Why Do Dual Class Companies Unify Their Stocks? Abstract

Transcription

1 Why Do Dual Class Companies Unify Their Stocks? Abstract I examine a sample of U.S. firms that switched from a dual class structure to a single class structure and test some hypotheses on the reasons for this recapitalization. First, dual class managers may wish to sell their firm or a control block. In order to signal this to the market, I posit, they unify their dual class shares. This hypothesis implies increased transfer of ownership for dual class firms following the unification. Second, dual class companies may suffer from agency costs and face higher costs of equity than comparable firms in their industries. In order to reduce their cost of equity, these companies reclassify their multiple classes into a single class of stock. This hypothesis implies that these companies will engage in more mergers and acquisitions, increase their investment activity and raise more equity once they have a lower cost of equity. Third, dual class managers may wish to increase the liquidity of their stock by reclassifying their shares into a single class. They may think that illiquidity is depressing the stock price. They also may prefer a more liquid market if they wish to reduce their shareholdings.

2 1. Introduction One of the most popular topics in corporate finance is the optimal security design. This issue merits examination because the design of securities has many implications related with corporate ownership and control. In general, shareholders of a company have claims to both the firm s cash flow and voting rights. If all shares are entitled votes that are in the same proportion as their claim to equity, we have a one share-one vote structure. In other cases there may exist several classes of common stock that have differing voting rights. This kind of capital structure is called dual class equity (multiple class equity, more generally). Academics have studied security design issues both theoretically and empirically. Models have been developed to set out conditions under which one share-one vote is optimal. There are also models that aimed at finding out the mechanism that determines the value of a vote. Empirical research consists of studies that have examined many aspects of dual class shares. Yet other aspects remain to be investigated. Managers and/or controlling shareholders may have different motivations for adopting multiple class shares. Pursuing the private benefits of control is one of the most cited reasons for dual class capitalization. When the private benefits of control are large, managers will insulate themselves from the market for corporate control making their removal difficult if not impossible. Among these private benefits are to be able to maintain family control, employ relatives, friends, and business associates to senior management positions, and pass control to offspring. Departures from one share-one vote might affect firm value and efficiency. It might also have an effect on firms decisions to go public. Corporate control, for example, becomes a major

3 issue once the new shareholders come into place. One specific aspect is the increased likelihood of acquisition. Firms foreseeing possible hostile takeover attempts may go public with multiple classes of shares and use this structure as an antitakeover mechanism. No matter what the motivations for multiple class shares are, one might wish to focus on the consequences of such a system. A dual class system that grants managers a high vote ownership and a low equity ownership may entrench them making value-increasing takeovers impossible and hence decreasing the value of the firm. Having a small amount of wealth tied to the firm, these managers will not suffer the wealth consequences of their actions as much as a manager whose equity ownership is proportional to her voting rights. In one of the earlier theoretical studies, Grossman and Hart (1988) analyze the impact of a firm s security-voting structure on the market for corporate control. They distinguish between the private benefits of control that accrue to the incumbent management and public benefits that are received by all shareholders. They show that one share-one vote maximizes firm value when only one of the parties (incumbent management or rival management) in a control contest has significant private benefits of control. On the other hand, if both parties to the contest have significant private benefits of control then deviations from one share-one vote may help outside shareholders extract some of those benefits and hence may be optimal. Their theory is consistent with the evidence in DeAngelo and DeAngelo (1985) for a sample of 78 dual class companies. The deviations from one share-one vote in that study were primarily for the purpose of maintaining family control, which is consistent with large private benefits of control for the incumbent only. With a very similar model, Harris and Raviv (1988) investigate the optimal corporate governance rules. They find that the simple majority rule and one share-one vote is an optimal

4 governance scheme in order to choose the best management (social optimality). However, if the aim is to maximize the value of the securities issued (private optimality), then it is optimal to issue two securities: voting rights with no equity claims and cash flow claims with no voting rights. These securities are the extreme deviations from one share-one vote suggested by Grossman and Hart (1988) as a means to capture private benefits of control from the winning candidate in a control contest. One thing to keep in mind, however, is that this dual class system does not ensure that the best management is elected which may result in an efficiency reduction. Harris and Raviv (1988) predict an increase in share value if a firm announces dual class capitalization. On the contrary, the agency literature suggests that dual class capitalizations may increase potential agency problems and entrench managers. These managers may hold a high proportion of superior-voting shares and a low proportion of inferior-voting shares. Securing the control of the company without having a large amount of wealth at stake may give managers incentives to pursue private benefits of control at the expense of outside stockholders who hold inferior-voting shares. Investors with this expectation would believe that these capitalizations are value decreasing. Consistent with this, Jarrell and Poulsen (1988) find significant negative announcement effects for 94 U.S. corporations that have gone through a dual class capitalization. Dann and DeAngelo (1988) also provide evidence that defensive recapitalizations are value decreasing. However, Partch (1987) finds no negative effects for 44 dual class capitalizations. With a sample of Canadian capitalizations, Jog and Riding (1986) also report no abnormal stock price response. On the other hand, Cornett and Vetsuypens (1989) report positive returns around recapitalization announcements. One problem with these studies is that their analyses do not control for managerial ownership before the recapitalization. Chang and Mayers (1992) document mixed results controlling for managerial vote ownership prior to the recapitalization.

5 Specifically, they report that recapitalizations are value increasing for firms with high managerial vote ownership before the recapitalization and value decreasing for firms with low managerial ownership of votes before the restructuring. This makes sense since moral hazard problems are expected to increase more for recapitalizations in which the pre-restructuring managerial vote ownership was low. Announcement effects aside, Jog, Srivastava and Panangipalli (1996) analyze the performance of 213 Canadian firms that have performed a dual class reclassification. They observe a superior performance prior to the reclassification and a significantly worse performance in the post-classification period. They also find the performance of restricted-voting shares and superior-voting shares to be very similar. Overall, they conclude that dual class capitalization is a disappointing experience for shareholders of those firms. Another line of research has studied the value of voting rights. Researchers have shown for different countries that superior-voting shares are worth more than inferior-voting shares. Jog, Riding, and Seguin (1985) and Robinson and White (1990) for Canada, Levy (1983) for Israel, Zingales (1994) for Italy, Rydqvist (1988) for Sweden, Horner (1988) for Switzerland, Lease, Ang and Megginson (1989) for U.K., and McConnell and Mikkelson (1983) for U.S. are some of those studies. Given the international evidence, Zingales (1995) investigates the determinants of the value of voting rights. He considers a model along the lines of Grossman and Hart (1988) and Harris and Raviv (1988) in order to determine the optimal bidding strategy when there are multiple classes of common stock with differential voting rights. Then, he tests his model with a sample of 94 dual class U.S. corporations. He finds that the value of a vote is determined by the expected additional payments that will be received by vote holders if there is a control contest.

6 He also finds that the magnitude of this payment depends on the probability that a vote is pivotal in case of a control contest and the size of the private benefits of control. Hauser and Lauterbach (2000) also study the value of voting rights. They consider 67 dual class unifications that took place in the Tel-Aviv Stock Exchange during They use the compensation received by superior vote stockholders to give up their superior voting status for computing the value of voting rights. They find that the median price of 1 percent of the vote is about 0.1 percent of the value of equity. They also find that unifications are accompanied by positive excess returns. Amoako-Adu and Smith (2001) analyze a sample of Canadian dual class companies over time. Among other questions, they seek answers for the following: How frequently do dual class firms reclassify back into a single class and what are the reasons behind the consolidation? They posit the following as possible reasons for unification: increase liquidity of shares, increase the appeal of the shares to outside investors, prepare for a seasoned offering, meet the terms of a debt restructuring agreement and hence reduce the leverage of the firm, reduced need to a takeover defense (and possible desire to sell the firm or a control block), increase institutional investor interest, adoption of a shareholder rights plan, scheduled expiry of the dual class structure as established by prior shareholders agreement. Stock unifications have become increasingly popular among U.S. dual class companies. 1 In this paper, I propose and test three hypotheses to explain these unifications. First, dual class managers might get ready to sell their firm and in order to signal this to the market, they might go through dual class unification. If this hypothesis is true, one would observe dual class 1 Recent examples are Symbollon Corp. (1997), Vion Pharmaceuticals Inc. (1998), PacifiCare Health Systems Inc. (1999), Dairy Mart Convenience Stores Inc. (2000), J M Smucker Co. (2000), Waddell & Reed Financial Inc. (2000), Continental Airlines (2001), Raytheon Corp. (2001) and Seacoast Banking Corp of Florida (2002).

7 companies change owners more frequently after the unification. I call this hypothesis control transfer hypothesis. Second, I argue that dual class companies suffer from agency costs, which lead to high costs of equity. In order to reduce their cost of equity, I posit, dual class companies unify their stocks. Once they have reduced their cost of equity, I expect these companies to increase their acquisition activity, investments and equity issuance. I call this the cost of equity hypothesis. Easley and O Hara (2001) investigate the link between a firm s information structure and its cost of capital. One implication of their analysis is that firms can affect their cost of capital by changing the precision and quantity of information available to investors. To the extent that dual class managers disclose less information than their otherwise identical single class counterparts, dual class companies may face relatively higher costs of capital. Himmelberg, Hubbard, and Love (2002) relate the cost of capital to investor protection and inside ownership in an agency context. They find a negative relation between investor protection and inside ownership. They also report a positive relation between inside ownership and implied cost of capital. These results would imply a higher cost of capital for dual class firms since their managers are more likely to expropriate outside investors. Third, liquidity concerns might trigger the decision to unify the dual class stock. 2 Under the liquidity hypothesis, dual class managers wish to increase the liquidity of their stock and unify their multiple classes of shares to that end. This hypothesis is difficult to test since in some cases higher voting shares are not publicly traded. In other cases there is less demand for superior voting shares due to lower cash flow rights. Recent studies have focused on the effects of corporate finance on asset pricing. For example, Holmstrom and Tirole (2000) develop an asset pricing model based on liquidity. They 2 There is anecdotal evidence that dual class companies in the U.S. have illiquidity concerns.

8 derive an explicit formula for the liquidity premia associated with corporate demand for liquidity. This demand has implications for the equity premium puzzle among other things. Holmstrom and Tirole s model is consistent with Bhide s (1993) argument that corporate control monitoring actions and liquidity interact and jointly affect a stock s price. 3 In their study of long-run IPO returns, Eckbo and Norli (2000) use liquidity as a risk factor in a Fama-French type of model. They find this liquidity factor to have a significant influence on expected returns. Gardiol, Gibson-Asner and Tuchschmid (1997) study the affects of liquidity and corporate control on the pricing of Swiss dual class shares. In conclusion, they argue that recent governance changes in Swiss corporations can be attributed to their desire to increase the liquidity of their stocks and thus to decrease their cost of capital. 4 The paper is organized as follows: Section 2 describes the sample. Section 3 tests the control transfer hypothesis. Section 4 tests the cost of equity hypothesis. Section 5 tests the liquidity hypothesis. Section 6 performs an event study to investigate the market reaction to the unification. Section 7 summarizes the paper and presents concluding remarks. 2. Sample and Sample Data All monthly share class (SHRCLS) data (1,567,056 observations) were obtained using all permanent numbers (PERMNO). Then, I focused on the 73,142 observations with non-blank share class data. I eliminated observations with share classes E, H, N and/or share codes (SHRCD) 12, 14, 15, 18, 30, 31, 48, 70, 71 and was left with 65,457. Then, for each permanent number, I took the earliest observation and 458 were remaining. Sorting by permanent company 3 Conoco, Inc. reclassified its dual class shares into a single class in September The management argued that the dual class structure was hurting the stock price and the easier single class structure was deemed better by the investors. 4 A U.S. corporation, Mitchell Energy and Development Corp., went from dual to single in June The stated reasons for the unification were to increase liquidity and eliminate confusion in the marketplace over the dual class structure.

9 number (PERMCO), 75 companies had two classes issued (two different PERMNOs). So, there were =383 unique firm observations. These are all the firms that issued dual class shares during of them are from January 1925 through December Table 1 Panel A reports a breakdown of dual class firms by year and exchange. There is a positive trend in the number of dual issues, especially after I also note that duals tend to be listed more often on NASDAQ (225) than on NYSE (40) or AMEX (82). 5 This sample includes all the companies that issued dual class shares some time before Thus, it is possible that some of them might have gone back to unitary or may have been delisted. I go through the SEC filings of these firms in order to check whether they still persist as dual class public companies as of the beginning of This test identifies 228 public companies that have multiple classes of shares with different voting rights as of January Next, Panel B reports 34 dual class companies that reclassified their multiple classes of stock into a single class over Out of these unifications, 2 were due to the firms being acquired by other companies. 5 companies unified their stock as part of bankruptcy proceedings. I test my hypotheses on dual class unification using a final sample that consists of the remaining 27 companies. Before moving on to hypothesis testing, Panel C breaks down the dual class unifications by year. There were no unifications in 1997 or Seventy percent of the unifications were undertaken during the first four years of the sample period, i.e., from 1993 through Control Changes 5 NYSE did not allow the listing of dual class shares prior to However, there were 4 dual class issues in the period on NYSE.

10 The first reason I propose for unification is the controlling shareholder s desire to sell the company or her control block. Since dual class structure is the most effective anti-takeover protection, elimination of this structure would signal to the market that the control of the company is up for grabs. Therefore, I expect to observe a high frequency of ownership transfer following the unification. Consistent with this, percent of my sample firms (8 out of 27) were subsequently acquired by other firms within 5 years after the unification. 6 Table 2 reports a breakdown of the sample by whether they were acquired and whether the controlling shareholder was a family. There were a total of 12 family firms (44.44%) 7 and 4 of them were subsequently acquired. 8 For the remaining 8 companies the controlling family did not relinquish control even after the unification. The proportion of family firms that were acquired is slightly higher than the proportion of non-family firms that were acquired (33.33 percent vs percent). Therefore, there is some evidence that the desire to sell may be a more prominent reason for the unification of dual class firms that are controlled by families. Table 2 also reports evidence that out of the 19 sample firms that were not acquired following the unification, one of them experiences a change in management and in seven cases management reduces their vote ownership which may further signal their intent to transfer the control of the firm. In sum, the control transfer hypothesis seems to be a potential explanation for the unification of 16 of my sample firms. 4. Cost of Equity 6 For 8 sample firms (including 2 that were acquired) 5 years haven t elapsed between the unification date and the cut-off date, April 1 st This high percentage is consistent with the evidence in DeAngelo and DeAngelo (1985). 8 There was one firm in which the family sold their shares before the unification. I did not identify this firm as a family firm since the family was not in control at the time of unification.

11 Agency literature suggests that dual class companies suffer from agency costs. One dimension of these costs may be a higher cost of equity faced by these companies. To the extent that elimination of the dual class structure alleviates, if not eliminate, the agency costs, the unified firm will face a lower cost of equity. Once a lower cost of equity is attained, these firms will increase their mergers and acquisitions activities. They will also boost their investments and raise more equity. In order to test the cost of equity hypothesis, I first compare the acquisition activity of the sample firms before and after the unification. For thirteen of my sample firms I report acquisitions five years before and after the unification. For the remaining fourteen the window is the time between unification month and April 2002 or acquisition month, whichever comes first. Table 3 Panel A summarizes the acquisition activity for the whole sample. My sample firms perform a total of 42 acquisitions before the unification as compared to 118 after the unification. So, in terms of numbers the acquisition activity almost triples after the unification. Moreover, the value of the acquisitions increases to a total of $8 billion from $3 billion. Both are consistent with the hypothesis that these firms were able to increase their acquisition activity once they lowered their cost of capital by unifying their stock. Also in Panel A is the number of firms that issued equity before and after the unification. 11 sample firms issued equity during the pre-unification period and 13 sample firms issued equity during the post-unification period. The difference is not significant. The cost of equity hypothesis would imply a significant increase in the number of firms issuing equity following the unification. Next, I focus on the sub-sample of 19 firms that were not acquired subsequent to the unification. Panel B shows that the results for the sub-sample are similar to the results for the

12 whole sample. 37 acquisitions are undertaken by the sample firms before the unification, which is almost one-third of the number (102) after unification. The total value of acquisitions again increases from around $3 billion to $5 billion. Number of equity issuing firms is nine for both before and after the unification. The third implication of the cost of equity hypothesis is an increase in the investment activities following the recapitalization. I use two measures of investment activity: capital expenditures and property, plant, and equipment. I scale both measures by total assets. The comparisons are reported in Table 4 Panel A. Median capital expenditure (property, plant, and equipment) is 5.82% (6.93%) before and 4.66% (4.74%) after. However, the 3-digit SIC median adjusted ratio is 0.11% (1.31%) before and 0.13% (0.15%) after. So, there is some evidence that firms increase their capital expenditures and property, plant, and equipment after the unification. The results for growth are stronger. Median growth in capital expenditures increases from 6.56% (4.47%) to 6.88% (5.85%) and industry-adjusted median growth goes up from 21.29% ( %) before the reclassification to 39.33% (39.33%) after the reclassification. 9 In sum, firms not only seem to be investing more heavily in the post-unification period but also the rate at which they increase their investments is also higher. Both results are consistent with the cost of equity hypothesis. Once again, I repeat the analysis after eliminating the acquired sample firms. As Panel B demonstrates, the results are qualitatively unchanged. Industry-adjusted capital expenditures (0.09% vs. 0.34%) and their growth (-12.39% vs %) are higher in the period after the unification. Industry-adjusted property, plant, and equipment figures are lower (0.78% vs. 0.20%) 9 Capital Expenditures (COMPUSTAT Data Item 128) and Property, Plant, and Equipment (COMPUSTAT Data Item 30) are identical for most firms in recent years.

13 in line with the whole sample. Their growth, however, is again higher after the recapitalization ( % vs %). Overall, the results are consistent with the cost of equity hypothesis. Firms do increase their acquisition and investment activity following the unification. There is also an insignificant increase in the number of firms raising equity. 5. Liquidity The final reason I consider for the unification of dual class companies is the desire to increase the liquidity of their stock. Dual class shares may have low liquidity and this may depress the stock prices. Moreover, if managers wanted to sell off their shares at large quantities, they would be able to do this more easily in a liquid market. In order to test the liquidity hypothesis, I compare liquidity of shares before and after the unification. I use trading turnover as a measure of liquidity. I define turnover as daily trading volume divided by the number of shares outstanding. The pre-unification period is 250 through 11 days before the unification announcement and the post-unification period is 11 through 250 days after the day unification actually takes place. Table 5 confirms that the liquidity increases following the unification. The average turnover is 2.79 before the unification and 3.02 after the unification. However, the difference is not statistically significant. The same is true for median turnover, 1.91 vs So, there is little evidence that the liquidity of shares increases following the unification of dual class shares into a single share class. I reported in Table 2 and mentioned in Section 3 that eight sample firms insiders reduced their shareholdings during the five years following the unification consistent with the cost of

14 equity hypothesis. This observation is also consistent with the liquidity hypothesis since with a more liquid market for their shares these managers were better able to sell their shares. 6. Market Reaction I implement a traditional event study to investigate the market s reaction to the reclassification. First, I identify the announcement dates of reclassifications as event dates. Second, I obtain daily returns for my sample firms and CRSP Value-Weighted Index returns. Third, I define the estimation period from 250 days prior to the announcement till 11 days prior to the announcement. I estimate the market model parameters for each firm during the estimation period: R it = α i + β Rmt + ε it, t = [-250, -11] i where R it is the return on stock i for day t, α i is the constant term for stock i, β i is the beta for stock i, R mt is the value-weighted index return on day t, and ε it is the error term. The three-day window (the day before, event day, and the day after) surrounding the event day is the event period. In order to find the abnormal returns during the event period, I use the estimated alpha and beta from the estimation period. AR it = Rit ˆ α ˆ i β Rmt, t = -1, 0, 1 i where AR it is the abnormal return on stock i for day t. I also compute the cumulative abnormal return over the event period: CAR i = ARit. 1 t = 1 Table 6 reports the event study results. Cumulative abnormal return over the event period is 5.32% and it is significantly positive at the 5% level. This means that on average market reacts positively to dual class unification announcements. In other words, investors perceive the reclassifications to be consistent with shareholder wealth maximization.

15 Finally, I would like to check the robustness of my results to event-induced variance. I utilize the standardized cross-sectional test developed by Boehmer et al. (1991). First, I compute the standardized residuals: SR it = sˆ i T i AR T i ( R k = 1 it ( R mt R mk m R ) 2 m ) 2 t = -1, 0, 1 where SR it is the standardized residual of stock i on day t, ŝ i is the estimated standard deviation of abnormal returns to stock i during the estimation period, T i is the number of days in the estimation period of stock i, and R m is the average market return over the estimation period. Then the test statistic is given by: t BMP = 1 N SR N i= 1 1 N ( SR N( N 1) i 1 N SRi) N i i = 1 = 1 i 2 where N is the number of firms in the sample. Boehmer et al. (1991) show that this statistic is not affected by event-induced variance changes. For my sample this statistic is 2.37, which indicates that the significant positive abnormal returns observed upon announcements of dual class unifications are robust to changes in the variance induced by the announcement. 7. Summary and Conclusions I examine a sample of dual class companies that go through a stock unification. I test three hypotheses regarding the motivations for these recapitalizations. Change of control hypothesis puts forward the desire to sell the firm or a control block as the primary motive for unifying the dual class shares. Cost of equity hypothesis posits that the reason for unification is

16 to decrease the cost of equity. Liquidity hypothesis conjectures that dual class managers hope to increase the liquidity of their stock by eliminating the dual class structure. First, I find evidence consistent with the change of control hypothesis. Almost 30 percent of the sample firms experience control changes. Another 26 percent observe their managers shareholdings decrease in time which may be interpreted as them relaxing their grab on control and signaling this to the market for corporate control. This reduction in managerial shareholdings is consistent with a liquid market after the unification as implied by the liquidity hypothesis. Second, I report evidence generally consistent with the implications of the cost of equity hypothesis. Following the unification, dual class firms not only increase their acquisition activities but also expend more on investments. However, there is no evidence that more firms issue equity in the post-unification period as compared to the pre-unification period. Third, I find that there is an insignificant increase in liquidity following the unification. This lack of evidence casts doubt on the liquidity hypothesis. However, there may still be some scope for this hypothesis. There is the fact that managers were able to decrease their shareholdings after the unification. There is also evidence that the market, on average, reacts positively to dual class recapitalization announcements. The increase in stock price is consistent with the investor perception that the illiquidity caused by the dual class structure was depressing the share price. One last thing to keep in mind is that the liquidity of the superior voting shares definitely increased for most cases since there was no public market for them prior to the reclassification. In summary, my evidence is more consistent with the change of control and cost of equity hypotheses than with the liquidity hypothesis. Controlling shareholders signal their intent to sell the firm by eliminating the most effective anti-takeover measure, dual class structure. They also

17 signal a commitment to lower agency costs, which will hopefully be reflected in a lower cost of equity. The quest for enhanced liquidity on the part of the dual class managers, however, does not seem to be successful.

18 References Amoako-Adu, Ben and Brian F. Smith, Dual Class Capitalization and Its Capital Structure Effect, Working paper, Wilfrid Laurier University, Waterloo, Ontario, Canada. Amoako-Adu, Ben and Brian F. Smith, 2001, Dual Class Firms: Capitalization, Ownership Structure and Recapitalization Back into Single Class, Forthcoming in Journal of Banking and Finance. Ang, James S. and William L. Megginson, 1990, Restricted Voting Shares, Ownership Structure and the Market Value of Dual Class Firms, Journal of Financial Research 12, Bhide, A., 1993, The hidden costs of stock market liquidity, Journal of Financial Economics 34, Boehmer, Ekkehart, Jim Musumeci, and Annette B. Poulsen, 1991, Event-study methodology under conditions of event-induced variance, Journal of Financial Economics 30, Chang, S. and D. Mayers, 1992, Managerial vote ownership and shareholder wealth, Journal of Financial Economics 32, Cornett, M. and M. Vetsuypens, 1989, Voting rights and shareholder wealth, Managerial and Decision Economics 10, Dann, L. Y. and H. DeAngelo, 1988, Corporate financial policy and corporate control: A study of defensive adjustments in asset and ownership structure, Journal of Financial Economics 20, DeAngelo, Harry and Linda DeAngelo, 1985, Managerial Ownership of Voting Rights: A Study of Public Corporations with Dual Classes of Common Stock, Journal of Financial Economics 14, Easton, Peter, Gary Taylor, Pervin Shroff, and Theodore Sougiannis, 2000, Empirical Estimation of the Expected Rate of Return on a Portfolio of Stocks, Ohio University Working Paper. Eckbo, B. Espen and Oyvind Norli, 2000, Leverage, Liquidity, and Long-Run IPO Returns, SSRN Working Paper. Fama, Eugene F. and Kenneth R. French, 1993, Common risk factors in the returns on stocks and bonds, Journal of Financial Economics 43, Fernandez, Pablo, 2001, Valuing Companies by Cash Flow Discounting: Eight Methods and Six Theories, IESE Business School Working Paper, Madrid, Spain.

19 Gardiol, Lucien, Rajna Gibson-Asner, and Nils S. Tuchschmid, 1997, Are liquidity and corporate control priced by shareholders? Empirical evidence from Swiss dual class shares, Journal of Corporate Finance 3, Garrod, Neil and Aljosa Valentincic, 2001, Empirical Estimation of the Term Structure of Implicit Discount Rates in Security Valuation, University of Glasgow Working Paper. Gebhardt, William R., Charles M. C. Lee, and Bhaskaran Swaminathan, 2000, Toward an Implied Cost of Capital, Forthcoming in Journal of Accounting Research. Gode, Dan and Partha Mohanram, 2001, What Affects the Implied Cost of Equity Capital?, NYU Working Paper. Grossman, Sanford J. and Oliver D. Hart, 1988, One share-one vote and the market for corporate control, Journal of Financial Economics 20, Hanson, Robert C., Moon H. Song, 1995, Managerial Ownership Change and Firm Value: Evidence from Dual Class Recapitalizations and Insider Trading, Journal of Financial Research, Vol. 18, Issue 3, Harris, Milton and Artur Raviv, 1988, Corporate Governance: Voting rights and majority rules, Journal of Financial Economics 20, Hauser, Shmuel and Beni Lauterbach, 2000, The Value of Voting Rights: Evidence from Dual Class Stock Unifications, Working paper, Bar-Ilan University, Ramat Gan, Israel. Himmelberg, Charles P., R. Glenn Hubbard, and Inessa Love, 2002, Investor Protection, Ownership, and the Cost of Capital, SSRN Working Paper. Holmstrom, Bengt and Jean Tirole, 2000, LAPM: A Liquidity-Based Asset Pricing Model, SSRN Working Paper. Horner, M., 1988, The Value of the Corporate Voting Right, Journal of Banking and Finance 12, Jarrell, Gregg A., and Annette B. Poulsen, 1988, Dual-Class Recapitalization as Antitakeover Mechanisms: The Recent Evidence, Journal of Financial Economics 20, Jog, Vijay M. and Allan L. Riding, 1986, Price Effects of Dual Class Shares, Financial Analysts Journal 42, Jog, Vijay M. and Allan L. Riding, 1990, A Note on Insider Trading and Issuances of Restricted-Voting Common Shares, Journal of Business Finance & Accounting, Vol. 17, Issue 3,

20 Jog, Vijay M., Allan L. Riding, and P. Seguin, 1985, Shareholders Reactions to Issuances of Restricted Voting Common Stock, Administrative Sciences Association of Canada Proceedings, Finance Division, Jog, Vijay M., Ashwani Srivastava and Madhuri Panangipalli, 1996, Evidence from Dual Class Firms Did the Shareholders Win?, Canadian Investment Review, Winter, Lease, Ronald C., John J. McConnell, and Wayne H. Mikkelson, 1983, The Market Value of Control in Publicly Traded Corporations, Journal of Financial Economics 11, Levy, H., 1982, Economic Valuation of Voting Power of Common Stock, Journal of Finance 38, Partch, Megan, 1987, The Creation of a Class of Limited Voting Common Stock and Shareholder Wealth, Journal of Financial Economics 18, Robinson, Chris and Alan White, 1990, Empirical Evidence on the Relative Valuation of Voting and Restricted Voting Shares, Canadian Journal of Administrative Sciences 7, Rydqvist, K., 1987, Empirical Investigation of the Voting Premium, Northwestern University, Working paper 35. Zingales, Luigi, 1994, The Value of the Voting Right: A Study of the Milan Stock Exchange Experience, Review of Financial Studies 7, Zingales, Luigi, 1995, What Determines the Value of Corporate Votes?, Quarterly Journal of Economics,

21 Table 1 Summary Sample Statistics This table reports summary statistics for sample firms. I identify dual class companies from CRSP and crosscheck with their SEC filings. Panel A: Breakdown of Dual Class Companies by Year and Exchange Period Number of Dual Issues NYSE AMEX NASDAQ * ** Total 349* Dual as of January * Two of them were not identified. ** All of them were after 1986 due to NYSE listing rules. Panel B: Reasons for Unification Unifications over Due to Acquisitions 2 Due to Bankruptcy 5 Other 27 Panel C: Breakdown of Unifications by Year Year Total Unifications

22 Table 2 Acquisitions of Sample Firms This table gives information regarding the acquisitions of sample firms. Family denotes whether a family is in control of the firm. Management Change denotes whether the controlling shareholder changed over the five years after unification. Decrease in Management Ownership is again over the five years following the unification. Whole Sample (27 firms) Family Non-Family Total Acquired Not Acquired Total Not Acquired Sub-Sample (19 firms) Family Non-Family Total Management Change Decrease in Management Ownership Total 3 5 8

23 Table 3 Acquisitions by Sample Firms This table compares acquisition activity for sample firms before and after unification. Panel A Whole Sample (27 firms) Before Unification After Unification Acquiror Firms Acquiror Firms (only after) - 8 Acquiror Firms (only before) 2 - Acquisitions (number) Acquisitions (dollar amount, in millions) Number of Acquisitions higher 3 15 Dollar Amount of Acquisitions higher 4 16 Issue Equity Panel B Not Acquired Sub-Sample (19 firms) Before Unification After Unification Acquiror Firms Acquiror Firms (only after) - 4 Acquiror Firms (only before) 1 - Acquisitions (number) Acquisitions (dollar amount, in millions) Number of Acquisitions higher 1 11 Dollar Amount of Acquisitions higher 3 11 Issue Equity 9 9

24 Table 4 Investment Activity This table reports comparisons of median Capital Expenditure (scaled by total assets), median growth in Capital Expenditure, median Property, Plant, and Equipment (scaled by total assets) and median growth in Property, Plant, and Equipment before and after the unification. Panel A Whole Sample (27 firms) Before Unification After Unification Median Capital Expenditure (raw) (ITEM128/ITEM6) 5.82% 4.66% Median Capital Expenditure (industry-adjusted) 0.11% 0.13% Median Growth in Capital Expenditure (raw) 6.56% 6.88% Median Growth in Capital Expenditure (industry-adjusted) 21.29% 39.33% Median Property, Plant, and Equipment (raw) (ITEM30/ITEM6) 6.93% 4.75% Median Property, Plant, and Equipment (industry-adjusted) 1.31% 0.15% Median Growth in Property, Plant, and Equipment (raw) 4.47% 5.85% Median Growth in Property, Plant, and Equipment (industry-adjusted) % 39.33% Panel B Not Acquired Sub-Sample (19 firms) Before Unification After Unification Median Capital Expenditure (raw) (ITEM128/ITEM6) 5.78% 4.57% Median Capital Expenditure (industry-adjusted) 0.09% 0.34% Median Growth in Capital Expenditure (raw) 2.37% 6.67% Median Growth in Capital Expenditure (industry-adjusted) % 32.49% Median Property, Plant, and Equipment (raw) (ITEM30/ITEM6) 7.03% 4.77% Median Property, Plant, and Equipment (industry-adjusted) 0.78% 0.20% Median Growth in Property, Plant, and Equipment (raw) 0.41% 5.03% Median Growth in Property, Plant, and Equipment (industry-adjusted) % 32.49%

25 Table 5 Trading Turnover This table compares trading activity 250 through 11 days before the unification announcement date and 11 through 250 days after the actual unification date. Turnover is defined as daily volume divided by number of shares outstanding. t-test is the t-value from a t-test of unequal variances. Before Unification After Unification t-test Average Turnover Median Turnover

26 Table 6 Market Reaction This table reports the results of an event study performed to find out the market reaction to unifications. Estimation period is 250 through 11 days before the unification announcement date. Event period is the day before the announcement date (Day -1), announcement day and the day after the announcement date (Day +1). t BMP is the standardized cross-sectional test statistic from Boehmer, Musumeci, and Poulsen (1991). Abnormal Return N Mean Median Std Error t p t BMP Day % 0.23% 4.09% Announcement Day % 1.68% 2.19% Day % -0.03% 1.53% Cumulative % 0.87% 2.47%