Print Email Download Reference This Send to Kindle Reddit This
submit to reddit

Merger Wave

The American economy experienced two great takeover waves in the postwar period, first in the 1960s and the second in the 1980s. Both waves had a deep affect on the structure of corporate America. The main trend in the '60s was diversification and conglomeration. In contrast the 1980s takeover reversed the previous process and brought US corporations back to specialization. In this respects, the last thirty years were a roundtrip for corporate America. This paper is an overview of the salient features of the two takeover waves.

1.1 The 1960's Conglomerate Merger Wave

The merger wave of the 1960s was the major since the turn of the century (Stigler, 1968). A typical characteristic of the 1960s transaction was a friendly acquisition, frequently for stock, of a smaller private or public firm which was outside the acquiring firms main line of business. During this period unrelated diversification was widespread among the large companies. Rumelt (1974) has reported that the fraction of single business companies in the Fortune 500 decreased from 22.8% in 1959 to 14.8% in 1969. Further, the portion of conglomerates with no dominant businesses increased to 18.7% from 7.3%. There was also a considerable move to diversification among companies that retained their core business. The driving force behind the 1960s wave was high valuations of company stocks and large corporate cash flows. However the management was unwilling to pay out the high cash flows as dividends, and on the other hand able to issue equity at attractive terms therefore, turned their attention to acquisitions (Donaldsoni. 1984).Dividends were considered as a complete waste, and acquisitions as a very attractive way to conserve corporate wealth.

There are two sets of arguments used to explain why companies diversify. The first set argues that firms diversify to increase shareholder wealth. A number of authors have discussed different aspects of diversification that can potentially raise shareholder wealth. Williamson (1970), suggest that firms diversify to beat imperfections in external capital markets. Through diversification, managers create internal capital markets, which are less prone to asymmetric information problems. Lewellen (1971), argues that conglomerates can carry on higher levels of debt since corporate diversification reduces earnings variability. if conglomerate firms are more valuable than companies operating in a single industry If the tax shields of debt increase. Shleifer and Vishny (1992), state that conglomerates may have a higher debt capacity since they can sell assets in those industries that suffer the least from liquidity problems in bad states of the world. Finally, Teece (1980) argues that diversification leads to economics of scale. The second set of arguments states diversification as a product of the agency problems between shareholder and managers. Amihud and Lev (1981) argue that managers follow a diversification strategy to protect the value of their human capital. However, Jensen (1986) suggests that companies diversify to increase the private benefits of managers. Similarly, Shleifer and Vishny (1989) suggest that managers diversify because they are better at managing assets in other industries. Thus, diversifying will make skills more indispensable to the firm.

1.2 The 1980's Merger Wave

Form a longer historical perspective, Golbe and White (1988) presented time series evidence of U.S. takeover activity from the late 1800s to the mid-1980s. Their findings have suggested that takeover activity above 2 to 3 percent of GDP is unusual. However, the greatest level of merger activity occurred around 1980s, at roughly 10 percent of GNP. By this measure, takeover activity in the 1980s is historically high.

The size of the average target in the 1980s had increased extremely from the modest level of the '60s. By 1989 28%, of Fortune 500 companies were acquired and many transactions, particularly the large ones, were hostile. Further the medium of exchange in takeovers was cash rather than stock, they were characterized by heavy use of leverage. Firms were purchased by other firms by leveraged takeovers by borrowing rather than by issuing new stock or using solely cash on hand. Other firms restructured themselves, borrowing to repurchase their own shares. The '80s was also characterized by latest forms of control changes, which included 'bustup' takeovers. Bustup takeovers involved the sell off of a substantial fraction of the target's assets to other firms. (Bhagat, Shleifer, and Vishny, 1990; Kaplan, 1997).

2 Merger Motives

The following sections will explain the motive behind the two merger waves.

2.1 Managerial Motives

Agency theory predicts that unless managers are strictly monitored by large block of shareholders they will certainly act out of self-interest. Amihud and Lev (1981) have provided proof that unless closely monitored by large block shareholders managers will attempt to reduce their employment risk through diversification. Lane et al.(1998) in this study have reexamined Amihud and Lev findings about agency theory Using a sample of 309 US firms that diversified between 1962 & 1970, from the Federal Trade Commission (FTC) Statistical Report on Mergers and Acquisitions (1976). This study falls in the third broad category[1] of agency studies. However this analysis only examines the strategic behaviors of managers when they are not under siege and are also not in a situation, in which their interests are clearly in conflict with those of shareholders. Specifically, firms without large block shareholders are expected to engage in more unrelated acquisitions and show higher levels of diversification than firms with large block shareholders (Jensen and Meckling (1976))

Using Multiple Regression, the study found no evidence for the standard agency theory predictions that management controlled firms are linked with strategically lower levels of diversification and lower levels of returns than are firms with large block shareholders. It was found that Ownership structure and diversification are largely independent constructs. Thus, managers may be are worthy of more trust and autonomy than what the agency theorists have prearranged for them. Rather than seeking to restrict managerial discretion through extreme oversight, a more balanced approach by principals is needed. Some safeguards are essential as conflicts of interests between managers and shareholders do arise in certain situations, therefore, the assumption that such conflicts dominate the day-to-day management is not realistic.

Matsusaka,(1993) takes a deep look at the astonishingly high pre-merger profit rates of target companies during the conglomerate merger wave. The main goal of the study is to assess how important was "managerial discipline" as a takeover motive.

The analysis uses an extensive data set of 806 manufacturing sector acquisitions that took place in 1968, 1971 and 1974. The sample was collected from New York Stock Exchange listing statements. Sample of 609 observations was taken from 1968, 117 from 1971, and 129 from 1974. The results did not differ in any vital way by year, so observations from the three periods were pooled. Because antitrust enforcement was strict in the late 1960s and early 1970s, it was safely assumed that the sample mergers were not motivated to increase market power Ravenscraft and Scherer (1987). This allowed the investigation to focus on a narrow set of merger motives. Profitability[2] throughout the study was measured as a rate of return on assets.

The theory identified two basic characteristics of mergers motivated to discipline target management. First it wsa observed that the target was underperforming its industry and the only reason to discipline the managers was that they were not maximizing profit. It could be because of incompetence that they were pursuing their own objectives. The second, the target company had publicly traded stock and the only posibility to discipline management was by electing an appropriate board of directors. In this situation a takeover was necessary to effect a change as the diffused stock ownership resulted in free-rider problems. Owners can remove bad managers of privately owned firms, as they are closely held. The problem occurs in large publicly traded firms with diffuse ownership.

The statistical results revealed that both public and private targets had extremely high profit rates prior to acquisition compared to their size classes and industries. Therefore, takeovers were not motivated to discipline target managers during the conglomerate merger wave. The second finding of the study is that public targets were not as particularly profitable as private targets. It was also found that the largest public targets had the lowest profit rates. A credible interpretation of the evidence is that managerial discipline may have been significant for just a small set of acquisitions that involved large publicly-traded targets. Matsusaka (1993) leaves the bigger question unexplained. Why buyers time and again sought high profit targets during the merger wave. There is a simple clarification, that high quality assets are generally favored to low quality assets, as high quality assets are more expensive. In addition to explaining why firms seek high-profit targets, an asset complementarity theory implies that firms tend to divest their low-profit divisions

Palmer and Barber (2001) have determined the factors that led large firms to participate in the1960s wave. The theoretical approach, of the study conceptualizes corporate elites (managers and directors) as actors. However it is assumed that these actors have interests which have arisen from positions held in organizational and institutional environments, and from multidimensional social class structure. Often Acquisitions are deviant and innovative ways by which corporate these elites can increase their status and wealth. Corporate elite diversify to the extent that their place in the class structure provides them with the capacity and interest to augment their wealth and status in this way. The authors have examined how the firm's top directors and managers class position influenced its tendency to employ diversification in the 1 960s. More specifically the following arguments on social status[3] have been tested empirically. Firstly, "Firms run by top managers who attended an exclusive secondary school or whose family was listed in a metropolitan social register were less likely than other firms to complete diversifying acquisitions in the 1960s." Secondly, " Firms run by top managers who were Jewish were more likely than other firms to complete diversifying acquisitions in the 1 960s." Thirdly, "Firms run by top managers situated in the South or west were more likely than other firms to complete diversifying acquisitions in the 1960s."

The study selected a sample of the largest 461 publicly traded U.S. industrial corporations from the Federal Trade Commission's Statistical Report on Mergers and Acquisitions (1976), between January 1, 1963, and December 31, 1968. This particular time period was chosen because as the merger wave took off at the end of 1962 and crested in 1968. The results of the study were found through count and binary regression models.

The findings of the study are consistent with that of Zeitlin (1974). According to him top managers' capacities and interests are shaped by their social class position. Corporate elite members differ in their social class position. It is this variation that influences the behavior of the firms they command. The results indicate that social club memberships and upper-class background influenced a firm's propensity to complete diversifying acquisitions in the 1960s. Network embeddedness and status influenced acquisition likelihood in opposite directions. Corporations that were run by chief executives who were central in social networks but marginal with respect to status were more likely than other firms to complete diversifying acquisitions in the 1960s. Therefore, individuals with high status had small interest in adopting innovation. Corporate elites can inhibit the spread of an innovation when it threatens their interests. As observed by Hayes and Taussig (1967), "One must never underestimate the moral suasion that the business and financial communities can bring to bear on those who engage in practices of which they disapprove." In this respect, the analysis provides additional evidence that intraclass conflict shaped corporate behavior during the 1960s merger wave. It seemed that in the 1960s, it was not concentrated ownership but, ownership in the hands of capitalist families that reduced a firm's tendency to complete diversifying acquisitions. Further, as predicted by agency theory , concentrated ownership would lower acquisition rates most when in the hands of the CEO or other top managers, as opposed to outsiders, However it was found the reverse to be the case. Overall, there was very little support for any of the agency theory in the 1960s merger wave. Further, the results provided no support for several of the class-theory hypotheses. Firms headquartered in the South or West run or by Jewish CEOs did not have a greater propensity to complete diversifying acquisitions during the 1960s.

The process of diversification of American firms reached its height during the merger wave of the late 1960s. Matsusaka(1993)evaluated the 1960s merger wave. In an attempt to do so the author has proposed a number of explanations that drove managers to diversify during the conglomerate merger wave.

There are reasons to suspect that managers may have pursued a diversification strategy even when it impaired the shareholder. They may have entered new lines of business to protect their organization-specific human capital or establish themselves. On the other hand, they may have been pursuing size as an end and because of strict antitrust opposition to horizontal and vertical mergers they had to expand by buying into unrelated industries.

The study has evaluated whether manager were diversifying for their own advantage or in the interest of shareholders returns .To do so the author inspected the effect of diversification on the value of his firm's equity. Thus, if the value of a firm declined upon announcement of an acquisition, then its management was not acting to maximize shareholder wealth.

One explanation for conglomeration stated in the study, stems from Managerial-Discipline theory. Firstly, Firms were taken over to discipline or replace their bad managers ie “Managerial-Discipline". Secondly, "Managerial Synergy" theory states that the bidder management wanted to work with target management, not replace it. In this case the acquirer management believed that the target management would complement to their skills. Therefore firm that had Managerial-discipline problem were likely to have had low profits, and on the other hand managerial-synergy targets were likely to have had high profits.

Another explanation is that buyers were motivated by earnings-per- share (EPS) manipulation. This explanation states that conglomerates have a high price-earnings ratio (P/E). [4] Therefore the bidder management was bootstrapping, by buying firms with low P/Es.

Construction of the dataset began with a list of mergers from the sample of 1968, 1971 and 1974 .The sample was identified from the takeovers from New York Stock Exchange listing statements and the results were presented through regression.

The announcement-period return to the bidders' shareholders was measured through "dollar return," [5] .Regression of the dollar-return measure found that the return to a diversification acquisition was significantly positive. On average their shareholders enjoyed an $11.0 million value increase in value when bidders made a diversification acquisition,. This rejects the hypothesis that diversification hurt shareholders and is thus inconsistent with the idea that diversification was driven by managerial objectives. On the other hand, bidders who made related acquisitions cost their shareholders $6.4 million on average. Thus, the hypothesis that the market's reaction was the same to related acquisitions and diversification is rejected, suggesting that there was a market "premium" to diversification.

Using descriptive statistical summaries it was found that both diversifying and horizontal buyers preferred to buy firms that were profitable. For both type of acquisitions the average operating profit was more than 5% in excess of the target's industry average. Therefore fame of high-profit targets argues against the importance of a managerial-discipline motive for both types of acquisition and in favor of a managerial-synergy motive. This is because Managerial-discipline takeovers should have been directed at low-profit firms, whose profitability needed improved. The motive was Managerial-synergy as the targets were takeovers were high- profit firms, this is because synergy-motivated managers were looking for good partners Matsusaka(1993).

Another factor linked to the managerial theories is whether or not the target's management was retained.Top management is said to have been "retained" if it meet the following criteria. Firstly It was reported in the Wall Street Journal that the acquired firm's management would continue to operate under the new management. Secondly, it was indicated in the buyer's listing statement that the target's management would be retained. Lastly, when the merger took place at least one of the top three executives of the target firm was still managing the firm three years later from when the merger took place. According to the above mentioned definitions, 61.8% of the managers in the sample were retained and only 3.5% of the acquisitions fell in the "Replaced" category.

The main finding is that buyers earned significantly positive announcement-period returns during the conglomerate merger wave when they made diversifying acquisitions. The hypothesis that conglomerates were driven by empire building or some other managerial objective can be rejected because such explanations imply value decreases to unrelated acquisitions.

Another explanation of the conglomerate merger wave is that mergers were driven by an accounting trick rather than expected efficiencies. Therefore, investors watched EPS; when the EPS went up they bid up the price of the stock. According to this argument, Conglomerates, tended to buy companies with lower P/E ratios than their own in order to increase their EPS and boost their stock prices. There was no evidence that firms earned positive returns which inflated EPS in this way.

The study indicated that early conglomerators earned significantly positive returns simply because they were first. They may have gained some rents to organizational innovation. Possibly the men who built the first conglomerates had a unique talent for diversification, which the market rewarded.

Hubbard, & Palia (1999), have examined the likelihood that internal capital markets were formed to alleviate the information costs associated with the less well-developed external capital markets of the time; that is, whether they were expected to create value by the external capital markets in the 1960s.In this paper, the authors have inspected a form of cross-subsidization that occurs when a financially unconstrained bidding firm takes over a financially constrained target firm and as a result forms an internal capital market.The study examined whether the external capital markets expected that the formation of internal capital markets in the 1960s were value-maximizing for the bidding firm. However, existing research has argued that internal capital markets can be value-enhancing. As argued by Geneen(1997), the financing and budgeting expertise that a firm possesses is not necessarily related to its degree of diversification. Accordingly, the internal capital market hypothesis for all acquisitions is tested.

The study also tests the "bootstrapping" explanation for conglomeration in the 1960s, which takes place when firms with a high price-earnings ratio (P/E) took over low P/E target firms and fooled the stock market with an increased combined earnings-per-share.

In the 1960s, external capital markets were less developed in terms of company-specific information production than in later years. The authors have classified "company-specific information" into two general categories. Firstly, production information; and secondly, financing and budgeting expertise. However, in this study information-intensive activities were introduced. This was because; it assists the manager to internally allocate capital across divisions of a diversified firm. It was suggested that diversified firms were perceived by the external capital markets to have an informational advantage, because external capital markets were less well developed at that time. Comparing it to the current decade, there was less access by the public to computers, data- bases, analyst reports, and other sources of company-specific information. Not only this there was less large institutional money managers and the market for risky debt was illiquid.

The authors selected a sample of 392 acquisitions that occurred during the period from 1961 through 1970. Diversifying acquisitions were defined as those in which the bidder and target do not share any two- digit SIC code Matsusaka(1993), and related acquisitions as those in which they do share a two-digit SIC code. Further the Wall Street Journal was used for announcement date as the event date. Four measures of abnormal returns to the conglomerate bidding firm were calculated. These measures are as follows. Firstly, the usual "percentage returns" or the cumulative abnormal returns from five days before to five days after the event date. Secondly the "percentage returns until date of last revision" or the cumulative abnormal returns from five days before to five days after the date of the last revision (Lang et al. (1991)). Thirdly, the "dollar returns" or the percentage return times the market value of the bidder six days before the announcement (Malatesta(1983); Matsusaka(1993)). Lastly, the "investment return" defined as the change in the value of the bidder divided by the purchase price (Morck et al. (1990)). Tobins r ratio[6] is used as a proxy for a firms capital market opportunities.

The evidence from these measures is mixed. Positive abnormal returns for all four measures were shown for related acquisitions. On the other hand, two of the four measures had shown statically significant positive abnormal returns for diversifying acquisitions in. Not only that diversifying acquisitions do not significantly earn less than related acquisitions in two of the four measures. Thus, evidence suggests, the capital markets believed acquisitions to be generally good for bidder shareholders during the 1960s.

More significantly, it was found that when financially unconstrained buyers acquired constrained target firms, highest bidder returns were earned. Further, bidders generally retain target management, signifying that management may have provided company- specific operational information and the bidder on his part also provided capital budgeting expertise. Therefore, external capital markets expected information benefits from the formation of the internal capital markets.

The study found no evidence in support of the bootstrapping hypothesis, as the coefficient on the dummy variable[7] was not statistically different from zero. This result is consistent with Matsusaka, (1993), who also finds no evidence for bootstrapping.Therefore, firms merged to form their own internal capital markets as there was a deficiency of well-developed external capital markets in the 1960s. Some firms apparently had an information advantage over the external capital markets and were expected to produce value in an internal capital market. In the 1960s diversified acquisitions were rewarded by financial markets, the informational advantage that acquiring firms appeared to possess was likely to be in the capital budgeting, allocation process and operational aspects of each division. Bidder firms generally retained the target management as it would facilitate them running the operational part of each target firm.

The Motives discussed in the above mentioned articles are appealing; however evidence from the stock market suggests that shareholders preferred their firms to diversify. Using a data set from the '60s and early '70s, Matsusaka (1993) reported that, when the company announced an unrelated acquisition, the stock price of the bidder increased on average of $8 million. However, on the announcement of a related acquisition, the bidding firm's stock price fell by $4 million. The difference between the two returns is quite significant. Thus it appears that investors fully believed that unrelated acquisitions benefited their firms relative to the alternatives. Thus the managers just did what the stock market told them to do that is to diversify. Evidence from 1980s stock market suggested that shareholders, again, liked what was happening. Shleifer, and Vishny (1992) found that in the 1980s, stock prices of the bidding firms rose when they bought other firms in the same industry, and fell with unrelated diversification. It is clear that the market disapproved unrelated diversification. Therefore it does not astonish that, in light of such market reception, managers stopped diversifying and did what the stock market directed them to do.

2.2 Legal Motives

Matsusaka (1996) investigated whether the antitrust enforcement of the 1960s led firms to take on the diversification goal, by preventing them from expanding within their own core industries. If correct, diversification should have occurred more less frequently when small firms merged than when large firms merged since small mergers were less likely to have attracted antitrust attention. Further the author examined the diversification patterns in the United Kingdom, Canada, Germany, and France in the late 1960s and early 1970s, where none of these countries had legal restrictions on horizontal growth similar to those in the Unites States.

The US Clayton Antitrust Act was the antitrust legislation in the postwar period (1950 Celler-Kefauver amendment to Section 7). The act, prohibited mergers that would substantially "lessen competition, or tend to create a monopoly." This new law was used by the antitrust authorities and the courts to limit the number of mergers between vertically related and firms in the same lines of business. The strictness of the antitrust environment in 1968 is illustrated by the observation that in the earlier 12 years, all antitrust cases that reached the Supreme Court had been resolved in support of the government. The study indicates the following two implications. Firstly, large horizontal mergers were more liable to have been challenged on antitrust grounds than small horizontal mergers. Secondly mergers between unrelated firms were unlikely to have been blocked, regardless of size. Firms diversified in 1960s, since antitrust authorities prevented them from expanding in their home industries. Later when antitrust policy became less rigid in the 1980s, firms expanded horizontally, leading them to refocus on their core business. Stigler (1966) was perhaps the first to present evidence on the antitrust hypothesis, concluding that, "the 1950 Merger Act has had a strongly adverse effect on horizontal mergers by large companies."

The author selected a sample of 549 mergers (that took place in 1968) from the New York Stock Exchange. Results of the study were reported through Logit regressions .It was found that bidders were as likely to have entered new industries when they made small acquisitions as when they made large acquisitions, and small buyers were as likely to have diversified as large buyers. Further the total number of diversification acquisitions concerning small companies was high.Though, according to the antitrust hypothesis; diversification should have been widespread primarily in large mergers where same industry acquisitions were prohibited by tough antitrust enforcement.

Secondly assembled international evidence indicated that diversification took place in many industrialized nations in the 1960s and 1970s, although restrictions against horizontal combinations were unique to the United States. Yet, most other industrialized Western nations[8] experienced diversification merger waves and general movements toward diversification in their largest companies (Chandler (1991)).Thus most of the evidence, is not consistent with the antitrust hypothesis, signifying that other explanations for corporate diversification should be emphasized not the anti trust hypothesis.

Scholes and Wolfson (1990) state, that the changes in U.S. tax laws[9] in the 1980s had obvious affect on the desirability of mergers and acquisitions. However such transactions were not only motivated by tax factors but also non tax factors[10]. Tax laws can have number of affects on mergers and acquisitions , which can include the following "capital losses, presence of tax-attribute carry forwards such as net operating losses , investment tax credits, and foreign tax credits, among others, that might be 'cashed in' more quickly and more fully by way of a merger; the desire to 'step up the tax basis' of assets for depreciation purposes to their fair market value; the desire to sell assets to permit a change in the depreciation schedule to one that is more highly accelerated." The authors in this study have examined the effect of changes in tax laws passed in 1980s on merger and acquisition activity in the United States.

The authors selected the annual values of mergers and acquisitions from 1968 through 1987 in nominal dollars. The data source for nominal values was W. T. Grimm and Company for 1968-85 and Mergers & Acquisitions (1987-88, rev. quarterly) for 1986 and 1987. Using time series analysis it was found that the dollar volume of merger activity between 1980-1981 increased from $44.35 billion to $82.62 billion (86%) in nominal terms. The percentage increase was approximately twice as large as the next largest percentage increase in annual merger and acquisition activity over the 1970-86 periods. There was spectacular increase in merger activity that began with the passage of the Economic Recovery Tax Act of 1981, however this was not the only "merger wave" that occurred in that time frame. Unusual merger activity was also witnessed in the 1960s. The termination of 1960s wave was accompanied by quite a few regulatory events that depressed such transactions. Firstly, the Williams Amendments had enlarged the cost and difficulty of effecting tender offers. Secondly the issuance of Accounting Principles Board Opinions 16 and 17, forced many acquiring firms to boost depreciation expense, goodwill amortization and cost of goods sold. Thirdly the Tax Reform Act of 1969, made transferability of tax attributes (net-operating-loss carry forwards) more restrained. Therefore there was a sudden decline in merger activity from the peak in 1968. Relative to the tax benefits when the non tax benefits of the transaction were small, current management were the most efficient purchasers, as they had an advantage along the hidden information dimension. Therefore 1981 act had increased the incidence of cases in which non tax benefits were less than the common tax benefits of mergers and acquisitions. As a result, there was an increase in the number of transactions involving management buyouts. The annual dollar value of unit management buyouts between 1978-80 increased by a factor of 3, and by a factor in excess of 20 for the period 1981-86.

The antitrust proposition mentioned above is appealing as one of the most important reason for diversification, during the '60s and '70s, which simply disallowed mergers of firms in the same industry, regardless of the effects of these mergers on competition. In a couple of cases, antitrust authorities even challenged mergers of two unrelated firms. The stringency of the antitrust environment in 1968 is illustrated by the observation that in the preceding 12 years, every antitrust case that reached the Supreme Court had been resolved in favor of the government. Faced with this strict policy, corporate managers who wanted to acquire firms in the 1960s basically had to diversify or face a costly antitrust challenge. However in the 1980s the Regan administration implemented hands off policy. The antirust authorities stopped challenging mergers of firms in the same industry, therefore mergers in the same industry increased sharply. Thus, when antitrust policy became less stringent in the 1980s, firms were able to expand horizontally, leading them to de-diversify and refocus on their core business.

2.3 Wealth Gains And Losses

Hostile Takeovers attract strong positive and negative reactions. Bhagat et al. (1990) in this study have tried to identify the driving force behind these takeovers, which typically involve major shareholders wealth gains of the target firms. The examined question is where do these wealth gains come from, further the possibility that wealth losses by bidding firms' shareholders as the explanation for target firms shareholder gains is also analyzed. Hence, the role of wealth changes in bidding firms' shareholders are first dealt with in the study, and then it explains some potentially important changes that can justify takeover premium in hostile takeovers

There are potential sources of takeover gains that have been identified and tested in this study. First, target shareholders can gain a major chunk of the wealth, however less is known about bidding firms shareholders. If they benefit then to account for the wealth gains, operational changes analysis must come up with superior savings. However, on the other hand, if they lose; a slight shareholder wealth increase explains the phenomena. Second, often hostile takeovers involve acquisitions of closely related firms. In this case gains from related acquisitions are expected to come from enlarged market power and superior operating efficiencies or from. Combined operating efficiencies can come from combined research and development, marketing, procurement, headquarters operations and distribution. These gains might be augmented if the target is not run efficiently and is acquired by a firm with better managers who find ways to decrease costs. Third, Labor costs are one of the major costs in most corporations. Therefore a reduction these cost is an effective ways to enhance cash flow. Such savings can be done number of ways, including layoffs, e hiring freezes, early retirements; reductions in future pension benefits, wage reductions and reduce employment. Forth, a large amount of hostile takeovers are followed by considerable divestitures, as the sales of divisions of the target companies to other firms. However, the authors describe the degree of divestitures to note how they should be interpreted. Bidding firms might sell off divisions of the target firms simply to pay off some of the debt incurred in the acquisition. Fifth, gains can be realized from tax savings. A merger of a profitable company with one that has tax losses can result in a tax save Auerbach and David (1988). Another source of tax savings is to convert the target into a partnership. As a result, double taxation can be avoided. Investment cut is the last source of gain. Jensen(1986) argues that takeovers prevent target firms from investing their surplus cash in projects that have a negative net present value . There are takeover gains since the money formerly wasted would now distributed as dividends.The above mentioned gains are possible sources of target shareholders' wealth gains and on the other hand are bidding shareholders' losses.

The author seclected a sample of 62 firms that were targets of hostile takeover offers of more than $50 million, as reported by W. T. Grimm for the period 1984-86. Like most others, W. T. Grimm[11] classifies hostile takeovers as those in which the target's board at least initially expressed opposition, if only to raise the price.The sample stops in the year 1986 so that changes in the coming two to three years following the takeover can also be analyzed. Out of the total sample 50 target firms were acquired and 12 remained independent. To avoid using proof from friendly acquisitions and to judge hostile ones only the study has selected the sample for only hostile takeovers.

The post-takeover analysis is complex because once merger takes place; it becomes near to impossible to attribute to the target the changes recorded in data that is jointly accounted. As a result, the study does not use the above mentioned data source. all through the analysis, the authors rely on numerous data sources which include the following, the Center for Research and Security Prices (CRSP) stock price data; 10 K forms; bidder and target annual reports;; Moody's Industrial Manual; Value Line Investment Survey; the Wall Street Journal; the New York Times; Business Week and trade publications; and DATEXT.the study than

To gauge wealth changes of shareholders the study has calculated the premium, paid to the target firm's share- holders. The authors have used the following formula to calculate the premium. First, the study takes the period from twenty days before the first bid for the target was announced, to the day when the target accepted or defeated the bid. the 1st date is stated as Date 1 and the 2nd Date2. Next the market model is estimated for each target firm from two hundred and sixty trading days before Date l to 60 trading days before Date 1. The market model is than used to forecast the price of the target that would be on Date2 given its real value on Date l, and the market return between Date l and Date2. The premium is than, the difference between the price paid for the target firm and the predicted price of the target firm on Date2 i.e. The difference between what was paid or offered for the target firm and what the price would have been on the resolution date if no takeover activity had occurred. However, calculating wealth changes of the bidding firm's shareholders is complex because there is no equal of the price paid. Consequently, to assess the wealth change of the bidding shareholders, the study has defined Date3. Date 3 is the date when the first bid was made by the actual acquirer. The market model is than estimated for the bidding firm from two hundred and sixty days before to sixty days before Date3. The model is than used to work out the abnormal change in bidding firm's shareholders' wealth from 3 days before and after Date3. The authors have defined this abnormal wealth change as the return to bidding shareholders. However the study does not estimate bidding shareholders' returns in unsuccessful bids since the authors found no satisfactory procedures. Hence, in these cases the bidding shareholders are ignored the gain offered to the target shareholders are in terms of changes implemented after the takeover bid. Secondly, layoffs have been measured as the sum of early retirements and layoffs and from the retained divisions of the target company. The majority of the information on layoffs came from the Wall Street Journal, even though other sources were also sometimes used. Thirdly to calculate the value of divestitures the authors observe at all post takeover divestitures during 3 calendar years from the start of the takeover. The study followed the same approach as with layoffs, making sure that the sold off firm was from the target. However, when it proved unworkable to attribute the divestiture to the target, the study counted it as zero. This was the main setback when the target and the bidder belonged to the same industry and had fairly homogeneous assets. The key sources of information on divestitures were the Wall Street Journal, and Moody's.In nearly all cases the study found prices of divestitures; otherwise, the divestiture was not counted.

The study found a number of factors that drove the hostile takeovers in the 1980s. First, hostile takeovers mostly allocate businesses to firms that own related businesses. These related businesses were acquired by not only in the takeover but also in sell offs after the takeover. Of the $69 billion in assets that changed hands in our sample 72% of firms in the sample ended up in the hands of firms that were managing similar assets. Second, MBO teams and Raiders mainly served the temporary function of brokering, which involved the transfer of assets to the related acquirers. Concentrated ownership and High debt levels gave organizations a powerful incentive to execute a bustup, but the task of following management was left to others. Third, layoffs were a significant but not a leading source of hostile takeover gains. Layoffs were excessively targeted at white collar employees. Fourth, sell offs were result of hostile takeovers, and in a lot of cases they resulted in a liquidation of the target. Fifth, tax savings were significant in some cases, mainly in LBOs, however the benefits of the debt tax shield were considerably reduced by the speedy repayment of debt. Conversion to partnerships and Tax losses were not that common, nevertheless they provide great benefits when they did take place. Sixth, takeover gains that were significant in some cases were from the bidding firm's shareholder losses and reduction of wasteful investment by the target firm. Therefore takeover wave of the 1980s was the wave of related acquisitions. A great deal of this wave was reflected in the increased speed of friendly related acquisitions. However in addition, the speed of hostile takeovers in the 1980s skyrocketed. Many of the hostile takeovers intended to deconglomerate big corporations and to allot their various divisions to related acquirers. MBO organizers and Raiders bought diversified firms and than sold off the parts to related acquirers, profiting substantially. Even though the takeover process was driven by this expansion into related businesses, numerous opportunities to increase the value of firms must have also encouraged hostile takeovers. Tax subsidies to debt undoubtedly helped the MBO organizers and the raiders to realize some gains.

The study clearly indicates what hostile takeovers were not in the 1980s. They were not a typical indication of transformation in the firms' internal organization. Acquisitions by raiders and Management buyouts in the 1980s were often a temporary step in asset reallocation. The eventual holders of assets were the big public corporations.

The Motives discussed in the above mentioned article leaves one question open ; where do the value gains come from in strategic acquisitions? The fact that in a typical strategic acquisition the combined wealth change of the target and bidding shareholders is positive hints that the market believed these gains did exist. The study identified some sources of efficiency improvements, such as tax savings and headquarters layoffs, but they clearly are not the whole story. There may be many efficiency gains in distribution, procurement, production, that the study has not captured.

Shleifer and Vishny (2003) have developed a theory concerning the causes of mergers. It provides an explanation for why acquirers' shareholders appear to lose from the merger. The authors have assumed that the shares of a number of companies become overvalued at some point in time. Hence the acquirers' managers increase their shareholders' wealth through trading their overvalued shares for the real assets of target companies. The market correction of the initial overvaluation is than the post-merger decline in acquirers' share prices, not a caused of the merger. Mueller and Yurtoglu (2007) evaluate whether mergers on average boost wealth or decrease it. To do so it is essential to find out whether the post-merger losses experienced by acquirers' shareholders were caused by the acquirers' overvaluation at the time of the acquisitions or by the merger itself. Accordingly, the authors expect that all mergers in all countries are not profitable, while others are. Hence shareholders of not only the acquired company gain from profitable merger but also shareholders of acquiring companies gain.

The authors have divided the hypotheses about the causes of mergers into two broad categories, the Neoclassical[12] and Behavioral theories. The neoclassical hypothesis is further divided into two categories. First, the Synergy Hypothesis (SH), all mergers that generate synergy result in an increase in market power and both shareholders of the acquiring and acquired firms gain from the merger. Second, the market for corporate control (MCCH) is tested. Under this hypothesis, a merger is a way to replace managers who are not maximizing the firm's value. This can be owing to agency problems or incompetence. Thus all gains occur from replacing the target's management. Next, the behavioral hypothesis is also divided in the following three categories. First, The managerial discretion hypothesis (MDH) it assumes that managers are growth maximizers, and tend to undertake mergers to expand their firm while destroying shareholder wealth. Second, the overvaluation hypothesis (OH), it has the same predictions as the MDH, however for different reasons. Acquiring companies' shareholders experience positive returns in the short run but negative abnormal returns over long windows after the merger has taken place. When the shares become overvalued, meaning the market value of the firm is greater than the present discounted value of its earnings stream, its managers take on the acquisition in which they swap their overvalued shares for the seemingly correctly valued shares of another firm. Consequently, their shareholders benefit from the merger, since they obtain properly valued real assets by trading the overvalued paper. The authors assume that the target's managers are eager to accept these overvalued shares, because they desire to ‘cash in' their stakes in their company. Later when the merger has taken place, the market rectifies its error about the acquirer's overvaluation, and than the share price decreases.

The authors have selected a total sample of 9733 acquisitions that took place in the US, Great Britian, Australia, Japan, Canada, and Scadinavia. The main source of data was Global Mergers and Acquisitions database from Thompson Financial Securities. The database covers all worldwide transactions valued at $1 million or more. The sample period of the study was 1981- 2002.The study calculated and than regressed the abnormal returns for every acquisition in the sample.Abnormal returns (AR) for an aquiring company over a t+n day window was calculated by the following formula.

Where is the return of A over the n days window (n =20 for the short term window, n=250,500 and 700 for long windows).Is the return on a portfolio of non- acquiring firms in the same size and same industry of the acquiring companies' country.

During the 1980s merger wave, the study found similar patterns of returns across all types of countries with exception of Japan. The mean gain over a short window (21 days) was 0.6%. However this picture changes considerably as the market gets additional time to assess the mergers and the acquiring firms. After 3 years, acquirers' shareholders lost market value. However, these post-merger negative returns might not be caused by the merger itself ,rather to some degree a consequence of the companies' being overvalued at the time of the mergers. This overvaluation was reason of mergers under the Overvaluation hypothesis, and was consistent with both the managerial discretion and Hurbris hypothesis. Further the study found that some mergers boost efficiency and market powers. Whereas others fit the behavioral hypotheses that assume agency problems or overvaluation cause a merger

The objectives discussed in the above mentioned article seem pleasing. The picture one gets of the returns from mergers changes radically as the observation window extends following the mergers. Although acquirers appear to experience humble gains over short windows, they experience big losses over long windows.As post-merger negative returns may not only be caused by the mergers, but to some degree as a result of the companies' being overvalued at the time of the mergers.

2.4 Value of Diversification

Servaes (1996) examined the benefits of diversification during the 1960s and early 1970s when firms started to diversify. Throughout this time, corporate America went through the conglomerate merger wave. Ravenscraft and Scherer (1987) documented that 36 percent of all the acquisitions during 1964-1972 and 32 percent of all acquisitions during 1973-1977 were conglomerates. The author examined whether diversification led to higher market values during this period, and if so, what were the sources of this valuation differential.

The authors have identified the following three factors that explain the change in the discount over time and the valuation discount of diversified firms. First profitability, firms that have low profitability are most probably to trade at a discount to similar firms with higher levels of profitability. Therefore, if diversified firms are systematically less profitable than single segment firms, and if this difference changes over time, it is possible that the trend in relative profitability can explain the trend in valuation. Secondly, differences in capital structure can also lead to valuation differences if leverage and firm value are related. Finally, corporate value can be related to a firm's investment level. As suggested by Meyer et al (1992), firms operating in more than one business unit may over invest because the profits of units with positive cash flows are used to finance poor investments in units with negative cash flows

To answer the above mentioned questions, the study examined whether the benefits of diversification were greater or less than the costs. The author selected a sample of US firms in three-year intervals from 1961 to 1976 and examined whether the benefits of diversification outweigh the costs. The year 1961 was when the conglomerate merger wave started, thus was chosen as the starting point. The study compared the mean and median Q ratios[14] of single segment firms with those of multiple segment to measure whether diversification was valuable during the time when it was fashionable.

The author found no proof that diversified firms were valued more than single segment firms in the 1960s and early 1970s. On the other hand, for some years when compared to single segments firms diversified firms sell at a considerable discount. This discount is significant and large over the 1961-1970 period, but it becomes insignificant and small in 1973-1976. These results hold after controlling for industry effects and for differences between diversified and undiversified firms in leverage, investment policy and profitability. When the penalty imposed by capital markets is small it is than that a major increase in diversification takes place over the 1970-1976 period. Therefore, the firms that diversified at that particular time did not impose a cost on their shareholders. Further clues are provided by the pattern of insider ownership over the sample period that can explain the behavior of corporations. Diversified firms had lower insider ownership than single segment firms when they were selling at a discount to single segment firms .however, when the discount was eliminated, it made slight difference in insider ownership between the two groups. These findings imply that insider ownership was a useful prevention to diversification when it was costly to shareholders. On the other hand when the cost to shareholders was insignificant, the firms with high ownership were the first to diversify, perhaps because insiders wanted to lower their exposure to firm-specific risk. This does not suggest, however, that agency costs caused the discount or led to the change in discount over time.

The study found that at the start of the diversification wave, diversified firms were also valued at a discount compared to single segment firms during the 1960s.Though, the diversification discount fell over time and is not considerably different from zero in the early and mid 1970s. It is during the latter years that the firms in the sample increased diversification the most. If managers obtain private benefits from diversification, the fall in the discount might explain why more firms started diversifying. Evidence on the relation between insider ownership and diversification is consistent with this assumption. When the diversification discount was large firms with high insider ownership remained focused (1961-1970). These firms were the first to diversify when the discount declined (1973-1976). Thus, managers did not inflict costs on their shareholders by diversifying because the diversification penalty was close to zero when diversification peaked. However, firms that were already diversified in the 1961-1970 period did impose substantial costs on their shareholders by not increasing their focus.

The authors also investigated what caused the discount to not only occur but also change over time. The differences in capital structure, profitability, and investment policy were insufficient to explain the level and change of the discount. However, what caused the diversification discount to change over time remained a puzzle.

The evidence presented in this article, joint with the evidence of Lang and Stulz (1994) and Berger and Ofek (1995), suggests that, in general, diversification has not been beneficial for U.S. corporations.

The above mentioned proposition is engaging, as the diversification of the 60s did not on average lead to profitability improvements and to a substantial extent reversed in the coming years therefore was not beneficial to the US corporate world.

Montgomery and Wilson (1986) evaluated the penalty of the 1960s merge wave. However the consequences of such macro-level phenomena are numerous. Thus study focused on one of the issues: the extent to which acquisitions made within this period were later resold. The study examined whether unrelated acquisitions were resold at a higher rate than related acquisitions. The authors selected a sample of 434 US acquisitions from the Federal Trade Commission's Statistical Report on Mergers and Acquisitions (1975)[15].Using chi-square statistic it was found that, at most, unrelated acquisitions were resold at a moderately higher rate than related acquisitions. One reason for this could have been that firms used both unrelated as well as related acquisitions as a strategic planning tool. In this case it may have be attempt by the acquiring firm to refocus defensively, as it planned to move away from core businesses which had limited future potential. Secondly, not finding that a significantly higher proportion of unrelated acquisitions were resold could indicate that related acquisitions were less successful than anticipated. However by the end of 1982 the greater parts of mergers of the 1960s were still un reversed. A slightly higher proportion of unrelated, than related acquisitions were resold, but the difference is not significant statistically.

The motives discussed in the articles above are interesting. However Contrary to the expectations of the 1960s wave, the experience with diversification has been disappointing. As widely documented by Ravenscraft and Scherer (1987), profitability of acquired companies did not, on average, improve. Furthermore, starting in the 1970s, many of the acquisitions were reversed through divestitures. Porter (1987) reports that half of the unrelated acquisitions made were later divested. Ravenscraft and Scherer (1987) estimate that one-third of all acquisitions made in the '60s and '70s were later divested, and their sample stopped before divestitures became really massive. Kaplan and. Nevertheless, the fact that diversification of the 1960s did not on average; lead to profitability improvements can not be refuted.

It has been extensively argued by many authors that the purpose of Merger wave of the 1980s was to create of more competitive and industry specialized firms. This was done in response to the increased global competition[16]. A number of studies demonstrate that corporate focus enlarged during the 1980s; this increase in corporate focus was often achieved through divestiture as it was associated with better corporate performance[17]. Hatfield et al. (1996) in this study has examined whether the corporate restructuring of 1980s really increased the degree to which incumbent firms within individual industries were specialized to that industry. As it has not been identified if aggregate industry specialization increased during the 198Os, nor is it known whether corporate restructuring of the 1980s was a determinant of change in aggregate specialization or not. Therefore this analysis is an attempt to answer the following two questions. First, “Did the average firm in any given industry become more or less specialized to that industry?” Second, “Was corporate restructuring a significant determinant of any change in aggregate industry specialization during the 1980s?” Industry level competitive conditions can be changed through an increase in industry specialization. There are two explanations. Firstly, the more focused the the better it can compete, as it has lower costs. Further it can also produce more differentiated products as they are industry specialized (Demsetz (1973); Porter (1985). Second factors of production which specialize in a particular industry can create high exit costs for incumbent firms, therefore industry specialized firms are more dedicated than diversified firms to staying in the industry and competing assertively Ghemawat (1991); Sutton (1991). Finally, an increase in the industry specialization of incumbent firms may decrease their organization costs, intensifying cost rivalry in the industry.

The study estimated aggregate industry specialization in two stages. First, the specialization is measured of each incumbent firm in a given industry. Firm i; industry j , the specialization S of i to j is measured as the ratio of all employees E of firm i employed in industry j to all the employees of firm i at time t:

The above equation produces an estimate of each firm i's specialization to each industry j which has a potential value between 0 and 1. Second, the weighted mean of Si for all i in j at t ,was used to calculate aggregate industry specialization. This weighted mean is used to avoid bias in industries where there are a lot of very small perfectly specialized firms however a small number of large diversified firms. Therefore, aggregate industry specialization for any industry j at time t is measured by the following equation:

This second measure of aggregate industry specialization has a potential value of between 0 and 1. Further, the Change in aggregate industry specialization was calculated as the absolute change between the years 1981 and 1989. Thus, in total the authors used four measures of changes in aggregate industry specialization in which are as follows. First, at the four-digit level for all incumbent firms between 1981 and 1989 the absolute change in aggregate industry specialization; Second, at the four-digit level for the largest four incumbent firms in each four-digit industry between 1981 and 1989 the absolute change in aggregate industry specialization. Third, the two-digit level for all incumbent firms between 1981 and 1989 the absolute change in aggregate industry specialization and lastly the absolute change in aggregate industry specialization at the two-digit level for the largest 40 incumbent firms in each two-digit industry between 1981 and 1989 were calculated. The results were than calculated through regressions. The author selected a sample of 750 firms from TRINET's database from 1989 &1981.

The study found that there was a slight decreased in the aggregate industry specialization during the 1980s. There was no proof found that ownership change of industry assets were a determinant of change in aggregate industry specialization. Therefore previous arguments that the merger boom of the 1980s was aggravated by the need to improve US industrial efficiency are refuted by this analysis. A possible explanation for the above mentioned findings is that other changes in industry assets during the 1980s very much outweighed ownership transfers in importance. Therefore, restructuring through plant addition, plant closure, industry entry, played a major role in altering industry level competitive conditions during the 1980s than did corporate control transactions.

Hostile takeovers were accepted actions during the1980s, most probably because during the 1960s firms had become over diversified from conglomeration, thus value could be created from buying these firms and then breaking them up. Therefore, a hostile takeover attempt may perhaps be viewed as a ‘wake-up call' as the target firm had been managed unproductively and that value could be produced by a change in management and strategy could produce value Denis, and Sarin, (1997). Chatterjee et al. (2003),in their study have attempted to answer the question that What happened to the firms that effectively repelled the takeover offer during the 1980s. Did these firms follow a restructure strategy to avoid future takeover offers or do they continue to follow the old strategy? In doing so the study has tested the following hypothesis:" In the aftermath of a rejected takeover offer, target firms that have boards with independent governance characteristics are less likely to refocus than firms without independent board characteristics. Therefore, board independence will be related negatively to the incidence of refocusing". The authors did not analyze the firms that right away accepted the initial take over offer since it would beat the purpose of trying to identify the exact point where the choice to realign a firm's strategy ought to be strongly considered by the firms in the study. Instead, the authors studied the firms that reject the first offer and then followed them to determine their succeeding actions. If these firms eventually decide to follow a refocus strategy, than the study had a strong basis for identifying the exact point in time when they were forced to think about altering their existing strategies to create shareholder value. In particularly, the authors were interested in the board's role for these two groups.

The author's selected a sample of 76 firms reported in Mergers and Acquisitions between 1981 and 1991. Data were collected from the Trinet Data Base, Compact Disclosure, the CRSP and Compustat tapes, Lexis/Nexis, Dow Jones News Retrieval, Security Data Corporation database. logistic regression was used to report the following results.

The study found that governance influenced the way in which a firm responds to a failed takeover attempt in the 1980s. Particularly, board vigilance explained this phenomenon. An independent board was more liable to be watchful on an ongoing basis with regard to evaluating firm strategies. Thus, the takeover attempt was less likely to make the board believe that the firm's diversification strategy was unsuitable. In other words, directors not only thought that their firm's strategy was suitable, but they were likely to be correct. The authors did not say that the board was correct ‘ex post' but that they believed they are correct ex ante and consequently they refused to change the firm's strategy.

This study also demonstrated the significance of a takeover attempt as a signaling mechanism in the 80's. Safieddine and Titman(1999) reported significant boost in leverage in targets that had terminated takeovers, though impact of governance characteristics were not accounted for on the actions taken by the firms.

During the 80s a takeover attempt acted as a ‘wake-up call' to cause the executives to rethink the strategies of the firm. Firms that had not been sufficiently governed most likely gained the most from refocusing. Hence, the takeover attempt served to lessen agency costs by exposing suboptimal diversification, which provided incentive to management and the board to alter the strategy to prevent additional takeover attempts and decrease employment risk. Boards that acted decisively and were not sensitively attached to the strategy of the firm were more likely to be objective about the indication from the market and respond to it Byrd and Hickman (1992). However, even the vigilant boards may have paid some concession to the takeover offer by increasing their leverage. This act may be a signal to the capital market that there is no reason to suspect that free cash flow is guiding the firm to take actions that are contrary to shareholder interests. Further, as the proportion of shares held by outsiders decreased and insiders increased, firms were more likely to refocus. Therefore during the 1980s wave target firms that were managed by independent directory boards were probable to disregard the takeover attempt and not refocus their firm's strategy. On the other hand, target firms that were managed by non independent boards were more likely to view the unsuccessful takeover attempt as a ‘wake-up call' resulting in a refocused strategy to survive.

The above mentioned proposition is interesting as it demonstrated that governance can make a difference with regard to how firms respond to a failed takeover attempt. Therefore, takeover attempt shock was unlikely to cause the board to think that the firm's diversification strategy is inappropriate.

2.5 Financial & Economic Rationale

The extraordinary growth in corporate merger activity of the 1960s revitalized interest and effects relating to corporate mergers. Shrives et al. (1979) have analyzed the evidence relating to the existence and, the extend of bankruptcy avoidance rationale for mergers of the 1960s. Lewellen (1971) gave the "pure financial rationale" for mergers. Combined debt capacity of the two firms as separate entities was less than the debt capacity of the firm resulting from the merger. The resulting capitalized tax subsidy from the merger would be beneficial to the shareholders of the merging firms. The authors selected a sample of 224 firms from the Federal Trade Commission's Large Mergers in manufacturing and mining during the period 1948-1971.Through Z Scores it was found that 15.2% of the firms in the acquired sample were found to be close to bankruptcy at the time of the acquisitions. Many instances or severe financial crisis among firms were resolved through the merger process, as recourses are reallocated to more productive use. Thus the merger served as a valuable function in the merger wave.

Melicher(1983) has analyzed the structural relationships between aggregate merger activity and macroeconomic factors. One explanation is that Merger activity may reflect both changes in business activity and changes in the capital markets. There is an increase in merger activity with expectations of economic growth and capital market conditions as they are favorable to financing mergers. On the other hand worsening capital market[18] conditions and declining economic expectations are expected to be associated with decreased merger activity. During the 1960s A substantial number of mergers did not involve the exchange of common stocks, acquiring firms used borrowed funds and relied heavily on convertible securities. In addition to stock prices the author explored changes in interest rate levels to better reflect the significance of a capital market conditions argument for changes in merger activity. Time series analysis was used to relate changes in aggregate merger activity to industrial activity, business failures, stock prices and interest rate levels. Quarterly data from the Federal Reserve Board's Index of Industrial Production and Dun and Bradstreet's record of failed firms provided the measures of economic activity. The time period chosen for the study is 1947 to 1977 which covers the conglomerate merger wave.

There was weak relationship between economic conditions and merger activity. However during the 1960s, there was considerable support between changes in aggregate merger activity as a capital market conditions phenomenon. Thus, changes in stock prices and bond yields could be used to forecast future changes in recorded merger activity. Merger negotiations had begun about two quarters before consummation. In the 1960s increased merger negotiation activity seemed to reflect the expectation of more receptive capital market conditions in the form of lower interest rates and higher stock prices.

The above mentioned article made an interesting proposition. As mostly it is expected that general "economic prosperity" provides the basis for explaining changes in aggregate merger activity over time. However the study has provided deep insight in this general belief, indicating a fragile relationship between merger activity and economic conditions and stating change in merger activity as a capital market phenomenon.

There are a number of motives behind the merger and accusation activity that has taken place in the past decades. Managers of bidding firms can be motivated by stock market undervaluation, by synergy, by corporate tax savings ,by the wish to restrict competition, ,by empire building motive, or by any other reasons. DeBondt and Thompson(1992) in this study have made an attempt to identify which mergers are socially desirable and which are not. However the answer depends on whether wealth is created or merely redistributed.The authors have examined the "efficiency story ", do efficiency motives account for a significant part of the merger motive. The study has discussed two classes of efficiency improvements. First of improvements is synergistic and creates value through the combination of businesses. The second improvement is focused on the target company only. Where by the acquiring company aims to discipline the target company.

The authors selected the sample all companies that were delisted from the New York Stock Exchange for reason of merger between 1926 and 1988 from the Center for Security Prices at the University of Chicago (CRSP).All companies were classified into industries. This approach enabled the authors not only looks at the volume of M&A activity in this time period but also, and the linked long- term price fluctuations in the stock market and the types of firms involved.

Through regression the analysis identified four distinct surges in merger activity: in the late 1920s, the mid-1950s, the late 1960s, and from the late 1970s onward through the 1980s.During the sample time period takeover booms were spanned almost all industries. They were not associated with bankruptcy and recession. However a boom was most likely to take place when the stock market was up, when the cost of capital was high, and when investors were generally more willing to trade. Thus aggregate merger activity was not linked to macroeconomic variables that can easily be interpreted in terms of economic efficiency. The company-by-company findings likewise suggested that efficiency motives played a minor role and other motives played a major role .during the 1960s there were long-term stock market losses of bidder , this suggested that the conglomerate merger wave of the 1960s was a failure Scherer (1988). Lastly it was found that, not all regulation was socially desirable but neither is all regulation socially negative.

The 1980s acquisition wave was a form of unwinding of the 1960s conglomerate merger wave. Most studies have reported poor performance of acquired business units of the 1960s wave and improved performance in the 1980s wave (Browne and Rosengren (1987); Ravenscrafi and Scherer(1987); Mueller (1985). Brush (1996) has examined the post acquisition performance of the merged business units of the 1980s.The study has analyzed performance changes which would be expected for each acquired business unit since opportunities to share resources increased Further these performance expectations were also compared with the actual changes in performance that were achieved in the 1980s.The author has used intraindustry analysis to control for industry effects on performance, and in so doing it allows the option that the effects of shared activities on performance will be unique to the industry under investigation.

The author has selected a sample of 356 manufacturing industries, between 1980 and 1984 from the FTC data base. Competitive performance, as measured by market share in an industry, is regressed. There were two important findings of the study. Firstly, operational synergy resulted in a net gain in shareholder wealth in acquisitions of the 1980s. Secondly this research has confirmed the findings, that the ex post performance of acquired businesses in the 1980s was superior than that of the first wave of 1960s.Superior performance of the 1980s wave is explained in terms of operational synergy. Therefore, the following alternative theories as causes for the 1980s acquisitions are not directly rejected. Agency problems may be more in conglomerates than in single business firms. These agency problems may also be reduced in related diversified firms that follow the strategic planning model as they focus on ‘interbusiness and interdivisional opportunities and dependencies', Chandler (1991). Other theories concerning antitrust, excess industry capacity and tax policy, as causes for the 1980s acquisitions are not directly rejected.

The proposition mentioned in the above articles is appealing. Unlike the 1960s takeovers, 1980s leveraged buyouts improved profitability, productivity of plants after the control change. There are, however, some reasons for optimism. First, there is evidence that takeover targets in the '80s were poor performers (Servaes, 1989), which suggests that they had room for improvement. Second, that divisions of diversified firms did not perform as well as similar businesses that stand alone or are part of undiversified firms. In sum, the evidence powerfully indicates that the takeovers of the 1980s were different from those of the '60s mainly because they undid what the previous wave had created. In the '60s, conglomerates were created; in the '80s, many of them were destroyed.

A first-mover advantage is a significant concept in the strategic management. By acting early relative to others, a pioneer can possibly gain a competitive advantage that enables it to reap positive economic profits Lieberman and Montgomery (1988). The resource based view states that an early mover can develop a resources that are uncommon, precious, hard to imitate and non substitutable[19]. Despite the wide recognition of first-mover advantages, few empirical studies have examined whether being an early mover in certain organizational practices materially affects performance. Carow et al. (2004) in their study have investigated affects on the share holder of an early mover in the acquisition process affects of the 1980s. Thus, the stock market reactions are compared to the early-mover acquisitions to evaluate whether pioneering acquirers gain larger shareholder gains from their acquisitions. In doing so the study has tested the following hypothesis. First," Early-mover acquisitions within an industry wave of acquisitions will realize higher total shareholder returns than will acquisitions made later in the acquisition wave. "Second," Early-mover acquirers within an industry wave of acquisitions will realize higher shareholder returns than acquirers that make acquisitions later in the acquisition wave". Third, " Early-mover acquirers within an industry wave of acquisitions that pay for their acquisitions primarily with cash conduct their acquisitions in growth industries, and acquire related businesses, will realize returns that are significantly higher than the returns realized by other acquirers."

The authors have selected a sample of 520 acquisitions of publicly traded U.S. companies between 1979 and 1998 from the Securities Data Company's (SDC) Mergers and Acquisitions database. Since the test relies on financial data, the authors have excluded all acquisitions whose financial data from Compustat and stock return data from the Center for Research in Security Prices (CRSP) for both the acquirer and the acquired firm were not obtained. The analysis is focused on industrial firms. Cornett and De (1991) has argued that the regulated nature of the financial services industry affects the stockholder reactions to acquisitions in this particular sector. Owing to this, transactions in financial services sectors are excluded. The primary step in the test for first-mover advantage was to identify industries that experienced a wave of merger activity. Second, the industries that experienced the merger wave were classified a according to whether the acquisition occurred in the beginning portion of the merger wave or later. This process, categorized acquirers as early movers for further analysis. Abnormal returns were calculated as indicators of successes .They were defined by differencing the returns earned by the shareholders of the firms involved in the acquisitions, and the returns to a value weighted portfolio of all firms in the same industry.

As an additional test, the acquirers long-term industry-adjusted stock returns were analyzed. Early movers differ significantly from the long-term industry-adjusted stock returns for the remainder of acquirers in the sample. The long-term industry-adjusted stock returns were determined for each acquiring firm as the compound return differential between the return on a value-weighted portfolio and the acquirer's stock returns of firms in the same industry. Compound returns were measured over one year period (250 trading days), a two year period (500 trading days), and a three year period (750 trading days) succeeding to the completion of the acquisition., the process of determining the industry-adjusted stock return performance for each of the acquiring firm For the 3-year period was represented by the following formula:

The hypotheses were tested, using regression and the following results were found. First, During the 1980s pioneer acquirers who acted strategically and had an informational advantage did in fact experience significantly higher stock returns than later acquirers in the 1980s acquisition wave. Thus there is a strong link between early movers and performance. In the acquisition Wave of 1980s the shareholders of the target firms captured the most of acquisition gains which were measured by announcement returns. On average, the announcement returns of acquirers within industry acquisition waves were negative. Moreover, strategic pioneers were those who had superior information experience positive acquisition announcement returns and outperformed other acquirers in terms of long-term stock price performance as well.

The above mentioned market reward of First mover advantage during the 1980s wave is appealing as it relies on the notion of information asymmetry, where a pioneer is able to benefit from on its superior information in order to identify and act upon some initiative to gain a head start over peer.

3 Conclusion & Discussion

There can be two alternative theories of the two takeover waves. According to the first theory 1960s conglomerate mergers were a good idea back than, but were no longer a good idea in the coming decades. Consequently, the cause for the roundtrip is that what efficient changed is over time. Williamson (1975) conglomerates were a product of the Multi Divisional form of corporate organization where central office allocate resources between divisions. The M-form organization is more efficient when divisions are independent firms, since the central office has a better understanding of the entire market. In the uncompetitive 1960s, when aggressive hands on management was not essential, this benefit in capital allocation exceeded the disadvantages may have occurred from the fact that headquarter's management was not closely familiar with the businesses of the divisions. In contrast the 1980s, the business environment became more competitive thus conglomerates were no longer an efficient organization form, as central offices could not understand the local market conditions. The markets response to these pressures was the break up of conglomerates to more focused and specialized companies. Further the aintitrust policy also accommodated this essential move to efficiency through specialization. This view of the two takeover waves is very appealing. As it is consistenit with stock market efficiency, as the market correctly approved of both the diversificatioin in the 1960s and the return to specialization in the 1980s.

Another interpretation of the experience is that diversification was a mistake from the start. If it were not for antitrust policy, managers would have pursued growth and survival objectives through related acquisitions as was the case in the '20s and '40s. Managers got away with diversification since shareholder control mechanism in the 1960s was weak. It is also probable that shareholders were puzzled about the benefits of unrelated diversification, as mentioned before the stock market evidence indicated that shareholders approved of unrelated diversification. Years later, it became obvious that diversification failed. Further, the antitrust environment had changed radically so diversification could be successfully reversed in the 1980s through divestitures, takeovers, and LBOs. Managers were quick to identify their past mistake; they divested and preserve the core businesses. On the other hand when managers who failed to recognize their past mistake, hostile raiders brokered the unrelated businesses to other firms in the same line of business.

In choosing between the two views, the review paper is inclined to give up the first view as it more logical. As profitability improvements under conglomerates were low. The market's confidence was misplaced, as unrelated diversification was a new experience and no one could really predict what would happen. The market not only made one mistake, but it made it consistently across a great number of transactions. In sum, I conclude that diversification was a mistake from the beginning.

[1] The first category: a context in which managers are clearly threatened. The second category :large block shareholder monitoring in settings where managers are not under siege, but their interests clearly conflict with those of shareholders.

[2] The profit rate of firm i is defined to be ri where ri = (NET OPERATING INCOME)/(BOOK VALUE OF ASSETS)

[3] Domhoff (1970) argued that social status in the U.S. is reflected in attendance at exclusive secondary schools (such as Groton) and inclusion in restricted metropolitan social registers (such as the Cleveland Blue Book). Only descendents of families of longstanding wealth and social prominence gain admission to such schools and listing in such registers. Baltzell (1958, 1964) maintained that social status in the U.S. has been segregated along religious and regional lines. He characterized the upper class as a Protestant and northeast- ern phenomenon. Thus, it is not surprising that Espeland and Hirsch (1990: 84) described the corporate elite members who pursued diversifying acquisitions in the 1 960s merger wave as "self-made men." Hirsch (1986) and Espeland and Hirsch (1990) also described these elite members as being disproportionately Jewish and from the South or West.

[4] It is an algebraic fact that when one firm acquires another with a lower P/ E, its EPS rises-called "bootstrapping." If the market sets the price of a stock at a constant multiple of EPS, then a firm can increase its stock price by buying firms with lower P/Es than its own. "The basic idea behind EPS manipulation can be seen with an example. Suppose firm A has $1 million in earnings, 1 million shares outstanding, and a price of $20. Its EPS is $1.00. Firm B has $1 million in earnings, 1 million shares outstanding, and a price of $10. Firm A has a P/E of 20 and firm B has a P/E of 10. Firm A plays the game by buying firm B. Firm A exchanges one of its shares for two of firm B's, increasing its outstanding shares to 1.5 million. However, its total earnings rise to $2 million, for a new EPS of $1.33. If investors watch EPS, so the argument goes, they observe a 33% increase and bid up the price of firm A

[5] The cumulated residual rates of return from a market model weighted by the value of the firm (Malatesta, 1983).

[6]Tobins ratio is defined as ratio

[7]The dummy variable that equals the value of unity if the bidder's price-earnings ratio is greater than the target's price-earnings ratio (and equals zero otherwise); in the bootstrapping explanation, the variable should be positively related to bidder returns.

[8] France , Germany, Canada & Britain

[9] "The Economic Recovery Tax Act (ERTA) of 1981 encouraged asset sales not only to effect a step-up in depreciable asset basis but also to effect a change in depreciation schedule to one that was much more accelerated (the so-called accelerated cost recovery system or ACRS). The Installment Sales Revision Act, passed in October of 1980, also promoted asset sales by making installment sales a more effective way of reducing the present value of the tax costs to the seller of assets from capital gains and from the recapture of past depreciation as ordinary income. The Deficit Reduction Act (DRA) of 1984 withdrew some of the tax advantages of asset sales by extending modestly the depreciable lives of certain assets and by removing the opportunity to post- pone the recapture of past depreciation through the use of installment sales. The Tax Reform Act (TRA) of 1986 put nails in the tax-induced asset-sales-and-mergers coffin along all of the tax dimensions that are important in motivating mergers and acquisitions involving domestic buyers. The 1986 act introduced less generous depreciation schedules; it increased the capital gains tax rate at both the corporate and personal levels; it eliminated the so-called General Utilities doctrine, which had provided an ability to avoid a corporate-level capital gains tax on the difference between the market value and the adjusted tax basis of corporate assets sold or distributed in a planned corporate liquidation."

[10] For example, a desire to exploit economies of scale or scope in operations or a desire on the part of the incumbent management to engage in empire building or further managerial entrenchment.'

[11] "Grimm classifies hostile takeovers as those in which the target's board at least initially expressed opposition, if only to raise the price."

[12] The neoclassical theories assume that (1) managers maximize their shareholders' wealth, (2) consequently mergers increase shareholder wealth, and (3) the capital market is efficient

[13] All variations are the same as used in the study

[14] Q ratios are computed using the Lindenberg and Ross (1981) algorithm and the specific assumptions of Hall, et al (1990). Q is often used as a measure of growth opportunities. This interpretation of Q would suggest that single segment firms have better growth prospects than multiple segment firms during the period 1961-1970. Such an interpretation is not dramatically different from a corporate value interpretation. Firms with better growth prospects should also be more valuable. It is nevertheless useful to examine whether my results are robust to alternative valuation specifica- tions.

[15] The F.T.C. classified each acquisition according to the primary economic relationship between the acquiring and acquired companies.

The five categories, listed below, are described by the F.T.C. as mutually exclusive. 1. Horizontal. An acquisition is horizontal when the companies involved produce one or more of the same, or closely related, products in the same geographic market. 2. Vertical. An acquisition is vertical when the two companies involved had a potential buyer-seller relationship prior to the merger. 3. Product extension. An acquisition is considered to be product extension in type when the acquiring and acquired companies are functionally related in production and/or distribution but sell products that do not compete directly with one another. An example of a product extension merger would be a soap manufacturer acquiring a bleach manufacturer. 4. Market extension. An acquisition is considered to be market extension in type when the acquiring and acquired companies manufacture the same products, but sell them in different geographic markets. An example of a market extension merger would be a fluid milk processor in Washington acquiring a fluid milk processor in Chicago. 5. Unrelated. This category involves the consolidation of two essentially unrelated firms. An example would be a shipbuilding company buying an ice cream manufacturer. Statistical Report on Mergers and Acquisitions (1975: 131-132.)

[16] (Jensen, 1991; Paulus and Gay, 1987; Shleifer and Vishny, 1992)

[17] (Comment and Jarrell, 1992; Hoskisson and Johnson, 1992; Kaplan and Weisbach, 1992; Lichtenberg, 1992; Lang and Stulz, 1993; Liebeskind and Opler, 1993).

[18] falling stock prices and increasing borrowing cost

[19] (Barney,1986; Conner, 1991; Makadok, 1998; Wernerfelt,1984)


More from UK Essays

Doing your resits? We can help!