European corporate governance - Semantic Scholar

European Economic Review 43 (1999) 1071}1083

European corporate governance: Trading o! liquidity against control Marco Becht ECARE and Solvay School of Business, Universite& Libre de Bruxelles, 21 avenue F.D. Roosevelt, 1050 Bruxelles, Belgium

Abstract Ownership dispersion is a pre-requisite for liquid stock markets, but it entails a collective action problem: individual investors have no incentives to engage in direct monitoring. Legal devices can provide solutions along three dimensions. One, they can concentrate or dilute voting power. Two, they can a!ect liquidity. Three, they can give the right or wrong monitoring incentives. This paper shows how these devices are used and how they can depress liquidity. Legal constraints aimed at strengthening minority protection can reduce the scope for monitoring, destroy liquidity and even create incentives for minority abuse: for example one-share-one-vote restrictions encourage the formation of pyramidal holding companies. The search for solutions that concentrate voting power, provide liquidity and protect minorities continues. 1999 Elsevier Science B.V. All rights reserved. JEL classixcation: G10; G32; G34 Keywords: Stock market liquidity; Corporate governance; Shareholder voting power; One-shareone-vote

1. Introduction Corporate governance is currently a subject of great international concern and debate. The rising interest is market driven. Funds want to invest in liquid stock markets and to diversify internationally. Governments need to capitalise their pension systems and "nd investors who are willing to fund their 0014-2921/99/$ } see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 4 - 2 9 2 1 ( 9 8 ) 0 0 1 1 5 - 9

1072

M. Becht / European Economic Review 43 (1999) 1071}1083

privatisation programmes. Companies with global ambitions want to access global equity capital. Banks and insurance companies are subject to global solvency requirements; unless they have su$cient retained earnings, they must seek outside equity "nance to expand. A certain degree of ownership dispersion is a pre-requisite for liquid stock markets. Stocks must be held by a diversity of investors and they must be willing to trade. However, the spread of ownership gives rise to problems in taking collective action. Individual investors are highly unlikely to have su$cient voting power to exert control over the company they help to "nance. Investors could exert control jointly, but they do not usually have the incentives to do so. The ideal solution to this problem has three elements. One, it must provide a mechanism that concentrates voting power and, thereby, allows for direct monitoring. Two, the concentration of voting power must not tie up large share blocks and destroy liquidity. Three, those who command the voting power must have the right incentives. It appears that a focus of the current corporate governance discussion is the world-wide search for the best solution to this problem. The European Corporate Governance Network (1997) has collected evidence on voting power concentration for 7 European countries and the United States showing that there are di!erences in blockholding patterns both between the United States and Europe and within Europe (Barca and Becht, 1999). Voting power in continental Europe is often highly concentrated, but the associated legal arrangements and incentives di!er. We observe blocks where substantial cash-#ow stakes give substantial voting power; blocks where small cash-#ow stakes give high levels of voting power; and blocks where large cash-#ow stakes give little voting power (for example through voting caps). When we can observe its e!ect (it is generally not disclosed), proxy voting power substitutes for voting power in blocks. In some cases, voting power concentration does not adversely a!ect liquidity. These facts need to be explained. Formal theoretical models provide little guidance as yet, but there are informal theories that might explain the observed di!erence in blockholding patterns. La Porta et al. (1997, 1998, 1999) argue that di!erences in shareholder protection can explain ownership concentration rates (in terms of cash-#ow rights), control arrangements (in terms of voting rights) and the size of debt and equity markets around the world. Roe (1994) argues that (over-)regulation can make it very expensive to hold blocks or prevent institutions from holding blocks at all. He documents that this is the case in the United States. Bhide (1993) shows how such regulation is linked to corporate governance and liquidity. Co!ee (1991) argues that many institutional investors in the United States deliberately trade o! control against liquidity. The purpose of this paper is to add to this debate by linking ownership, voting power, liquidity and the legal system from a European perspective. The legal


1073

system creates a choice set for founders and blockholders. The set is de"ned, and constrained, by company law, securities regulation, listing requirements and business practice. For example, when company law or listing requirements impose &one-share-one-vote', shareholders have to rely on ownership blocks permanently to concentrate voting power and on pyramids to separate cash#ow rights from voting rights. The degree of separation between ownership and voting power matters for blockholder incentives. The legal instruments used to separate ownership and voting power matter for liquidity. Successful proxy solicitations by executive directors do not a!ect liquidity, but concentrated voting power is not used for direct monitoring. Large ownership blocks provide the right monitoring incentives for blockholders but they have a negative impact on liquidity. Dual class shares with multiple voting rights have a small negative impact on liquidity and provide the wrong blockholder incentives. Pyramids have a large negative e!ect on liquidity and they provide the wrong incentives. The theoretical literature, so far, has not investigated the interaction of all the associated trade-o!s. Section 2 motivates the analysis using the ownership and control structure of the Dutch ING Groep N.V. Section 3 investigates the link between voting power, ownership and liquidity with German and Belgian data. Section 4 concludes and discusses potential policy implications.

2. Ownership, voting power and liquidity The Netherlands has low levels of minority protection, a large stock market and sizeable domestic pension funds. The country shares a French legal origin and one o$cial language with neighbouring Belgium, and a two-tier board system with Germany. The Netherlands is also one of the few countries where we can observe voting blocks and cash-#ow stakes simultaneously. ING Groep N.V. is a representative example of the way large Dutch "rms combine liquidity

For example, in the control chain A}51%PB}51%PC the voting power of A in C is 51% but the cash-#ow rights of A in C is 26.01%. Thus pyramiding makes it possible to concentrate control rights without concentrating cash-#ow rights even when &one-share-one-vote' is imposed on direct ownership claims. In the previous example A's incentives are misaligned with those of minority shareholders of C and the direct stake of B in C ties up 51% of C's capital. If 100% of C's capital is listed, 51% of C's shares are unavailable for trading. For example, Burkart et al. (1997) and Bolton and Von Thadden (1998) analyse monitoring and liquidity trade-o!s respectively, but they do so in a &one-share-one-vote' setting. On the other hand, the literature that investigates deviations from &one-share-one-vote' has been formulated in the context of takeovers, which are mostly absent in continental Europe; see for example Grossman and Hart (1988). For a general analysis of the Dutch corporate governance system see De Jong et al. (1999).

1074


and concentrated voting power. ING Groep N.V. was founded in 1991 through the merger of two prominent Dutch "nancial institutions, Nationale Nederlanden N.V. and NMB Postbank Groep. It is one of the most dynamic "nancial institutions in Europe and has been expanding rapidly over the last few years. From 1993}1997 the company's capital base has grown from NLG 21.5 billion to 46.1 billion and its market capitalisation from approximately NLG 25 billion in 1993 to just over NLG 100 billion in March 1998 (ING 1997, Annual Report). On 30 June 1998, ING was the world's 56th largest public company in terms of market capitalisation, up from 73rd in 1997 (as ranked by The Wall Street Journal Europe on 28 September, 1998), and the third largest company on the Amsterdam Stock Exchange. Fig. 1 shows the control structure of the ING Group. ING has issued four classes of registered voting shares that are deposited in an &Administratiekantoor' or trust o$ce. The trust o$ce is bound by a trust agreement that de"nes its &"duciary duties'. The management board of the trust has seven members, of whom two are appointed by the supervisory board of ING and "ve are appointed by the board of the trust itself, subject to approval by the ING executive board (SEC disclosure, Form 20-F, 1997). The supervisory board of ING NV appoints itself and the executive board. The trust o$ce has issued non-voting depositary receipts for ordinary shares which are traded in Amsterdam, Brussels, Frankfurt, Paris, Switzerland and the New York Stock Exchange. Depositary receipts for preference shares are traded in Amsterdam. Depositary receipts are held by four institutional investors with 5}7% each and by small shareholders; 48% of the depositary receipt holders reside in the Netherlands, 18% in the United Kingdom, 16% in the United States and Canada and 18% in other countries (ING 1997 Annual Report). The ING control structure has several features that do not "t well into our usual measurement framework and informal theories of law and "nance: E The sum of the ownership holdings of the largest "ve shareholders is 23.1%. Four blocks of 5}7% are held by institutional investors who include a Dutch pension fund. Hence, the ownership structure of ING is very similar to that of U.K. and U.S. listed companies. To use the voting rights &in such a manner that (i) the interests of ING Group N.V2 are served (ii) the interests of ING2 and all parties concerned are safeguarded as well as possible; and (iii) in#uences which could violate the independence, the continuity or the identity of ING2 are barred to the greatest possible extent' (ING 1997, SEC disclosure Form 20-F, p. 60). In response to the 40 recommendations of the Peters Committee (1997), ING issued a 10-page position paper on corporate governance. A full page is devoted to the depositary receipts issue: &&Certain companies operating in the same "eld as ING Group have adopted a di!erent shareholding structure in order to ensure continuity and balanced decision making at shareholders' meetings. ING does not have such a structure and therefore places great value on its ability to issue depositary receipts for shares''.


1075

Fig. 1. Ownership and voting power in ING Groep NV. Source: Articles of Association (May 1998) and Het Financieele Dagblad. Note: Ownership (cash-#ow) blocks and ultimate voting blocks are disclosed separately in the Netherlands. The company has a two tier board. Members of the executive board and employees of the group cannot be members of the supervisory board (Article 26). The composition of the board of the trust o$ce is not disclosed. However, the administration company can only issue depositary receipts with the approval of the executive board and the supervisory board (Article 18). The members of the supervisory board appoint themselves. The company is managed by an executive board (Article 19). The members of the executive board are appointed and dismissed by the supervisory board (Article 20). Share transfers to legal or natural persons holding more than 1% (before or after the transfer) are subject to approval by the supervisory board (Articles 8 and 9). A pledge or and usufruct can be established on shares (Articles 11 and 12). Many motions and resolutions by the executive board, including a motion to modify the articles of association are subject to approval by the supervisory board (Article 23). The remuneration of the supervisory board is determined by the General Meeting of Shareholders (Article 23; in practice by the supervisory board).

1076


E The 100% voting block controlled by the ING trust o$ce is, obviously, larger than most voting blocks in the rest of continental Europe (see Becht and RoK ell, 1999). E Ownership and voting power are completely separated through the trust o$ce. The owners of the company, when de"ned as those who hold the depositary receipts, have no voting power. E There is a revealed preference for control. However, because of the trust company instrument, liquidity does not have to be traded o! against control. E ING is controlled by a self-appointing board that represents the perpetuated will of the founders. The board of the trust company has clearly de"ned &"duciary duties', but it cannot be removed by the certi"cate holders. Also, like in the United States, managers and directors cannot be sued for incompetence. E There have been no indications of minority abuse at ING. The company is well managed and investors from the United States, the United Kingdom and Canada do not appear to be deterred by the ING control structure. How can we explain these facts? The minority protection argument predicts the opposite of what we observe. In La Porta et al. (1998) the anti-director index for the Netherlands, the author's indicator of shareholder protection, is very low (2 out of 6). Poor shareholder protection should be associated with concentrated ownership, large outside voting blocks, and low global market capitalisation. Yet, ownership concentration is low, there are no outside voting blocks, and the Amsterdam stock market is large. Roe's (1994) political economy explanation could be made to "t. He might argue that the powerful insiders successfully lobbied for the current system. The founders of ING could have chosen for one-share-one-vote, but they did not. They have chosen the trust company route. Concentrated voting power does not come at the price of less liquidity. If the Netherlands had imposed one-share-one-vote by law we might have observed concentrated voting power, concentrated ownership and low levels of liquidity. How well the self-appointing supervisory boards represent the interests of minority shareholders, in this case all the shareholders, and who removes them when they err, is a central issue in the Dutch corporate governance debate (see Peters Committee, 1997). The average for English legal origin countries is 4 and for French legal origin countries 2.33; the index ranges from 0}6. RoK ell (1996, p. 1072) reports a market capitalisation to GNP ratio of 0.76 for the Netherlands, as large as the NYSE, NASDAQ and AMEX combined. The Netherlands shares this feature with Chile (La Porta et al., 1997), a French legal origin country with strong minority protection (5/6) and a capitalised pension system. ING was the third largest company on the Amsterdam Exchange in 1997. Indeed, the political economy explanation could probably explain any governance arrangement, a serious de"ciency of this approach.


1077

3. Blocks and liquidity This section empirically investigates the link between blockholdings and liquidity in Belgium and Germany. For the most liquid segment of both markets, there is a negative relationship between the number of shares tied up in blocks and liquidity. German companies use a variety of legal devices that allow them to concentrate voting power without concentrating ownership. Belgian companies do not. German liquidity levels are higher than in Belgium. In Belgium, the separation of ownership of ultimate cash-#ow rights and voting power is small for companies with only "rst level blocks and large for companies that are controlled through pyramidal structures. The interests of blockholders and minority shareholders are always misaligned in Germany. In Belgium they are mainly misaligned when control is exerted via pyramids. 3.1. Liquidity regressions For the United States, there is extensive empirical evidence going back as far as Demsetz (1968) that the number of shareholders is positively related to liquidity. I provide recent evidence from Germany and Belgium. Table 1 reports regressions of liquidity, measured by annual turnover divided by market capitalisation, on the percentage of traded shares that are tied up in voting blocks (to be precise, the sum of all disclosed voting blocks of 5% or over). The German sample includes the 100 largest and most liquid German listed companies, the DAX100. The Belgian sample includes the 70 prominent companies listed in Banque Degroof (1996). The regression analysis con"rms that blocks absorb liquidity. More voting power through blocks must be paid for with a loss in liquidity. What are the magnitudes? For Germany, the 1996/97 regression predicts that a block of 100% is associated with a turnover/market capitalisation ratio of 0%; a block of 4.99% with a liquidity level of 319%. A decrease in block size from 60% to 10% doubles the liquidity level (from 128% to 260%). The relationship in 1998/97 is similar, but liquidity levels during 97/98 were much lower; between 6}127%, as opposed to 32}527% the year before.

Deutsche BoK rsen AG and the Brussels Exchange have di!erent turnover de"nitions. The German exchanges take a &Regulated Environment View', the Brussels Exchange takes a &Trading System View'. The Copenhagen, Madrid and Paris exchanges publish turnover "gures taking both views. From January}July 1997 the "gures for the REV are about 3}4 times higher than for the TSV (Federation of European Stock Exchanges, European Stock Exchange Statistics, July 1997). Extrapolating from these "gures, the di!erent turnover de"nitions do not account for the observed di!erences in turnover.

1078


Table 1 Liquidity and blockholding regressions Germany 96/97

Logit (Sum Non-Traded Shares %) Constant term

Number of observations Prob'F R (MSE

Germany 97/98

Coef. (Std. Err.)

t (P'"t")

Coef. (Std. Err.)

!0.271

!8.56

!1.926

(0.032) 0.362 (0.049)

(0.00) 7.39 (0.00)

(0.041) !1.024 (0.640)

99 0.00 0.40 0.43

99 0.00 0.20 0.58

t (P'"t")

Belgium 95/96 Coef. (Std. Err.)

t (P'"t")

!4.69

!0.914

!5.43

(0.00) !16.0 (0.00)

(0.168) !1.972 (0.119)

(0.00) !16.5 (0.00)

70 0.00 0.38 0.78

Note: The German data set, provided by Deutsche BoK rsen AG, contains blockholdings, turnover and market capitalisation for all traded stock for the DAX100 companies. Deutsche BoK rsen AG uses this data to determine who is included in the DAX indices. The composition of the DAX100 index changed from 1996/97 to 1997/98. The original explanatory variable was the percentage of traded shares tied up in blocks. In the German case, when non-voting stock is traded but voting stock is not, only the liquidity and blockholding for the non-voting stock was included in the data set. The disclosure threshold is 5% for voting stock. Blockholding information is generally not disclosed for non-voting stock which is assumed to be dispersed. All observations smaller than 5% were set to 4.99%. This was the case for 10 observations in the 1996/97 German data and 24 times in the 1997/98 data. No other attempt was made to address this censoring problem. The observations were transformed using a standard logit transformation [ln(X/(100!X))]. The dependent variable is turnover in DM (7/96}6/97; 7/97}6/98) divided by market capitalisation (30/6/97, 30/06/98). For Belgium the data was obtained from Banque Degroof (1996). The dependent variable is yearly turnover divided by market capitalisation in June 1996. Almost no Belgian companies issue non-voting stock. There were two inconsistencies in the Belgian data. Becht et al. (1997) found that there are no known 5%#blockholders for Delhaize. Banque Degroof reports that 38% of the shares are tied up in blocks held by the family. This is well known in Belgian investor circles and this observation was left unchanged. For CONFINIMMO the &free #oat' "gure was 100%, yet 43.9% of the shares were noti"ed as tied up in blocks. The latter value was used. The regressions were performed using the Stata procedure reg with the robust standard errors option. The range of the German (Turnover/Market Capitalisation) variable was 32.1}526.8% (mean 185.7%); 4}52.4% (mean 21.4%) for Belgium. The range of the percentage of shares tied up in blocks was 4.9}82% (mean 42.4%) for Germany, after the censoring correction; 22.5}87.4% (mean 60.5%) for Belgium. Dependent Variable: ln(Turnover/Market Capitalisation), ordinary least squares regression with robust standard errors.

For Belgium the same regression was run for 1995/96 data. Again, liquidity is found to be inversely related to blockholdings: when 90% of the shares are tied up in blocks, the predicted liquidity level is 1.87%; when 20% are tied up, the predicted liquidity level is 49.3%.


1079

3.2. How DAX companies obtain liquidity The DAX includes the 30 largest and most liquid companies in the DAX100; they typically account for 60}70% of market capitalisation and turnover on the German markets (see http://www.exchange.de for a description of the indices). Companies that do not meet the size or turnover criterion are excluded from the index when there are substitutes available that do. Since index tracking funds will not buy the stock of companies that are not in the DAX or DAX100, there is considerable competition for inclusion. Of the 30 DAX companies, 19 have issued voting bearer shares exclusively. The mean voting block for the remaining 11 companies, which have issued non-voting or other stock, is signi"cantly higher than for the former (43% vs. 29.1%). Liquidity considerations might o!er an explanation. By issuing traded non-voting stock, the company can have larger voting blocks and still achieve the amount of turnover needed to stay in the index. However, this does not mean that voting power in the &one-share-one-vote' bearer share group is less concentrated. 16 of the 19 companies in this group rely on bank proxies to leverage voting power. Indeed, Commerzbank and Deutsche Bank combine dispersion (liquidity) and concentrated voting power through the default proxy voting power process and e!ectively control themselves. It is not clear what price the banks &charge' the companies for lending them their proxy votes, or whether they engage in active monitoring, as has been frequently suggested in the corporate governance literature. A second means of securing insider control is used by Allianz AG and Munich Re, who have issued registered shares that are subject to transfer restrictions. All share transfers are subject to approval by the management. The two companies also have a 25% cross-shareholding and rely on additional friendly blocks held by banks to prevent statute changes. On the downside, they each had to tie up more than 45% of the traded shares in order to secure control. Other devices include voting caps, multiple voting rights (RWE and Siemens), non-voting stock (8 companies, often family controlled) and two special cases (Lufthansa and Volkswagen, both subject to special laws safeguarding German control and circumscribing bank voting, respectively). Pyramids are not common amongst DAX companies, possibly due to their liquidity-absorbing e!ect. 3.3. How Belgian companies fail to obtain liquidity Belgian blockholders do not have the German proxy voting instrument at their disposal (Becht et al., 1999). Otherwise their legal choice set is similar.

More detailed evidence on the control of DAX30 companies can be found in Becht and BoK hmer (1998).

1080


In practice listed companies comply with one-share-one-vote and rely on voting pacts or large cash-#ow blocks to concentrate voting power. First-level ownership and voting power are very concentrated and the liquidity cost of this arrangement is high. In the case of highly leveraged pyramids, blockholders' incentives are also misaligned. It is not clear why blockholders do not make more use of the non-voting stock instrument.

4. Implications This paper has provided evidence that voting power concentration through blocks has a negative e!ect on liquidity in the German and Belgian stock markets. In Belgium, blocks are often large and voting power is associated with cash-#ow rights of roughly the same size. The interests of blockholders and minorities are aligned (except in the case of pyramidal control structures), but liquidity is very poor. In Germany the liquidity cost is mitigated because blockholders deviate from one-share-one-vote or rely on bank proxies for concentrated voting power. In these cases, the interests of the blockholders are not aligned with those of minority shareholders. One lesson to be drawn from this analysis is that ownership statistics, even when available, do not tell us everything there is to know about voting power. In the extreme case of the Netherlands, ownership can be completely divorced from voting power. Neither do voting block statistics tell us everything about voting power. In the United States and the United Kingdom, voting power concentrations are often temporary. In the United Kingdom voting at annual meetings is commonly done by hand and not proportional to share ownership (&x-sharesone-hand', see Goergen and Renneboog, 1999). These phenomena are not disclosed and we do not measure them. If we were able to measure these e!ects, we might discover that U.K. and U.S. boards command considerable voting power. The direction of causality in the trade-o!s between outside monitoring, liquidity and minority protection is subject to debate. Does statutory minority protection allow controlling shareholders to reduce their blocks and does this cause stock markets to grow in size and make them more liquid? Or does pension reform cause a demand for stock markets, which creates a demand for liquid stocks, which leads to the search for a solution to the collective action problem, which leads to changes in company law, securities regulation, listing requirements and company statutes? La Porta et al. (1997, 1998, 1999) argue the former and Easterbrook (1997), RoK ell (1996) and others the latter. In itself, minority protection is desirable. However, minority protection laws and regulations can have (at least) two side e!ects. One, they can deter oversight by large blockholders. Two, they can be harmful to liquidity. Two European draft directives are cases in point. The "rst draft 13th Company Law Directive


1081

sought to strengthen minority protection by introducing a mandatory takeover bid requirement for all listed companies; a requirement that was dropped in recent drafts. The mandatory takeover bid requirement would have weakened direct monitoring throughout Europe by reducing voting block sizes. The draft 5th Company Law Directive (now abandoned completely) proposed the imposition of one-share-one-vote on all European corporations. Blockholders would have been prevented from using non-voting stock and similar devices, and would have had to rely on large cash-#ow blocks, voting pacts or pyramids to achieve a desired level of voting power. Unless blockholders had been willing to give up voting power, markets would have become less liquid. If blockholders had adjusted by opting for pyramids, monitoring incentives could have become more distorted than with non-voting stock. Poor liquidity and poor oversight would have resulted. The search for the best way to overcome the &liquidity and monitoring' collective action problem is far from over. In Europe, the single currency, privatisation and pension reform are creating a demand for liquid stock markets. Due to the Investment Services Directive, competition between stock market rulebooks and regulators is a reality. If and when the European Company Statute and/or the 10th Company Law Directive arrive, European companies will be able to freely choose their statutes as well. Nearly all European voting power concentration devices will be at their disposal. The relative merits of legal devices need to be investigated further. Highly leveraged pyramids seem particularly undesirable, as they usually absorb liquidity and provide the wrong monitoring incentives. For founders, the Dutch voting trust device might be a particularly attractive device. However, it bypasses the voting mechanism, usually considered an important feature of the corporation. Nonvoting stock is an intermediate solution. Large ownership blocks absorb liquidity. As cross-border equity investment increases, proxy voting is starting to play a more important role, also in Continental Europe. The proxy voting instrument does not absorb liquidity. But, depending on who controls it, proxy voting can be &a tremendous force for good or evil in our economic system' (Loss, 1988). It appears that the ideal solution to the liquidity-control problem has not yet been found. But at least it is quite clear what we should be looking for: a liquidity neutral monitoring device that protects the interests of all shareholders.

Acknowledgements This paper is part of a research project undertaken by the European Corporate Governance Network, supported by the Directorate General for Industry of European Commission (DGIII) and Fondazione Eni Enrico Mattei. I am

1082


grateful to Patrick Bolton, Mathias Dewatripont, LeH o Goldschmidt, Colin Mayer, Ailsa RoK ell and Eddy Wymeersch for their encouragement and comments, and to Abe de Jong, Banque Degroof, Karsten El-Nemr and Stefanie Heun of Deutsche BoK rsen AG and Andrew Baker of Wedlake Bell for information, explanations and data. ING Group NV, the DAX companies and many other German and Belgian listed companies kindly provided copies of their annual reports and articles of association.

References Banque Degroof, 1996. Belgian Company Pro"les. June, Brussels. Barca, F., Becht, M., (Eds.), 1999. Ownership and control: A European perspective. Forthcoming. Becht, M., 1997. Strong blockholders, weak owners and the need for European mandatory disclosure. In: ECGN (1997), The separation of ownership and control: A survey of 7 European countries. Preliminary Report to the European Commission. Brussels. Becht, M., BoK hmer, E., 1998. Ownership and voting power in Germany. In: Barca, F., Becht, M. (Eds.), Ownership and control: A European perspective. unpublished; available from http: //www.ecgn.org. Becht, M., RoK ell, A., 1999. Blockholdings in Europe: An international comparison. European Economic Review, this issue. Becht, M., Chapelle, A., Renneboog, L., 1999. Shareholding cascades: The separation of ownership and control in Belgium. In: Barca, F., Becht, M. (Eds.), Ownership and control: A European perspective. unpublished; available from http: //www.ecgn.org. Bhide, A., 1993. The hidden costs of market liquidity. Journal of Financial Economics 34, 31}51. Bolton, P., Von Thadden, E.-L., 1998. Blocks, liquidity and corporate control. Journal of Finance 53 (1), 1}25. Burkart, M., Gromb, D., Panunzi, F., 1997. Large shareholders, monitoring and the value of the "rm. Quarterly Journal of Economics 111, 693}728. Co!ee, J.C., 1991. Liquidity versus control: The institutional investor as corporate monitor. Columbia Law Review 91, 1277}1366. De Jong, A., Kabir, R., Marra, T., RoK ell, A., 1999. Ownership and control in the Netherlands. In: Barca, F., Becht, M. (Eds.), Ownership and control: A European perspective. unpublished; available from http://www.ecgn.org. Demsetz, H., 1968. The cost of transacting. Quarterly Journal of Economics 82 (1), 33}53. Easterbrook, F.H., (1997) International corporate di!erences: Market or law? Journal of Applied Corporate Finance 9 (4) 23}29. European Corporate Governance Network (ECGN), 1997. The separation of ownership and control: A survey of 7 European countries. Preliminary Report to the European Commission. Brussels, available from http://www.ecgn.org. Goergen, M., Renneboog, L., 1999. Strong managers and passive institutional investors in the U.K. In: Barca, F., Becht, M. (Eds.), Ownership and control: A European perspective. unpublished; available from http://www.ecgn.org. Grossman, S., Hart, O., 1988. One-share-one-vote and the market for corporate control. Journal of Financial Economics 20, 175}202. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1997. Legal determinants of external "nance. Journal of Finance 52, 1131}1150. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1998. Law and "nance. Journal of Political Economy 106 (6).


1083

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Ownership around the world. Journal of Finance, forthcoming. Loss, L., 1988. Fundamentals of Securities Regulation. 2nd ed., Little, Brown and Company, Boston and Toronto. &Peters' Committee on Corporate Governance, 1997. Corporate Governance in the Netherlands } Forty Recommendations. Amsterdam Stock Exchange, Amsterdam. Roe, M.J., 1994. Strong Managers, Weak Owners. The Political Roots of American Corporate Finance. Princeton University Press, Princeton, NJ. RoK ell, A., 1996. The decision to go public: An overview. European Economic Review 40, 1071}1081.