formal Date PieT 1 Inc. Pogo Producing Co. Prentice Hall Inc. Puritan Fashions Corp. R E D M Inds. Inc. Real Estate Invt Tr. Amer. Revkm Inc. Richardson Co. Richardson Vicks Inc. Rio Grande Inds. Inc. S C A Services Inc. S C M Corp. S S P Industries St. Joe Minerals Corp. St. Regis Corp. Sav A Stop Inc. Schlitz Jos. Brewing Co. Schrader Abe Corp. Scovill Inc. Searie G. D. & Co. Signal Cos. Inc. Southland Rty. Co. Spectro Inds. Inc. Speed O Print Bus. Mach. Sta Rite Inds. Inc. StauHer Chemical Co. Stoker/ Van Camp Inc. Suburban Propane Gas Corp. Sunbeam Corp. Supron Energy Corp. Technicolor Inc. Texas Gas Res. Corp. Texasgulf Inc Thiokol Corp. Torin Corp. Transway Intl. Corp. Tymshare Inc. Unidynamics Corp. Uniroyal Inc. United Energy Res. Inc. United Rlty Invs. Inc. United States Inds. Inc. Unocal Corp. Volume Merchandise Inc. Vulcan Inc. Walbar Inc. WieboHt Stores Inc REFERENCES Ball, P.., and P. Brown "An Empirical Evaluation of Accounting Income Numbers," 6 Journal of Accounting Research 159.

24 248 / JOURNAL OF LAW, ECONOMICS, AND ORGANIZATION V:2, 1989 Behr, P "Spotlight on the Insiders," Washington Post, 21 September, Dl. Bradley, M "Interfirm Tender Offers and the Market for Corporate Control," 53 Journal of Business 345. Comment, R "Price and Volume Before Tender Offen: Market Anticipation Activity or Inside Trading," Working Paper. Comment, R., and G. Jarrell "Two-Tier and Negotiated Tender Offers: The Imprisonment of the Free-Riding Shareholder," 21 Journal of Financial Economics 283. Crudele, J "How Investors Sniff Out News," New York Times, 12 August, D8. Dann, L., and H. DeAngek) "Corporate Financial Policy and Corporate Control: A Study of Defensive Adjustments in Asset and Ownership Structure," 20 Journal of Financial Economics 87. Fama, E Foundations of Finance. New York: Basic Books. Grossman, S., and O. Hart "Allocational Role of Takeover Bids In Situations of Asymmetric Information," 36 Journal of Finance 253. Harris, L "Cross-Security Tests of the Mixture of Distributions Hypothesis," 21 Journal of Financial and Quantitative Analysis 39. Holderness, C, and D. Sheehan "Raiders or Saviors? The Evidence on Six Controversial Investors," 14 Journal of Financial Economics 555. Jarrell, G "The Wealth Effects of Litigation by Targets: Do Interests Diverge in a Merge," 28 Journal of Law and Economics 151. Jarrell, G., J. Brickley, and J. Netter "The Market for Corporate Control: The Empirical Evidence Since 1980," 2 Journal of Economic Perspectives 49. Jensen, M., and R. Ruback "The Market for Corporate Control: The Scientific Evidence," 11 Journal of Financial Economics 5. Keown, A., and J. Pinkerton "Merger Announcements and Insider Trading Activity: An Empirical Investigation," 36 Journal of Finance 855. Larcker, D., and T. Lys "An Empirical Analysis of the Incentives to Engage in Costly Information Acquisition: The Case of Risk Arbitrage," 18 Journal of Financial Economics 111. Macey, Jonathan R., and Jeflry M. Netter "Regulation 13D and the Regulatory Process," 65 Washington University Law Quarterly 131. Markey, E "Congressional Study Finds Persistent Run Ups in Target Company Stock, Indicating Possible Pervasive Insider Trading." Washington, D.C.: News release, July 15. Mikkelson, W, and R. Ruback "An Empirical Analysis of the Interfirm Equity Process," 14 Journal of Financial Economics 523. Netter, J., A. Poulsen, and P. Hersch "Insider Trading: The Law, the Theory, the Evidence," 6 Contemporary Policy Issues 1. Office of the Chief Economist, Securities and Exchange Commission The Economics of Any-or-All, Partial and Two-Tier Tender Offer. Washington, DC. Office of the Chief Economist, Securities and Exchange Commission Stock Trading Before the Announcement of Tender Offers: Insider Trading or Market Anticipation? Washington, D.C. Scholes, M., and J. Williams "Estimating Betas From Nonsynchronous Data," 3 Journal of Financial Economics 309. Stern, R., and L. Jereski "My Cousin Works For This Guy Who...," Forbes, 10 February, 33. Stewart, J., and D. Hertzberg "Post-Levine Trauma Ends, But Arbitragers Are Hurt by Fallout," Washington Post, 28 October, 1.