# Postacquisition performance of acquiring firms.

Claudio Loderer is a Professor at the Institut fur Finanzmanagement,
Universitat Bern, Bern, Switzerland. Kenneth Martin is an Assistant
Professor of Finance at the College of Business Administration,
University of Iowa, Iowa City, Iowa.

This paper analyzes the postacquisition stock-price performance of acquiring firms. Between 1966 and 1986, there were more than 10,000 domestic acquisitions in the U.S. of legally independent firms in which the acquirer traded on the New York or American Stock Exchange. An investment in each acquiring firm on the date the acquisition was completed, held for 500 trading days, would have yielded an equally weighted return of 21%. The return on a share in the market portfolio, by comparison, would have been 36%, more than 50% higher. The difference in investment performance is likely to reflect, at least in part, differences in systematic risk. But after risk is controlled for, acquiring firms are often found to underperform the market.(1) This raises questions about the profitability of corporate acquisitions and about the notion of market efficiency, since it seems that rational investors could benefit by shorting the stock of bidding firms after the acquisition has been completed. Most researchers appear to attribute negative postacquisition performance to estimation bias (see the surveys of Jensen and Ruback |11~ and Roll |18~). Not surprisingly, various performance measures are sensitive to the estimation technique (Mandelker |16~).

Franks, Harris, and Titman |8~ confirm the suspicion of estimation bias by showing that, with the appropriate benchmark, the negative postacquisition performance disappears. But, their paper is not the last word on the issue. Agrawal, Jaffe, and Mandelker |1~ reopen the Pandora's box by showing that the Franks, Harris, and Titman |8~ findings are specific to the 1975-1984 period and cannot be replicated for the years before and after. Agrawal, et al |1~ report significant postacquisition underperformance for mergers completed in the 1950s, 1960s, and 1980s, although not in the 1970s.(2)

Since the evidence on postacquisition performance appears to be sensitive not only to the estimation technique but also to the sample investigated, we reexamine the issue with a comprehensive sample of domestic acquisitions by firms traded on the New York Stock Exchange (NYSE) or the American Stock Exchange (Amex) during the years 1966-1986. We find abnormal performance in three years but not in five years following the acquisition. Negative performance in the second and third years after the acquisition is most prominent in the 1960s, and to a lesser extent in the 1970s, but not in the 1980s.

The first section of the paper discusses our data and investigative design. The second section presents our results, and the third section summarizes our conclusions.

I. Data and Test Design

Our sample includes domestic acquisitions by bidding firms listed on the Center for Research in Security Prices at the University of Chicago (CRSP) daily tapes starting at least 200 trading days before the acquisition is completed. This sample is compiled from three sources. The first source is a sample used in a companion paper (Loderer and Martin |14~) that includes mergers, acquisitions of privately held corporations, and comprehensive acquisitions of the assets of individual firms completed during the 1965-1984 period, as reported in various issues of Mergers and Acquisitions (M&A). The second source consists of tender offers from the same period in Martin and McConnell |17~, and the third source consists of all domestic acquisitions completed in the 1985-1986 period according to M&A. The three sources yield a sample of 11,020 acquisitions. For all |of the acquisitions~, we gather information on the date they are first mentioned in the Wall Street Journal Index (WSJI) (the announcement date, AD) and the date they are reported by M&A as having been completed (the effective date, ED).(3)

Using the full sample to gauge postacquisition performance could result in double-counting and lead to estimation bias if all acquisitions by any firm, no matter how close to each other, were treated as separate events. We might, for example, see an average negative postacquisition performance as a result of counting the occasional negative experience of frequent acquirers several times instead of just once. To limit the potential for double-counting, we restrict our attention to the following acquisitions. For each firm, we start in 1966 and search for the completion date of the first acquisition in excess of ten percent of acquiring firm size. We then skip 500 days and resume the search to identify the completion date of the next acquisition in excess of ten percent of firm size. The focus on acquisitions larger than ten percent is to avoid acquisitions too small to affect firm value. Smaller acquisitions are less likely to be the cause of any abnormal postacquisition performance. This search is conducted through the end of 1986. Since the search procedure involves the use of only contemporaneous information, no obvious selection bias is introduced into the analysis. These selection criteria reduce the sample from 11,020 to 1,298 acquisitions.(4)

The crux of the experiment is to test whether modified market model measures of negative abnormal performance become insignificant when risk is appropriately accounted for. Unfortunately, when abnormal returns are computed using market model parameters estimated with preacquisition data, there are several possible sources of downward estimation bias. One source can be illustrated assuming expected returns are set according to the capital asset pricing model (CAPM). Then, computing abnormal returns with a preacquisition estimate of the market model intercept assumes stationarity of the risk-free rate. But, since corporate acquisitions are not uniformly distributed in time,(5) this assumption could be inappropriate. One possible scenario is that during times of intense acquisition activity, the risk-free rate is driven up by firms that want to fund their acquisitions. When the activity subsides, the risk-free rate could decline. If so, in a cross-section of acquisitions, the risk-free rate could be higher before the completion date and lower afterward. Preacquisition estimates of the market model intercept could therefore be correspondingly too high, which could lead to downward bias in the computation of postacquisition abnormal returns.

Downward bias can also occur if the systematic risk of the acquiring firm is lower after the acquisition than before, which is not unlikely, since acquiring firms typically experience positive abnormal returns before an acquisition. This growth tends to reduce both financial leverage and systematic risk and could therefore lead to downward bias in the measurement of postacquisition performance.

Finally, downward bias can occur because the market model ignores firm size as a proxy for risk. Small firms tend to have higher market model intercepts than large firms. Since acquiring firms tend to grow before the acquisition, they will have smaller intercepts afterward, and using a preacquisition estimate of the intercept to compute abnormal returns once again introduces downward bias.(6)

To simultaneously take into account nonstationarity of the risk-free rate, changes in systematic risk, and changes in firm size, we adopt the following procedure for estimating abnormal returns. We first determine the size decile to which a particular acquiring firm belongs, according to the CRSP files. For each year, CRSP assigns each firm to a size decile using the market value of the firm's equity at the end of the preceding year as the basis for the assignment. We use these assignments to form ten size portfolios with equally weighted daily returns. Each acquiring firm is then matched with its appropriate size portfolio to determine the relevant size-based daily return. Abnormal performance is then measured as follows.

We assume that the market model can be written in terms of deviations from the risk-tree rate for the returns to firm i and to firm i's relevant size portfolio. Then, the return to firm i (|R.sub.i~) net of the return to its corresponding size portfolio (|R.sub.Si~) can be set equal to (time subscripts are omitted for simplicity):

|Mathematical Expression Omitted~

where |alpha.sub.i~ reflects both abnormal performance and possible size effects, whereas |alpha.sub.Si~ reflects only possible size effects (of course, under the CAPM, this term is zero). Both |beta.sub.i~ and |beta.sub.Si~ measure systematic risk, |R.sub.m~ is the value-weighted CRSP index, |R.sub.f~ is the yield-to-maturity on a three-month Treasury bill, ||upsilon~.sub.i~ and |xi.sub.i~ are i.i.d. error terms independent of each other, and |eta.sub.i~ = |(|upsion~ - |xi~).sub.i~.

Any abnormal performance of firm i will be reflected in the intercept term of this equation, ||gamma~.sub.0i~, since that term captures the systematic firm-specific performance net of any possible size effect. Consequently, to investigate the impact of an acquisition on firm performance, one can test for the sign and significance of the estimate of ||gamma~.sub.0i~. If negative postacquisition performance is a genuine phenomenon, ||gamma~.sub.0i~ should be negative and significant.

For more detail, we analyze the five years separately with a modified version of Equation (1):

|Mathematical Expression Omitted~

The variables D. are defined as follows:

|D.sub.0~ = 1 over the interval (ED + 1, ED + 250);

|D.sub.0~ = 0 otherwise;

|D.sub.1~ = 1 over the interval (ED + 251, ED + 500);

|D.sub.1~ = 0 otherwise;

|D.sub.2~ = 1 over the interval (ED + 501, ED + 750);

|D.sub.2~ = 0 otherwise;

|D.sub.3~ = 1 over the interval (ED + 751, ED + 1,000);

|D.sub.3~ = 0 otherwise:

|D.sub.4~ = 1 over the interval (ED + 1,001, ED + 1,250);

|D.sub.4~ = 0 otherwise.

II. The Evidence

A. Overall Results

We measure postacquisition abnormal performance by estimating the intercepts ||gamma~.sub.ji~, j = 1,...,4, of Equation (2) above. We estimate the regression with OLS for the acquiring firm in each of the 1,298 acquisition events in the sample over the period (ED + 1, ED + 1,250). For a random sample of events, we also use the consistent estimator proposed by White |19~ for the covariance matrix of the regression coefficients in the presence of an unknown form of heteroskedastic residuals. This alternative procedure yields the same conclusions.

Our results are shown in Panel A of Exhibit 1. In interpreting them, one should keep in mind that the coefficients ||gamma~.sub.ji~, j = 1,...,4, are measures of daily abnormal performance. The number of observations reported for the coefficient estimates differ because some acquiring firms drop out of the CRSP tapes before the end of year five. Four firms leave the tapes during the first postacquisition year, 22 drop out during year two, 51 during year three, 53 during year four, and 57 during year five. The main reason for these firms' disappearance is takeover (140 of 187 cases). In ten cases, the acquiring firm is liquidated or declared bankrupt, and in the remaining 37 cases, it is simply delisted.

The average intercept estimates are not significantly different from zero with 0.95 confidence, except in the second and third years after the acquisition. In those two years, acquiring firms experience significant negative abnormal returns. This phenomenon is also observed in a TABULAR DATA OMITTED sample of 765 mergers during the years 1955-1987 by Agrawal, Jaffe, and Mandelker |1~, using a method(7) which also controls for firm size, cross-sectional correlation, and changes in systematic risk and the risk-free rate. A nonparametric sign test confirms these findings. Panel A of Exhibit 1 also reports the estimates of the slope coefficient ||gamma~.sub.5~. The average estimate is positive and significant, an indication that acquiring firms have higher systematic risk than firms of similar size; more than 58% of the firms are riskier than those in their control portfolios.

In comparison to Agrawal, et al, we also find some negative performance in the first three years, but do not find supporting evidence for the five-year period. As reported in Panel B of Exhibit 1, the sum of the five intercept coefficient estimates (a measure of the overall abnormal performance) equals 0.00006 with a Z-statistic of 0.41, the median is 0.00003, and the number of events with positive overall abnormal performance essentially equals the number with negative performance. Summed over the 1,250 days under analysis, the 0.00006 average overall performance implies a total abnormal return of 1.5% -- the overall performance is a five-day performance, and there are 1,250/5 = 250 five-day intervals. Even though significant underperformance is reported in the parametric tests for the first two and three postacquisition years, the sign tests for these years are only marginally significant.

The parametric significance tests in the exhibit treat each observation as a random drawing from a distribution with finite mean, variance, and third moment about the mean. Given our large sample, the Liapounov central limit theorem implies that the average intercept estimate is asymptotically normally distributed even if the individual estimates do not come from identical distributions.(8) It is not clear, however, that all the assumptions necessary for this theorem to apply are satisfied.

As an alternative, the ||gamma~.sub.ji~, j = 1,...,4, coefficient estimates are standardized to make sure that the resulting variables have the same variance. This allows us to replicate the tests by invoking the Lindeberg and Levy variant of the central limit theorem (see DeGroot |5, p. 228~). Panel C of Exhibit 1 reports the distribution characteristics of a standardized measure of abnormal performance, namely the Z-statistics of the |g.sub.ji~, j = 1,...,4, estimates, all of which are approximately unit normally distributed. A look at the average standardized performance reveals that, if anything, the results are more positive than in Panel A: abnormal performance is negative and significant only during the second year after the acquisition. Otherwise, it is zero or positive (in the first and the fourth years).

One other assumption made in the test procedure, random drawings, warrants even closer scrutiny. This assumption is questionable for two reasons. First, some events are contemporaneous. Second, the periods used to estimate the individual regressions overlap in several cases (the regressions are estimated for the first 1,250 days following each acquisition completion date, and there are 1,298 events in the 21 sample years). As far as the first problem is concerned, the number of simultaneous acquisitions is very small, and the results are unchanged when these cases are excluded. To gauge the seriousness of the estimation-period overlap, one can analyze the time-series properties of the error terms in the regression equation. If the coefficient estimates are significantly correlated because of overlap, the error term for an individual firm in Equation (2) should be correlated with current and past error terms of other firms. Consequently, an equally weighted index of individual regression residuals should display significant serial correlation.

As it turns out, there is little evidence of serial correlation of the regression residual index for lags 1 through 20 (not shown here). No correlation coefficient is significantly different from zero, except for lags 6 and 12. Since there is no reason to suspect seasonality in a cross-sectional average of residuals, the statistical significance at lags 6 and 12 is more likely a fortuitous event than the result of any systematic pattern, especially since the numerical value of the two coefficients in question is small (-0.12 and 0.11).

Overall, there is only weak evidence of negative postacquisition performance. Acquiring firms do not perform worse than their comparison benchmarks over the first five postacquisition years. They do worse, on average, during the first three postacquisition years, but this underperformance is due to less than 53% of the sample firms; and there are marginal gains during years four and five.

B. Postacquisition Performance by Relative Acquisition Size

Of course, whatever negative performance there is could be unrelated to the acquisitions in question. If corporate acquisitions are responsible for negative performance, such performance should be observed with acquisitions that are large enough to affect share prices, but not with smaller acquisitions. To explore this conjecture, Exhibit 2 sorts the estimation results by relative acquisition size (amount paid for the target divided by acquiring-firm value, defined as the market value of the acquiring firm's common stock 20 days before the bid announcement). As shown there, performance over the full five years after the acquisition is zero, regardless of acquisition size. Interestingly, however, whereas the middle three quintiles of acquisition size (those with size comprising between 13% and 55% of firm value) experience negative returns (with 0.95 confidence) during the second postacquisition year, the lowest and highest quintiles never experience statistically significant returns. The fact that the negative performance observed in the full sample does not originate in the lowest quintile suggests that that performance could indeed TABULAR DATA OMITTED be related to the acquisition and not to some statistical artifact.

C. Postacquisition Performance Over Time

The trouble with negative postacquisition performance is that it is inconsistent with market efficiency. As shown in Exhibit 1, the average sum of the daily postacquisition performance during the first three postacquisition years is -0.02% (Z= -3.08); summed over the 750 days in question, this implies abnormal performance of -5%. This underperformance is large enough that investors willing to short the stock of acquiring firms would earn abnormal returns that would more than cover their transaction costs. The puzzle could be solved if it turned out that the negative performance was concentrated in only some of the calendar years considered. In that case, negative performance would not really be systematic and would thus be consistent with market efficiency.

To see whether postacquisition performance exhibits time patterns, possibly similar to those found for announcement effects, we sort the sample into the three decades of the 1960s, 1970s, and 1980s, and repeat the regression estimation. The number of acquisition events falling into these decades is 261,598, and 439, respectively. Exhibit 3 documents the results of this investigation.

We find that performance in the five years following the acquisition is reliably negative only in the 1960s: the average sum of the daily performance measures for that decade is -0.05% with a Z-statistic of -2.7, and only 44% of the firms in the sample show positive abnormal returns. By contrast, overall performance in the two subsequent decades is zero. A sign test confirms our results.

TABULAR DATA OMITTED

When examined from a three-year perspective, we find abnormal returns of -0.05% and -0.02% for the 1960s and 1970s, respectively. The later performance, however, is only marginally significant and involves only 52% of the firms. No underperformance occurs in the 1980s. Exhibit 3 confirms that, as in the full sample, the negative returns of the 1960s arise in years two and three after the acquisition. The 1970s and the 1980s do not have regression intercept estimates different from zero, except in years three and four in the 1970s, when the average intercepts are negative and positive, respectively. The negative performance in year three in the 1970s, however, cannot be confirmed by a sign test. Consequently, there is little evidence of negative performance after the 1960s, and none in the 1980s.

The main message of Exhibit 3 is that negative postacquisition performance is not inconsistent with informationally efficient markets. In fact, our evidence seems to buttress the case for efficient markets, since the negative returns observed in the 1960s disappear in the next two decades, as if the capital markets had realized a valuation mistake. Interestingly, the 1960s is the decade with the largest offer announcement effects for acquiring firms (see, among others, Bradley, Desai, and Kim |4~, Jarrell and Poulsen |10~, and Loderer and Martin |14~). One scenario consistent with these findings is that the market was too enthusiastic about corporate acquisitions in the 1960s, necessitating negative postacquisition price adjustments, and that, as a result of this experience, it valued the profitability of acquisitions more realistically thereafter, when both offer announcement effects and postacquisition performance are essentially zero.

D. Postacquisition Performance by Acquisition Form

Previous researchers find that postacquisition performance differs markedly depending on acquisition form. Whereas acquiring firms engaged in mergers underperform their benchmarks, those engaged in tender offers exhibit the same performance as their benchmarks. To investigate this result, we partition our sample into mergers (304), tender offers (155), comprehensive acquisitions of assets (93), and other acquisitions (746) that cannot be assigned to any of the three preceding categories on the basis of information reported in M&A or in the WSJI.(9)

The results of partitioning postacquisition performance by acquisition form and decade are shown in Exhibit 4, with each panel referring to a different form of acquisition. In contrast to the results of previous researchers,(10) we find that overall five-year performance is zero regardless of acquisition form: the sum of intercepts for the full sample period of 1966-1986 is -0.003% for mergers, 0.004% for tender offers, and 0.005% for acquisitions of assets. None of these figures is statistically different from zero. We find significant negative abnormal returns for the first three years for only the mergers and the other acquisitions samples. The negative returns are mostly found in the 1960s (years one and two in the mergers sample, year three in the tender offers sample, year two in the acquisition of assets and the other acquisitions samples), and to a lesser extent, in the 1970s (year two of the other acquisitions sample and year three of the mergers and the acquisition of assets samples). However, we do find significant positive returns for year four in both the mergers and the other acquisitions samples. Moreover, the median intercept estimates (not shown) are negative in years two and three not only for mergers, but also for acquisitions of assets.

More importantly, examination of individual postacquisition years by decade shows that all acquisition forms, including tender offers, underperform their benchmark at least once in the 1960s, and that mergers are not the only acquisition form to underperform their benchmark in the 1970s (they are, however, the only acquisition form to overperform in that decade). Sign test Z-values (not shown) confirm these results. Consequently, there is a lack of consistent evidence to buttress the claim that postacquisition performance is worse for mergers than for other forms of acquisition. As in the case of the full sample, whatever underperformance there is in the first three years, it occurs especially in the 1960s and 1970s, and disappears completely in the 1980s. No acquisition form, in any of the first five years after the acquisition, displays negative returns during the 1980s.(11)

III. Conclusions

We investigate whether negative postacquisition performance is a genuine phenomenon or a statistical artifact. The question seemed to have been settled by Franks, Harris, and Titman |8~, but Agrawal, Jaffe, and Mandelker |1~ argue that the Franks, et al |8~ findings of no abnormal performance are period-specific and cannot be generalized. Agrawal, et al |1~ report negative abnormal performance both before and after the Franks, et al |8~ sample period.

Our experiment controls for size effects, changes in the risk-free rate, and changes in systematic risk. We conduct TABULAR DATA OMITTED the experiment with a comprehensive sample of domestic acquisitions of publicly and privately held firms by NYSE- and Amex-listed firms during the 1966-1986 period. The sample selection procedure is designed to substantially limit the problem of double-counting postacquisition performance and to focus on acquisitions that can have a measurable impact on firm value.

We find that, on average, acquiring firms do not underperform a control portfolio during the first five years following the acquisition. They simply earn their required rate of return, no more or no less. There is some negative performance for the first three years, especially during the second and the third years after the acquisition, but it is most prominent in the 1960s, it diminishes in the 1970s, and disappears completely in the 1980s. Thus, in the later years, the postacquisition years do not provide strong enough evidence that corporate acquisitions are wasteful nor do they provide evidence contradicting market efficiency.

TABULAR DATA OMITTED

1See Langetieg |13~, Asquith |2~, Malatesta |15~, and Dodd and Ruback |7~.

2They explain the results in the 1970s by showing that the only five-year period with significant overperformance occurs during 1975-1979. Furthermore, they report that the 1980-1984 period exhibits significant underperformance.

3One hundred ninety-five acquisitions by 85 firms have a common announcement date, and 601 acquisitions by 201 firms have a common completion date. The results do not change when we remove these acquisitions.

4Note that the resulting sample is not a sample of infrequent acquirers. Acquisition frequency is not a sample selection criterion. In fact, the firms in our sample are responsible for about 67% of all acquisitions in our entire database (7,417 out of 11,020 acquisitions). We simply make sure that we limit double-counting of postacquisition performance and that we focus on events that can have a measurable impact on firm value.

5The time series of domestic mergers and acquisitions compiled by M&A shows large increases in acquisition activity in the 1967-1969 and 1980-1985 period (see Golbe and White |9~).

6The use of daily returns may lead to an upward bias in computing abnormal returns (see Blume and Stambaugh |3~).

7The method is adopted from Dimson and Marsh |6~, and Lakonishok and Vermaelen |12~.

8A complete set of sufficient assumptions also includes the assumption that the sum of the standardized third moments about the mean is zero in the limit (see DeGroot |5, pp. 228, 229~).

9The reason we have fewer mergers than Agrawal, Jaffe, and Mandelker |1~ is that our sample period does not include the 1950s and the full decade of the 1960s. Moreover, to limit double-counting and exclude potentially irrelevant events, we focus only on acquisitions that are 500 trading days apart and are in excess of 10% of acquiring firm size.

10For example, Agrawal, Jaffe, and Mandelker |1~ report a significant average abnormal return of -10% for their sample of mergers during the first five postacquisition years. However, similar to ours, they report significant negative returns for their mergers sample for the three-year period and zero abnormal returns in their tender offers sample.

11In contrast, Agrawal, Jaffe, and Mandelker |1~ report significantly negative average abnormal returns for mergers during the 1980s.

References

1. A. Agrawal, J.F. Jaffe, G.N. Mandelker, "The Post-Merger Performance of Acquiring Firms: A Re-examination of an Anomaly," Journal of Finance (forthcoming).

2. P. Asquith, "Merger Bids, Uncertainty, and Stockholders Returns," Journal of Financial Economics (April 1983), pp. 51-83.

3. M.E. Blume and R.F. Stambaugh, "Biases in Computed Returns: An Application to the Size Effect," Journal of Financial Economics (November 1983), pp. 387-404.

4. M.A. 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.

5. M. DeGroot, Probability and Statistics, Reading, MA, Addison Wesley, 1975.

6. E. Dimson and P. Marsh, "Event Study Methodologies and the Size Effect: The Case of UK Press Recommendations," Journal of Financial Economics (September 1986), pp. 113-142.

7. P. Dodd and R. Ruback, "Tender Offers and Stockholder Returns: An Empirical Analysis," Journal of Financial Economics (December 1977), pp. 351-373.

8. J. Franks, R. Harris, and S. Titman, "The Postmerger Share-Price Performance of Acquiring Firms," Journal of Financial Economics (March 1991), pp. 81-96.

9. D.L. Golbe and L.J. White, "A Time-Series Analysis of Mergers and Acquisitions in the U.S. Economy," in Corporate Takeovers: Causes and Consequences, A.J. Auerbach (ed.), Chicago, IL, University of Chicago Press, 1988, pp. 265-302.

10. G.A. Jarrell and A.B. Poulsen, "The Returns to Acquiring Firms in Tender Offers: Evidence From Three Decades," Financial Management (Autumn 1989), pp. 12-19.

11. M.C. Jensen and R.S. Ruback, "The Market for Corporate Control: The Scientific Evidence," Journal of Financial Economics (April 1983), pp. 5-50.

12. J. Lakonishok and T. Vermaelen, "Anomalous Price Behavior Around Repurchase Tender Offers," Journal of Finance (June 1990), pp. 455-477.

13. T. Langetieg, "An Application of a Three-Factor Performance Index to Measure Stockholder Gains From Merger," Journal of Financial Economics (December 1978), pp. 365-384.

14. C. Loderer and K. Martin, "Corporate Acquisitions by NYSE and AMEX Firms: The Experience of a Comprehensive Sample," Financial Management (Winter 1990), pp. 17-33.

15. P.H. Malatesta, "The Wealth Effect of Merger Activity and the Objective Functions of Merging Firms," Journal of Financial Economics (April 1983), pp. 155-181.

16. G. Mandelker, "Risk and Return: The Case of Merging Firms," Journal of Financial Economics (December 1974), pp. 303-335.

17. K.J. Martin and J.J. McConnell, "Corporate Performance, Corporate Takeovers, and Management Turnover," Journal of Finance (June 1991), pp. 671-687.

18. R. Roll, "The Hubris Hypothesis of Corporate Takeovers," Journal of Business (April 1986), pp. 197-216.

19. H. White, "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity," Econometrica (May 1980), pp. 817-838.

This paper analyzes the postacquisition stock-price performance of acquiring firms. Between 1966 and 1986, there were more than 10,000 domestic acquisitions in the U.S. of legally independent firms in which the acquirer traded on the New York or American Stock Exchange. An investment in each acquiring firm on the date the acquisition was completed, held for 500 trading days, would have yielded an equally weighted return of 21%. The return on a share in the market portfolio, by comparison, would have been 36%, more than 50% higher. The difference in investment performance is likely to reflect, at least in part, differences in systematic risk. But after risk is controlled for, acquiring firms are often found to underperform the market.(1) This raises questions about the profitability of corporate acquisitions and about the notion of market efficiency, since it seems that rational investors could benefit by shorting the stock of bidding firms after the acquisition has been completed. Most researchers appear to attribute negative postacquisition performance to estimation bias (see the surveys of Jensen and Ruback |11~ and Roll |18~). Not surprisingly, various performance measures are sensitive to the estimation technique (Mandelker |16~).

Franks, Harris, and Titman |8~ confirm the suspicion of estimation bias by showing that, with the appropriate benchmark, the negative postacquisition performance disappears. But, their paper is not the last word on the issue. Agrawal, Jaffe, and Mandelker |1~ reopen the Pandora's box by showing that the Franks, Harris, and Titman |8~ findings are specific to the 1975-1984 period and cannot be replicated for the years before and after. Agrawal, et al |1~ report significant postacquisition underperformance for mergers completed in the 1950s, 1960s, and 1980s, although not in the 1970s.(2)

Since the evidence on postacquisition performance appears to be sensitive not only to the estimation technique but also to the sample investigated, we reexamine the issue with a comprehensive sample of domestic acquisitions by firms traded on the New York Stock Exchange (NYSE) or the American Stock Exchange (Amex) during the years 1966-1986. We find abnormal performance in three years but not in five years following the acquisition. Negative performance in the second and third years after the acquisition is most prominent in the 1960s, and to a lesser extent in the 1970s, but not in the 1980s.

The first section of the paper discusses our data and investigative design. The second section presents our results, and the third section summarizes our conclusions.

I. Data and Test Design

Our sample includes domestic acquisitions by bidding firms listed on the Center for Research in Security Prices at the University of Chicago (CRSP) daily tapes starting at least 200 trading days before the acquisition is completed. This sample is compiled from three sources. The first source is a sample used in a companion paper (Loderer and Martin |14~) that includes mergers, acquisitions of privately held corporations, and comprehensive acquisitions of the assets of individual firms completed during the 1965-1984 period, as reported in various issues of Mergers and Acquisitions (M&A). The second source consists of tender offers from the same period in Martin and McConnell |17~, and the third source consists of all domestic acquisitions completed in the 1985-1986 period according to M&A. The three sources yield a sample of 11,020 acquisitions. For all |of the acquisitions~, we gather information on the date they are first mentioned in the Wall Street Journal Index (WSJI) (the announcement date, AD) and the date they are reported by M&A as having been completed (the effective date, ED).(3)

Using the full sample to gauge postacquisition performance could result in double-counting and lead to estimation bias if all acquisitions by any firm, no matter how close to each other, were treated as separate events. We might, for example, see an average negative postacquisition performance as a result of counting the occasional negative experience of frequent acquirers several times instead of just once. To limit the potential for double-counting, we restrict our attention to the following acquisitions. For each firm, we start in 1966 and search for the completion date of the first acquisition in excess of ten percent of acquiring firm size. We then skip 500 days and resume the search to identify the completion date of the next acquisition in excess of ten percent of firm size. The focus on acquisitions larger than ten percent is to avoid acquisitions too small to affect firm value. Smaller acquisitions are less likely to be the cause of any abnormal postacquisition performance. This search is conducted through the end of 1986. Since the search procedure involves the use of only contemporaneous information, no obvious selection bias is introduced into the analysis. These selection criteria reduce the sample from 11,020 to 1,298 acquisitions.(4)

The crux of the experiment is to test whether modified market model measures of negative abnormal performance become insignificant when risk is appropriately accounted for. Unfortunately, when abnormal returns are computed using market model parameters estimated with preacquisition data, there are several possible sources of downward estimation bias. One source can be illustrated assuming expected returns are set according to the capital asset pricing model (CAPM). Then, computing abnormal returns with a preacquisition estimate of the market model intercept assumes stationarity of the risk-free rate. But, since corporate acquisitions are not uniformly distributed in time,(5) this assumption could be inappropriate. One possible scenario is that during times of intense acquisition activity, the risk-free rate is driven up by firms that want to fund their acquisitions. When the activity subsides, the risk-free rate could decline. If so, in a cross-section of acquisitions, the risk-free rate could be higher before the completion date and lower afterward. Preacquisition estimates of the market model intercept could therefore be correspondingly too high, which could lead to downward bias in the computation of postacquisition abnormal returns.

Downward bias can also occur if the systematic risk of the acquiring firm is lower after the acquisition than before, which is not unlikely, since acquiring firms typically experience positive abnormal returns before an acquisition. This growth tends to reduce both financial leverage and systematic risk and could therefore lead to downward bias in the measurement of postacquisition performance.

Finally, downward bias can occur because the market model ignores firm size as a proxy for risk. Small firms tend to have higher market model intercepts than large firms. Since acquiring firms tend to grow before the acquisition, they will have smaller intercepts afterward, and using a preacquisition estimate of the intercept to compute abnormal returns once again introduces downward bias.(6)

To simultaneously take into account nonstationarity of the risk-free rate, changes in systematic risk, and changes in firm size, we adopt the following procedure for estimating abnormal returns. We first determine the size decile to which a particular acquiring firm belongs, according to the CRSP files. For each year, CRSP assigns each firm to a size decile using the market value of the firm's equity at the end of the preceding year as the basis for the assignment. We use these assignments to form ten size portfolios with equally weighted daily returns. Each acquiring firm is then matched with its appropriate size portfolio to determine the relevant size-based daily return. Abnormal performance is then measured as follows.

We assume that the market model can be written in terms of deviations from the risk-tree rate for the returns to firm i and to firm i's relevant size portfolio. Then, the return to firm i (|R.sub.i~) net of the return to its corresponding size portfolio (|R.sub.Si~) can be set equal to (time subscripts are omitted for simplicity):

|Mathematical Expression Omitted~

where |alpha.sub.i~ reflects both abnormal performance and possible size effects, whereas |alpha.sub.Si~ reflects only possible size effects (of course, under the CAPM, this term is zero). Both |beta.sub.i~ and |beta.sub.Si~ measure systematic risk, |R.sub.m~ is the value-weighted CRSP index, |R.sub.f~ is the yield-to-maturity on a three-month Treasury bill, ||upsilon~.sub.i~ and |xi.sub.i~ are i.i.d. error terms independent of each other, and |eta.sub.i~ = |(|upsion~ - |xi~).sub.i~.

Any abnormal performance of firm i will be reflected in the intercept term of this equation, ||gamma~.sub.0i~, since that term captures the systematic firm-specific performance net of any possible size effect. Consequently, to investigate the impact of an acquisition on firm performance, one can test for the sign and significance of the estimate of ||gamma~.sub.0i~. If negative postacquisition performance is a genuine phenomenon, ||gamma~.sub.0i~ should be negative and significant.

For more detail, we analyze the five years separately with a modified version of Equation (1):

|Mathematical Expression Omitted~

The variables D. are defined as follows:

|D.sub.0~ = 1 over the interval (ED + 1, ED + 250);

|D.sub.0~ = 0 otherwise;

|D.sub.1~ = 1 over the interval (ED + 251, ED + 500);

|D.sub.1~ = 0 otherwise;

|D.sub.2~ = 1 over the interval (ED + 501, ED + 750);

|D.sub.2~ = 0 otherwise;

|D.sub.3~ = 1 over the interval (ED + 751, ED + 1,000);

|D.sub.3~ = 0 otherwise:

|D.sub.4~ = 1 over the interval (ED + 1,001, ED + 1,250);

|D.sub.4~ = 0 otherwise.

II. The Evidence

A. Overall Results

We measure postacquisition abnormal performance by estimating the intercepts ||gamma~.sub.ji~, j = 1,...,4, of Equation (2) above. We estimate the regression with OLS for the acquiring firm in each of the 1,298 acquisition events in the sample over the period (ED + 1, ED + 1,250). For a random sample of events, we also use the consistent estimator proposed by White |19~ for the covariance matrix of the regression coefficients in the presence of an unknown form of heteroskedastic residuals. This alternative procedure yields the same conclusions.

Our results are shown in Panel A of Exhibit 1. In interpreting them, one should keep in mind that the coefficients ||gamma~.sub.ji~, j = 1,...,4, are measures of daily abnormal performance. The number of observations reported for the coefficient estimates differ because some acquiring firms drop out of the CRSP tapes before the end of year five. Four firms leave the tapes during the first postacquisition year, 22 drop out during year two, 51 during year three, 53 during year four, and 57 during year five. The main reason for these firms' disappearance is takeover (140 of 187 cases). In ten cases, the acquiring firm is liquidated or declared bankrupt, and in the remaining 37 cases, it is simply delisted.

The average intercept estimates are not significantly different from zero with 0.95 confidence, except in the second and third years after the acquisition. In those two years, acquiring firms experience significant negative abnormal returns. This phenomenon is also observed in a TABULAR DATA OMITTED sample of 765 mergers during the years 1955-1987 by Agrawal, Jaffe, and Mandelker |1~, using a method(7) which also controls for firm size, cross-sectional correlation, and changes in systematic risk and the risk-free rate. A nonparametric sign test confirms these findings. Panel A of Exhibit 1 also reports the estimates of the slope coefficient ||gamma~.sub.5~. The average estimate is positive and significant, an indication that acquiring firms have higher systematic risk than firms of similar size; more than 58% of the firms are riskier than those in their control portfolios.

In comparison to Agrawal, et al, we also find some negative performance in the first three years, but do not find supporting evidence for the five-year period. As reported in Panel B of Exhibit 1, the sum of the five intercept coefficient estimates (a measure of the overall abnormal performance) equals 0.00006 with a Z-statistic of 0.41, the median is 0.00003, and the number of events with positive overall abnormal performance essentially equals the number with negative performance. Summed over the 1,250 days under analysis, the 0.00006 average overall performance implies a total abnormal return of 1.5% -- the overall performance is a five-day performance, and there are 1,250/5 = 250 five-day intervals. Even though significant underperformance is reported in the parametric tests for the first two and three postacquisition years, the sign tests for these years are only marginally significant.

The parametric significance tests in the exhibit treat each observation as a random drawing from a distribution with finite mean, variance, and third moment about the mean. Given our large sample, the Liapounov central limit theorem implies that the average intercept estimate is asymptotically normally distributed even if the individual estimates do not come from identical distributions.(8) It is not clear, however, that all the assumptions necessary for this theorem to apply are satisfied.

As an alternative, the ||gamma~.sub.ji~, j = 1,...,4, coefficient estimates are standardized to make sure that the resulting variables have the same variance. This allows us to replicate the tests by invoking the Lindeberg and Levy variant of the central limit theorem (see DeGroot |5, p. 228~). Panel C of Exhibit 1 reports the distribution characteristics of a standardized measure of abnormal performance, namely the Z-statistics of the |g.sub.ji~, j = 1,...,4, estimates, all of which are approximately unit normally distributed. A look at the average standardized performance reveals that, if anything, the results are more positive than in Panel A: abnormal performance is negative and significant only during the second year after the acquisition. Otherwise, it is zero or positive (in the first and the fourth years).

One other assumption made in the test procedure, random drawings, warrants even closer scrutiny. This assumption is questionable for two reasons. First, some events are contemporaneous. Second, the periods used to estimate the individual regressions overlap in several cases (the regressions are estimated for the first 1,250 days following each acquisition completion date, and there are 1,298 events in the 21 sample years). As far as the first problem is concerned, the number of simultaneous acquisitions is very small, and the results are unchanged when these cases are excluded. To gauge the seriousness of the estimation-period overlap, one can analyze the time-series properties of the error terms in the regression equation. If the coefficient estimates are significantly correlated because of overlap, the error term for an individual firm in Equation (2) should be correlated with current and past error terms of other firms. Consequently, an equally weighted index of individual regression residuals should display significant serial correlation.

As it turns out, there is little evidence of serial correlation of the regression residual index for lags 1 through 20 (not shown here). No correlation coefficient is significantly different from zero, except for lags 6 and 12. Since there is no reason to suspect seasonality in a cross-sectional average of residuals, the statistical significance at lags 6 and 12 is more likely a fortuitous event than the result of any systematic pattern, especially since the numerical value of the two coefficients in question is small (-0.12 and 0.11).

Overall, there is only weak evidence of negative postacquisition performance. Acquiring firms do not perform worse than their comparison benchmarks over the first five postacquisition years. They do worse, on average, during the first three postacquisition years, but this underperformance is due to less than 53% of the sample firms; and there are marginal gains during years four and five.

B. Postacquisition Performance by Relative Acquisition Size

Of course, whatever negative performance there is could be unrelated to the acquisitions in question. If corporate acquisitions are responsible for negative performance, such performance should be observed with acquisitions that are large enough to affect share prices, but not with smaller acquisitions. To explore this conjecture, Exhibit 2 sorts the estimation results by relative acquisition size (amount paid for the target divided by acquiring-firm value, defined as the market value of the acquiring firm's common stock 20 days before the bid announcement). As shown there, performance over the full five years after the acquisition is zero, regardless of acquisition size. Interestingly, however, whereas the middle three quintiles of acquisition size (those with size comprising between 13% and 55% of firm value) experience negative returns (with 0.95 confidence) during the second postacquisition year, the lowest and highest quintiles never experience statistically significant returns. The fact that the negative performance observed in the full sample does not originate in the lowest quintile suggests that that performance could indeed TABULAR DATA OMITTED be related to the acquisition and not to some statistical artifact.

C. Postacquisition Performance Over Time

The trouble with negative postacquisition performance is that it is inconsistent with market efficiency. As shown in Exhibit 1, the average sum of the daily postacquisition performance during the first three postacquisition years is -0.02% (Z= -3.08); summed over the 750 days in question, this implies abnormal performance of -5%. This underperformance is large enough that investors willing to short the stock of acquiring firms would earn abnormal returns that would more than cover their transaction costs. The puzzle could be solved if it turned out that the negative performance was concentrated in only some of the calendar years considered. In that case, negative performance would not really be systematic and would thus be consistent with market efficiency.

To see whether postacquisition performance exhibits time patterns, possibly similar to those found for announcement effects, we sort the sample into the three decades of the 1960s, 1970s, and 1980s, and repeat the regression estimation. The number of acquisition events falling into these decades is 261,598, and 439, respectively. Exhibit 3 documents the results of this investigation.

We find that performance in the five years following the acquisition is reliably negative only in the 1960s: the average sum of the daily performance measures for that decade is -0.05% with a Z-statistic of -2.7, and only 44% of the firms in the sample show positive abnormal returns. By contrast, overall performance in the two subsequent decades is zero. A sign test confirms our results.

TABULAR DATA OMITTED

When examined from a three-year perspective, we find abnormal returns of -0.05% and -0.02% for the 1960s and 1970s, respectively. The later performance, however, is only marginally significant and involves only 52% of the firms. No underperformance occurs in the 1980s. Exhibit 3 confirms that, as in the full sample, the negative returns of the 1960s arise in years two and three after the acquisition. The 1970s and the 1980s do not have regression intercept estimates different from zero, except in years three and four in the 1970s, when the average intercepts are negative and positive, respectively. The negative performance in year three in the 1970s, however, cannot be confirmed by a sign test. Consequently, there is little evidence of negative performance after the 1960s, and none in the 1980s.

The main message of Exhibit 3 is that negative postacquisition performance is not inconsistent with informationally efficient markets. In fact, our evidence seems to buttress the case for efficient markets, since the negative returns observed in the 1960s disappear in the next two decades, as if the capital markets had realized a valuation mistake. Interestingly, the 1960s is the decade with the largest offer announcement effects for acquiring firms (see, among others, Bradley, Desai, and Kim |4~, Jarrell and Poulsen |10~, and Loderer and Martin |14~). One scenario consistent with these findings is that the market was too enthusiastic about corporate acquisitions in the 1960s, necessitating negative postacquisition price adjustments, and that, as a result of this experience, it valued the profitability of acquisitions more realistically thereafter, when both offer announcement effects and postacquisition performance are essentially zero.

D. Postacquisition Performance by Acquisition Form

Previous researchers find that postacquisition performance differs markedly depending on acquisition form. Whereas acquiring firms engaged in mergers underperform their benchmarks, those engaged in tender offers exhibit the same performance as their benchmarks. To investigate this result, we partition our sample into mergers (304), tender offers (155), comprehensive acquisitions of assets (93), and other acquisitions (746) that cannot be assigned to any of the three preceding categories on the basis of information reported in M&A or in the WSJI.(9)

The results of partitioning postacquisition performance by acquisition form and decade are shown in Exhibit 4, with each panel referring to a different form of acquisition. In contrast to the results of previous researchers,(10) we find that overall five-year performance is zero regardless of acquisition form: the sum of intercepts for the full sample period of 1966-1986 is -0.003% for mergers, 0.004% for tender offers, and 0.005% for acquisitions of assets. None of these figures is statistically different from zero. We find significant negative abnormal returns for the first three years for only the mergers and the other acquisitions samples. The negative returns are mostly found in the 1960s (years one and two in the mergers sample, year three in the tender offers sample, year two in the acquisition of assets and the other acquisitions samples), and to a lesser extent, in the 1970s (year two of the other acquisitions sample and year three of the mergers and the acquisition of assets samples). However, we do find significant positive returns for year four in both the mergers and the other acquisitions samples. Moreover, the median intercept estimates (not shown) are negative in years two and three not only for mergers, but also for acquisitions of assets.

More importantly, examination of individual postacquisition years by decade shows that all acquisition forms, including tender offers, underperform their benchmark at least once in the 1960s, and that mergers are not the only acquisition form to underperform their benchmark in the 1970s (they are, however, the only acquisition form to overperform in that decade). Sign test Z-values (not shown) confirm these results. Consequently, there is a lack of consistent evidence to buttress the claim that postacquisition performance is worse for mergers than for other forms of acquisition. As in the case of the full sample, whatever underperformance there is in the first three years, it occurs especially in the 1960s and 1970s, and disappears completely in the 1980s. No acquisition form, in any of the first five years after the acquisition, displays negative returns during the 1980s.(11)

III. Conclusions

We investigate whether negative postacquisition performance is a genuine phenomenon or a statistical artifact. The question seemed to have been settled by Franks, Harris, and Titman |8~, but Agrawal, Jaffe, and Mandelker |1~ argue that the Franks, et al |8~ findings of no abnormal performance are period-specific and cannot be generalized. Agrawal, et al |1~ report negative abnormal performance both before and after the Franks, et al |8~ sample period.

Our experiment controls for size effects, changes in the risk-free rate, and changes in systematic risk. We conduct TABULAR DATA OMITTED the experiment with a comprehensive sample of domestic acquisitions of publicly and privately held firms by NYSE- and Amex-listed firms during the 1966-1986 period. The sample selection procedure is designed to substantially limit the problem of double-counting postacquisition performance and to focus on acquisitions that can have a measurable impact on firm value.

We find that, on average, acquiring firms do not underperform a control portfolio during the first five years following the acquisition. They simply earn their required rate of return, no more or no less. There is some negative performance for the first three years, especially during the second and the third years after the acquisition, but it is most prominent in the 1960s, it diminishes in the 1970s, and disappears completely in the 1980s. Thus, in the later years, the postacquisition years do not provide strong enough evidence that corporate acquisitions are wasteful nor do they provide evidence contradicting market efficiency.

TABULAR DATA OMITTED

1See Langetieg |13~, Asquith |2~, Malatesta |15~, and Dodd and Ruback |7~.

2They explain the results in the 1970s by showing that the only five-year period with significant overperformance occurs during 1975-1979. Furthermore, they report that the 1980-1984 period exhibits significant underperformance.

3One hundred ninety-five acquisitions by 85 firms have a common announcement date, and 601 acquisitions by 201 firms have a common completion date. The results do not change when we remove these acquisitions.

4Note that the resulting sample is not a sample of infrequent acquirers. Acquisition frequency is not a sample selection criterion. In fact, the firms in our sample are responsible for about 67% of all acquisitions in our entire database (7,417 out of 11,020 acquisitions). We simply make sure that we limit double-counting of postacquisition performance and that we focus on events that can have a measurable impact on firm value.

5The time series of domestic mergers and acquisitions compiled by M&A shows large increases in acquisition activity in the 1967-1969 and 1980-1985 period (see Golbe and White |9~).

6The use of daily returns may lead to an upward bias in computing abnormal returns (see Blume and Stambaugh |3~).

7The method is adopted from Dimson and Marsh |6~, and Lakonishok and Vermaelen |12~.

8A complete set of sufficient assumptions also includes the assumption that the sum of the standardized third moments about the mean is zero in the limit (see DeGroot |5, pp. 228, 229~).

9The reason we have fewer mergers than Agrawal, Jaffe, and Mandelker |1~ is that our sample period does not include the 1950s and the full decade of the 1960s. Moreover, to limit double-counting and exclude potentially irrelevant events, we focus only on acquisitions that are 500 trading days apart and are in excess of 10% of acquiring firm size.

10For example, Agrawal, Jaffe, and Mandelker |1~ report a significant average abnormal return of -10% for their sample of mergers during the first five postacquisition years. However, similar to ours, they report significant negative returns for their mergers sample for the three-year period and zero abnormal returns in their tender offers sample.

11In contrast, Agrawal, Jaffe, and Mandelker |1~ report significantly negative average abnormal returns for mergers during the 1980s.

References

1. A. Agrawal, J.F. Jaffe, G.N. Mandelker, "The Post-Merger Performance of Acquiring Firms: A Re-examination of an Anomaly," Journal of Finance (forthcoming).

2. P. Asquith, "Merger Bids, Uncertainty, and Stockholders Returns," Journal of Financial Economics (April 1983), pp. 51-83.

3. M.E. Blume and R.F. Stambaugh, "Biases in Computed Returns: An Application to the Size Effect," Journal of Financial Economics (November 1983), pp. 387-404.

4. M.A. 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.

5. M. DeGroot, Probability and Statistics, Reading, MA, Addison Wesley, 1975.

6. E. Dimson and P. Marsh, "Event Study Methodologies and the Size Effect: The Case of UK Press Recommendations," Journal of Financial Economics (September 1986), pp. 113-142.

7. P. Dodd and R. Ruback, "Tender Offers and Stockholder Returns: An Empirical Analysis," Journal of Financial Economics (December 1977), pp. 351-373.

8. J. Franks, R. Harris, and S. Titman, "The Postmerger Share-Price Performance of Acquiring Firms," Journal of Financial Economics (March 1991), pp. 81-96.

9. D.L. Golbe and L.J. White, "A Time-Series Analysis of Mergers and Acquisitions in the U.S. Economy," in Corporate Takeovers: Causes and Consequences, A.J. Auerbach (ed.), Chicago, IL, University of Chicago Press, 1988, pp. 265-302.

10. G.A. Jarrell and A.B. Poulsen, "The Returns to Acquiring Firms in Tender Offers: Evidence From Three Decades," Financial Management (Autumn 1989), pp. 12-19.

11. M.C. Jensen and R.S. Ruback, "The Market for Corporate Control: The Scientific Evidence," Journal of Financial Economics (April 1983), pp. 5-50.

12. J. Lakonishok and T. Vermaelen, "Anomalous Price Behavior Around Repurchase Tender Offers," Journal of Finance (June 1990), pp. 455-477.

13. T. Langetieg, "An Application of a Three-Factor Performance Index to Measure Stockholder Gains From Merger," Journal of Financial Economics (December 1978), pp. 365-384.

14. C. Loderer and K. Martin, "Corporate Acquisitions by NYSE and AMEX Firms: The Experience of a Comprehensive Sample," Financial Management (Winter 1990), pp. 17-33.

15. P.H. Malatesta, "The Wealth Effect of Merger Activity and the Objective Functions of Merging Firms," Journal of Financial Economics (April 1983), pp. 155-181.

16. G. Mandelker, "Risk and Return: The Case of Merging Firms," Journal of Financial Economics (December 1974), pp. 303-335.

17. K.J. Martin and J.J. McConnell, "Corporate Performance, Corporate Takeovers, and Management Turnover," Journal of Finance (June 1991), pp. 671-687.

18. R. Roll, "The Hubris Hypothesis of Corporate Takeovers," Journal of Business (April 1986), pp. 197-216.

19. H. White, "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity," Econometrica (May 1980), pp. 817-838.

Printer friendly Cite/link Email Feedback | |

Title Annotation: | Special Issue: Corporate Control |
---|---|

Author: | Loderer, Claudio; Martin, Kenneth |

Publication: | Financial Management |

Date: | Sep 22, 1992 |

Words: | 4816 |

Previous Article: | Foreign takeover activity in the U.S. and wealth effects for target firm shareholders. |

Next Article: | Corporate investment decisions and corporate control: evidence from going-private transactions. |

Topics: |