Corporate investment decisions and corporate control: evidence from going-private transactions.
34. R. Morck, A. Shleifer, and R.W. Vishny, "Characteristics of Hostile and Friendly Takeover Targets," in Corporate Takeovers: Causes and Consequences, A.J. Auerbach (ed.), University of Chicago Press, 1988.
35. R. Morck, A. Shleifer, and R.W. Vishny, "Management Ownership and Market Valuation: An Empirical Analysis," Journal of Financial Economics (January/March 1988), pp. 293-316. David J. Denis is an Assistant Professor of Finance at the R.B. Pamplin College of Business, Virginia Polytechnic Institute and State University, Blacksburg, Virginia.
During the 1980s, going-private transactions became an increasingly important mechanism in the market for corporate control. Lehn and Poulsen |24~ report that the value of these transactions rose from $496.8 million in 1980 to $9.4 billion in 1987. Moreover, the average equity value of firms going private increased from $33.1 million to $224.3 million over the same period. The magnitude of shareholder gains resulting from the transactions is similar to the gains generated by other control-changing devices such as tender offers, mergers, and voluntary liquidations. Shareholders who sell their shares to the buyout group typically receive premiums on the order of 30 to 40%.(1)
The source of the premiums earned by shareholders in going-private transactions has been the subject of much debate. Potential sources of value include tax benefits, wealth transfers from other claimholders, and reduced agency costs. Both Kaplan |18~ and Lehn and Poulsen |23~ present evidence supporting the tax advantages of leveraged buyouts, but argue that these advantages cannot account for the total value created in an LBO. Similarly, Marais, Schipper, and Smith |28~ find no evidence of significant wealth transfers from the holders of senior securities. Asquith and Wizman |4~ find evidence of significant bondholder losses in leveraged buyouts, but the magnitude of these losses is insufficient to explain much of the shareholder gains.
Recent empirical evidence suggests that the stockholder-manager conflict may explain a number of observed phenomena in the corporate control market.(2) One way in which the separation of ownership and control can impose costs on shareholders is through investment in projects which reduce shareholder wealth. Jensen |14~, |15~ and Stulz |42~ argue that going-private buyouts may be particularly useful in controlling managerial discretion over investment decisions because of the large amount of debt and increased ownership concentration typically associated with these transactions. This paper provides evidence on this issue by documenting the wealth effects of the investment decisions of a sample of 192 candidates of going-private transactions over a five-year period prior to each transaction.
Previous evidence on the relationship between corporate control transactions and prior investment decisions has been inconclusive. Mitchell and Lehn |33~ examine a specific type of investment, i.e., acquisitions, and find that firms making value-reducing acquisitions are more likely to become targets of acquisitions themselves. However, Loderer and Martin |26~, using a larger sample of acquisition investments, find little support for the hypothesis that poor acquisitions increase the likelihood of being taken over.(3) This study provides further evidence on this issue, but differs in that it focuses on targets of going-private transactions and analyzes all announced investment decisions, both acquisitions and nonacquisitions.
The sample firms display significantly negative median abnormal returns of -0.33% (mean = 0.10%) during the two-day period surrounding the announcement of new investment decisions. Moreover, firms with low managerial shareholdings (5%) display the lowest median abnormal returns. These two findings are consistent with agency problems leading to poor investment decisions by going-private candidates. However, the relatively small magnitude of the abnormal returns and the fact that 42% of the sample firms have no investment decision announcements in the Wall Street Journal during the pre-buyout period suggest that poor investment strategies may be, at best, only a partial explanation for going private.
The empirical analysis also separates the sample into contested and uncontested transactions. Lehn and Poulsen |24~ note that one reason for the increased going-private activity in the 1980s may have been the threat of hostile takeovers. Absent a hostile takeover threat, there is little reason to expect managers to take their firm private for the purpose of reducing agency costs. However, if the firm is a takeover target, managers may propose a going-private transaction as a means of preempting a hostile takeover. Since hostile control changes are more likely to be disciplinary and motivated by target management inefficiency than friendly control changes (Morck, Shleifer, and Vishny |34~ and |36~), targets of contested transactions may be more prone to be characterized by poor investment choices than uncontested targets.(4)
The evidence presented here, using all types of investment decisions, is consistent with the hypothesis that contested going-private transactions are indeed more likely to be disciplinary. Targets of contested transactions exhibit negative median two-day abnormal returns of -0.55% (mean = -0.55%) when they announce new investment decisions in the pre-going-private period. These abnormal returns are significantly different from zero and from the abnormal returns surrounding the announcement of investment decisions by a control set of firms. Moreover, the likelihood of becoming a target of a contested going-private transaction is significantly negatively related to the wealth effects of a firm's investment decisions. In contrast, announcements of new investments by targets of uncontested transactions are met with a median abnormal change in share price of -0.07% (mean = 0.83%) which is not significantly different from zero or from that of a control sample. The difference in median abnormal returns between the contested and uncontested samples is significant at the 0.01 level. Further examination reveals that the difference in abnormal returns is due primarily to the tendencies of contested candidates to make value-reducing acquisitions and uncontested candidates to make acquisitions which increase shareholder wealth, similar to Mitchell and Lehn |33~. These findings suggest that there are different motivations for contested and uncontested going-private transactions.
The paper is organized as follows. Section I describes how agency problems can lead to suboptimal investment decisions. Section II details the sample selection process and the characteristics of the sample firms. Section III documents the wealth effects of investment decisions in the five-year period preceding the sample transactions and relates these wealth effects to the level of managerial ownership of shares and to the likelihood of becoming a going-private candidate. Section IV concludes.
I. Investment Decisions and Changes in Corporate Control
It is often argued that the separation of ownership and control can lead to large agency costs in the modern corporation. One such agency cost is the investment in projects that benefit managers at the expense of shareholders. However, literature dating back at least to Manne |27~ recognizes that when managers deviate from value-maximizing decisions, there are greater incentives for outside parties to launch an acquisition attempt for the firm.
Jensen |14~, |15~ argues that managers may have an incentive to expand the scope of the assets under their control beyond the point where shareholder wealth would be maximized. Conflicts of interest between managers and shareholders are likely to arise whenever the firm generates a substantial amount of cash flow in excess of that required to fund all positive net present value projects. Shareholders have a preference for the disbursement of this free cash flow, while managers may prefer to make investments even if they earn less than the firm's marginal cost of capital.
Agency problems associated with free cash flow are not the only reasons managers may make investment decisions that maximize their personal welfare at the expense of shareholders. For example, risk-averse managers may prefer a risk-reducing diversification strategy even if such a strategy has a negative net present value. Similarly, Shleifer and Vishny |39~ argue that managers have the incentive to invest in assets that are specific to their own skills and knowledge even if these investments are not ex-ante value-maximizing. Finally, managers may invest in unprofitable projects simply because they lack the skills required to evaluate the profitability of various projects.
For whatever reason, if a firm persists in accepting negative net present value projects, a transaction that constrains future discretionary investments or improves the incentive structure within the firm can increase firm value. Jensen |14~, |15~ and Stulz |42~ argue that going-private transactions may be particularly effective in this endeavor because of the increased debt and managerial ownership stakes typically associated with these transactions. Therefore, a testable proposition addressed in this paper is that targets of going-private transactions systematically misallocated resources by accepting negative net present value projects in the years prior to the going-private transaction.
A problem with the above hypothesis is that it is not clear what incentive managers have to take the firm private. If these firms are misallocating resources, why would managers suddenly pay out a premium and reduce their discretion over investment policy? One possibility, suggested by Lehn and Poulsen |24~, is that these firms are likely targets of hostile takeover attempts and that the going-private transaction effectively preempts the hostile bid. Thus, because of prior poor investments, firms may become good targets, as in Mitchell and Lehn |33~, but are not taken over if they take some action that preempts a hostile takeover. In this way, a going-private transaction may resemble a defensive leveraged recapitalization (Denis and Denis |11~) or a defensive share repurchase (Dann and DeAngelo |6~, and Denis |10~). In all three cases, Jensen |14~, |15~ and Stulz |42~ argue that the action may benefit shareholders because of the reductions in managerial discretion over future investments.
The above discussion yields a second testable hypothesis. Those going-private transactions associated with external takeover pressures should be more likely to be characterized by poor investment decisions prior to going private than those transactions not associated with takeover threats.
II. Data Description
A. Sample Characteristics
The sample is derived from Lehn and Poulsen's |24~ comprehensive sample of 278 going-private transactions between 1980 and 1987.(5) Their sample includes all successful going-private transactions announced in the Wall Street Journal during this time period, where a going-private transaction is defined as one that converts "a free-standing, publicly traded corporation into a privately held corporation." The sample used here is restricted to those firms found on the daily returns tape of the Center for Research in Security Prices (CRSP). This leaves a final sample of 192 going-private transactions over the sample time period. For each going-private candidate, a control firm is defined as that firm listed on the COMPUSTAT tape that has the same four-digit SIC code, is nearest in market value of equity to the sample firm as of the year ending just prior to the transaction, has available CRSP data, and is not taken over by the time of the going-private proposal.(6)
On the basis of Wall Street Journal reports, each sample transaction is classified as contested or uncontested. A transaction is classified as contested if (i) it is proposed by some outside party and publicly objected to by the incumbent management team (13 cases); (ii) it is proposed by the incumbent management team in response to an alternative offer made by an outside party (25 cases); (iii) it is proposed by the incumbent management team in response to published rumors of takeover activity (35 cases);(7) or (iv) a competing bid for the firm occurs after the proposed going-private transaction (12 cases). These requirements attempt to isolate those cases where it is more likely that the sample transaction is in response to external takeover pressures. Uncontested transactions are those for which there is no indication in the financial press of other potential suitors or of opposition to the going-private proposal. This classification scheme results in 85 contested and 107 uncontested transactions.(8)
Panel A of Exhibit 1 lists the time profile of the sample. Over the sample period of 1980 to 1987, the percentage of transactions classified as contested rises from 21% in 1980 to nearly 70% in 1986-1987. As Lehn and Poulsen |24~ note, this finding supports the argument that one reason for increased going-private activity in the 1980s is the threat of hostile takeovers. Using the standard market model procedure, cumulative abnormal returns (CARs) associated with the sample transactions were computed. These CARs are similar to those in previous studies. For the three days centered on the first announcement of the going-private proposal, (i.e., days |-1, 1~), cumulative abnormal returns average 12.0%. Over a wider event window, days |-40, 40~, cumulative abnormal returns average 22.3%. The CARs of the contested and uncontested transactions are statistically indistinguishable over both the |-1, 1~ and |-40, 40~ event windows.(9)
Panel B of Exhibit 1 indicates that there are substantial differences between contested and uncontested targets. Contested targets are significantly larger than uncontested candidates. The pre-transaction value of a contested target's equity averages $367.3 million, while that of uncontested targets averages only $177.6 million. Similarly, contested targets have significantly greater total assets and sales for the year ending just prior to the going-private proposal. Contested targets also have higher debt ratios than uncontested targets. Differences in average total assets, sales, and debt ratios between the contested and uncontested targets are all significant at the 0.05 level. Not surprisingly, given the matching procedure based on firm size, there are no significant differences between the characteristics of the sample and control firms.
Panel C documents summary measures of variables typically used as proxies for the likelihood of agency problems. Managerial ownership is obtained from Spectrum 6 and is defined as the beneficial ownership of all officers and directors as of the year ending just prior to the sample transaction. There are large differences in the ownership structures of contested and uncontested candidates. Managerial ownership averages 12.1% (median = 6.9%) for the contested targets and 34.2% (median = 32.8%) for the uncontested candidates. This difference is significant at the 0.01 level. While the differences in managerial ownership are partially driven by the well-known negative correlation between firm size and managerial ownership, the sample firms appear to have ownership structures that differ from their control firms. The managerial ownership of contested targets is significantly less (at the 0.05 level) than that of its control firms, while the ownership of uncontested targets is significantly higher (at the 0.10 TABULAR DATA OMITTED level) than that of its control firms.(10) Given their lower level of managerial ownership, one would expect greater agency problems in the sample of contested candidates. However, the findings are also consistent with the hypothesis that ownership structure determines the mood of the acquisition attempt. As managerial ownership increases, the likelihood of a successful hostile offer may be diminished. Several authors have recently suggested Tobin's q as an alternative measure of agency costs (i.e., Lang, Stulz, and Walkling |21~ and |22~, and Servaes |38~). Lacking complete data on replacement costs, q is estimated as the sum of the market value of equity and the book values of long-term debt, preferred stock, and current liabilities, all divided by the book value of total assets. These data are obtained from COMPUSTAT and are available for 181 of the 192 sample firms. To control for potential biases in this measurement of q, the estimated q is compared to the q of the control firms.(11)
The results in Panel C suggest that contested targets have marginally lower average q ratios than uncontested targets. However, the median q is nearly identical in the two samples. It is noteworthy that the sample firms, as a whole, significantly underperform their control firms when performance is measured by Tobin's q. This finding is largely driven by the sample of contested candidates. Both mean and median q ratios are significantly lower (at the 0.05 level) for the contested targets than their control firms. In contrast, the performance of the uncontested targets does not differ significantly from that of their control firms.(12) This evidence supports the view that managers of contested going-private candidates allocate the resources of their firms in a suboptimal fashion, thereby reducing the value of the firm relative to that of its industry peer. Alternatively, the evidence is consistent with systematic undervaluation of contested targets.
Finally, Panel C examines a specific type of agency problem, the free cash flow hypothesis proposed by Jensen |14~, |15~. Similar to Lehn and Poulsen |24~, undistributed cash flow is measured as operating income less taxes, interest payments, preferred stock dividends, and common stock dividends. Following Lang, Stulz and Walkling |22~, undistributed cash flow is standardized by the book value of total assets. The results do not indicate any significant differences in undistributed cash flow between the sample and the control firms. Moreover, the undistributed cash flow is actually lower for contested transactions than it is for uncontested transactions, although the difference is statistically insignificant.
B. Announcements of Investment Decisions
Announcements of investment decisions by the sample and control firms are identified by examining the Wall Street Journal Index for the five-year period ending with the going-private proposal. These decisions are defined as any expenditure of corporate resources for a purpose other than a disbursement to the claimholders of the firm. An announcement is not included if there are any other corporate announcements for that firm on either the day of or the day prior to the announcement of the investment decision.
On the basis of descriptions in the Wall Street Journal, each announcement is placed in one of nine separate categories. Panel A of Exhibit 2 classifies the sample announcements by type of investment decision. There are a total of 333 announcements made by the sample firms. Of these, 175 are made by the 85 contested candidates and 158 are announced by the 107 uncontested candidates. The control firms make 253 announcements, 126 by the contested control firms and 127 by the uncontested control firms. Acquisitions of other firms account for the greatest number of announcements in each group, perhaps because they are likely to be the transactions of greatest size and, therefore, more likely to be reported. In fact, acquisitions of firms account for 34% of the sample investment announcements. There is some overlap between the acquisition announcements studied here and those examined in Mitchell and Lehn |33~. Of the 250 acquisitions announced by the sample firms, 33 can also be found in Mitchell and Lehn.
The relative frequency of each type of investment decision is generally greater within the contested group. While one possible explanation for this is that the firms in this group are characterized by greater cash flow which is being invested in new projects, a more likely explanation is that because the firms in the contested group are, on average, larger firms, the Wall Street Journal simply reports a greater fraction of their announcements. However, note also that the frequency of announcements is generally greater in the sample firms than in their control firms. This is especially true for the contested candidates.
Panel B reports the frequency of investment decisions announced by each of the sample firms. The data indicate TABULAR DATA OMITTED that a substantial number of firms have no reported announcements over the five-year period preceding the going-private proposal. This is the case for 80 out of the 192 (42%) sample firms and 106 (55%) of the control firms. The percentage of firms with zero announcements is greater within the sample of uncontested targets (50%) than within the contested sample (32%). These numbers suggest that poor investment decisions can be only a partial explanation for going-private transactions. Alternatively, there may be a number of firms whose investment decisions are simply not reported in the Wall Street Journal.
II. The Wealth Effects of Investment Decisions
This section documents the average stock price reaction to and total wealth effects of announcements of new investments made by the sample firms. If poor investment decisions are a motivating factor for going-private transactions, there should be negative market reactions associated with the announcements of these firms. By the same token, if candidates of contested transactions are characterized by greater agency problems, market reactions to new investments should be more negative for the subsample of contested candidates. By way of comparison, the wealth effects of investment decisions made by the control firms are also documented. If the correction of poor investment strategies is a source of value in a going-private transaction, the sample firms should display more negative announcement effects than the control firms.
A. Summary Measures
Panel A of Exhibit 3 reports abnormal returns (ARs) for the two-day period including the day of and the day before the announcement of investment decisions in the Wall Street Journal.(13) Abnormal returns are computed using TABULAR DATA OMITTED standard event study methods where market model parameters are estimated over the 250-day period beginning 290 days before each announcement.(14) These abnormal returns will underestimate the magnitude of the wealth effects associated with new investments to the extent that the market anticipates investment announcements from the sample firms. For example, a firm with a history of poor investments will likely have some expectation of further poor investments built into its stock price. The announcement period abnormal return will capture only the unanticipated component of the full wealth effect.
For the full sample of 333 announcements, abnormal returns average an insignificant 0.10% (t = 0.59). However, this mean is apparently driven by outlier observations. The median abnormal return is -0.33% and 58.3% of the abnormal returns are negative. The median is significantly different from zero at the 0.07 level, while the percentage of negative observations is significantly different from 50% at the 0.01 level. In contrast, the 253 announcements made by the control firms result in mean and median abnormal returns of 0.45% and -0.03%, respectively. While the mean is significantly different from zero at the 0.09 level, neither the mean nor median abnormal return of the control firms is significantly different from that of the sample firms. These results offer little evidence that firms going private are characterized by poor investment decisions in the pre-transaction period. If the Wall Street Journal reports most significant investment decision announcements, this suggests that poor investments may just be one of potentially many motivations for the sample transactions.(15)
The results in Exhibit 3 reveal a difference in the market's reaction to announcements of new investment decisions across the contested and uncontested firms. Median abnormal returns are -0.55% (mean = -0.55%) and 64.6% of the abnormal returns are negative in the contested sample. Both the mean and median are significantly different from zero and from the ARs of the contested control sample (mean = 0.61%, median = 0.11%) at the 0.01 level. In contrast, median abnormal returns are -0.07% (mean = 0.83%) and 51.4% of the observations are negative in the uncontested sample.(16) These ARs are not significantly different from those of the uncontested control firms. The difference in medians across the contested and uncontested groups is significant at the 0.01 level. These findings suggest that there may be different motivations underlying contested and uncontested going-private transactions. The data are consistent with the hypothesis that one source of value in a contested transaction is the correction of poor investment strategies undertaken by the target firm.
Panel B of Exhibit 3 provides a measure of the wealth effects of investment decisions by summing the two-day abnormal returns associated with each firm's investments. Thus, if a given firm announced three investment decisions, each with an AR of -1%, the total wealth effect for that firm would be -3%. These total wealth effects are then averaged across all firms announcing investment decisions. The results in Panel B are largely consistent with those presented in Panel A. The overall sample provides weak evidence of negative wealth effects associated with investment decisions of going-private candidates. Neither the mean nor median total wealth effect is significantly different from zero. However, the percentage of firms with negative wealth effects, 60.7%, is significantly different from 50% at the 0.05 level.
Again, there appear to be differences between contested and uncontested transactions. The contested candidates display negative total wealth effects that are both significantly different from zero and from the total wealth effects of their control firms. Moreover, 68.4% of the contested targets exhibit negative total wealth effects associated with announced investment decisions. In contrast, uncontested targets exhibit total wealth effects which are generally not significantly different from zero or from those of the control firms.
B. Abnormal Returns By Type of Investment Decision
Exhibit 4 reports two-day announcement period abnormal returns when the sample is split into acquisition and nonacquisition investments.(17) Taken as a whole, acquisitions of all types announced by the sample firms display a median abnormal return of -0.24% (mean = 0.20%), which is not significantly different from zero or from that of the control firms. This finding indicates that, unlike the takeover targets in Mitchell and Lehn |33~, going-private candidates do not systematically engage in value-reducing acquisitions in the years leading up to the going-private proposal.
Interestingly, nonacquisition investments of the sample firms display median abnormal returns of -0.65% (mean = -0.19%). This median is significantly different from zero at the 0.05 level and significantly different from that of the control firms at the 0.04 level. Furthermore, the percentage of negative abnormal returns, 60.2%, is significantly different from 50% at the 0.05 level. These findings for going-private candidates contrast with those of McConnell and Muscarella |30~, who find that announcements of increases in capital expenditures in a random sample of firms are typically associated with share price increases.
Exhibit 4 also documents differences in the announcement effects of contested and uncontested candidates. The evidence indicates that a major source of the average negative wealth effects in the contested group is acquisitions made by the target firms prior to the going-private proposals. Acquisitions of all types result in negative median abnormal returns of -0.56% (mean = -0.74%) for contested candidates. These abnormal returns are significantly different from zero and from those surrounding the acquisition announcements of the control firms. In contrast, acquisitions result in a median abnormal share price increase of 0.21% (mean = 1.19%) for uncontested candidates. The difference between the contested and uncontested medians is significant at the 0.01 level. These results support the finding in Mitchell and Lehn |33~ that firms are more likely to become hostile targets of acquisition attempts if they have made value-reducing acquisitions themselves.(18)
Exhibit 4. Mean and Median Abnormal Returns for Acquisitions and Nonacquisitions for the Sample of 333 Investment Decision Announcements Made By the 112 Going-Private Candidates and 253 Announcements Made By the 86 Control Firms Announcing New Investments Over the Five-Year Period Immediately Preceding the Going-Private Proposal, 1980-1987(a,b) (Sample Size is Listed in Brackets) Panel A. All Transactions P-Value Sample Control (Diff.) Acquisition 0.20% 0.43% 0.628 -0.24% -0.25% 0.593 |N = 250~ |N = 193~ Nonacquisition -0.19% 0.53% 0.156 -0.65%(*) 0.43% 0.038 |N = 83~ |N = 60~ Panel B. Contested Transactions P-Value Sample Control (Diff.) Acquisition -0.74%(**) 0.68% 0.007 -0.56%(**) 0.00% 0.014 |N = 128~ |N = 96~ Nonacquisition -0.03% 0.37% 0.581 -0.42% 0.50% 0.212 |N = 47~ |N = 30~ Panel C. Uncontested Transactions P-Value Sample Control (Diff.) Acquisition 1.19%(*) 0.17% 0.193 0.21%(*) -0.67% 0.089 |N = 122~ |N = 97~ Nonacquisition -0.39% 0.69% 0.126 -0.88%(*) 0.39% 0.103 |N = 36~ |N = 30~ Notes: a Abnormal returns are computed using standard event study methods where market model parameters are estimated over the 250-day period beginning 290 days prior to the event date. b Acquisitions include acquisitions of firms, divisions, assets, and stakes in other firms. Nonacquisitions include joint ventures, product line extensions, plant expansions, capital expenditures, and promotion/advertising expenditures. ** and * denote significance at the 0.01 and 0.05 levels respectively.
It is puzzling that while there is no evidence of value-reducing acquisitions on the part of uncontested targets, there is some weak evidence that the internal investments of uncontested candidates result in share price reductions.(19) The median abnormal return associated with the 36 nonacquisition investments of uncontested candidates is -0.88% (mean = -0.39%), which is significantly different from zero and from that of the uncontested control firms at the 0.11 level. Abnormal returns associated with internal investments are also negative for contested candidates (median = -0.42%, mean = -0.03%). These abnormal returns are not significantly different from those of uncontested candidates, but are also not significantly different from zero.
C. Ownership Structure and the Wealth Effects of Investment Decisions
In order to further assess the role of agency costs in explaining the market's reaction to announcements of investment decisions, Exhibit 5 presents summary measures of abnormal returns when the sample is separated on the basis of managerial ownership. For each announcement, the beneficial ownership of officers and directors for the period ending just prior to the announcement of the going-private transaction is obtained from Spectrum 6.(20)
There is currently no theoretical consensus on whether there should necessarily be a monotonic relation between managerial ownership and the reduction of agency problems. In the Jensen and Meckling |16~ model, as managerial ownership increases, agency costs decrease. However, as Stulz's |41~ model demonstrates, increased managerial ownership can reduce the effectiveness of the corporate TABULAR DATA OMITTED control market as a means of monitoring managerial behavior. Hence, as managerial ownership increases, there is a tradeoff between improved managerial incentives for value maximization and the decreased probability of takeover.
Morck, Shleifer, and Vishny |35~ present evidence that is consistent with Stulz's model by documenting a nonmonotonic relation between firm performance, as measured by Tobin's q, and board ownership. They find that, for ownership levels below 5% and above 25%, firm performance is increased through an increase in board ownership. However, for ownership levels between 5% and 25%, firm performance declines as ownership increases. Wruck |44~ confirms these findings in a study of ownership changes induced by private equity financings. McConnell and Servaes |31~ find a curvilinear relationship between firm value and inside ownership with firm value maximized when managers own 30 to 40% of the firm's shares. One interpretation of their results is that agency problems are most severe in those firms with very low managerial ownership and in those firms with very high ownership, where managers are insulated from the corporate control market.
To examine the impact of ownership structure on announcement period abnormal returns in more detail, Exhibit 5 compares abnormal returns for various subsamples corresponding to the ownership categories used in Morck, Shleifer and Vishny |35~ and Wruck |44~. Each announcement is placed into either a low ownership (0% to 5%), medium ownership (5% to 25%) or high ownership (25% to 100%) group.(21) Results are presented for the full sample and the samples of contested and uncontested going-private transactions.
The results provide some evidence on the relation between managerial ownership and the likelihood of poor investment decisions. Both the full sample and the sample of contested transactions display median abnormal returns which are significantly negative in the low ownership category. The fraction of negative abnormal returns (0.64 and 0.65, respectively) is significantly different from 0.50 for both of these samples. For uncontested candidates, median abnormal returns are lowest in the low ownership category, but are not significantly different from zero.
All three subsamples display the highest median abnormal returns in the medium (5% to 25%) ownership category. None are significantly different from zero. Median abnormal returns are lower in the high (25%) ownership category, but again, are not significantly different from zero.
Overall, these findings provide some weak support for the agency cost rationale of poor investments. Firms making value-reducing investments tend to be those with relatively low managerial ownership of shares. This is consistent with the findings in Lewellen, Loderer, and Rosenfeld |25~ on the returns to acquiring firms in mergers and suggests that managers are less likely to make poor investments when they bear a larger fraction of the wealth consequences of these investments. On average, those firms with managerial ownership greater than 5% do not display significant abnormal returns around the announcement of investment decisions.
D. Wealth Effects and the Likelihood of Going Private
As a final test of whether poor investment decisions are a motivation for going-private transactions, Exhibit 6 presents the results of logit models where the dependent variable is the transformed probability of being a going-private candidate and the independent variable is the sum of the two-day abnormal returns associated with each firm's announced investments.(22) If poor investments are a motivation for the sample buyouts, there should be a negative relation between the likelihood of going private and the sum of the wealth effects of the firm's prior investment decisions.
Panel A of Exhibit 6 presents the estimated coefficients for each logit equation. For the full sample, the probability of being a going-private candidate is not significantly related to the wealth effects of investment decisions (coefficient = -0.016, chi-square = 0.85) in Model (1). This overall result masks substantial differences between the targets of contested and uncontested transactions. In Model (2), the logit equation is reestimated with the addition of an interaction term of the product of a binary variable denoting a contested transaction and the total wealth effects. The significance of the interaction term (coefficient = -0.112, chi-square = 6.45) indicates a significant difference in the relation between the likelihood of becoming the target of a going-private transaction and the wealth effects of investment decisions for contested and uncontested transactions. For contested transactions, the coefficient of -0.091 (0.021 - 0.112) indicates a significant negative relation (chi-square = 5.86, p-value = 0.01). In contrast, there is an insignificant positive relation between the likelihood of becoming an uncontested target and the wealth effects of investment decisions (coefficient = 0.021, chi-square = 0.81).
Exhibit 6. Logit Estimates Relating the Probability That a Firm Became a Candidate for a Going-Private Transaction Over the Period 1980-1987 to the Total Wealth Effect of the Firm's Announced Investment Decisions in the Preceding Five Years(a) Panel A. Coefficient Estimates (Chi-Square in Parentheses) Model Model Intercept Wealth C x Wealth(b) Chi-Square (1) 0.411 -0.016 0.86 (7.14)(**) (0.85) (2) 0.388 0.021 -0.112 8.83(**) (6.11)(**) (0.81) (6.45)(**) Panel B. Implied Probability of Going Private Wealth Full Contested Uncontested Effects Sample Buyouts Buyouts 10% 0.36 0.56 0.20 -5% 0.33 0.44 0.25 -3% 0.31 0.38 0.29 0% 0.29 0.25 0.31 3% 0.27 0.06 0.34 5% 0.25 0.00 0.35 10% 0.20 0.00 0.40 Notes: a Total wealth effect is defined as the sum of the two-day abnormal returns associated with a given firm's announced investment decisions. b C is a binary variable taking on the value of one if the going-private transaction is contested and zero if uncontested. ***, **, and * denote significance at the 0.01, 0.05, and 0.10 levels, respectively.
Panel B shows how variations in the total wealth effects of investment decisions affect the probability of becoming a going-private target. The results show that the probability of becoming a contested target is very sensitive to the wealth effects of the firm's investment decisions. In contrast, the probability of becoming the target of an uncontested transaction is relatively insensitive to the wealth effects of investment decisions.
Cross-sectional regressions relating the total wealth effects of investment decisions to the premiums paid in the sample buyouts were also estimated. However, no significant relations were found. This is perhaps not too
surprising for several reasons. First, for reasons discussed earlier (see footnote 9), even if agency problems are greater in contested firms, it is not clear that the going-private premiums will be higher for these firms. Second, as mentioned previously, the correction of poor investments is just one of many possible sources of value in going-private transactions. These other sources will add considerable noise to the regression. Along similar lines, it may be that constraints on future discretionary investments are more important than the correction of previous investments. If so, the relation between premiums and wealth effects of prior investments will be weakened. Finally, to the extent that the Wall Street Journal does not report all announced investment decisions, the correlation between the premium and total wealth effects will again be weakened.
IV. Summary and Conclusions
This paper examines the investment decisions of going-private candidates during the five-year period preceding each going-private proposal. Overall, the results provide evidence that going-private candidates invested in projects which were viewed by the market as negative net present value opportunities and were, thus, met with a median negative stock price reaction upon announcement. This finding suggests that one motivation for going-private transactions is the correction of poor investment strategies undertaken by the going-private candidates. However, this is probably only a partial explanation for the transactions since nearly half of the sample firms have no reported announcements of new investments in the five years preceding the going-private proposal.
The results also provide evidence that there may be different motivations underlying contested and uncontested going-private transactions. There are significant differences in the market's reaction to investment decisions of contested and uncontested candidates. These differences are due primarily to announcements of acquisitions made by contested candidates which reduce shareholder wealth. In contrast, acquisitions announced by uncontested candidates typically increase shareholder wealth. These findings suggest that a reduction of agency problems observed in the form of poor investment decisions is a more likely motivation for contested than for uncontested going-private transactions. This interpretation is supported by the fact that contested candidates are characterized by significantly lower managerial shareholdings than uncontested candidates and that poor investments are concentrated primarily in those firms with managerial ownership less than five percent. Further support is provided by logit models which document a significant inverse relation between the wealth effects of investment decisions and the likelihood of becoming a target of a contested going-private transaction, but no relation between these wealth effects and the likelihood of becoming an uncontested candidate.
This study's findings complement those found in Kaplan |19~ and Smith |40~, in that all three studies present evidence consistent with the hypothesis that going-private transactions are motivated by improved incentives and performance. Kaplan |19~ and Smith |40~ document significant improvements in operating efficiencies following management buyouts, while this study documents the potential for increasing value through improved investment policy. It should be noted that the lack of evidence of poor investment decisions by targets of uncontested transactions does not necessarily imply that reduced agency costs are unimportant in explaining the wealth effects of these transactions. When Smith |40~ classifies her sample management buyouts as defensive or nondefensive, she does not find any differences in post-buyout operating returns. Thus, there is no reason to believe that the uncontested targets studied here do not experience operating efficiencies similar to those documented in Kaplan |19~ and Smith |40~.
The findings for contested going-private transactions also complement Denis and Denis' |11~ study of leveraged recapitalizations. They find that firms proposing leveraged recaps are also characterized by unprofitable investments in the years preceding the proposed recapitalizations. Similar to the contested transactions studied here, Denis and Denis also find that almost all of their sample firms are takeover targets. Thus, it appears that poor investment decisions increase the probability of becoming a takeover target, as in Mitchell and Lehn |33~, but the takeover can be preempted through an alternative transaction, such as a going-private transaction or a leveraged recapitalization. As Jensen |14~, |15~ and Stulz |42~ argue, one benefit of these transactions may be the leverage-induced constraints imposed on future discretionary investments. An interesting question that remains is why certain firms choose to go private while others remain public through a leveraged recapitalization.(23)
1For the shareholder gains resulting from tender offers, mergers, and voluntary liquidations, see Bradley, Desai, and Kim |5~, Asquith |2~, and Kim and Schatzberg |20~, respectively. See DeAngelo, DeAngelo, and Rice |8~ and Lehn and Poulsen |23~ for the wealth effects of going-private transactions.
2See, for example, Denis and Denis |11~, Kaplan |19~, Lang, Stulz, and Walkling |21~, Lehn and Poulsen |24~, Mitchell and Lehn |33~, Morck, Shleifer, and Vishny |37~, and Smith |40~.
3Note that it is not possible to directly compare the Mitchell/Lehn and Loderer/Martin results. Mitchell and Lehn examine the likelihood of a firm becoming a takeover target, while Loderer and Martin examine the likelihood of a firm actually being acquired. Moreover, Mitchell and Lehn separate hostile and friendly acquisition attempts, while Loderer and Martin make no such distinction.
4Martin and McConnell |29~ find some evidence that is inconsistent with the hypothesis that hostile tender offers are more likely to be disciplinary. They find that the rate of top management turnover following successful tender offers is approximately the same for hostile and friendly offers in their sample.
5The analysis in Lehn and Poulsen |24~ is restricted to the 263 firms in their sample with available COMPUSTAT data. This restriction is not initially imposed here.
6This selection process does not eliminate control firms that subsequently received takeover bids. However, an examination of the Wall Street Journal revealed that only three of the control firms were acquired within two years after the corresponding sample transaction.
7This takeover activity can take many different forms and is not restricted to formal takeover offers. Examples include 13-d filings, discussions among management teams, and share accumulations.
8It is not immediately clear that the 12 transactions meeting requirement (iv) are in response to external takeover pressures. A reading of the Wall Street Journal suggests that this is the most likely scenario in the majority of cases. However, imposing requirement (iv) also labels as contested all transactions where managers proposed to go private with no prior knowledge of any control-related activity. To check the sensitivity of the paper's results to the contested/uncontested classification scheme, the results were replicated classifying the 12 transactions falling under requirement (iv) as uncontested. There were no statistically significant differences in the results.
9Note that the hypotheses in Section I make no prediction about whether the CARs should be different for contested and uncontested transactions. On the one hand, if agency problems are greater in contested firms or if the going-private transaction can be viewed as the winning bid in a multiple bid control contest, one might expect larger CARs for these transactions. However, other possible gains from going private may be larger in the uncontested firms (see footnote 15). Moreover, if the managers of the contested firms can take their firms private for less than what an outsider may have been willing to bid, the CARs for the contested transactions may be lower due to this entrenchment effect. Thus, it is not clear how the CARs of contested transactions would be expected to compare with those of uncontested transactions.
10Mikkelson and Partch |32~ provide a useful benchmark by which to compare the ownership of the sample firms with a random sample of firms of similar size. In their random sample, firms of comparable size to the contested targets have managerial ownership ranging from 13.5% to 18.5%, while firms comparable in size to the uncontested candidates have managerial ownership on the order of 18% to 25%. Thus, managerial ownership appears to be slightly below average for the contested candidates and slightly above normal for the uncontested candidates.
11There are two potential sources of bias in the computation of q. First, to the extent that changing prices cause the book value of assets to differ from their replacement cost, there will be measurement error in the estimate of q. Second, q will be a function of the overall level of stock prices. Hence, those transactions taking place in the latter part of the sampling period will be more likely to display higher q ratios because of the extended bull market of the 1980s. This bias will be particularly important in any comparison of contested and uncontested candidates since the fraction of contested transactions increases throughout the sampling period.
12Replacement cost data is available for 107 of the 192 sample firms from the NBER's Manufacturing Sector Master File. Using this data yields similar results to those presented in Exhibit 1. Both mean and median q ratios of the sample firms are significantly lower (at the 0.01 level) than those of the control firms for the full sample and the sample of contested candidates, but are insignificantly different for the uncontested candidates.
13To control for the possibility of information leakage prior to announcements, as well as further information being revealed about particular investments after the initial announcements, the tests conducted in the paper were replicated using varying windows surrounding the investment decision announcements. Specifically, the event window was extended as far as 40 days prior to and 40 days after the Wall Street Journal announcement. The results were qualitatively similar, although, in most cases, the statistical significance was lower for the wider event windows. Given the similarity of results, only those using the narrow two-day window are reported in the paper.
14Specifically, abnormal returns are averaged across securities to generate a portfolio abnormal return. T-statistics are then computed by dividing each portfolio by its standard deviation estimated from the market model estimation period. Significance of median abnormal returns and the percentage negative abnormal returns is computed using the Wilcoxon signed-ranks test and a normal approximation to the binomial distribution, respectively.
15Other possibilities, such as avoiding the costs of servicing shareholders, tax benefits, avoidance of potential takeover attempts, and improved managerial incentives, are discussed in DeAngelo and DeAngelo |7~. Note that these other possibilities and the correction of poor investment strategies are not mutually exclusive.
16The mean abnormal return in the uncontested sample is driven by one observation of 46%. This abnormal return is associated with the acquisition of a 40% stake in Don Sophisticates Inc., by Caressa Group Inc. Without this single observation, the mean abnormal return in the uncontested sample falls to 0.54%.
17Acquisitions include acquisitions of other firms, divisions, assets, and stakes in other firms. Nonacquisitions include joint ventures, product line extensions, plant expansions, capital expenditures, and promotion/advertising expenditures.
18Asquith, Bruner, and Mullins |3~ and Travlos |43~ show that acquisitions financed with new equity result in lower abnormal returns for the acquiring firm than those financed by cash or debt. Furthermore, Amihud, Lev, and Travlos |1~ show that firms with lower managerial shareholdings are more likely to finance acquisitions with equity. Since the contested buyout sample is characterized by lower levels of managerial ownership, contested candidates may display more equity-financed acquisitions. To check the impact of the form of financing, the analysis in Exhibit 4 was replicated after eliminating those transactions that were at least partially equity-financed. The results were qualitatively unaffected.
19Two possible explanations for this are that the nonacquisitions are either financed differently from the acquisitions or that the nonacquisitions are of larger relative size. Unfortunately, neither of these explanations can be tested since the nonacquisition announcements rarely contain any information about either the financing or the dollar magnitude of the investment.
20An alternative is to measure ownership as of the period ending just prior to each investment decision announcement. This information is available for only 273 of the 333 announcements. Using this definition of ownership does not alter the results in Exhibit 5.
21An alternative is to measure managers' dollar shareholdings relative to their total compensation or to their total wealth. However, as Mitchell and Lehn |33~ note, it is not clear that risk-averse managers with large percentage wealth stakes have a greater incentive to maximize shareholder wealth. Moreover, Jensen and Murphy |17~ argue that the fractional ownership stake is the best measure of the extent to which managers bear the wealth consequences of their actions.
22The logit models were also estimated, controlling for firm size with no qualitative differences in the results. This is not surprising given that the sample and control firms were matched on the basis of industry and firm size.
23See also Gupta and Rosenthal |12~ and Handa and Rhadhakrishnan |13~ for other differences between leveraged buyouts and leveraged recapitalizations.
1. Y. Amihud, B. Lev, and N.G. Travlos, "Corporate Control and the Choice of Investment Financing: The Case of Corporate Acquisitions," Journal of Finance (June 1990), pp. 603-616.
2. P. Asquith, "Merger Bids, Uncertainty, and Stockholder Returns," Journal of Financial Economics (April 1983), pp. 51-84.
3. P. Asquith, R.F. Bruner, and D.W. Mullins, "Merger Returns and the Form of Financing," Working Paper, Harvard Business School, 1986.
4. P. Asquith and T.A. Wizman, "Event Risk, Covenants, and Bondholder Returns in Leveraged Buyouts," Journal of Financial Economics (September 1990), pp. 195-214.
5. M. Bradley, A. Desai, and E.H. Kim, "Synergistic Gains From Corporate Acquisitions and Their Division Between the Stockholders of Target and Acquiring Firms," Journal of Financial Economics (May 1988), pp. 3-40.
6. L.Y. Dann and H. DeAngelo, "Corporate Financial Policy and Corporate Control: A Study of Defensive Adjustments in Asset and Ownership Structure," Journal of Financial Economics (January/March 1988), pp. 87-128.
7. H. DeAngelo and L. DeAngelo, "Management Buyouts of Publicly Traded Corporations," Financial Analyst's Journal (May-June 1987), pp. 38-49.
8. H. DeAngelo, L. DeAngelo, and E. Rice, "Going Private: Minority Freezeouts and Stockholder Wealth," Journal of Law and Economics (October 1984), pp. 367-402.
9. H. Demsetz and K. Lehn, "The Structure of Corporate Ownership: Causes and Consequences," Journal of Political Economy (December 1985), pp. 1155-1177.
10. D.J. Denis, "Defensive Changes in Corporate Payout Policy: Share Repurchases and Special Dividends," Journal of Finance (December 1990), pp. 1433-1456.
11. D.J. Denis and D.K. Denis, "Managerial Discretion, Organizational Structure, and Corporate Performance: A Study of Leveraged Recapitalizations," Journal of Accounting and Economics (forthcoming).
12. A. Gupta and L. Rosenthal, "Ownership Structure, Leverage and Firm Value: The Case of Leveraged Recapitalizations," Financial Management (Autumn 1991), pp. 69-83.
13. P. Handa, and A.R. Radhakrishnan, "An Empirical Investigation of Leveraged Recapitalizations With Cash Payout as Takeover Defense," Financial Management (Autumn 1991), pp. 58-68.
14. M.C. Jensen, "Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers," American Economic Review (May 1986), pp. 323-329.
15. M.C. Jensen, "The Takeover Controversy: Analysis and Evidence," Midland Corporate Finance Journal (Summer 1986), pp. 6-32.
16. M.C. Jensen, and W.H. Meckling, "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure," Journal of Financial Economics (October 1976), pp. 305-360.
17. M.C. Jensen and K.J. Murphy, "CEO Incentives -- It's Not How Much You Pay, But How," Journal of Applied Corporate Finance (Fall 1990), pp. 36-49.
8. S. Kaplan, "Management Buyouts: Evidence of Taxes as a Source of Value," Journal of Finance (July 1989), pp. 611-632.
19. S. Kaplan, "The Effects of Management Buyouts on Operating Performance and Value," Journal of Financial Economics (October 1989), pp. 217-254.
20. E.H. Kim and J.D. Schatzberg, "Voluntary Corporate Liquidations," Journal of Financial Economics (December 1987), pp. 311-328.
21. L.H.P. Lang, R.M. Stulz, and R.A. Walkling, "Managerial Performance, Tobin's q, and the Gains from Successful Tender Offers," Journal of Financial Economics (September 1989), pp. 137-154.
22. L.H.R. Lang, R.M. Stulz, and R.A. Walkling, "A Test of the Free Cash Flow Hypothesis: The Case of Bidder Returns," Journal of Financial Economics (June 1991), pp. 315-336.
23. K. Lehn and A. Poulsen, "Leveraged Buyouts: Wealth Created or Wealth Redistributed," in Public Policy Towards Corporate Takeovers, M. Weidenbaum and K. Chilton (eds.), New Brunswick, NJ, Transaction Publishers, 1988.
24. K. Lehn and A. Poulsen, "Free Cash Flow and Stockholder Gains in Going Private Transactions," Journal of Finance (July 1989), pp. 771-788.
25. W. Lewellen, C. Loderer, and A. Rosenfeld, "Merger Decisions and Executive Stock Ownership in Acquiring Firms," Journal of Accounting and Economics (April 1985), pp. 209-231.
26. C. Loderer and K. Martin, "Corporate Acquisitions by Listed Firms: The Experience of a Comprehensive Sample," Financial Management (Winter 1990), pp. 17-33.
27. H.G. Manne, "Mergers and the Market for Corporate Control," Journal of Political Economy (April 1965), pp. 110-120.
28. L. Marais, K. Schipper and A. Smith, "Wealth Effects of Going Private for Senior Securities," Journal of Financial Economics (June 1989), pp. 155-191.
29. K.J. Martin and J.J. McConnell, "Corporate Performance, Corporate Takeovers, and Management Turnover," Journal of Finance (June 1991), pp. 671-687.
30. J.J. McConnell and C.J. Muscarella, "Corporate Capital Expenditure Decisions and the Market Value of the Firm," Journal of Financial Economics (September 1985), pp. 399-422.
31. J.J. McConnell and H. Servaes, "Additional Evidence on Equity Ownership and Corporate Value," Journal of Financial Economics (September 1990), pp. 595-612.
32. W.H. Mikkelson and M.M. Partch, "Managers' Voting Rights and Corporate Control," Journal of Financial Economics (December 1989), pp. 263-290.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Special Issue: Corporate Control|
|Author:||Denis, David J.|
|Date:||Sep 22, 1992|
|Previous Article:||Postacquisition performance of acquiring firms.|
|Next Article:||A contingent claims analysis of trade credit.|