Corporate Governance, Economic Entrenchment, and Growth

sp05_Article 1 8/18/05 9:32 AM Page 655

Journal of Economic Literature Vol. XLIII (September 2005), pp. 655–720

∗ RANDALL MORCK, DANIEL WOLFENZON, and BERNARD YEUNG

Outside the United States and the United Kingdom, large corporations usually have controlling owners, who are usually very wealthy families. Pyramidal control structures, cross shareholding, and super-voting rights let such families control corporations without making a commensurate capital investment. In many countries, a few such families end up controlling considerable proportions of their countries’ economies. Three points emerge. First, at the firm level, these ownership structures, because they vest dominant control rights with families who often have little real capital invested, permit a range of agency problems and hence resource misallocation. If a few families control large swaths of an economy, such corporate governance problems can attain macroeconomic importance—affecting rates of innovation, economywide resource allocation, and economic growth. If political influence depends on what one controls, rather than what one owns, the controlling owners of pyramids have greatly amplified political influence relative to their actual wealth. This influence can distort public policy regarding property rights protection, capital markets, and other institutions. We denote this phenomenon economic entrenchment, and posit a relationship between the distribution of corporate control and institutional development that generates and preserves economic entrenchment as one possible equilibrium. The literature suggests key determinants of economic entrenchment, but has many gaps where further work exploring the political economy importance of the distribution of corporate control is needed.

1. Introduction Journal of Economic Literature, are relevant as points of departure. The first, reviewed growing body of work indicates that by Philippe Aghion, Eve Caroli, and Cecilia economic growth depends on the distri- A Garcia-Penalosa (1999), discusses reasons bution of control over capital assets. Two why economic inequality might impede vast literatures, already surveyed in the growth. The second, surveyed by Ross ∗ Morck: University of Alberta and National Bureau of Levine (1997), discusses the functional role Economic Research. Wolfenzon: Stern School of Business, of capital markets in stimulating growth. We New York University, and National Bureau of Economic build on these insightful contributions by Research. Yeung: Stern School of Business, New York University. We gratefully acknowledge helpful suggestions reviewing recent research relating the dis- from Mark Casson, Ron Daniels, Mara Faccio, Roger tribution of corporate control with the Gordon, Henry Hansmann, Campbell Harvey, Diana workings of capital markets and showing Kniazeva, Bruce Kogut, Ross Levine, Enrico Perotti, Andrei Shleifer, René Stulz, Paul Vaaler, Belén Villalonga, how these interactions affect economic Marina Whitman, and the referees. growth. 655 sp05_Article 1 8/18/05 9:32 AM Page 656

656 Journal of Economic Literature, Vol. XLIII (September 2005)

The key insight in Aghion, Caroli, and investments. Raghuram G. Rajan and Luigi Garcia-Penalosa (1999) is that perfect capital Zingales (2003), Art Durnev, Morck, and markets make wealth distribution irrelevant Yeung (2001, 2004), and others argue that by allocating capital to each investment capital markets are inimitably, though cer- opportunity until its marginal return equals tainly not perfectly, effective in this regard. the market clearing equilibrium interest This is because, by creating incentives for rate. Simple capital market frictions, like self-interested investors to gather and moral hazard problems, drive a wedge process information, as well as to monitor between internal and external cost of capital. and control top corporate insiders, capital This can lead the wealthy to invest more markets generate conditions amenable to than is first best optimal, while the impecu- efficient capital investment decisions. nious invest less. Assuming diminishing rates Corporate insiders needing outside capital of return, redistributing wealth from the rich must submit to this analysis, monitoring, and to the poor directly raises total output. control by outside investors. When official Moreover, wealth redistribution from rich to regulations, laws, and enforcement make poor lessens overall borrowing in the econo- outside investors sufficiently confident of my, easing these moral hazard problems, their ability to undertake these tasks, outside raising overall production, and reducing the investors hold equity. Pooling their capital cost of debt financing on the margin under the management of institutional (Aghion, Caroli, and Garcia-Penalosa 1999). investors allows small investors to attain Levine (1997) surveys a second literature economies of scale in accomplishing these on the functions of capital markets in eco- tasks. The resulting optimally efficient finan- nomic growth—the organization of informa- cial market directs capital to its highest value tion acquisition, allocation of resources, uses—a situation James Tobin (1982) monitoring of managers, and exercise of cor- defines as functional form stock market effi- porate control. If financial markets are large ciency. Note that functional form efficiency and liquid enough, and if investors’ property differs from the standard usage of the term rights are sufficiently well protected, individ- market efficiency in the finance literature, uals expend resources to acquire information which describes a situation where investors and conduct profitable informed risk arbi- cannot obtain abnormal returns after trans- trage, as modeled by Sanford J. Grossman actions costs. The two concepts are related, and Joseph E. Stiglitz (1980) and Andrei but not identical. Shleifer and Robert W. Vishny (1997), and Economists have little difficulty listing this trading capitalizes information into stock fracture points where actual capital markets prices as in Morck, Yeung, and Wayne Yu fail to satisfy functional form efficiency. The (2000). Richard Roll (1988) argues that this literature we survey below suggests that the mechanism is especially important for the functional efficiency of capital markets timely updating of stock prices with new depends on the distribution of corporate firm-specific information. Stock prices that control in an economy. In particular, this lit- closely track fundamental values aid outside erature views economic growth as critically investors and corporate insiders in capital dependent on institutions that restrain allocation decisions, simplify monitoring, and entrenched elites, who could otherwise facilitate the design of credible incentive con- come to dominate the capital investment tracts that induce managers to keep faith with decisions of an economy. outside investors and maximize firm value. This literature has its recent roots in cor- These two literatures converge on a cen- porate finance, perhaps because the extreme tral theme: economic growth requires that concentration of corporate control rights in savings be directed into value creating the hands of tiny elites observed in many sp05_Article 1 8/18/05 9:32 AM Page 657

Morck, Wolfenzon, and Yeung: Corporate Governance 657

countries raises concerns about corporate actual ownership. These sorts of structures governance in those economies. The corpo- can let a few wealthy families control the rate finance literature also considers greater part of a country’s large corporate entrenched management to be a corporate sector. They also leave the structure typical governance issue at the firm level in ways of large U.S. firms—stand alone firms with that are helpful in understanding diffuse ownership and professional manage- entrenched control in general. ment—the rarest of curiosities in most of the Our starting point is a curious empirical rest of the world. observation reported by Morck, David This concentrated control can lead to cor- Stangeland, and Yeung (2000). These porate governance concerns—a range of authors divide up the world’s U.S. dollar bil- agency problems. But, more importantly, lionaires into two categories—self-made bil- entrusting the governance of huge slices of lionaires and billionaires who inherited their a country’s corporate sector to a tiny elite wealth—and sum up the wealth owned by can bias capital allocation, retard capital each category of billionaire in each country. market development, obstruct entry by out- Unsurprisingly, they find that a country’s per sider entrepreneurs, and retard growth. capita GDP grows faster if its self-made bil- Furthermore, to preserve their privileged lionaire wealth is larger as a fraction of GDP. positions under the status quo, such elites What is more surprising is that per capita might invest in political connections to GDP growth is slower in countries where stymie the institutional development of cap- inherited billionaire wealth is larger as a ital markets and to erect a variety of entry fraction of GDP. barriers. These economywide implications The literature we review below points to can be most serious. possible explanations. These turn on wealthy Such an outcome is a suboptimal political individuals who magnify their already sub- economy equilibrium, which we dub eco- stantial wealth into control over multiple nomic entrenchment. While wealthy estab- corporations worth vastly more. Rafael La lished families are probably important to Porta, Florencio Lopez-de-Silanes, and economic entrenchment in many countries, Shleifer (1999) show that this magnification the problem is probably not restricted to is most commonly achieved using control countries whose corporate sectors are con- pyramids: structures in which a family firm trolled by small cliques of extremely wealthy controls several listed companies, each of families. Other sorts of elites can also which controls yet more listed companies, become entrenched in positions of control each of which controls yet more listed com- over corporate assets and behave much like panies, and so on. Family members are usu- elite families, generating a similar out- ally placed as executives of key firms come—functionally inefficient capital mar- throughout the structure. Other less com- kets, high barriers to entry, and a slow pace mon devices also used to this end include of innovation. superior voting shares and crossholdings. The layout of this article is as follows. First, Superior voting shares, distinct classes of we motivate this survey by discussing some stock with many more votes per share than empirical findings that suggest a link ordinary common stock, allow insiders to between economic growth and, not the con- control the majority of votes in shareholder centration of wealth per se, but the “hands” meetings even though they own only a small in which control over corporate assets is con- fraction of the firm’s equity. Crossholdings centrated. Second, we review the literature are structures in which firms own blocks of on corporate ownership, pyramidal groups, each others’ stocks. Since insiders vote these and family control. This literature shows that blocks, they exercise control beyond their many economies entrust the governance of sp05_Article 1 8/18/05 9:32 AM Page 658

658 Journal of Economic Literature, Vol. XLIII (September 2005)

large parts of their corporate sectors to tiny growth and variables plausibly reflecting elites of extremely wealthy families. Third, the distribution of control over an economy’s we discuss issues associated with such con- capital assets. centrated corporate governance and how Table 1 recapitulates these results. The these might affect economic performance. alternative specifications shown treat wealth Fourth, we evaluate the linkage between that is not unambiguously controlled by

concentrated corporate control on the one either a founder or an heir differently. H1 hand and overall capital allocation, capital includes only the wealth of billionaires who market development, creative destruction, are unambiguously heirs, politicians, or

and macroeconomic growth on the other. politicians’ relations. H2 also includes the Fifth, and we think this is the most important wealth of billionaires who are probably heirs.

component of the paper, we discuss H3 includes H1 plus fortunes jointly con- entrenchment as a political economy prob- trolled by a founder and his heirs. H4 lem and the determinants of economic includes all the above. H5 through H8 are entrenchment, which include investment analogous to H1, H2, H3, and H4 but do not opportunities, societal tradition, and initial include politician billionaires and their rela- endowments. Also, we examine the relation- tions. The positive coefficients on new ship between economic entrenchment and money billionaire wealth and the negative openness. Finally, we distill some general one on old money billionaire wealth remain conclusions out of the above. highly significant across all the different specifications. Previous work, surveyed by Aghion, 2. Inherited Wealth and Growth—A Caroli, and Garcia-Penalosa (1999), dis- Preliminary Reading cusses the relationship between general Our starting point is a curious observa- inequality and growth, but Morck, tion by Morck, Stangeland, and Yeung Stangeland, and Yeung (2000) add another (2000). Using Forbes’ listing of the one twist—inequality involving new money thousand wealthiest individuals in the wealth seems different from inequality world for 1993, and the brief biographies involving old money wealth. Perhaps econ- accompanying each, they distinguish new omists need to think less about concentra- money billionaires, entrepreneurs who tion of wealth per se and more about made their billions themselves, from old concentration of wealth in whose hands? money billionaires, who inherited their Why might inequality associated with wealth. They gauge the importance of each inherited wealth be fundamentally different type of billionaire in a given country by the from inequality associated with entrepre- sum of the wealth of that type of billionaire neurial wealth? To explore this, we must as a fraction of the country’s GDP. When first examine how highly concentrated they regress real per capita GDP growth wealth can translate into even more highly from 1994 to 1996 on these measures, con- concentrated corporate governance power. trolling for initial per capita GDP, physical capital accumulation, and education levels, 3. The Ubiquity and Purpose of the they find new money billionaire wealth to Control Pyramid be associated with faster economic growth, but old money billionaire wealth to be The extensive control wealthy families associated with slower growth. That is, in exert over the corporate sectors of most of standard growth theory regressions of the the world’s economies depends on a particu- sort used by Mankiw (1995), they report a lar type of ownership structure, the control highly significant relationship between pyramid, and other closely related structures. sp05_Article 1 8/18/05 9:32 AM Page 659

Morck, Wolfenzon, and Yeung: Corporate Governance 659

TABLE 1 THE CROSS-COUNTRY RELATIONSHIP BETWEEN GROWTH IN REAL PER CAPITA GDP AND CAPITAL OWNERSHIP, CONTROLLING FOR CURRENT PER CAPITA INCOME, CAPITAL INVESTMENT RATE, AND LEVEL OF EDUCATION

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Intercept 1.43 1.58 1.59 1.65 1.75 1.73 1.86 1.78 (0.32) (0.30) (0.27) (0.28) (0.22) (0.26) (0.20) (0.25)

Log of per capita –1.76 –1.77 –1.80 –1.79 –1.54 –1.66 –1.62 –1.69 GDP: ln(Y/L) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Capital Accumulation 0.21 0.22 0.21 0.21 0.17 0.20 0.18 0.20 Rate: I/K (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Average Total Years 00.23 00.20 00.25 00.21 00.24 00.20 00.26 00.21 of Education: ln(E) (0.27) (0.35) (0.23) (0.32) (0.24) (0.35) (0.21) (0.32)

Business Entrepreneur 0.44 0.37 0.42 0.37 0.50 0.38 0.45 0.37 Billionaire Wealth Over (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) GDP: B/Y Heir Billionaire Wealth –0.29 –0.17 –0.27 –0.16 –0.41 –0.19 –0.33 –0.17 Over GDP: H/Y (0.03) (0.10) (0.03) (0.09) (0.01) (0.09) (0.01) (0.08)

1 Definition of “heir” H1 H2 H3 H4 H5 H6 H7 H8 R squared 0.519 0.488 0.531 0.489 0.545 0.491 0.536 0.491

Note: Numbers in parenthesis are two tailed t-test probability levels for rejecting a zero coefficient. Coefficients in boldface are statistically significant at 90 percent confidence or more. Sample of thirty-nine countries consists of the countries listed in table 1 minus the United Kingdom and United States. See Morck, Stangeland, and Yeung. (2000), table 2 for further details. 1 = H1 includes only the wealth of billionaires known positively to be heirs, politicians or politicians relations. H2 also includes the wealth of billionaires who are probably heirs. H3 includes H1 plus fortunes jointly controlled by a founder and his heirs. H4 includes all the above. H5 through H8 are analogous to H1, H2, H3, and H4 but do not include politician billionaires and their relations.

While such structures are commonplace just a few firms, but over large groups of outside of the United States, they were not corporations. well known in the academic literature until This section surveys empirical evidence highlighted in recent work by La Porta, about the structure of corporate control Lopez-de-Silanes, and Shleifer (1999). That around the world. We first highlight the research inspired further empirical work ubiquity of concentrated family control out- confirming that (1) most large corporations side the United States. We then discuss con- around the world have controlling owners trol pyramids in detail and show they let a and (2) controlling owners use pyramidal very small number of wealthy individuals or structures (and other mechanisms) to amass families leverage substantial wealth into con- control over not just a single firm, or even trol over corporate assets worth vastly more. sp05_Article 1 8/18/05 9:32 AM Page 660

660 Journal of Economic Literature, Vol. XLIII (September 2005)

While this section primarily emphasizes large blockholders, though they are smaller concentrated control over large groups of and much less common in large U.S. firms firms using control pyramids, which is our than in most other countries. An important focus, it also briefly mentions other related distinction, however, is that block holders in structures. the United States seldom control more than one corporation. Thus, the Ford family con- 3.1 Wealthy Families Control Most trols Ford Motors, but not GM, IBM, and Companies around the World 3M as well. Indeed, the term “family firm” in 3.1.1 Corporations Have Dominant Owners the United States is often a synonym for “small firm.” That large corporations have individuals However, elsewhere in the world, the typ- and families as controlling shareholders con- ical large firm has a “controlling owner”— tradicts the view of the firm popularly attrib- usually a wealthy family that controls it as uted to Adolph A. Berle and Gardiner C. part of a large group of firms. The assump- Means (1932), which implicitly or explicitly tion of freestanding diffusely owned firms, underlies much of modern economics and typically justified with a cite to Berle and finance.1 In this view, a corporation is wide- Means (1932), is therefore of questionable ly held, in that its ownership is dispersed generality. This was first pointed out in stud- across a large number of small public share- ies of corporate governance in Germany and holders, and freestanding, in that listed com- Japan, such as Stephen D. Prowse (1992), panies generally do not control other listed Erik Berglöf and Perotti (1994), and Jeremy companies. Since there is no dominant Edwards and Klaus Fischer (1994). In a sys- owner, effective control resides in the hand tematic investigation, La Porta, Lopez-de- of the management team. In such corpora- Silanes, and Shleifer (1999) show that U.S. tions, corporate governance is about mitigat- style corporate ownership is quite exception- ing the divergence of interests, described by al. In most other countries, even very large Michael C. Jensen and William H. Meckling firms have controlling shareholders, and (1976), between utility maximizing profes- these are usually extremely wealthy families. sional managers and small public sharehold- La Porta, Lopez-de-Silanes, and Shleifer ers who would like the value of their shares (1999) investigate the ownership structures to be maximized. of large corporations in twenty-seven The modern U.S. large corporate sector mainly developed economies. Their sample loosely approximates this archetype. Of contains the top twenty firms in each coun- course, large blockholders and intercorpo- try, ranked by market capitalization of rate equity holdings occur even there. common equity at the end of 1995. They Shleifer and Vishny (1986), Morck, Shleifer, also collect information on ten firms of and Vishny (1988), Clifford G. Holderness similar size in each country. These are the and Dennis P. Sheehan (1988), Ronald C. ten smallest firms in each country with Anderson and David M. Reeb (2003b), market capitalization of common equity Anderson, Sattar A. Mansi, and Reeb (2003), greater than U.S.$500 million at the end of and others all find numerous instances of 1995. They allow for five types of ultimate 1 Berle and Means (1932, chapter 5) actually describe controlling owners: (1) a family or an indi- various forms of pyramiding, nonvoting shares, and voting vidual, (2) the state, (3) a widely held trusts. In their sample of the largest 200 firms in the United States in 1930, they find that 44 percent are wide- financial institution such as a bank or an ly held (they call them “management controlled”) and 22 insurance company, (4) a widely held cor- percent are controlled through a “legal devise” (i.e., a pyra- poration, and (5) miscellaneous, an entity mid). In any case, the finance literature has for the most part taken Berle and Means as demonstrating that large such as a cooperative, a voting trust, or a U.S. firms are largely widely held. group with no single controlling investor. sp05_Article 1 8/18/05 9:32 AM Page 661

Morck, Wolfenzon, and Yeung: Corporate Governance 661

La Porta, Lopez-de-Silanes, and Shleifer 3.1.2 Concentrated Control of Corporation (1999) argue that stakes exceeding fifty per- is in the Hand of Wealthy Families cent are not necessary to lock in control in These controlling shareholders are gener- most cases. This is because most small share- ally wealthy families. La Porta, Lopez-de- holders do not vote at annual meetings. Silanes, and Shleifer (1999) report that, Consequently, voting stakes in the ten to using a 20 percent control threshold, 30 per- twenty percent range are generally sufficient. cent of large firms are family-controlled in They report widely held corporations pre- the average country. This increases when a dominating among large firms only in the sample of smaller firms is used or when the United States and United Kingdom. threshold of control is 10 percent. Again, Elsewhere, particularly in countries with subsequent work, like Claessens, Djankov, weak legal and regulatory protection for and Lang (2000), Faccio and Lang (2002), public shareholders, even the largest corpo- and Najah Attig, Yoser Gadhoum, and Lang rations usually have controlling sharehold- (2003), confirms these findings. ers—predominantly wealthy families and Diffuse corporate ownership of the sort the State. Using a 20 percent control thresh- generally associated with Berle and Means old, they find that 36 percent of large corpo- (1932) for the United States is thus highly rations on average are widely held. The exceptional. Table 2 summarizes the findings distribution shifts further away from widely of these various studies, and clearly displays held firms when the sample is based on a ten the rarity of widely held firms and the ubiq- percent control threshold, or on a set of uity of family control.3 Although table 2 con- smaller firms, or on countries with poorer tains a wealth of data, further work involving shareholder rights protection. In a few coun- additional countries and clarifying the deter- tries, notably Belgium and Germany, finan- minants of these differences would be very cial institutions exercise extensive voting useful. control over other large publicly traded companies. 3.2 The Use of Control Pyramids Studies extending La Porta, Lopez-de- The previous subsections present evi- Silanes, and Shleifer (1999), notably Stijn dence that large corporations throughout Claessens, Simeon Djankov, and Larry H. P. most of the world are not widely held and Lang (2000) on Asian countries, and Mara more often than not are controlled by very Faccio and Lang (2002) and Fabrizio Barca wealthy families. This subsection outlines and Marco Becht (2001) on European countries, expand the number of firms examined 2 and sometimes the number of countries. 3 The different studies in table 2 rely on very different They confirm the La Porta, Lopez-de- samples of firms. For example, La Porta, Lopez-de- Silanes, and Shleifer (1999) findings: out- Silanes, and Shleifer (1999) examine the twenty largest listed firms; Attig, Gadhoum, and Lang (2003) use the side of the United States and United 1,121 firms in the Toronto Stock Exchange; Claessens, Kingdom, a large proportion of large firms Djankov, and Lang (2000) use 2,980 listed financial and has controlling shareholders. nonfinancial firms; and Faccio and Lang (2002) study 5,232 corporations in thirteen Western European countries. These sample differences explain the occasional 2 Claessens, Djankov, and Lang (2000) find that in wide differences between their estimates. Also, Claessens, Japan a large proportion of firms are widely held. The high Djankov, and Lang (2000, p. 95) defined control slightly figures for widely held firms in Japan are perhaps mislead- differently by allowing a firm to have more than one sig- ing, in the sense that most large Japanese firms are organ- nificant owner. For example, if a family holds a 25 percent ized into keiretsu, groups of firms without controlling voting stake and a widely held corporation holds a 10 per- shareholders that collectively own control blocks in each cent stake, when defining control at the 10 percent thresh- other. These firms are thus not really widely held in the old, they classify the firm as one-half controlled by each sense of U.S. and U.K. firms, for public shareholders do type of owner. At the 20 percent level, the firm is fully not hold a majority of their stock. family controlled. sp05_Article 1 8/18/05 9:32 AM Page 662

662 Journal of Economic Literature, Vol. XLIII (September 2005)

TABLE 2 WIDELY HELD FIRMS VERSUS FAMILY CONTROLLED FIRMS AS PERCENT OF 1 LARGE CORPORATIONS IN VARIOUS COUNTRIES Control inferred at Control inferred at 10%2 20% Source Widely Family Widely Family Country held control held control Argentina 0 65 0 65 La Porta, Lopez-de-Silanes, and Shleifer (1999) Australia 55 10 65 5 La Porta, Lopez-de-Silanes, and Shleifer (1999) Austria 5 15 5 15 La Porta, Lopez-de-Silanes, and Shleifer (1999) Belgium 0 50 5 50 La Porta, Lopez-de-Silanes, and Shleifer (1999) Canada 50 30 60 25 La Porta, Lopez-de-Silanes, and Shleifer (1999) 17.54 46.67 36.24 26.18 Attig, Gadhoum, and Lang (2003) Denmark 10 35 40 35 La Porta, Lopez-de-Silanes, and Shleifer (1999) Finland 15 10 35 10 La Porta, Lopez-de-Silanes, and Shleifer (1999) France 30 20 60 20 La Porta, Lopez-de-Silanes, and Shleifer (1999) 8.92 70.44 17.79 64.83 Faccio, Lang (2002)a Germany 35 10 50 10 La Porta, Lopez-de-Silanes, and Shleifer (1999) 5.61 71.64 14.02 64.62 Faccio and Lang (2002) Greece 5 65 10 50 La Porta, Lopez-de-Silanes, and Shleifer (1999) Hong Kong 10 70 10 70 La Porta, Lopez-de-Silanes, and Shleifer (1999) 0.6 64.7 7 66.7 Claessens, Djankov, and Lang (2000) Indonesia 0.6 68.6 5.1 71.5 Claessens, Djankov, and Lang (2000) Ireland 45 15 65 10 La Porta, Lopez-de-Silanes, and Shleifer (1999) Israel 5 50 5 50 La Porta, Lopez-de-Silanes, and Shleifer (1999) Italy 15 20 20 15 La Porta, Lopez-de-Silanes, and Shleifer (1999) 7.83 64.87 15.86 59.61 Faccio and Lang (2002) Japan 50 10 90 5 La Porta, Lopez-de-Silanes, and Shleifer (1999) 42 13.1 79.8 9.7 Claessens, Djankov, and Lang (2000) Korea 40 35 55 20 La Porta, Lopez-de-Silanes, and Shleifer (1999) 14.3 67.9 43.2 48.4 Claessens, Djankov, and Lang (2000) Malaysia 1 57.5 10.3 67.2 Claessens, Djankov, and Lang (2000) Mexico 0 100 0 100 La Porta, Lopez-de-Silanes, and Shleifer (1999) Netherlands 30 20 30 20 La Porta, Lopez-de-Silanes, and Shleifer (1999) New Zealand 5 45 30 25 La Porta, Lopez-de-Silanes, and Shleifer (1999) Norway 5 25 25 25 La Porta, Lopez-de-Silanes, and Shleifer (1999) Philippines 1.7 42.1 19.2 44.6 Claessens, Djankov, and Lang (2000) Portugal 0 50 10 45 La Porta, Lopez-de-Silanes, and Shleifer (1999) Singapore 5 45 15 30 La Porta, Lopez-de-Silanes, and Shleifer (1999) 1.4 52 5.4 55.4 Claessens, Djankov, and Lang (2000) Spain 15 25 35 15 La Porta, Lopez-de-Silanes, and Shleifer (1999) 12.74 67.33 28.06 55.79 Faccio and Lang (2002) Sweden 0 55 25 45 La Porta, Lopez-de-Silanes, and Shleifer (1999) Switzerland 50 40 60 30 La Porta, Lopez-de-Silanes, and Shleifer (1999) Taiwan (China) 2.9 65.6 26.2 48.2 Claessens, Djankov, and Lang (2000) Thailand 2.2 56.5 6.6 61.6 Claessens, Djankov, and Lang (2000) United Kingdom 90 5 100 0 La Porta, Lopez-de-Silanes, and Shleifer (1999) 27.06 33.75 69.08 19.88 Faccio and Lang (2002) United States 80 20 80 20 La Porta, Lopez-de-Silanes, and Shleifer (1999) 38.95 23.37 69.26 6.13 Attig, Gadhoum, and Lang (2003) 1 The figures from La Porta, Lopez-de-Silanes, and Shleifer (1999) are for the twenty largest publicly listed firms. The Attig, Gadhoum, and Lang (2003) sample is comprised of the 1,121 firms in the Toronto Stock Exchange for which they could find ownership data. Faccio and Lang (2002) and Claessens, Djankov, and Lang (2000) also construct samples of all publicly traded firm for which ownership information is available. 2 The published version of Faccio and Lang (2002) does not have tables with the 10 percent cutoff. We report statistics in their earlier unpublished draft. sp05_Article 1 8/18/05 9:32 AM Page 663

Morck, Wolfenzon, and Yeung: Corporate Governance 663

A family firm controls a first tier of firms with dominant , voting stakes, in this case greater than. fifty percent. Each first tier firm controls several , second tier firms, each of which controls . yet more firms. The overall effect is to extend the family’s control to encompass assets worth substantially . more than its actual wealth.

> 50% Firm3,1

> 50% < 50% Firm2,1

< 50% > 50% Firm3,2

> 50% < 50% Firm1,1

< 50% > 50% Firm3,3

> 50% < 50% Firm2,2

< 50% > 50% Firm3,4

< 50% Family Firm

> 50% Firm3,5

> 50% < 50% Firm2,3

< 50% > 50% Firm3,6

> 50% < 50% Firm1,2

< 50% > 50% Firm3,7

> 50% < 50% Firm2,4

< 50% > 50% Firm3,8

< 50%

Public Shareholders

Figure 1. A Stylized Control Pyramid

how these families achieve such concentrat- than the family’s actual wealth. Figure 1 ed corporate control. The impression is that illustrates a simplified control pyramid. A the most important mechanism families use family firm, which is a holding company in is the control pyramid, sometimes aug- this simplified example, holds a majority mented by multiple voting shares and stake, fifty percent plus one vote, in each of

crossholdings, although the literature has a first tier of firms, in this case Firm1,1 and not generated estimates on the relative Firm1,2. The remaining fifty percent minus significance of these means. one vote stakes in each are held by small Control pyramids let these families con- public shareholders. Each of the first tier trol not just one large firm, but many large firms then holds fifty percent plus one vote firms collectively worth substantially more in a set of second tier firms. In this case, sp05_Article 1 8/18/05 9:32 AM Page 664

664 Journal of Economic Literature, Vol. XLIII (September 2005) α Firm1,1 holds control blocks in Firm2,1 and family with the firm. Since the ’s are all less Firm2,2 while Firm1,2 holds control blocks in than one, this product can be a very small Firm2,3 and Firm2,4. Again, the remaining number. Nonetheless, the family controls fifty percent, less one vote, in each tier two every firm in the chain of firms connecting

firm is held by public investors. Each second the family firm to Firm3,1 and so controls tier firm then controls a set of third tier firms, that firm too. Ultimately, the family can con- and so on, always with public shareholders trol many firms in this way even though its holding the remaining stakes. actual investment in each is small. Suppose each firm in figure 1 is worth one Actual pyramids are more complicated billion dollars. The control pyramid lets the than the diagram in figure 1. Firms may have family, through its one billion dollar family varying numbers of subsidiaries, and some firm, control fourteen other firms, also each firms in intermediate tiers of the pyramid worth one billion dollars. There is double may engage in actual production activities. counting in this tally since, in this simplified Privately held firms may be interspersed example, the assets of the higher tier firms among the publicly traded firms of the pyra- are shares of lower tier firms. But even if mid. Firms also need not be neatly arranged only the third tier firms contain actual phys- in successive tiers. Firms in lower tiers may ical assets, the family has, at a minimum, have cross-shareholdings. That is, they may leveraged one billion dollars of family wealth own shares in firms in their own tiers or in into control over eight billion dollars of real upper tiers. corporate assets. Contrast this with creating Cross shareholdings, differential voting eight freestanding one billion dollar firms or shares, and public shareholders’ low partici- a single eight billion dollar firm. In either pation rates in corporate votes generally allow case, the family contributes only one-eighth control through chains of even smaller α ’s, of the equity and controls only 12.5 percent resulting in even more extreme divergence of of the votes. In contrast, the pyramidal struc- ownership from control with each additional ture in figure 1 lets the family retain absolute layer. Since family controlled boards vote control of the eight firms in the third layer of cross holdings, these augment direct family the pyramid, but hold only a 12.5 percent voting power. Super-voting shares also allow cash flow stake [(50%)(50%)(50%) = 12.5%]. control with smaller values of α . Returning to To see this more generally, take the chain our example in figure 1, if super-voting shares of ownership from the family apex to any let the family exercise control with ten per-

firm in the third layer, say firm Firm3,1. cent at each layer, rather than fifty percent, α 0 Suppose the family holds 1,1 of Firm1,1 the family’s stake in the third layer would be α 1,1 = which holds 2,1 of Firm2,1 which, in turn, (10%)(10%)(10%) 0.1% instead of the 12.5 α 2,1 holds 3,1 of Firm3,1. The superscript, e.g., percent calculated above. Since voting rates (1,1), indicates first the “controlling” firm’s by public shareholders are generally quite level in the pyramid and its identity index at low, a ten to twenty percent stake is usually the level. The subscript, e.g., (2,1), indicates sufficient to control board elections, so the the “controlled” firm’s level in the pyramid last calculation is often relevant even in the and its identity index at that level. If, at each absence of multiple voting shares. link in the chain of control, the direct stake By way of illustration, we use an example is sufficiently large to control the next firm, in La Porta, Lopez-de-Silanes, and Shleifer

the family effectively controls Firm3,1 while (1999), although other examples are amply holding stock that entitles it to only available, e.g., in Morck, Stangeland, and α 0 × α 1,1 × α 2,1 1,1 2,1 3,1 of the firm’s dividends. This Yeung (2000). La Porta, Lopez-de-Silanes, formula can be straightforwardly expanded and Shleifer (1999, figure 8) describe the when there are more layers connecting the ABB pyramidal structure which we replicate sp05_Article 1 8/18/05 9:32 AM Page 665

Morck, Wolfenzon, and Yeung: Corporate Governance 665

Source: La Porta, Lopez-de-Silanes, and Shleifer (1999)

Figure 2. The ABB Control Pyramid of the Swedish Wallenberg Family Ownership stakes are represented with C and voting stakes with V. Ultimate control is assigned to the Wallenberg family through a chain of stakes of at least 20 percent.

in figure 2. The authors follow a chain of group. La Porta, Lopez-de-Silanes, and control by which ABB, Sweden’s fourth Shleifer (1999) show that this frequently largest firm by market capitalization, is con- gives controlling families direct executive trolled via a 32.8 percent stake held by decision-making power, as well as control via Incentive, Sweden’s seventeenth largest pyramids. In 69 percent of their sample of firm. Incentive, in turn, is controlled via a large firms (the top twenty firms ranked by 43.1 percent stake held by the Wallenberg market capitalization of common equity at Group, which is controlled by Investor, the end of 1995 in each of twenty-seven Sweden’s fifth largest firm, through a voting countries), the controlling families also par- arrangement that gives it control over 35.7 ticipate in management. Participation is percent of the Wallenberg Group’s 43.1 per- defined as a family member (sharing the cent stake in Incentive. Investor is the same family last name) being the CEO, the Wallenberg’s family run closed end fund and Chairman, the Honorary Chairman, or the serves as the apex firm for their family con- Vice-Chairman of the firm. The reported trol pyramid. In this way, a sharp divergence percentage underestimates the actual between control rights and cash flow rights involvement of family members in executive characterizes many Swedish firms: For position because executives married into the example, the Wallenbergs have voting con- family may not share the family name. trol over ABB, but actually have a cash flow Claessens, Djankov, and Lang (2000) report rights stake of only about 5 percent. that, in East Asian countries (Hong Kong, As we explained above, a control pyramid Indonesia, Japan, South Korea, Malaysia, allows a wealthy family or individual to con- the Philippines, Singapore, Taiwan, and trol a large number of firms with relatively Thailand), professional managers are rare, limited wealth. Pyramids permit this by and family members or trusted associates are creating large deviations between voting usually in charge. Top managers are family rights and cash flow rights; allowing control members in about 60 percent of firms that over many firms with a small cash flow are not widely held. These studies are about stake in each. controlling family participation in general, Controlling families also effectuate control not controlling family participation in the by placing family members in executive posi- management of pyramid member firms in tions in key firms throughout a pyramidal particular. Empirical exploration of this issue sp05_Article 1 8/18/05 9:32 AM Page 666

666 Journal of Economic Literature, Vol. XLIII (September 2005)

is needed. For Italy, however, Paolo F. as de facto apex firms. Since public share- Volpin (2002) presents data about family holders in German firms routinely sign over members’ participation in firms that belong their voting rights to the banks that manage to pyramidal groups. He finds that 50 per- their stock accounts, these control pyramids cent of top executives in the top layer of a can consist of nominally widely held firms pyramid are members of the controlling held together by banks voting the holdings family. For lower layers, the figure is small- of small shareholders. The large German er: 26 percent for layer two firms and 7.4 banks are all widely held, so the bank man- percent for layers three and below. agers collectively vote majorities of their The discussion in this subsection leaves own stock.4 Recent reforms require German us with the following important points. By banks to advise investors of their right to using pyramids and, to a lesser extent, vote their own shares, and this may effec- crossholdings, super-voting shares, and the tively dismantle such structures over time. appointment of family members as execu- Of course family control pyramids are also tives, a wealthy family can secure control of important in all of these countries. a corporation without making a commensu- We use the term corporate group to rate equity investment. Pyramids in partic- refer to all of these structures collectively, ular allow a divergence of voting rights and reserve the term family control pyra- from cash flow rights much greater than is mid for business groups with a basically typically possible through direct owner- pyramidal structure of intercorporate own- ship. Pyramids thus allow a family with a ership and a family firm at the apex. given level of wealth to control corporate Corporate groups other than family con- assets worth considerably more than direct trolled pyramidal groups are economically ownership would permit. important in some countries. La Porta, Lopez-de-Silanes, and Shleifer (1999) find 3.3 Other Types of Group Control that family controlled pyramidal groups La Porta, Lopez-de-Silanes, and Shleifer predominate in the large corporate sectors (1999) show that control pyramids are by far of most countries, so they are the focus of the most important corporate group struc- section 3.2. However, other corporate tures, but publicly traded companies can groups are lumped together with them in control each other in other ways. These gen- some studies, introducing an ambiguity erate other group structures that are impor- that is perhaps unavoidable at this early tant in particular countries. For example, stage in the literature. Further work deter- Japanese firms are organized into groups mining the economically important differ- called keiretsu, in which firms with no con- ences, if any, between different sorts of trolling shareholders each hold small stakes corporate groups would be welcome. in one another, but these stakes collectively 3.4 A Few Wealthy Families Control a amount to control blocks. Thus, each firm is Large Fraction of Many Economies controlled by the professional managers of all the other firms in the group (see, e.g., That family control is widespread around Berglöf and Perotti 1994). France, Canada, the world is perhaps interesting, but hardly Germany, and other countries also contain unsettling. What makes this situation control pyramids without family firms at the important to our understanding of apex. In some cases, the apex firm is itself economies outside the United States and widely held. In France and Canada, some United Kingdom is that control pyramids control pyramids of publicly traded firms have state-owned enterprises as their apex 4 Along with the managers of large insurance companies firms. In Germany, banks sometimes serve affiliated with the banks. sp05_Article 1 8/18/05 9:32 AM Page 667

Morck, Wolfenzon, and Yeung: Corporate Governance 667

concentrate a country’s corporate decision Switzerland to only 1.10 percent in the making in remarkably few hands. United Kingdom. For the ten largest fami- The articles surveyed above are filled with lies, the figures are 19 percent for Austria, representative examples. 30 percent for Belgium, 22 percent for Morck, Stangeland, and Yeung (2000) Finland, 29 percent for France, 21 percent describe how the Canadian Bronfman fami- for Germany, 14 percent for Ireland, 20 per- ly extends its considerable wealth into con- cent for Italy, 23 percent for Norway, 34 per- trol over corporate assets worth vastly more. cent for Portugal, 11 percent for Spain, 13 Attig, Gadhoum, and Lang (2003) add percent for Sweden, 29 percent for descriptions of the Manix, McCain, and the Switzerland, and only 4 percent for the Bentley families. La Porta, Lopez-de- United Kingdom. Silanes, and Shleifer (1999) discuss yet other East Asian economies tend to resemble examples of Canadian control pyramids. Sweden more than Canada. Claessens, While these examples show how wealthy Djankov, and Lang (2000) find that the top Canadian families greatly magnify their eco- fifteen family control pyramids in typical nomic control, Morck, Stangeland, and East Asian economies hold corporate assets Yeung (2000) report that widely held firms worth a large fraction of GDP—84 percent are also commonplace in that country. Thus, of GDP in Hong Kong, 76.2 percent in no single family controls a predominant slice Malaysia, 48.3 percent in Singapore, 46.7 of the national economy. percent in the Philippines, and 39.3 percent Contrast this to Sweden, where Jonas in Thailand (these are the largest numbers Agnblad, Berglöf, Peter Högfeldt, and reported). They state (p. 109) that “these Helena Svancar (2001) report that one fam- results suggest that a relatively small number ily, the Wallenbergs, controls roughly half of of families effectively control most East the market capitalization of the Stockholm Asian economies.” They also find that the Stock Exchange. This is accomplished single wealthiest family controls 17.1 per- through a huge and complicated control cent of the market capitalization in the pyramid. What distinguishes Sweden from Philippines, 16.6 percent in Indonesia, and Canada is that control rights in Swedish 11.4 percent in South Korea. firms are substantially in the hands of a sin- La Porta, Lopez-de-Silanes, and Shleifer gle family, rather than a handful of families (1999) report that Huchison Whampoa, and numerous professional managers. the third largest listed company in Hong Other countries typically fall between the Kong, is 43.9 percent controlled by extremes represented by Canada and Cheung Kong Holdings, the fifth largest Sweden. For example, La Porta, Lopez-de- listed company, which is 35 percent owned Silanes, and Shleifer (1999), Marco Bianchi, by Li Ka Shing family. Li Ka Shing, Magda Bianco, and Luca Enriques (2001), through a pyramidal arrangement, controls and Faccio and Lang (2002) describe the three of the twenty largest companies in control pyramid that lets the Agnelli family Hong Kong, including the eleventh largest, control corporate assets worth vastly more Hong Kong Electric Holdings. Claessens, than their actual family wealth in Italy. Djankov, and Lang (2000) provide further Faccio and Lang (2002) further report that description of the Li Ka Shing control the Agnellis control 10.4 percent of that pyramid. country’s total market capitalization. Claessens, Djankov, and Lang (2000) Overall, across Western Europe, the value of describe the extensive corporate control corporate assets controlled by the leading exercised by the Ayala control pyramid in family, measured as a fraction of market cap- the Philippines. The Ayala Corporation, the italization, ranges from 18 percent in second largest listed company on the sp05_Article 1 8/18/05 9:32 AM Page 668

668 Journal of Economic Literature, Vol. XLIII (September 2005)

Australia Inited Kingdom India South Africa Japan Turkey United States Canada Israel Western Europe Latin America South East Asia 0102030405060708090100 110 120 130 140

Source: Morck, Stangeland, and Yeung (2000).

Figure 3. Billionaire Wealth in millions per Billion Dollars of GDP: Billionaire nationalities are as listed in Forbes Magazine’s annual list of U.S. dollar billionaires for 1996.

Manila Stock Exchange in terms of market billionaire wealth per thousand dollars of capitalization, controls Ayala Land, the gross domestic product. largest. The Bank of the Philippine Islands, the fifth largest, also belongs to the Ayala 3.5 Summary control pyramid. The principal owner of the Ayala Corporation is the privately held The findings in this section condense into Mermac Inc. The Tokyo Mitsubishi Bank three crucial points. First, the prevalence of controls 23 percent and others control less widely held firms in the large corporate sec- than 5 percent each. Mermac is 100 percent tors of the United States and United controlled by the Ayala family. Kingdom is highly exceptional. Second, the Kee-Hong Bae, Jun-Koo Kang, and Jin- large corporate sectors, excluding state- Mo Kim (2002) describe the importance of owned enterprises, of most countries are Korean family controlled pyramids, or chae- predominantly controlled by very wealthy bols. Chaebols, while basically pyramidal in families. Third, the number of wealthy fam- nature, have extensive reciprocal holdings ilies exercising this control is typically quite as well as simple pyramidal intercorporate small because control pyramids translate ownership links. The top thirty chaebols substantial family wealth into command of “contribute to 62.5 percent of the total corporate assets worth vastly more. In some assets and 72.6 percent of the gross sales of countries, individual families use this tech- all listed firms” (p. 2699) and each is diver- nique to control considerable, and even sified across many unrelated industries. predominant, slices of the country’s market Each is also controlled by a single wealthy capitalization. family (p. 2702). Further careful empirical work is clearly Figure 3 shows the wealth of billionaires needed to verify the actual extents to as a fraction of GDP, expressed as dollars of which these points genuinely characterize sp05_Article 1 8/18/05 9:32 AM Page 669

Morck, Wolfenzon, and Yeung: Corporate Governance 669

different countries. But assuming they are plus individual and family ownership and correct, they raise a curious question: Why dummies for the presence of a founder or are pyramidal firm groups so much heir as CEO.5 They report that public share- rarer in the United States and United holders assign substantially higher values to Kingdom? Or, are pyramidal groups a firms with a large block held by an individual heretofore neglected aspect of corporate or family. They report an especially large governance in the United States and value premium for firms whose founders United Kingdom? retain equity holdings and a smaller, but still statistically significant, value premium for firms with other large individual or family 4. Potential Benefits of Control Pyramids shareholders. Anderson and Reeb (2003a) use the same sample and find that family We now examine the impact of such dis- controlled firms are less diversified, after proportionate control over a country’s cor- controlling for size, outside directors, and porate assets by tiny elites of extremely block shareholders. They conclude that wealthy individuals and families—both at “family firms perform as well as, if not better the firm level and at the economy level. We than, nonfamily firms.” first review work consistent with dominant Morck, Shleifer, and Vishny (1988) reach shareholders exerting a positive effect on the same conclusion for U.S. firms in which firm performance in freestanding firms. the founder or a member of the founding We then show that findings from free- family is present in management and the firm standing firms with dominant shareholders was incorporated within the previous thirty need not carry across to pyramidal groups. years. But for older firms with founders or Pyramidal groups may have important pos- their relatives in management, they report a itive effects in economies with underdevel- significant value discount. Holderness and oped factor markets and institutions. Sheehan (1988) report no systematic per- However, asking why family controlled formance difference associated with large pyramidal groups persist raises a range of possible negative effects, discussed in 5 section 5. Tobin’s average q is the market value of all the firm’s financial obligations over the replacement cost of its physical and intangible assets. Note that this literature uses average q, not marginal q (the increase in market value 4.1 The Performance of Freestanding induced by a unit increase in replacement cost of assets), Family Controlled Firms which should be one absent taxes and transactions costs. In a very long run equilibrium, average Tobin’s q should also be one, but need not be in the shorter run. For example, if Several studies suggest that family control uncertainty about insiders’ stealing partially resolves after per se might often have a positive effect on shares were issued to the public, or if inframarginal and firm performance. Anderson and Reeb marginal financing have different costs, an individual firm’s average q might end up above or below one. Durnev, (2003b) study 403 firms in the Standard and Morck, and Yeung (2004, table 2) show empirically that Poor’s 500, an index of large U.S. firms. U.S. firms’ average qs and marginal qs differ markedly. In They find that about one-third of them have corporate finance, a long empirical tradition, beginning with Morck, Shleifer, and Vishny (1988) and justified the- individuals or families as substantial share- oretically by Stulz (1988), links average q (not marginal q) holders, holding altogether 18 percent of to proxies for corporate governance quality. Recent work equity value. Using data for 1992 through by Paul A. Gompers, Joy L. Ishii, and Andrew Metric (2003) and others explicitly ties average q to direct meas- 1999, excluding utilities and banks, they ures of corporate governance quality. The underlying intu- regress Tobin’s average q, a standardized ition is that rationally expected corporate governance measure of firm valuation and a well-accept- problems are revealed through time, changing individual firms’ total market values (the numerator of average q), but ed proxy for the quality of corporate gover- only affecting their replacement costs (the denominator of nance, on a standard set of control variables average q) on the margin. sp05_Article 1 8/18/05 9:32 AM Page 670

670 Journal of Economic Literature, Vol. XLIII (September 2005)

U.S. shareholders. However, Francisco Perez- substantial contributor to observed cross- Gonzalez (2002), using U.S. data, divides 162 country differences in productivity. CEO transitions where the retiring CEO is a Unfortunately, the empirical results cited member of the founding family into cases above, except Smith and Amoako-Adu where the succeeding CEO is a member of (1999), use U.S. data. Further studies using the founding family and cases where she is data from other countries are needed. As not. Stock prices, returns on assets, and section 3 shows, the United States is excep- market-to-book ratios fall sharply for firms tional for its almost complete absence of with inherited control, but not for firms control pyramids, and for the rarity and where the CEO is unrelated to the founding fleeting nature of large listed firms holding family. Moreover, these declines are partic- control blocks in other large listed firms.6 ularly prominent for firms that appoint fam- Elsewhere, family control pyramids, in ily CEOs who did not attend a selective which tiers of listed firms hold control blocks college. Indeed, the differences between in other listed firms, are commonplace. firms with outside succession and in-family The U.S. results described above allow succession by an offspring alumnus of a that, in freestanding firms, family control selective college are not statistically differ- might sometimes be a net advantage. ent. Clearly, these results suggest that large However, findings regarding freestanding block-shareholdings by a family, and family firms cannot be applied hastily to pyramid involvement in management, need not firms and empirical studies suggest impor- destroy value, and may even add value for tant differences. For example, Morck, public shareholders—especially when fam- Stangeland, and Yeung (2000), after control- ily control means control by an entrepre- ling for firm age, size, and main industry, neurial founder or a demonstrably report that Canadian heir controlled firms competent heir. Further studies along are less profitable than otherwise compara- these lines for other countries would be ble firms in the United States and in Canada. highly useful. One possible explanation is that typical Similarly, in an event study using Canadian heirs running large firms are Canadian family controlled firms, Brian F. worse managers than their U.S. counter- Smith and Ben Amoako-Adu (1999) report parts. Another is that many large Canadian share price drops when heirs take charge— firms belong to pyramidal groups. While we especially if the heirs are young. They propose that older heirs, with more business 6 experience, might reassure investors. Jeffrey W. Allen and Gordon M. Phillips (2000) report 227 instances of U.S. firms buying blocks of stock in one Francesco Caselli and Nicola Gennaioli another from 1980 to 1991. These equity stakes are usual- (2003) develop a model of dynastic manage- ly small (averaging 15 percent of the seller’s equity) and ment—the passing of control from father to are often preliminary to complete takeovers. They are sometimes referred to as toeholds. Indeed Wayne H. son. In their model, this is a potential source Mikkelson and Richard S. Ruback (1985) and Dosoung of inefficiency because heirs may inherit lit- Choi (1991) can explain positive announcement abnormal tle of their parents’ talent for managerial returns associated with such purchases by the anticipation of subsequent takeovers. In some cases, joint venture part- decision making. They explore the macro- ner firms may also exchange blocks of stock. Thus, U.S. economic causes and consequences of intercorporate equity blocks are usually between pairs of dynastic management, and derive that the otherwise freestanding firms. The only extant pyramidal structure in the United States, to our knowledge, is the incidence of dynastic management depends venture capital organization, Thermo Electron, which on the severity of asset market imperfec- retains control blocks in its high technology public spin tions, the saving rate, and the inheritability offs. Thus, family control and intercorporate blockholdings in the United States are quite atypical; for wealthy U.S. of talent. Using numerical simulations, they families seem virtually always to control but a single large show that dynastic management may be a publicly trade firm. sp05_Article 1 8/18/05 9:32 AM Page 671

Morck, Wolfenzon, and Yeung: Corporate Governance 671

cannot preclude the former reason, the lat- stock market measures of firm performance ter seems more immediately plausible. We for group affiliates have nonmonotonic rela- therefore consider ways in which pyramidal tionships with group diversification. Firms structures might affect firm performance— in somewhat diversified control pyramids positively and negatively. We discuss the perform worse than free-standing firms, positives in the remaining part of this section while firms in highly diversified control pyr- and the negatives in the next section. amids perform better than their freestanding peers. Apparently, a high level of 4.2 Could Control Pyramids Have Positive diversification ultimately benefits group Effects? affiliates. Control pyramids must be of value to Khanna and Palepu (2000a) and Khanna someone or they would not predominate as and Rivkin (2001) both explain their findings they do in many economies. An emerging by positing various arguments as to why the empirical literature attempts to identify group structure might be an advantage in advantages conferred by family control pyra- economies with poorly functioning markets mids. We review some of the most important and institutions that might cause unrelated contributions. firms to have difficulty writing and enforcing Tarun Khanna and Jan W. Rivkin (2001) contracts with each other. The idea that examine firms affiliated with business groups mitigate market frictions goes back to groups in Argentina, Brazil, Chile, India, Nathaniel H. Leff (1976, 1978). Indonesia, Israel, Korea, Mexico, Peru, the One set of arguments stems from the need Philippines, South Africa, Taiwan (China), to overcome capital market frictions. In Thailand, and Turkey. They define a busi- economies where information asymmetry is ness group as “a set of firms which, though severe and institutional devices to signal legally independent, are bound together by a trustworthiness are inadequate and ineffec- constellation of formal and informal ties and tive, external financing is expensive and lim- are accustomed to taking coordinated ited. A group structure can enhance actions.” Thus, they are examining control monitoring and overcome liquidity con- pyramids run by families as well as other cor- straints by letting group firms pool resources. porate groups; however, most of the groups The pyramid’s controlling shareholder could they examine appear to be family control potentially allocate capital more efficiently pyramids. They report higher average than the weak capital market could. This is returns on assets (ROA) among group firms essentially the same “internal versus external in all the countries they study save capital market argument” that Wilbur G. Argentina, Brazil, Chile, Mexico, Peru, and Lewellen (1971), Robert H. Gertner, David the Philippines. They also find lower varia- S. Scharfstein, and Jeremy C. Stein (1994), tion in ROA for group firms, except in and Stein (1997) develop in connection with Mexico and Turkey. From these findings, freestanding U.S. conglomerates. they conclude “not only that members of a In a similar vein, Mike Burkart, Fausto group incur costs and reap benefits, but also Panunzi, and Shleifer (2003) examine whether that the costs and benefits are shared.” entrepreneurs want to surrender control of Khanna and Krishna Palepu (2000a) com- firms they found. They accept the potential pare the performance of member firms of benefits of owner control, but pitch them Indian business groups, which are family against the utilization of outside capable pro- control pyramids, with that of freestanding fessional managers. When the benefits are less firms. Some of these business groups span than the forgone benefits of rendering control many distinct industries, while others are to capable outside professional managers, a more focused. They find that accounting and case may exist for surrendering control. sp05_Article 1 8/18/05 9:32 AM Page 672

672 Journal of Economic Literature, Vol. XLIII (September 2005)

Outside professional managers, however, may developing economies. In economies where expropriate firm value. Thence, the forces that external markets for professional managers factor in on the question of whether to sur- are thin and underdeveloped, a group’s render control are the amenities and benefits internal market structure can lead to more of running a family business verse the extra investment in recruiting, training, and firm value capable professional managers can greater incentives for employees to develop bring in net of their expropriation and the “group specific human capital.” Similar con- monitoring costs. The net extra firm value siderations apply to the development of capable professional managers can bring in is market skills and brand-names. less in an environment where managers can A separate argument turns on the rela- more readily disregard outside shareholder’s tionships between a firm’s stakeholders—the property rights. The lack of separation various claimants to its cash flows such as the between management and ownership control State, the public, the firm’s workers, credi- is evidence of underdevelopment in financial tors, suppliers, customers, and various types markets. of its shareholders. Mark J. Roe (2003) Most of the results which show that argues that family control pyramids serve to “groups” may have a positive effect on firm protect shareholders from powerful labor performance are based on developing or unions and other social interests. These newly developed countries data. It is not arguments are perhaps plausible as explana- clear whether the lack of reported results tions for the prevalence of control pyramids for developed countries indicates lack of in developed welfare state economies like empirical attempts using developed country Sweden and Canada, but, as Roe (2003) con- data or that it validates the theoretical argu- cedes, they seem inadequate at explaining ments like Burkart, Panunzi, and Shliefer their persistence in developing economies. (2003). (Journals seem to have a bias against In support of his argument, Roe (2003) publishing papers that report insignificant shows that countries with more socialistic statistical results.) politics and stronger labor rights have less More generally, the arguments proposed dispersed ownership. One difficulty we have by Khanna and Palepu (2000a) and others with the Roe (2003) argument is that causal- stem from the need to overcome frictions in ity is not easily determined. Leftist politics factor markets. In essence, markets for cap- and strong labor movements might just as ital, managerial talent, and intangibles are easily form to counter the power of wealthy internalized within groups because external families controlling multiple corporations. markets are poorly developed. When institu- Another difficulty with Roe’s argument is tions are weak, doing business with strangers that pyramids give controlling shareholders is dangerous and unreliable. This impedes interests that can sharply diverge from those the operation of labor, capital, knowledge, of public shareholders. Further work is and product markets. However, families clearly needed to clarify these issues. with reputations for fairness and good man- 4.3 But, Why Pyramids? agement practices are especially sought after as business partners in such environments. The literature positing positive effects of Control pyramids let families with good rep- pyramidal structures, whether derived from utations achieve greater economies of scale, Khanna and Palepu (2000a) or Roe (2003), and thus allow a higher degree of trust than begs the question: Why do pyramidal groups would otherwise be possible. This, in turn, form, rather than freestanding conglomer- facilitates economic activity. Control pyra- ates? Research on this question is very mids may well be an economically rational sparse, as far as we know. We can therefore response to poorly functioning markets in only offer several conjectures. sp05_Article 1 8/18/05 9:32 AM Page 673

Morck, Wolfenzon, and Yeung: Corporate Governance 673

First, separately registered firms might are worth more than a single conglomerate’s provide at least some limited liability pro- stock. They propose that distinct stocks might tection, so that a parent firm’s losses not allow readier resolution of information asym- exceed its initial investment. Limited liabili- metries and principal agent problems, and ty, however, also reduces outside investors’ Frank Gigler and Thomas Hemmer (2002) ability to recover their investments in case of provide an information theory model along a loss. In addition, limited liability need not these lines. Perhaps analogous models apply always provide full protection. For instance, to pyramidal and other corporate groups. Morck and Masao Nakamura (2004) show Fifth, Takeo Hoshi, Anil Kashyap, and that, when the Suzuki group’s Taiwan Bank Scharfstein (1991) argue that Japanese firms failed in 1930s Japan, all the other compa- provide capital to other firms in their corpo- nies in the Suzuki family’s pyramid failed rate groups that are experiencing financial simultaneously too. difficulties. They propose that this intercor- Second, a pyramid might create space for porate insurance enhances economic effi- the promotion of bright executives to top ciency by reducing bankruptcy costs. management positions. Perhaps these peo- Although Morck and Nakamura (1999) ple would be frustrated as middle managers present evidence that such transactions are in a big conglomerate. Also, running a legal- better viewed as bailouts, and are probably ly separated unit may give these executives not economically efficient, corporate groups more freedom and perhaps might create in other countries might still perform the better career incentives. However, this free- function Hoshi, Kashyap, and Scharfstein dom is likely to be quite illusory because the envision. controlling family often retains executive Sixth, Heitor Almeida and Wolfenzon decision power. (forthcoming-a) argue that capital market Third, many smaller firms, rather than frictions induce pyramids. In their model, one huge one, might conceivably allow bet- weak investor protection keeps firms from ter monitoring of professional managers by raising external finance unless internal funds the family. Pyramids might be less opaque are available as “seed money.” If a wealthy than a single conglomerate—to the control- family sets up a new freestanding firm, it has ling family, if not to public shareholders. only family wealth available as seed money, This point and the previous one are consis- which should include dividends paid from tent with George P. Baker (1992), who pro- firms it controls. But if it uses an existing poses that conglomerates are more firm to set up a new one, the accumulated efficiently run when head office delegates retained earnings of that firm are available as much power as possible to operational as “internal funds.” 7 units. It is possible that pyramidal groups Of course, this framework fails to explain insulate individual operations from the head why the wealthy family doesn’t simply office more effectively than conglomerate expand an existing firm by issuing equity or divisions. However, we are unaware of any debt. However, as David Landes (1949) evidence regarding this. points out in connection with French family Fourth, many U.S. conglomerates have firms, equity issues dilute the family’s con- recently issued tracking stocks. These are trol block and debt issues raise the risk of stocks whose dividends are tied to a single bankruptcy. Both modes of expansion thus division of the conglomerate. Julia D’Souza raise the risk of the family losing control. and John Jacob (2000) report a positive 7 abnormal stock return upon the creation of Martin Holmén and Högfeldt (2005) propose an alternative, but closely related, model of a “pecking order the- tracking stocks, indicating that many related ory of finance” induced by insiders’ extraction of private stocks controlled by the same top managers benefits of control. sp05_Article 1 8/18/05 9:32 AM Page 674

674 Journal of Economic Literature, Vol. XLIII (September 2005)

In essence, Almeida and Wolfenzon (forth- highly profitable grow slowly. If pyramidal coming-a), augmented by wealthy family’s groups prosper because they are especially desires to retain control, echo the explanation adept rent-seekers, they perhaps ought to of Yoshisuke Aikawa (1934), the founder of contain the best performing firms in the the Nissan pyramidal corporate group in pre- worst performing countries. Further work is war Japan, of why he opted for a pyramidal needed to clarify these issues. structure. He explains that pyramids are an For example, the determinants of pyra- ideal solution to what he calls the “capitalist’s miding are not well understood. Almeida quandary.” If a capitalist uses only his own and Wolfenzon (forthcoming-a) and Aikawa money or his family’s money, his scale of (1934) correctly emphasize that pyramids operations is too small. If he taps public equi- are formed because they magnify a given ty markets, he risks losing control. But, as dis- level of personal wealth into control over cussed above in connection with figure 1, a corporations worth vastly more. However, pyramidal group provides the best of both more work is needed to clarify how the pref- worlds—secure control and unlimited access erence of the controlling owner enters the to public capital. 8 picture and why firm units are sometimes Finally, Morck, Stangeland, and Yeung transferred between pyramidal groups and (2000) and Morck and Yeung (2004) pro- in and out of pyramids. pose that political influence is proportional For example, Berle and Means (1932) and to what one controls, not what one actually others argue that a pyramid’s lower-tier firms owns. As we explain in more detail below, are especially inefficiently run because of this might allow the controlling owners of the large gap between the controlling pyramidal groups unrivaled political influ- owner’s control rights and cashflow rights. ence in many countries. Morck and Yeung This inefficiency should attract takeovers. In (2004) argue that, by magnifying the wealth general, control should pass to the party who of an individual or family into control over values the cashflow rights plus private bene- corporate assets worth vastly more, pyra- fits of control the most. The private benefits mids magnify political influence in the same include tangible ones, such as money from proportion. The wealthy individual or fami- tunneling, and intangibles, like the power ly can use this amplified political clout to and influence derived from ruling a business alter the economy’s institutional framework empire—see Morck, Stangeland, and Yeung to favor themselves or their firms, to cap- (2000) and Morck and Yeung (2004). ture transfer payments, and so on. This If tunneling were costless, the controlling advantage might be considerable, especially owner could appropriate all the benefits of in economies with readily corruptible politi- enhanced efficiency in all her pyramid’s cians and officials. Indeed, it might be so firms. If intangible benefits were also unim- large as to explain apparently superior firm portant, the pyramid controlling owner level performance, as found by Khanna and capable of running firms most efficiently Palepu (2001a) and Khanna and Rivkin should pay the most for control blocks, and (2001), despite obvious evidence of poor maximally efficient management should pre- governance and overall poor economy per- vail. Arguably, the small personal wealth formance. Anne O. Krueger (1993), Kevin needed to acquire control via a pyramid and M. Murphy, Shleifer, and Vishny (1991, the subsequent extraction of all the benefits 1993), and many others argue that of enhanced efficiency via costless tunneling economies in which political rent-seeking is should make efficiency enhancing corporate takeovers loom larger as attractive investments than they would be in an economy of 8 See Morck and Nakamura (2004) for details. freestanding firms. sp05_Article 1 8/18/05 9:32 AM Page 675

Morck, Wolfenzon, and Yeung: Corporate Governance 675

Alternatively, if tunneling were costly or Second, in family-controlled pyramids, intangible benefits important, control family members often retain top management would pass to the controlling shareholder positions. Succession by inheritance can have most adept at tunneling or most enthusias- a negative effect because top management tic about consuming intangible private positions go, not to the most capable, but, at benefits of control. The former considera- best, to the most capable member of the con- tion leads Lucian Arye Bebchuk, Reinier trolling family, as modeled by Caselli and Kraakman, and George G. Triantis (2000) Gennaioli (2003). to postulate that control in an economy Third, even were resources efficiently with pyramiding might pass to the most managed and allocated within a group con- efficient thief. If the controlling sharehold- trolled by single family, corporate groups er consumes intangible benefits directly from can still be undesirable. Almeida and all the firms she controls, but consuming Wolfenzon (forthcoming-b) argue that the pecuniary benefits require costly tunneling, efficient allocation of resources within a the small personal wealth needed to acquire group can actually exacerbate the ineffi- control magnifies the importance of intangible cient allocation of resources across groups. private benefits in allocating corporate con- Their arguments apply to corporate groups trol. Control then might pass to the most in general, not just to family controlled power-hungry oligarch. pyramids. Much additional theoretical and empiri- Finally, the concentrated control of pro- cal work is needed to solidify hypotheses in ductive assets by a few families gives them this area and to distinguish presumptions market power, both in the goods and capital from facts. markets, leading to potentially undesirable economic consequences. Morck and 5. Control Pyramids and the Separation Nakamura (2004) explain how this was the of Ownership from Control explicit rationale cited by the United States Occupation Force for dismantling the main The discussion above suggests that, even Japanese family control pyramids after were control pyramids economically salubri- World War II. ous, for example as devices for circumvent- Widespread problems of these types, ing inadequate institutions, there are especially those that impair capital mar- nonetheless reasons for taking a more skep- kets, can create inefficiencies that hamper tical view. Certainly, Aikawa’s (1934) enthusi- outside financing, discourage innovation, asm for control pyramids demonstrates scant and retard economic growth. These ineffi- concern for the rights of public shareholders. ciencies can arise at both the firm level and The first reason for skepticism is that a overall economy level. In this section, we pyramidal corporate ownership structure focus on inefficiencies at the firm level. allows the controlling shareholder to These stem primarily from the separation secure control rights without commensu- of ownership and control in the pyramid— rate cash flow rights. Secured control rights the first two points enumerated above. The protect the controlling owner from losing next section focuses on inefficiencies at the power—that is, they allow her entrench- economy level which stems from the con- ment. Furthermore, as we show below, trol of a very large number of firms by a once control is secured, the controlling single family—the last two points listed shareholder’s low cash flow rights lead to above. This section is deliberately brief, for agency problems, including non-value max- other surveys, notably Shleifer and Vishny imizing investment and incentives to divert (1997) and Diane K. Denis and John J. resources. McConnell (2003), cover some of this sp05_Article 1 8/18/05 9:32 AM Page 676

676 Journal of Economic Literature, Vol. XLIII (September 2005)

material, though not from the perspective only a small fraction of their cash flow we emphasize here. rights. Indeed, we presented examples in which the cash flow rights of the controlling 5.1 Divergence of Interests Agency family in some of the pyramid member Problems firms are comparable to the stakes of the The problems caused by the separation of managers of the most diffusely held of U.S. ownership and control in pyramidal groups corporations. By allowing cash flow rights are similar to those described in Berle and and voting rights to diverge, control pyra- Means (1932) and Jensen and Meckling mids permit the same divergence of interest (1976) in connection with widely held firms. problems as dispersed ownership, even Indeed, Berle and Means explicitly describe though the firms in the pyramid are not pyramids as having problems like those of widely held. widely held firms. Jensen and Meckling This divergence of interests can lead to argue that the managers of firms with dis- inefficient investment in firms in which a persed shareholders have substantial discre- controlling owner has small cash flow tion as to the actions they take because rights. This is because the controlling fami- individual shareholders cannot coordinate ly earns only a small part, corresponding to to share monitoring and control costs. This its small cash flow rights in such a firm, of lets managers take actions with benefits any investment’s monetary payoff but can that, by their nature, are not shared with retain all of any private benefits the invest- shareholders. In addition to indirect finan- ment generates. Only in the special case of cial benefits, like on-the-job consumption or an investment that maximizes the total sur- shirking, and direct financial benefits, like plus also maximizing the controlling owner’s redirecting corporate assets into a personal private benefits should we observe efficient account, control also provides intangible investment decisions. benefits, like status, political influence, and A related reason for inefficient invest- power over people. The literature denotes ment arises from the high cost of capital that all of these benefits as private benefits of this divergence of interest entails. This high control. cost stems from minority shareholders’ Managers pursue private benefits, but rational expectation that this divergence of balance this against their losses when the interest distorts corporate decisions to ben- price of their shares in the firm falls plus the efit controlling shareholders. Such problems consequences of getting “caught and pun- are defined in the Corporations Law of ished,” such as losing their jobs. This bal- many countries as oppression of public ance means managers might choose an shareholders by the controlling shareholder. action with lower total value over one with Oppression problems are thus a variant of higher value, as long as the former generat- “perks” consumption, where the perks are ed sufficient private benefits. This diver- diversions to benefit the controlling share- gence of interests is greater the lower the holder rather than professional managers. managers’ equity stake and the lower the La Porta, Lopez-de-Silanes, and Shleifer likelihood of her getting caught and punished (1999) document countries varying substan- for non-value-maximizing behavior. tially regarding the protection their laws As we described in section 3, firms out- afford external investors against oppres- side the United States and United Kingdom sion. Shleifer and Wolfenzon (2002) show are owned, not solely by dispersed small that, when a controlling shareholder has a shareholders, but by wealthy families via low cash flow stake, as with firms at the control pyramids. Pyramids allow a family to base of a pyramid, the cost of capital to the retain control of many firms while holding firm is high since other investors anticipate sp05_Article 1 8/18/05 9:32 AM Page 677

Morck, Wolfenzon, and Yeung: Corporate Governance 677

expropriation and pay depressed prices for entrenchment is especially difficult to count- any securities that firm sells. A consequence er, for if the controlling shareholder used his of this high cost of capital is severely distort- control rights to enrich himself pecuniarily ed investment decisions because of those at the expense of public shareholders, this firms’ reluctance to raise external funds. could be proscribed as “oppression.” Of course, this picture is overly simplified. 5.2 Entrenchment Agency Problems In principle, without maintaining a domi- The corporate finance literature notes a nant voting share, the CEO of a widely held second type of agency problem, called firm could also dominate the board through entrenchment. While Jensen and Meckling sheer force of personality and push through (1976) argue that insiders with larger stakes an array of antitakeover defenses, such as have less incentive to misallocate corporate poison pills, staggered boards, and the like, resources, Stulz (1988) argues that higher to deter raiders and activist shareholders. equity stakes also give insiders more free- Managerial entrenchment can occur in dom to misallocate resources. This is freestanding firms. Indeed, Almazan and because top executives who own large blocks Suarez (2003) argue that a degree of of equity are effectively tenured. Non-value- entrenchment might be part of an optimal maximizing managers with little stock can be compensation contract for top managers. removed in proxy fights by disgruntled insti- Nonetheless, as Nissan founder Aikawa’s tutional investors or ousted by hostile (1934) unwittingly emphasized, a fundamen- raiders attracted by a depressed share price. tal raison d’être of control pyramids is pre- Top executives with large blocks of equity cisely that they lock in control. That is, a cannot be cast out in these ways. Thus, a pyramidal structure is itself a means to dominant equity stake lets corporate insid- entrench the controlling shareholder with- ers enjoy private benefits of control robustly, out the cost of maintaining a large equity whereas a smaller stake means they must stake nor the machinations of having to limit their inhalation to keep their share establish poison pills or staggered boards. prices high enough to ward off raiders and Control pyramids per se are simple and placate institutional investors. highly effective antitakeover devices. Moreover, entrenchment problems can To see this, return to figure 1 and recall take on a qualitatively distinct air, for the discussion in section 3.2. The family entrenchment can lock in control by honest appoints the boards of directors of compa- but inept insiders as well as clever self-serv- nies in the first tier and so can place family ing insiders. For example, entrenchment members in dominant positions on those problems occur if a firm’s controlling boards and as top executives. The first tier founder bequeaths her stock to an egre- firm’s boards then appoint the boards of giously incompetent, but power hungry son. companies in the second tier, which then The incompetent son cannot manage the appoint the boards of companies in the third firm well, but cannot be removed by the tier. This locks in family control over all the other shareholders because his share of companies in the pyramid, even though pub- votes is sufficiently large to control the lic shareholders contribute most of the board. The power hungry son gains such financing for most of its firms. In contrast, utility from controlling the firm that the had the family established a single firm with effects of his blunders on the value of his eight billion dollars worth of physical assets stock are an acceptable cost to him. While (a union of the firms in the third layer), its this might even be Pareto efficient, it clearly billion dollars in family wealth would have raises important distributive questions to constituted a 12.5 percent stake and public which we return below. This sort of shareholders could outvote the family in sp05_Article 1 8/18/05 9:32 AM Page 678

678 Journal of Economic Literature, Vol. XLIII (September 2005)

shareholder meetings. This would render between controlled firms via transfer pric- family control contestable. ing, the provision of capital at artificial Note further that entrenchment via family prices; or via inflated payments for intangi- control pyramids is intimately linked to the bles such as patents, brand names, and appointment of family members to the boards insurance. Although the firm at the receiving of the apex firm and other firms throughout end of the tunnel benefits from the wealth the pyramid. The example above—allowing transfer, public shareholders of the wealth “an egregiously incompetent, but power hun- donating firm understandably might view gry son” to retain control—underscores the the transfer as a governance problem. potential for damage to firm value due to Johnson, La Porta, Lopez-de-Silanes, and entrenchment. But the potential for damage Shleifer (2000) provide detailed examples of goes beyond the controlling manager having tunneling. excessive freedom to pursue her self-interest The transfer of funds from one group firm at the expense of firm value. It also includes to another is integral to many of the socially the possibility that an incompetent manager desirable functions of corporate groups might retain her job precisely because she posited above. Tunneling undertaken with can extract more personal gains from the job such ends in mind perhaps ought to be than a competent manager could. Indeed, encouraged, not condemned as a gover- Bebchuk, Kraakman, and Triantis (2000) sug- nance problem. Indeed, such transfers gest that control might naturally pass to “the might well balance out over time for any most efficient thief” because he is willing to given group firm. pay the most for a control block. However, the controlling owners of pyramids benefit from the systematic percolation 5.3 Tunneling of wealth from firms near the pyramid’s base The firms low in control pyramids, in which upward to firms near its apex. To see this, the controlling shareholder’s cash flow rights return again to figure 1. Suppose that firm3,1 are small, are thus vulnerable to concurrent generates new wealth worth one million dol- divergence of interests and entrenchment lars. As was shown above, this translates into problems. One consequence of this juxtaposi- additional wealth of $125,000 for the ulti- tion is that the controlling shareholder has an mate owners. The remaining $875,000 incentive to transfer money aggressively out accrues to public shareholders in the various of these firms and into firms near the pyra- tiers of the pyramid. But if firm overpays mid’s apex—or even into her personal hold- 3,1 firm1.1 by one million dollars for some good ings. Of course, these transactions can or service, the new wealth is “tunneled” to require complex schemes to be effective. firm1,1. This means the controlling owner Nonetheless, when the controlling family benefits by $500,000 and the public share- runs many firms, they can take the form of holders of firm3,1 and firm2,1 get none of the intercorporate business transactions. new wealth. In the most extreme case of The act of transferring value from one tunneling, the wealth is transferred directly pyramid firm to another is dubbed tunneling to the family firm at the apex or to another by Johnson, La Porta, Lopez-de-Silanes, and 10 9 firm that is fully owned by the family. In Shleifer (2000). Value can be transferred these ways, tunneling lifts assets and income

9 Tunneling is of course another type of agency problem, but we treat it separately here due to its importance. An agency problem is one where an agent, in this case con- 10 In practice, the process of tunneling might well trolling family, can take actions that negatively affect the destroy part of the wealth being transferred, so the con- utility of the principal, in this case the minority sharehold- trolling owner might net less than the full million dollars ers. In this sense, both entrenchment and tunneling are because of lawyers fees, payments for discrete financial agency problems. services, and the like. We return to this point below. sp05_Article 1 8/18/05 9:32 AM Page 679

Morck, Wolfenzon, and Yeung: Corporate Governance 679

from lower to higher tier firms, and dumps less able managers who particularly value losses and liabilities from higher to lower control are especially likely to be locked into levels of the pyramid. positions of corporate power, and that ineffi- Tunneling does not require a control pyra- cient investment might plausibly be associat- mid of firms. For example, the Enron scan- ed with control pyramids.11 If these dal in the United States apparently involved problems are isolated incidents, any macro- the transfer of wealth from Enron, a widely economic effects are marginal. Indeed, pyra- held company, to private companies con- mid firms with poor corporate governance trolled by Enron’s professional managers. ought to be out-competed, and their neg- However, tunneling is plausibly a more lected investment opportunities ought to be everyday concern in economies dominated appropriated by better-run firms. However, by control pyramids. if control pyramids allow a very small num- In passing, tunneling should be familiar to ber of controlling shareholders to amass con- students of multinationals, which are known trol over so many firms that they govern a to engage in an analogous shifting of earn- substantial part of the corporate sector, their ings and assets to avoid declaring income in suboptimal behavior might aggregate to a high tax countries. The key differences are: macroeconomics problem. (1) tunneling is about hiding money from Whether these problems outweigh any public shareholders, while income shifting positive effects associated with pyramidal by multinationals is about hiding money control is unclear a priori. Corporate gover- from the tax authorities, and (2) tunneling nance problems might plausibly be expected can involve solely domestic firms. In both to dominate in situations where the control- cases, however, the shifting of assets and lia- ling shareholder is less able, entrepreneurial, bilities among a multitude of related corpo- or honest. The beneficial effects listed earli- rations complicates the workings of the er in this section might plausibly dominate firms in question and renders them less where entrepreneurial talent is scarce, mar- transparent to investors. ket institutions are weak, bankruptcy costs In summary, freestanding firms are most high, and so on. The actual balance is thus an vulnerable to divergence of interests agency empirical question. problems if widely held and to entrenchment problems if narrowly held, though this 5.4 Empirical Evidence is an oversimplification. In contrast, firms in We now review a large literature that lower tiers of control pyramids are vulnera- helps us answer this question. While some of ble to both problems simultaneously. Their this work focuses on freestanding firms, it dual presence is mutually reinforcing. nonetheless allows us to make carefully Entrenchment allows a more fulsome diver- guarded inferences about control pyramid gence of interests agency problem, which firms. While a rapidly growing body of work raises the value of becoming entrenched, including maintaining the pyramidal struc- 11 ture and keeping family members in key One might expect poor corporate governance to raise a firm’s outside capital costs. The issue, however, is com- executive positions and on key boards. plicated. A firm’s cost of outside capital is affected by many Tunneling further exacerbates both prob- factors. Pyramidal groups might have capital market power lems by allowing an entrenched controlling so that their firms’ costs of outside capital might even be lower than those of stand alone firms. For example, pyram- shareholder to appropriate corporate idal firm groups might be more favored clients of banks resources and to raise a corporate veil than stand alone firms because of their sheer size or the against outsider monitoring. pyramids themselves might contain banks. The prevalence of pyramidal groups has implications on capital market This unhappy confluence of firm-level conditions and hence on the overall allocation of outside corporate governance problems means that financing in an economy. We revisit these issues below. sp05_Article 1 8/18/05 9:32 AM Page 680

680 Journal of Economic Literature, Vol. XLIII (September 2005)

also focuses specifically on pyramid firms, Bruce Johnson, Robert P. Magee, Nandu J. further effort along these lines is likely to be Nagarajan, and Henry A. Newman (1985), highly productive. who report that stock prices often rise significantly upon the announcement of CEO 5.4.1 Firm-Level Studies of Performance sudden deaths. This suggests that CEOs often stay in office when they might better As mentioned above, Khanna and Palepu contribute to their firm’s value by retiring. (2000a) and Khanna and Rivkin (2001) both They presumably do this because they enjoy find that firms belonging to pyramidal private benefits of control. Morck, Shleifer, groups in developing countries outperform and Vishny (1988) also report large and sig- independent firms on average. This could be nificant value discounts for U.S. firms with due to any of the explanations listed above large shareholders and for U.S. firms estab- for business groups having an advantage. lished more than thirty years ago and run by Eric Friedman, Simon Johnson, and Todd founders or their relative. They interpret Mitton (2003) present evidence consistent both findings as evidence of managerial with transfers of wealth from controlling entrenchment. Myron B. Slovin and Marie shareholders benefiting public shareholders E. Sushka (1993) relate the abnormal return in pyramid firms. They describe both specif- upon CEO sudden deaths to the departed ic instances of pyramid controlling owners manager’s equity stake and confirm the “propping up” firms to the benefit of their importance of entrenched management in public shareholders during the Asian Crisis narrowly held U.S. firms. of the 1990s and econometric results consis- While these studies use U.S. data and tent with this occurring on an economically therefore refer to freestanding firms, it important scale. They propose that, under seems likely that similar entrenchment prob- some circumstances, controlling owners lems exist elsewhere. This is because the credibly promise to prop up distressed group United States has a well developed corporate firms in order to attract external debt financ- takeover market, sophisticated institutional ing. This is essentially a special case of the investors, and litigious shareholders, all of internal capital markets hypothesis whom are likely to work hard at dislodging described above. Nonetheless, Jae-Seung entrenched corporate insiders. La Porta, Baek, Jun-Koo Kang, and Kyung Suh Park Lopez-de-Silanes, Shleifer, and Vishny (2004) report that pyramid firms tightly con- (1998) present evidence that these pressures trolled by wealthy Korean families exhibit are almost surely much weaker in most other worse stock market performance than inde- countries—especially Civil Law countries, in pendents during the crisis. Indeed, Graham which pyramids and dominant family share- (2003), Kang (2002), and many others docu- holders are generally quite important. ment the history of scandal, political rent- A low sensitivity of top manager turnover seeking, and bailouts that characterizes to firm performance may also indicate Korean pyramidal groups. entrenchment. Here, evidence more directly related to family control pyramids is 5.4.2 The Empirical Importance of available. Volpin (2002) shows that, when Entrenchment Problems the controlling shareholder is also the CEO, CEO turnover is largely unrelated to Entrenchment problems are clearly impor- firm performance. Interestingly, within tant in the United States and are likely to be control pyramids, he finds that family even more important in other countries. The members are replaced less often than first empirical evidence now recognized as professional managers and that their sensitivi- indicative of managerial entrenchment is W. ty of turnover to performance is significantly sp05_Article 1 8/18/05 9:32 AM Page 681

Morck, Wolfenzon, and Yeung: Corporate Governance 681

lower than that of nonfamily CEOs. Giorgio believe that the controlling party obtains Brunello, Clara Graziano, and Bruno M. benefits over and above those that are Parigi (2003) report a similar result—CEO shared with public shareholders. This is evi- turnover is elevated in poorly performing dence of entrenchment. In the absence of Italian firms only if the CEO is not affiliated entrenchment, these benefits would be with controlling shareholder. ephemeral (and their value close to zero), for Although it seems plausible that entrench- public shareholders would react by selling to ment is more pervasive in countries with a raider or supporting dissidents at a share- more extensive family control pyramids, fur- holders’ meeting. Persistent large private ther work is needed to make this conclusion benefits of control imply both a divergence unequivocal. At present, we must rely on of interest and insider entrenchment. country and case studies that may not be There are two ways to measure the value directly comparable. For example, Igor of control. One is to infer it from the block Filatotchev, Mike Wright, and Michael premium, defined as the difference between Bleaney (1999) document Russian managers’ the price per share paid in a block transac- remarkably blatant efforts to entrench them- tion and the market price after the transac- selves. Abe de Jong and Chris Veld (2001) tion. This premium is a lower bound on the argue that Dutch managers are profoundly value the acquirer places on control. An entrenched around old-boy networks, thanks early example of this methodology is to explicit protective arrangement depriving Michael J. Barclay and Holderness (1989), shareholders of their voting rights. Ugurlu who study block premiums in the United (2000) argues that managerial entrenchment States. Alexander Dyck and Zingales (2004) is pervasive in Turkey. apply the same methodology across a sample Unfortunately, comprehensive cross- of thirty-nine countries by examining 412 country studies that could decide the matter control transactions between 1900 and 2000. are not yet available. Studies across large Their results, summarized in table 3, show numbers of countries comparing the deter- that the value of control ranges from –4 per- minants of CEO turnover would be highly cent to 65 percent, and averages 14 percent useful. Studies using data for other countries of firm value. In countries where block pre- that rerun the Perez-Gonzales (2002) analy- miums are larger, Dyck and Zingales (2004) sis of the stock price reactions to family ver- find that capital markets are less developed, sus professional CEO succession would also minority shareholders have fewer legal be highly useful. rights, and law enforcement is uncertain. Interestingly, they also find that nonlegal fac- 5.4.3 Valuing the Private Benefits of tors, such as more diffuse ownership of the Control press, stronger product market competition, We argued above that entrenchment frees and higher levels of tax compliance are relat- corporate insiders to extract private benefits ed to lower block premiums. Indeed, these of control more energetically, and that pyra- factors have at least as much explanatory mids may magnify the available benefits. It is power as the legal measures. therefore useful to asses the magnitude of The other method for estimating the pri- private benefits of control in different coun- vate value of control is based on the differ- tries and for different sorts of firms and ence between the market prices of different owners. classes of stock. In many countries, firms A way to measure corporate insiders’ pri- have more than one class of common shares. vate benefits from control is to estimate the Public shareholders invest in a class of value the market attaches to control rights. If shares, called restricted voting shares, with this value is positive, market participants one or sometimes no votes per share. The sp05_Article 1 8/18/05 9:32 AM Page 682

682 Journal of Economic Literature, Vol. XLIII (September 2005)

TABLE 3 ESTIMATED PRIVATE BENEFITS OF CONTROL IN DIFFERENT COUNTRIES MEASURED AS BLOCK AND VOTING PREMIUMS AND EXPRESSED AS PERCENTAGE PREMIUM OVER MARKET VALUE

Block Voting Block Voting Country Premium1 Premium2 Country Premium1 Premium2 Argentina 27 Netherlands 2 Australia 2 23 New Zealand 3 Austria 38 Norway 1 6 Brazil 65 23 Peru 14 Canada 1 3 Philippines 13 Chile 15 23 Poland 11 Colombia 27 Portugal 20 Czech Republic 58 Singapore 3 Denmark 8 1 South Africa 2 7 Egypt 4 South Korea 16 29 Finland 2 5 Spain 4 France 2 28 Sweden 6 1 Germany 10 10 Switzerland 6 5 Hong Kong 1 –3 Taiwan (China) 0 Indonesia 7 Thailand 12 Israel 27 Turkey 30 Italy 37 29 United 2 10 Kingdom Japan –4 United States 2 2 Malaysia 7 Venezuela 27 Mexico 34 36

1 Block premium is average across control transactions of the difference between the price per share paid for the control block and the exchange price two days after the announcement of the control transaction, dividing it by the exchange price two days after the announcement and multiplying the ratio by the proportion of cash flow rights represented in the controlling block and expressed as a percentage premium. See Dyck and Zingales (2003), table 2, for details. 2 Voting premium is average of estimated total vote value as a percent of firm value. See Nenova (2003), table 5, for details.

controlling owner holds super-voting shares, This methodology has been applied to dif- which bestow superior voting rights. Such ferent single countries in different studies, structures allow the controlling owner to however, this makes the results difficult to command a majority of votes in the share- compare.12 Tatiana Nenova (2003) uses this holder meeting while owning relatively few technique for eighteen countries using shares. Both classes of shares sometimes comparable 1997 data on 661 firms with such trade in public equity markets. Because only the shares with superior voting rights 12 See Kristian Rydqvist (1996) and Clas Bergstrom and need be purchased in a sale of control, Rydqvist (1990, 1992), for Sweden; Levy (1992) for Israel; Ronald C. Lease, McConnell, and Mikkelson (1983) and they ought to command a value premium if Zingales (1995b) for the United States; Zingales (1994) for control provides private benefits. Italy, etc. sp05_Article 1 8/18/05 9:32 AM Page 683

Morck, Wolfenzon, and Yeung: Corporate Governance 683

dual class shares. In doing this, she controls for very useful approaches to determining pri- the probability of a sale of control, block-hold- vate benefits of control. Much more work ing costs, and dividend and liquidity differ- could be done along these lines. For ences between the share classes to estimate a instance, both studies are cross sectional. lower bound on the private benefits of control. Work on how these measures vary over The estimates she reports, also shown in table time in response to institutional changes, 3, vary from near zero for Finland, and less legal reforms, and the like would be of than 4 percent of firm value in Canada, great value. Denmark, Hong Kong, Sweden, and the United States, to over 50 percent of firm value 5.4.4 Entrenchment, Private Benefits of in Brazil, Chile, France, Italy, South Korea, Control, and Pyramids and Mexico. Nenova’s (2003) regressions show that the control premium is significantly lower We argued above that controlling owners in countries with stronger protection for out- of pyramidal groups of listed companies are side investors’ property rights, and finds that most likely to sacrifice the interests of public 68 percent of the cross-country variation in shareholders in firms low in their pyramids, the value of control-block votes is explained by for the controlling owner’s cash flow rights in her measures of investor rights. Franco these firms are minimal. Several empirical Modigliani and Perotti (2000), using country studies present evidence consistent with level average from Canada, Israel, Italy, such an effect. Sweden, Switzerland, the United Kingdom, Claessens, Djankov, Joseph Fan, and and the United States, show that high control Lang (2002) examine 1996 data on 1,301 premium countries have smaller equity mar- publicly traded corporations in eight East kets and larger proportion of bank loans, Asian economies—Hong Kong, Indonesia, which are characteristics of poor shareholder South Korea, Malaysia, the Philippines, rights countries. Singapore, Taiwan, and Thailand. Their Not unexpectedly, private benefits of con- sample is 37 percent of the 3,544 publicly trol are significantly higher in countries traded firms in those countries. The study whose large corporate sectors are more com- shows that firm value, measured by market pletely controlled by wealthy families. The to book ratio, rises with the cash-flow rights simple correlation coefficient of the control of the largest shareholder and falls as the block premium, as reported in table 3, with control rights of the largest shareholder the proportion of top ten firms whose con- decline relative to her cash-flow rights. The trolling shareholders are families with 20 latter result is especially pertinent, for the percent blocks, as reported in La Porta, primary reason for control rights to diverge Lopez-de-Silanes, and Shleifer (1999) and from cash flow rights in these countries is reproduced in table 2, is 0.41. If family con- pyramidal groups. This result is thus consis- trol is inferred from a 10 percent largest tent with controlling owners having less block, the correlation coefficient is 0.37. concern for maintaining the value of firms Both are statistically significant at 5 percent. lower in their control pyramids. Unsurprisingly, the correlation between the Sung Wook Joh (2003) reexamines this control block premium and the proportion issue using data from Korea, a country of large firms that are widely held is –0.45 if whose corporate sector is heavily dominated widely held is defined as having no 10 per- by large family controlled pyramidal groups, cent blockholder, and –0.61 if a 20 percent or chaebol. Using data on 5,829 publicly threshold is used. traded and private firms from 1993 through The techniques of Dyke and Zingales 1997, Joh finds a higher excess of control (2004) and Nenova (2003) are potentially rights over cash flow rights associated with sp05_Article 1 8/18/05 9:32 AM Page 684

684 Journal of Economic Literature, Vol. XLIII (September 2005)

lower profitability. Again, this is consistent voting rights. However, Michael L. Lemmon with controlling owners having less regard and Lins (2003) find a stronger result. They for the financial performance of firms lower use a dataset of 800 firms in eight East Asian in their pyramids. countries during the 1997–98 financial crisis. Karl V. Lins (2003) uses 1995 data on After controlling for size, beta, debt to assets 1,433 firms from eighteen emerging markets ratio, and market-to-book ratio, they find to study large equity blockholders.13 He stock returns of firms in which ultimate own- reports that two thirds of insider controlled ers’ control rights exceed their cash flow firms belong to control pyramids and that right to be 10 to 20 percentage points lower insiders’ voting rights average 2.7 times their than those of other firms. For firms in which cash flow rights in those firms. Regressions insiders have below median control stakes or controlling for firm size, investment rate, in which the largest block holder is not an leverage, and country show firm values to be insider, performance is independent of depressed in proportion to the excess of insider’s control rights relative to their cash insider blockholders’ control rights over flow rights. their cash flow rights.14 This is consistent The above empirical studies, which pro- with more pronounced agency problems in vide evidence on the relationship between firms in lower tiers of control pyramids. entrenchment and private benefits of con- Johnson, Peter Boone, Alasdair Breach, trols within pyramidal structures, are mostly and Eric Friedman (2000) point out that the based on East Asian countries that have rel- controlling owners’ diversion of resources atively weak and corruption-prone institu- may depend on the anticipated future need tions, at least according to La Porta, for outside financing in each firm in a con- Lopez-de-Silanes, Shleifer, and Vishny trol pyramid or other group. When future (1998). Unfortunately, comparable investi- investment opportunities are unexpectedly gations using data from countries with more curtailed, trimming the need for internal developed legal and other institutions, such funds, a controlling owner may prefer to as Western European countries, are lacking. siphon off the funds rather than pay them This is an exciting area for further work. out to investors in that firm. This is a specif- 5.4.5 Tunneling ic empirical prediction that identifies insiders’ tendency to pursue self-interest while We argued above that tunneling might balancing against repercussions, and there- serve as an especially effective way for fore is a rather convincing empirical test. entrenched controlling shareholders to Mitton (2002) compiles a sample of 398 extract private benefits of control. The stud- firms in five East Asian countries and finds ies discussed above that link the excess of a that firm performance during the 1997–98 controlling owner’s control rights over his East Asian financial crisis—a period when cash flow rights to depressed financial per- investment opportunities in that region formance or firm value are consistent with abruptly shrank—is only marginally affected either tunneling or a simple neglect of the by the divergence between cash flow and interests of public shareholders in lower tier pyramid firms. Direct evidence of the rela- 13 These countries include Argentina, Brazil, Chile, tive cross-country importance of tunneling is Czech Republic, Hong Kong, Indonesia, Israel, Malaysia, yet to be gathered. However, country studies Peru, Philippines, Portugal, Singapore, South Africa, South Korea, Sri Lanka, Taiwan (China), Thailand, and again provide evidence that tunneling can be Turkey. economically important. 14 Firm value creation or destruction is measured by The findings of Marianne Bertrand, Paras high or low Tobin’s average q, respectively. Average q is approximated by market value of equity plus book value of Mehta, and Sendhil Mullainathan (2002) liabilities scaled by total assets. concerning Indian family control pyramids, sp05_Article 1 8/18/05 9:32 AM Page 685

Morck, Wolfenzon, and Yeung: Corporate Governance 685

which they call business groups, are consis- well as simple pyramidal intercorporate own- tent with tunneling and are difficult to ership links. They ask whether chaebol firms explain with any other hypothesis. They benefit from acquisitions they make, or divide firms into free-standing firms, many whether such benefits appear instead in firms of which have controlling families, and firms in higher tiers of the pyramid. Using data on belonging to family control pyramid groups. 107 mergers involving eighty-seven firms Their focal dependent variable is cash flow between 1981 and 1997, they find that when which is defined as earnings before depreci- one firm buys another in the same chaebol, ation, interest, and taxes over assets. Using the acquirer’s stock price falls but the stocks data from 1989 to 1999 and regressions con- of other firms in the same chaebol rise. The trolling for the controlling owner’s cash flow authors argue that the controlling owner typ- rights and firm size, they find that control ically has lower cash flow rights in the acquir- pyramid firms’ cash flows are less sensitive to er than in the target, for which the acquirer own industry factors than are those of free- typically overpays. Intra-chaebol acquisitions standing firms15 This sensitivity to own thus pass wealth away from firms whose cash industry factors is lower among firms in flows accrue more to public shareholders and which the controlling owner’s cash flow up toward firms whose cash flows accrue rights are lower (relative to the group medi- more to controlling families. This is consistent an). In addition, group firms’ performance is with tunneling. positively associated with the industry fac- A few papers do not find evidence for tun- tors of other affiliates in which the control- neling in pyramids. Holmén and Högfeldt ling owner’s cash flow rights are low. (2005), using the procedure of Bertrand, Interestingly, positive cash flow shocks in the Mehta, and Mullainathan (2002), find no evi- lower tier firms are echoed by positive dence of tunneling in Swedish pyramidal shocks in upper tier firms’ cash flows, but groups. They argue Sweden’s high account- that the converse is not true. A positive ing and judicial standards make tunneling shock to a low cash flow rights firm’s indus- difficult. However, they find that pyramid try increases the cash flow for high cash flow member firms overinvest significantly. One rights affiliates about twice as much as for interpretation of this is that Swedish control- low cash flow rights affiliates. Thus, firms in ling families extract private benefits via cor- which the controlling owner’s cash flow porate empire-building but not via tunneling. rights are highest—that is, firms located 5.5 Summary and Implications nearest the apex of the pyramid—benefit the most from a positive shock to firms else- In this section, we showed that family where in the group. This is consistent with control pyramids allow the simultaneous the controlling owner transferring net presence of divergence of interests agency resources from lower to higher tier firms in problems, of the sort discussed by Jensen the control pyramid—that is, with tunneling. and Meckling (1976), and entrenchment Bae, Kang, and Kim (2002) also report agency problems, of the sort discussed by direct evidence of tunneling in Korean fam- Stulz (1988). This means family control pyr- ily controlled pyramids, or chaebols. amids are potentially subject to worse Chaebols, while basically pyramidal in agency problems than freestanding firms. nature, have extensive reciprocal holdings as This is because divergence of interest agency problems are most likely at their 15 Industry performance is an assets weighted average worst in widely held freestanding firms, cash flow across each industry. Cash flow is profits before while entrenchment problems are most depreciation, interest and taxes. Each firm’s industry factor is inner product of this ratio times a vector reflecting the likely at their worst in narrowly held free- distribution of its assets is across industries. standing firms. Lower tier firms in family sp05_Article 1 8/18/05 9:32 AM Page 686

686 Journal of Economic Literature, Vol. XLIII (September 2005)

control pyramids potentially suffer from and transition economies are in the first both problems simultaneously. The control- stages of establishing patterns of ownership ling owners of pyramid firms can easily have over large corporations. Intriguingly, control cash flow rights as small as those of the pro- pyramids seem quite popular in these coun- fessional managers of widely held firms, tries. For example, Robert Cull, Jana while exercising voting rights sufficient to Matesova, and Mary Shirley (2002) describe deter any control challenge, thereby Czech control pyramids. However, research entrenching their control. on the development of pyramidal group While divergence of interest and the dominance is only beginning and much entrenchment problems are well understood more work is needed. in the corporate finance literature, pyramids This section considers possible repercus- potentially aggravate agency problems by sions at the economywide level of entrusting conferring both on their lower-tier firms the governance of large swaths of a country’s simultaneously. corporate assets to small elites. Its focus is This unwelcome reinforcement of agency on established wealthy families, but other problems might be especially unpropitious if elites—professional managers or others— less able managers who particularly value also merit study. “control” become the controlling owners of We shall first describe capital allocation pyramidal groups. Since effectuating family problems associated with established pyra- control over pyramidal groups frequently mids. We shall argue that, even if established involves assigning family members to key pyramidal groups allocate capital efficiently executive and board positions and to corpo- internally, the overall allocation of capital rate boards, corporate governance power across pyramidal group boundaries can still may often be based on family ties and ability be inefficient. Moreover, this problem can to protect the controlling family’s interests, be exacerbated if pyramidal groups have rather than ability. The overall consequence capital market power. is that these firms may under perform from Second, extending the above arguments an asset utilization viewpoint. Finally, the leads to the possibility that an economy with lack of effective external monitoring in pyra- extensive established pyramidal structures mid firms whose governance is dominated may undertake less innovation. Because of by a powerful family may deprive even biased capital allocation, upstarts may have scrupulous managers of effective investor difficulties raising finance and thus invest less feedback regarding investment decisions.16 in innovation. Also, pyramidal group firms might invest less than freestanding firms in 6. Implications of Control Pyramids innovations that might compete against the current products of established group firms. So far, we have analyzed problems that Intriguingly, tunneling might actually miti- arise in individual firms in a pyramid by gate this last effect, for the controlling owner focusing on the controlling shareholder’s can benefit from innovation if she can tunnel behavior. In this section, we broaden the the gains toward firms in which her cash flow scope of the survey and review the literature stake is large while leaving the negative on the implications for the economy as a effects of creative destruction to public whole of a few old, established families con- shareholders. Further work on net welfare trolling a substantial fraction of the economy implication of these effects would be useful. through large pyramidal groups. This is an Finally, we speculate that the prevalence especially timely topic, for many emerging of pyramidal structures is associated with an economywide undersupply of external 16 See Durnev et al. (2004). capital. This speculation is based on the sp05_Article 1 8/18/05 9:32 AM Page 687

Morck, Wolfenzon, and Yeung: Corporate Governance 687

premise that a lack of corporate transparen- but expects only a fraction of the return cy reduces public investors’ expected over- when it invests in a firm controlled by some- all return. If this depresses overall domestic one else. Thus, the family may decide to allo- savings as well as the supply of capital from cate the capital within the group even if an abroad, upstart firms’ liquidity constraints outsider’s project were more productive. could be more binding than otherwise. We Almeida and Wolfenzon propose that the hasten to reiterate that this is speculative internal capital market in corporate groups and that further research is needed. may not even be a second best alternative All these issues point to pyramidal groups from a social perspective. They deduce that controlled by established families being the more efficiently capital is allocated with- associated with slow economic growth. in a conglomerate or group, the greater the likelihood that capital is misallocated overall. 6.1 Capital Allocation Problems Their argument parallels the view that We now consider how the allocation of regional free trade blocs augment welfare by capital across the boundaries of pyramidal increasing trade between member states but groups might be distorted even though cap- reduce welfare by diverting trade away from ital is efficiently allocated within each group. globally most efficient producers outside the While this problem is most likely aggravated bloc. Likewise, corporate groups make cred- if pyramidal group controlling owners have it more available to member firms but also capital market power, it can occur under divert credit away from efficient uses outside competitive capital markets too. the group. In both situations, the overall welfare effects are uncertain. 6.1.1 External Allocation Problems 6.1.2 Capital Market Power Almeida and Wolfenzon (forthcoming-b) propose that physical capital might be over- Section 3.4 showed that the large corpo- used within conglomerates and business rate sectors of many countries are made up groups in that more productive applications of a few large pyramidal groups and scatter- outside of the groups are forgone. Their ings of independent firms. Morck, argument suggests that, even if physical cap- Stangeland, and Yeung (2000) suggest that, ital is efficiently allocated within a group, due to their sheer size relative to the rest of the economywide allocation of capital might the economy, large pyramidal groups might still be inefficient in that groups have less attain a degree of price setting power in their incentive to relinquish productive assets to domestic capital markets. outside users than is socially optimal. To avoid confusing the current argument Their intuition is that, in an economy with with the Almeida and Wolfenzon (forthcom- poor investor protection, public shareholding-b) argument above, we assume away all ers rationally expect insiders to appropriate information and agency problems between part of the return on corporate investments. lenders and users of capital. This eliminates Consider an economy in which some existing any possibility of capital market underdevel- projects must be liquidated. A family owning opment due to institutional inadequacy. such a project, and controlling no other Now, assume that a pyramidal group has a firms, has no option but to lend the capital surplus of capital so large that it has a con- released in this liquidation to the outside siderable “market share” in the supply of entrepreneur with the highest productivity. capital. The group would supply capital for However, if the family controls other firms, internal and external use until the two alter- capital allocation can be distorted. This is natives generated identical marginal because the family earns the full return returns. For an internal application, the when it reinvests the capital within the group marginal return is the simple marginal sp05_Article 1 8/18/05 9:32 AM Page 688

688 Journal of Economic Literature, Vol. XLIII (September 2005)

return on capital. However, the marginal wealthy families in the typical country. The return for the external alternative reflects a State is the controlling owner 19 percent of downward sloping capital demand sched- the time. ule, which is the external users’ marginal Sharing a controlling owner presumably return on capital. The marginal return of lessens information asymmetry problems capital in an internal application is thus less between borrower and lender firms, and so than the external users’ marginal return. In allows firms in a pyramidal group preferen- other words, the group provides too little tial access to bank financing. This is thus a capital to external users at too dear a price. further extension of the above idea of busi- Along this vein, additional considerations ness groups allocating capital among their arise. First, outside investment might erode member firms at low cost relative to alterna- the controlling owner’s capital market power tive allocation mechanisms. Having a bank as as outsiders accumulate wealth from their an integral part of the group’s structure is not investments. Hence, when supplying capital always necessary, for the apex firm can serve to external users, a thoughtful controlling as a de facto bank in any event. Regardless, owner considers this negative effect, which it is instructive to consider the implications is absent when supplying capital to internal of group banks within the framework of the users. This should induce group firms to relationship banking literature. charge external users of capital more than A borrower having a relationship with his internal users. Second, gradations might bank mitigates information asymmetry prob- arise in the cost of capital within the pyramid lems and thus eases liquidity constraints, as related to the controlling owner’s cash flow in Peterson and Rajan (1994). However, rights in each firm. Rajan (1992) also shows that such a relation- In the alternative case, where the pyrami- ship locks a borrower into a long-term bond dal group has insufficient capital for its with its bank. This is because, once the rela- internal needs, its sheer size could give it tionship is established, switching banks sends monopsony “market power” in the sourcing a negative quality signal about the borrower of external capital. This, through a precisely to other prospective lenders. This makes the symmetric set of arguments to those out- borrower vulnerable to “hold-up problems.” lined above, induces the group to use too lit- If the bank and borrower firm are in the tle capital and to pay too low a rate of return same control pyramid, and so have the same to the providers of external capital. ultimate owner, both information asymmetry and hold-up problems might be mitigated. 6.1.3 Relationship Banking Problems Even if the bank is not a member firm of Pyramidal groups often contain major the pyramid, these considerations might still commercial, investment, merchant, or uni- remain. The share of the rent arising from versal banks as member firms. For example, the relationship that the bank captures one branch of the Canadian Bronfman fami- depends on the outside opportunities and ly controls Hees International Bancorp, a respective market power of both the bank large merchant bank, and the Swedish and the borrower. For example, a larger Wallenberg family controls SEB, a large borrower with more outside financing alter- Swedish universal bank. Gerard Caprio, Luc natives should be able to keep a greater Laeven, and Levine (2003) examine the ten share of the relationship rent. Likewise, if largest banks of forty-four countries in 2001. the bank’s profits are more dependent on They find that most banks are not widely retaining the pyramid firms as borrowers, its held. On average, only 25 percent have no 10 share of relationship rent falls. Family con- percent voting blockholder. More than half trol pyramids containing more corporate of the controlling shareholders of banks are assets ought thus to be better able to retain sp05_Article 1 8/18/05 9:32 AM Page 689

Morck, Wolfenzon, and Yeung: Corporate Governance 689

relationship rents than smaller stand alone Consequently, further empirical investiga- firms. tion along these lines is likely to be quite In either case, pyramid firms would enjoy interesting. The relative capital intensity of cheaper access to capital, all else equal. pyramidal group versus freestanding firms in Indeed, banks might appear to discriminate countries other than Canada is not known. If against freestanding firms. pyramidal group members do have better access to capital than stand alone firms, they 6.1.4 Empirical Evidence should be more concentrated in more capi- At present, little empirical work examines tal intensive industries. Such differences capital allocation within and between family should presumably be mitigated by capital controlled pyramidal groups, or capital allo- market openness. cation across group firms and freestanding Likewise, studies comparing levels of inter- firms. Morck, Stangeland, and Yeung (2000) nal utilization of retained earnings in pyrami- report that Canadian firms controlled by old dal groups versus freestanding firms would money families have elevated capital–labor be highly useful. Finally, and perhaps most ratios after controlling for industry, firm difficult, evidence on macroeconomic capital size, and firm age. Consistent with this, intensity and reinvestment of earnings across Attig, Fischer, and Gadhoum (2003) find economies with different incidences of old, that Canadian firms affiliated with pyrami- established family groups would be desirable. dal groups have greater capital expenditures Each of these exercises would, of course, than comparable nonaffiliated firms. require an “all other things equal” baseline. Moreover, the higher the controlling This means further thought is needed regard- owner’s cash flow stake, the more elevated ing necessary controls beyond the obvious the capital expenditure. ones, such as industry, and time fixed effects. The monopsony case discussed above can 6.2 Depressed Investment in Innovation probably be eliminated as a description of the Canadian economy, for it predicts, at Innovation is an especially interesting type least in its most naive form, that group firms of investment in an economy of control pyra- ought to use less capital than other firms, mids. Joseph A. Schumpeter (1912, 1942) is which is not observed. However, any or all of now generally acknowledged as correct in the other capital allocation distortions might arguing that growth is a process of creative be occurring in that country. destruction. Growth requires the continual Further work on other countries is also creation of innovations based on new technol- clearly needed for potentially valid a priori ogy or a recombination of old technologies, arguments can lead in either direction. For with technology broadly defined to include example, Khanna (2000) argues cogently scientific, managerial, and other practices that the Tata pyramidal group of India is important to a firm. Yet, innovations also social welfare enhancing because it substi- cause the destruction of firms built around tutes for imperfect capital markets in that activities rendered obsolete. Schumpeter country and thus allows entrepreneurs (1912) points out that innovators seldom have inside the group to obtain financing that access to large flows of retained earnings and would otherwise be unavailable. This has to that innovation consequently usually requires be pitched against the Almeida and outside financing. This suggests that innova- Wolfenzon (forthcoming-b) argument that tion ought to depend on the existence of func- groups may inefficiently concentrate capital. tional capital markets and institutions, a The nature of pyramidal groups’ effects on hypothesis empirically verified by Robert G. the allocation of capital is thus, in the final King and Levine (1993). Alternatively, analysis, an empirical issue. Schumpeter (1942) proposes that large firms sp05_Article 1 8/18/05 9:32 AM Page 690

690 Journal of Economic Literature, Vol. XLIII (September 2005)

with ample cash flows from previous invest- letting her plastics company disrupt her ments might best be able to finance innova- copper company. Morck and Yeung (2004) tion. More broadly, innovation requires that call this problem creative self-destruc- innovators have access to capital. tion.17 Even if the plastics firm is not part If pyramidal groups limit unrelated par- of a group, the controlling owners of ties’ access to capital, the latter might inno- pyramidal groups containing steel compa- vate less. Pyramidal groups might form nies might exert whatever capital market corporate venture capital to finance the power they command to deny financing to unrelated parties’ innovations, as Khanna the upstart plastics maker. Certainly, the and Palepu (forthcoming) document controlling owner has little incentive to regarding the software investments of cooperate with the innovator to form a India’s Tata group. However, if the pyrami- commercial alliance and might even seek to dal groups have capital market power, they expropriate the innovation to slow the pace presumably extract rents from the innova- of its application. tors. Freestanding firms accepting such In essence, corporate groups internalize financing must pay such rents and cede the disruptive effects of innovation, and this their independence. Gompers and Josh presumably can reduce group firms’ invest- Lerner (1999) show that freestanding inno- ment innovation relative to that of otherwise vators pay American venture capital firms comparable freestanding firms. The overall rents and cede independence too, but in welfare effect is unclear, for how the welfare exchange for critical technological and man- enhancing effects of innovation offset its dis- agerial expertise. The Tata group may be ruptive, and presumably welfare reducing quite exceptional, for Gompers and Lerner effects, is poorly understood at present. describe venture capital investment, includ- That governments commonly subsidize ing that by established nonfinancial firms, as “innovation” reinforces the conventional surprisingly concentrated in the United wisdom that, in much of the world at least, States. investment in innovation is not excessive. Morck, Stangeland, and Yeung (2000) Intriguingly, tunneling within pyramidal study pyramidal groups’ incentives to inno- groups might actually mitigate the negative vate. They argue that the destruction impact of creative self-destruction on invest- intrinsic to Schumpeter’s (1912) “creative ment on innovation. This is because creative destruction” is typically an externality to a destruction might proceed apace if the con- freestanding firm but is less likely to be so trolling shareholder can tunnel away profits for a large control pyramid. In an economy from the creative innovation and leave public of freestanding firms, a maker of superior shareholders to absorb the costs of the plastic pipes, who takes business away from “destruction” of firms with obsolete technolo- copper pipe makers, bears none of the costs gy. Thus, stronger public shareholder rights associated with the declining copper indus- might impede innovation by curtailing tunnel- try. Even if some investors hold stocks in ing, but might also accelerate innovation by both businesses, competition between deepening financial markets, which King and investors insures that the plastic pipes ven- Levine (1993) argue lets entrepreneurs estab- ture can be financed as a stand alone entity. lish new firms to develop their innovations. In an economy of control pyramids, where Thus, while family control pyramids ought to the copper pipe maker and plastics compa- be related to a slower pace of innovation, the ny have the same controlling shareholder, financing for the plastic firm’s innovation 17 In a somewhat similar vein, Daron Acemoglu and Robinson (2002) argue that elites historically sometimes may be less forthcoming. The controlling blocked new and efficient technologies to avoid changes shareholder may see little net advantage in that threatened their political power. sp05_Article 1 8/18/05 9:32 AM Page 691

Morck, Wolfenzon, and Yeung: Corporate Governance 691

interaction of this effect with stronger public accounting in large groups of firms. It might shareholder rights is difficult to predict. also be indirect, in that the controlling own- Finally, it is important to note that the ers of large pyramidal groups might press above need not apply to all pyramidal or government officials to maintain ineffective other groups. For example, J. Bradford de shareholder protection and weak disclosure Long (1990) argues that the J. P. Morgan rules to better facilitate tunneling. La Porta, group in the late 19th and early 20th centu- Lopez-de-Silanes, Shleifer, and Vishny ry United States was instrumental in financ- (1997b, 1998) demonstrate convincingly that ing numerous innovative ventures and in weak institutions of this sort are detrimental lending its financial credibility to entrepre- to capital market development. Morck, neurs. Empirical work is needed to deter- Yeung, and Yu (2000) speculate that such mine which outcome is more economically weak institutions hinder informed risk arbi- important and under what circumstances. trage, and thus the incorporation of informa- Only the most preliminary empirical work tion into equity prices, reducing the tests these ideas. Morck, Stangeland, and functional efficiency of equity markets. See Yeung (2000) report that firms controlled by also Durnev, Morck, Yeung (2004) and old, established Canadian families spend less Durnev, Kan Li, Morck, and Yeung (2004). on R&D than other Canadian firms and than As section 6.1 describes, pyramidal group similar U.S. firms, controlling for industry, firms often appear to have preferential firm size, and firm age. They also report that access to bank capital. Sometimes this is countries with larger billionaire inherited because a bank is part of the pyramidal wealth over GDP have lower private sector group, and having a common controlling R&D spending. Given the importance of shareholder perhaps reduces information innovation to economic growth, further work asymmetry problems. Also, large corporate on the investment policies, R&D strategies, groups’ sheer size makes them less vulnera- capital intensity, and liquidity constraints of ble to outside banks’ extraction of rents. Part firms with different ownership structures is of the reason stems from groups’ complexity. clearly a priority. Se-Jik Kim (2004) shows that banks have more difficulty using observed loan repay- 6.3 Overall Capital Market Effects ments to gauge the quality of members of a So far, this section has focused on prob- business groups than of freestanding firms. lems associated with capital access and This is because intergroup loan guarantees investment decisions. In economies where a prevent an outside bank from knowing few old families control great swathes of the whether the payment is from the borrower or corporate sector, such governance prob- other group firms. Consequently, such a lems, which economists typically think of as bank is more likely to liquidate a freestanding firm-level effects, might easily attain macro- firm than an otherwise identical group firm. economic proportions. Also, by affecting However, from our perspective, group the nature of capital markets, extensive firms’ most important preferential access to pyramiding might more directly affect capital arises from banks being affiliated macroeconomic outcomes. with groups of nonfinancial firms. La Porta, For example, a prevalence of pyramidal Lopez-de-Silanes, and Guillermo Zamarripa groups in a country’s corporate sector might (2003) argue that, in many countries, banks render the average firm more opaque to out- lend to related customers, including firms side investors, both domestic and foreign, controlled by the bank’s controlling share- and this might induce both to place their sav- holders. Caprio, Laeven, and Levine (2003) ings elsewhere. This might be a direct effect, examine the ten largest banks of forty-four arising out of the increased complexity of countries in 2001 and find that 75 percent sp05_Article 1 8/18/05 9:32 AM Page 692

692 Journal of Economic Literature, Vol. XLIII (September 2005)

have a controlling shareholder, often a intermediation could thus retard the devel- wealthy family. Using Mexican data, Lopez- opment of both equity markets and the de-Silanes, and Zamarripa (2003) show that banking sector; with concomitant negative related lending accounts for 20 percent of macroeconomic implications. commercial loans. Related borrowers pay Further work testing these speculative lower interest rates than unrelated borrow- hypotheses might involve testing whether or ers. This cannot reflect lower risk in related not a high prevalence of pyramidal groups borrowing, for they find related borrowers relates to underdeveloped equity capital more likely to default than unrelated bor- market, underperforming banking systems, rowers—and less likely to pay back their or generally misallocated capital. Comparing debts eventually. Such findings raise the how severely liquidity constraints impede possibility that banks in some countries the growth of upstarts might also be instruc- might inefficiently channel capital toward tive. If pyramids per se render their member other companies controlled by the banks’ firms more opaque, freestanding firms controlling shareholders. might be at an advantage in raising external James R. Barth, Caprio, and Levine capital (though not capital from related (2004) note that regulations facilitating the group banks). But if controlling owners press private monitoring of banks are associated for weak institutions, this might especially with greater development in the banking constrain high quality upstarts by making it sector, lower costs of bank loans, and a small- harder for them to signal investors about er ratio of nonperforming loans. In contrast, their quality. direct government monitoring and regula- Arguing along these lines, Shleifer and tions are unrelated to the banking sector’s Wolfenzon (2002) stress that the misappro- development and stability. If the sort of bank priation of wealth by controlling sharehold- behavior documented by La Porta, Lopez- ers raises costs of capital to the whole de-Silanes, and Zamarripa (2003) in Mexico corporate sector and thus impedes growth. is commonplace elsewhere, the ubiquity of Davide Lombardo and Marco Pagano (2002) family controlled banks throughout the show that aggravated information asymme- world, documented by Caprio, Laeven, and try problems in economies with weak insti- Levine (2003), pervasive capital misalloca- tutions move demand and supply schedules tion by banks might be a major impediment for capital to reduce both the aggregate to development. Since governments com- amount of capital used and social welfare. monly act as “lenders of last resort” to trou- Rui Castro, Gian Luca Clementi, and Glenn bled banking systems, chronic capital MacDonald (2003) propose that strength- misallocation by banks might also destabilize ened property rights for public investors can government finances. have a two-fold effect. Stronger legal rights Thorsten Beck, Asli Demirgüç-Kunt, and instill trust in public investors, lowering Levine (2005) show that empowering offi- entrepreneurs’ cost of capital. But stronger cial supervisory agencies to monitor, disci- investor property rights also reallocate pline, and influence banks directly leads to wealth from controlling families to the mid- perverse results: it actually engenders bank dle classes. If middle class investors have corruption and also negatively affects lower propensities to save, this might actual- growth. On the other hand, well-governed ly reduce overall savings. Castro, Clementi, banks, monitored by legally empowered pri- and MacDonald (2003) develop an overlap- vate sector investors and depositors, seem to ping generations model that magnifies the constitute best practice at this point. latter effect by positing that stronger Opaque family pyramidal groups’ direct investor rights reallocate capital from the control of banks and involvement in capital younger generation of entrepreneurs to the sp05_Article 1 8/18/05 9:32 AM Page 693

Morck, Wolfenzon, and Yeung: Corporate Governance 693

older generation of retirees, and that the lat- 7. Political Economy Considerations ter have a minimal propensity to save. This highlights the ambiguous impact of While inequality exists in virtually every strengthened property rights on overall costs country, we showed above that control pyr- of capital and welfare. amids can magnify large inequalities in Empirical work is needed to test these wealth into truly enormous inequalities in theories and to develop stylized facts within corporate control. Control pyramids effec- which future theories in this area must fit. tively entrust the corporate governance of the greater parts of the corporate sectors of 6.4 Summary many countries to handfuls of elite, estab- This section and the previous one show lished families, who can quite reasonably be that the effects of control pyramids on corpo- described as oligarchs. The extent to which rate governance might well be especially this concentration of corporate governance injurious in countries that provide public power occurs can be used to arrange capi- shareholders ineffective legal rights against talist economies along a metric with two malfeasance by corporate insiders. These stylized endpoints. One endpoint, diffuse governance problems, allowed free reign, capitalism, best approximated by the plausibly retard macroeconomic perform- United States and United Kingdom, is char- ance for three reasons. First, a single insiders’ acterized by numerous professionally man- poor governance becomes a macroeconomic aged freestanding large corporations, each problem if she controls a pyramid that owned by multitudinous small investors includes a substantial fraction of a country’s whose property rights are well protected by corporate assets. The structure of a control effective laws. The other endpoint, oli- pyramid also leads to a variety of capital allo- garchic capitalism, approximated by most cation distortions. By altering the overall East Asian and Latin American economies, investment level and by skewing the distribu- is characterized by public investors holding tion of capital expenditure across groups, minority voting stakes in large corporations firms, and projects, these distortions can controlled by a few families through control compromise economic growth. Second, con- pyramids. trol pyramids might adversely affect invest- This section considers this metric more ment in innovation because they internalize closely in a framework of political economy, the disruptive effects of one firm’s innovation the discussion of which we have postponed upon another. This, especially, might be a until now. We propose that the two ends of powerful growth retardant. Third, the con- the metric, diffuse capitalism and oligarchic centration of capital within opaque pyramidal capitalism, represent two different political groups might raise public investors’ required economy outcomes. Our essential point is returns by raising the levels of risk they per- that institutional development is endoge- ceive. The distribution of returns to different nous. On the one end, oligarchs wield their classes of investor might also affect the over- political influences to attain policies that all supply of savings because of difference in preserve and expand their corporate gover- their propensities to save. nance power and thus sustain concentrated The regulation of capital markets and insti- corporate control. This in turn preserves tutions in general, and of corporate gover- and expands the resources at their disposal nance in particular, are ultimately functions for further political lobbying. The objects of of a country’s political process—see e.g., this lobbying presumably are policies to Douglass C. North (1991). We therefore turn limit public investors’ rights, impede entry to the interaction of control pyramids with by upstarts, and raise international trade political economy in the next section. and capital barriers. These institutional sp05_Article 1 8/18/05 9:32 AM Page 694

694 Journal of Economic Literature, Vol. XLIII (September 2005)

features associated with oligarchic capitalism 7.1 A Political Economy and Institutional have been linked to poor long-run macroeco- Focus nomic performance. Indeed, they may well explain why Morck, Stangeland, and Yeung Our conceptual framework lets economic (2000) find that economies with larger inher- agents use resources to influence “institu- ited wealth post slower growth.18 The concur- tional development.” This allows a political rence of oligarchic capitalism, weak economy framework for understanding the institutions, and slow growth in a stable, but relationship between the distribution of cor- Pareto-inferior, equilibrium, we call economic porate control and the development of eco- entrenchment.19 nomic institutions. We first clarify the Clearly economic entrenchment cannot “institutions” we have in mind. be the only possible outcome. There must be One is public shareholders’ rights. La circumstances in which the elites fail to Porta, Lopez-de-Silanes, Shleifer, and secure this outcome, or in which they actual- Vishny (1997b, 1998) construct an index of ly prefer—and hence lobby for—sound public shareholders’ rights by adding up dis- institutions. tinct legal rights—the right to vote by mail, The concepts we develop in this section the right to sell shares prior to meetings, the are quite general. Although we focus on right to cumulative voting, the right to sue family control pyramids as the means to directors, preemptive rights to new issues, concentrate control, other institutional and the right to call extraordinary sharehold- arrangements can serve a similar goal. For er meetings. This index is used in many example, the professionally managed empirical studies. For example, Jeffrey keiretsu groupings of Japanese corpora- Wurgler (2000) shows that sound capital tions might be another arrangement, as institutions and markets, especially share- Morck and Nakamura (1999) argue. Yet holder legal rights, effectively check capital another might involve entrenched elites of misallocation—presumably including that by bureaucrats presiding over inefficient corporate insiders running control pyramids. state-owned enterprises, as in Krueger The “Law and Finance” literature, which (1993). See also William L. Megginson and descends from La Porta , Lopez-de-Silanes, Jeffrey M. Netter (2001). Shleifer, and Vishny (1997b, 1998), argues more broadly that how well capital markets function depends on a range of customs, 18 There are many possible reasons why oligarchic cap- rules, laws, and regulations, and on how vig- italism might allocate resources more or less efficiently orously they are enforced. Thus, statutory than diffuse capitalism. If the oligarchs are especially adept at business, they may excel at placing resources shareholder rights are dead letters if the where their values are highest. This might underlie the judiciary, police, and regulators are all finding of Morck, Stangeland, and Yeung (2000) of high deeply corrupt. While legal rights for public growth in countries in which great wealth is concentrated in the hands of self-made billionaires. However, if the oli- shareholders are clearly an important deter- garchs were especially poor managers, a similar concentra- minant of how well capital is allocated, other tion of corporate governance in oligarchic capitalism could considerations factor in as well. These magnify the mistakes of a few individuals into macroeconomic problems. This might underlie the Morck, include a general absence of corruption, Stangeland, and Yeung (2000) findings of slow growth in transparent accounting and disclosure, and countries in which great wealth in concentrated in the many other dimensions of institutional hands of old, established families. 19 Note that this is deliberately distinct from manageri- development. Ultimately, a pervasive respect al entrenchment in corporate finance, where suboptimal for outsiders’ private property rights seems managers cannot be ousted and the performance of the basic to an efficient allocation of capital. firm suffers. We view economic entrenchment as a macroeconomic counterpart to this, and a potentially much more This respect must engender not only important problem. rights for public investors, but also rights for sp05_Article 1 8/18/05 9:32 AM Page 695

Morck, Wolfenzon, and Yeung: Corporate Governance 695

aspiring entrepreneurs. Djankov, La Porta, 7.2 Political Influence in Oligarchic Lopez-de-Silanes, and Shleifer (2002) docu- Capitalist Economies ment the difficulties entrepreneurs confront in becoming registered business firms in There are several reasons for believing many poor countries. These include expen- that oligarchic families are likely to be high- sive, arbitrary, and repeated inspections, ly effective political lobbyers, especially taxes, and fees relating to “safety and when they control large pyramidal groups of health,” the environment, labor, and “screen- companies. In particular, Morck and Yeung ing.” They show that the time and direct cost (2004) argue that the apex shareholders of these procedures as a fraction of 1999 per of control pyramids have particularly low capita GDP, as well as their sheer number, is political lobbying costs. significantly positively correlated with meas- One reason they stress is that the control- ures of government corruption. Such arbi- ling owners of great pyramidal groups com- trary entry barriers deter prospective mand vast economic resources from which entrepreneurs and lock in insider control. they can readily make up-front side-pay- We count such impediments as infringe- ments to public officials. In contrast, new ments on entrepreneurs’ private property entrants, usually short of capital, must prom- rights. See also Hernando de Soto (1989). ise future side-payments for advantageous For expository convenience, we refer to all government policies now. Since many new of these rights together as private property ventures fail, or are taken over by others, rights. The degree to which these private such promises are risky investments for a property rights are protected, both from the public official. Thus, lobbying costs are lower government and from more powerful indi- for the controlling owners of established viduals, depends on government policies— great pyramidal groups than for the principals their stipulation and enforcement. Our or professional managers of new entrants. proposition is that economic agents, oli- A second set of reasons has to do with garchs, and others can use their resources to lower transactions costs. A pyramid’s con- influence the government’s choices regard- trolling owner can pool the resources of ing the enactment and enforcement of prop- many companies to make side payments to erty rights. Although economics long politicians without the coordination costs a regarded government policies as exogenous similar array of freestanding firms would except in limited areas—such as taxes, mon- incur. An array of freestanding firms would etary policy, and fiscal policy—recent work, also surely experience some free-rider prob- such as Krueger (1993) and Mancur Olson lems, while an array of pyramid firms with a (1982, 2000), favors treating government common controlling owner need not. policies as thoroughly endogenous; in partic- Olson (1982) argues that larger or more ular, they depend on political influence, or diverse groups of people have higher trans- lobbying.20 action costs in organizing effective lobbying. Section 3.4 shows that control pyramids can entrust corporate control over vast reaches 20 For example, Krueger (1993) stresses dynamic interactions between political and economic factors. She argues of a country’s economy to handfuls of that import substitution programs divert wealth to certain wealthy families. As the number of oligarchs bureaucrats and influential individuals whose fortunes is small, the transaction costs of coordinating grow under the programs. This wealth makes them formi- dable interest groups and gains them political power, with their actions are correspondingly low. This which they defend the programs that enrich them. The makes coordinating the lobbying of different result is a “vicious cycle” of ever slower economic growth. oligarchs easier. She pitches this against a “virtuous cycle” in which export oriented programs enrich defenders of export oriented Yet another reason oligarchs’ lobbying policies, and economic growth steadily rises. costs might be low is that politicians might sp05_Article 1 8/18/05 9:32 AM Page 696

696 Journal of Economic Literature, Vol. XLIII (September 2005)

perceive them as more trustworthy partners firm or a firm in an industry with little in favor trading. Favor trading between government business. This increases oli- political and corporate insiders often garchs’ flexibility as to the specific form involves trading actions today for promised favors to officials can take. future responses. The beneficiary of today’s Finally, corporate insiders and political action has an ex post incentive not to pay up. insiders may be the same people, or at least The CEOs of widely held firms in the from the same families. This may be the United States serve an average seven year result of effective past lobbying, but it sure- term—an eye blink compared to the perma- ly also facilitates further cooperation nence of the old money families controlling between oligarchic families and the State. pyramidal corporate groups in many coun- Such ties are surprisingly prevalent. tries. One CEO of a widely held firm may Claessens, Djankov, and Lang (2000, p. 109) feel little obligation to uphold promises cite a New York Times (September 8, 1998, made by a predecessor. In contrast, a dynas- p. 2) report in which Imelda Marcos, the tic family can better keep faith over time. widow of former Philippines president For similar reasons, an oligarch should dis- Ferdinand Marcos, explains that her rela- count promised future consideration from a tives “practically own everything in the public official at a lower rate than that the Philippines from electricity, telecommunica- professional manager of a widely held free- tions, airlines, banking, beer and tobacco, standing firms would use. Public officials newspaper publishing, television stations, should also be less willing to break faith with shipping, oil and mining, hotels and beach the controlling owner of a great pyramidal resorts, down to coconut milling, small group, for she would have means to “pun- farms, real estate and insurance.” They also ish” the defecting official, given her exten- report that “the business empire of the sive corporate governance powers. Thus, Suharto (the previous strong man in the controlling owners of pyramids have Indonesia) family is thought to control 417 lower lobbying costs than the professional listed and unlisted companies through a managers of widely held freestanding firms. number of business groups led by children, A third set of reasons why oligarchs’ lob- other relatives, and business partners, many bying costs are likely to be especially low of whom also have held government offices” involves their use of tunneling to transfer (Financial Times, December 8, 1998, p. 16). resources among controlled firms. If the Christian Leuz and Felix Oberholzer-Gee controlling owner paid politicians from (2003) show that Indonesian firms with close lower tier firms for favors that benefited connection to former President Suharto are higher tier firms, public shareholders would significantly less likely than unconnected pay most of the group’s lobbying costs while firms to have publicly traded foreign securi- the controlling family would reap most of ties. This is presumably either because they the benefits of the lobbying. Moreover, by have special access to lower cost domestic coordinating where the costs and benefits capital or because they are disinclined to of political rent-seeking fall among pyramid comply with foreign disclosure require- firms, the controlling owner can effectively ments. The first chief executive of Hong tunnel wealth from one firm to the other Kong is the offspring of a shipping tycoon without directly transferring assets or and Kashing Li, who controls some large income between firms. Furthermore, to pyramidal groups, allegedly maintains a con- render favors to government officials, the siderable stake in the chief executive’s fami- controlling family of a pyramid can tunnel ly business. The Lee family provided resources to where the payment can be Singapore a head of state for decades, while made most discretely—perhaps a private also managing many of its businesses. The sp05_Article 1 8/18/05 9:32 AM Page 697

Morck, Wolfenzon, and Yeung: Corporate Governance 697

appendix to Johnson and Mitton (2003) announcements of directors or dominant describes the close ties of the Mahathir fam- shareholders entering politics or of politi- ily, Daim Zainuddin (a previous finance min- cians joining boards result in significant ister), and Anwar Ibrahim (a purged vice increases in firms’ value in highly corrupt prime minister) with Malaysian businesses. countries. De Soto (1989), Shleifer and Former Russian Prime Minister Viktor Vishny (1993, 1994), and others argue that Chernomyrdin is a large shareholder of such political connections give firms prefer- Gazprom. Faccio (2003) provides other ential access to government subsidies, abundant examples. financing from government owned enter- Such connections are not restricted to prises and banks, tax breaks, and exemptions emerging market countries. Faccio and from burdensome regulations. Lang (2002) report that two members of the This is not to suggest that political lobby- Agnelli family, which controls the largest ing is an unimportant corporate investment Italian industrial group, serve Italy as mem- in diffuse capitalist economies like the bers of parliament. Silvio Berlusconi, who United Kingdom and United States. A huge controls another of the same country’s literature, in its modern form descending largest pyramidal groups, serves Italy as from George Stigler (1971), attests to the Prime Minister. And Francois Pinault, the near universal importance of political rent- controlling shareholder of Pinault- seeking. Indeed, many of the arguments list- Printemps-Redoute, Grand Bazar de Lyon, ed above suggest that large family firms Rexel, and Zodiac, is reputedly a close friend might also have a political influence advan- of French President Jacques Chirac.21 Paul tage in diffuse capitalist economies. For Desmarais, the controlling shareholder of example, the Financial Times (February 26, one of Canada’s largest pyramidal groups, is 2004, p. 8) quotes a “former U.S. govern- aclosely related in-law of former Prime ment trade official” as being impressed with Minister Jean Chrétien and the longtime Cargill, a huge private U.S. family agribusi- business associate of the current Prime ness because “They weren’t simply looking Minister, Paul Martin. in a short sighted way at the bottom line, but Faccio (2003) examines these relation- had an eye on the longer term. It made me ships formally by carefully documenting more and more inclined to listen to them for family ties between the political and corpo- trade negotiations.” Econometric evidence rate leaders of forty two different countries. on which sorts of firms are best at political She reports that firms representing 7.76 per- rent-seeking is scant.22 cent of the world’s market capitalization are Our point here is not that diffuse capital- run by relatives of their countries’ political ism is free of political rent-seeking, but that leaders. While there is no hint of corruption oligarchic capitalism, because of pyramidal in many of these family ties, others presum- corporate control and the other factors ably result from political influence and lower described above, narrowly concentrates and the costs of further lobbying. greatly magnifies political rent-seeking Indeed, political influence might be power in the hands of a tiny elite. instrumental to oligarchs’ private benefits of Morck and Yeung (2004) find a strong control. Faccio (2003) shows that politically correlation between corruption and family connected companies enjoy easier access to control. Countries with high incidences of debt financing, lower taxation, and higher family control over large firms have low market shares. She also demonstrates that compliance with tax laws, high official

21 See “Heat on Chirac for Public Inquiry,” Financial 22 Though see Stefanie Lenway, Morck, and Yeung Times, December 6, 2003. (1996). sp05_Article 1 8/18/05 9:32 AM Page 698

698 Journal of Economic Literature, Vol. XLIII (September 2005)

TABLE 4 MEASURES OF THE RETURN TO POLITICAL RENT SEEKING AND THE INCIDENCE OF FAMILY FIRMS

Regression Coefficients Simple Correlation Coefficients controlling for log of 1995 per capita GDP Incidence of Family Control in 1995 in Incidence of Family Control in 1995 in Twenty Largest Ten Middle-size Twenty Largest Ten Middle-size Firms Firms Firms Firms 20% 10% 20% 10% 20% 10% 20% 10% Threshold Threshold Threshold Threshold Threshold Threshold Threshold Threshold Sample Tax System Corruption Higher scores –0.470 –0.444 –0.472 –0.270 –0.889 –0.588 –0.732 0.470 25 indicate general (0.02) (0.03) (0.02) (0.19) (0.32) (0.54) (0.40) (0.57) compliance with tax laws [0.28] [0.26] [0.27] [0.26] Political System Corruption Higher scores –0.414 –0.438 –0.526 –0.523 0.367 0.188 –1.05 –0.980 27 indicate a general (0.03) (0.02) (0.00) (0.01) (0.73) (0.87) (0.31) (0.33) absence of official corruption [0.49] [0.49] [0.51] [0.51] Judicial System Corruption The efficiency and –0.340 –0.375 –0.457 –0.426 0.501 –0.691 0.292 –0.036 27 integrity of the (0.08) (0.05) (0.02) (0.03) (0.56) (0.55) (0.78) (0.97) judicial system, particularly as it [0.55] [0.55] [0.55] affects business Civil Service Corruption High scores indicate –0.663 –0.685 –0.722 –0.630 –2.64 –2.59 –2.83 –1.82 27 bureaucrats have (0.00) (0.00) (0.00) (0.00) (0.01) (0.02) (0.00) (0.07) “autonomy” and the “strength and [0.70] [0.69] [0.73] [0.66] expertise to govern” Regulatory Barriers to Entry Estimated 0.521 0.501 0.578 0.424 0.195 0.160 0.218 0.080 27 regulatory (0.01) (0.01) (0.00) (0.03) (0.12) (0.23) (0.06) (0.50) compliance cost of starting a new [0.36] [0.33] [0.39] [0.30] business, as % of GDP Numbers in parenthesis are probability levels for the null hypothesis of a zero simple correlation or a zero coefficient on oligarchic family control in regressions of each rent-seeking variable of that variable and the logarithm of 1995 per capita GDP. Numbers in square brackets are regression adjusted R2 statistics.

corruption, low judicial efficiency and Extensive lobbying power in the hands of integrity, inefficient bureaucrats with low a country’s leading business families might autonomy, and high regulatory barriers to seem a recipe for strong private property entry. Table 4 reproduces their results. rights, but Morck, Stangeland, and Yeung sp05_Article 1 8/18/05 9:32 AM Page 699

Morck, Wolfenzon, and Yeung: Corporate Governance 699

(2000), Glaezer, Sheinkman, and Shleifer value of their existing capital, discussed in (2003), Rajan and Zingales (2003), the previous section. And the lobbying for Konstantin Sonin (2003), and others argue entry barriers and the like that lock in oli- that this is far from clear. Glaezer, garchs’ governance powers can also have Sheinkman, and Shleifer (2003) posit that side effects that further distort capital and the very wealthy can prefer weak property other resource allocation. rights if they benefit from the transfer pay- 7.3 Formalizing the Flavors of Capitalism ments weak property rights allow. Shleifer (1997) points out that, in economies with In this section, we formalize the reason- weak general property rights, the wealthy ing that oligarchs may use their political can invest in private protection of their prop- influence to sustain institutional arrange- erty rights. This results in private enforce- ments that preserve their status quo which ment of their property rights and an absence benefits them at the expense of the rest of of property rights protection for others. the economy. Entrepreneurial ability, like other dimen- Wealthy families need not always lobby for sions of intelligence, is likely at most only weak private property rights, which are detri- partially hereditary. On average, the entre- mental to financial development.23 A more preneurial ability of oligarchic family scions developed, or more functionally efficient, ought to regress steadily toward the popula- financial system better allocates capital to its tion mean with each successive generation, highest value uses. Greater financial develop- as modeled by Caselli and Gennaioli ment would improve the terms on which (2003). Yet an oligarch with little entrepre- wealthy families could sell out or raise outside neurial ability still commands all the politi- capital. But wealthy families might lobby for cal lobbying advantages enumerated above. institutions that enhance and protect their Thus, older billionaire families should be private benefits of control, weakening the pri- progressively more drawn toward using vate property rights of others and retarding political influence, rather than entrepre- financial development. Public investors, in neurship, to sustain their economic posi- contrast, unambiguously prefer strong private tions. Alternatively, Sonin (2003) argues that property rights protection since the main a first generation of entrepreneurs in many effect of weak rights is more predation by oli- transition economies became wealthy garchs. Potential entrepreneurial entrants through adept rent-seeking, and favors weak likewise prefer strong private property rights general property rights enforcement because this lets them raise funds from public because this facilitates their continued rent investors on better terms. seeking. Genuine entrepreneurship never enters the picture. 23 By using their political influence to lock We deliberately adopt a narrow view of property rights related to financial development; however, a broad- in their own corporate governance power, er view may also be sensible. For instance, a class of top economically entrenched oligarchs thus professional managers might become “entrenched” by also lock in on-going suboptimal capital setting up takeover defenses and undermining shareholder democracy. Or, as Krueger (1993) argues, government allocation. This is partly because hereditary activism might create a class of bureaucrats who oligarchs are often simply less able entre- “entrench” to preserve their influence and economic preneurs. In part, it also reflects the corpo- power. Of course, large interventionist governments, pur- suing social objectives other than economic efficiency, rate governance problems intrinsic to the might deliberately sacrifice investors’ property rights and control pyramids that provide oligarchs financial development to those ends—even without with vast rent-seeking influence. It also entrenched bureaucrats. If these other forms of entrenchment, or even just big government per se, corre- reflects the incentives oligarchs have to late with family control pyramids, this could confound the suppress innovations that threaten the framework we develop below. sp05_Article 1 8/18/05 9:32 AM Page 700

700 Journal of Economic Literature, Vol. XLIII (September 2005)

Whether economic entrenchment ensues foreigners. It also abstracts away from salient or not thus depends upon which way the oli- aspects of real political processes, like voting garchs opt. This jibes with the discussion mechanism and constitutions. We return to above of oligarchs’ magnified lobbying power. these issues in subsection 7.5 below. This can be formalized along lines devel- For simplicity, suppose there are only two oped in Shleifer and Wolfenzon (2002). Public levels of private property rights: strong and shareholders expect some “stealing” by corpo- weak. The maximum side-payment the rate insiders. Thus, an entrepreneur seeking corporate insiders are willing to pay to the external financing must compensate investors political insiders is the difference between up front for her expected thievery, in the form their wealth in these two states. An analo- of a low share price in the initial sale of equity gous constraint applies to each outsider. to the public. Better private property rights This is a minimally binding constraint for reduce actual and expected thievery by insid- the corporate insiders, for pyramids might ers, and so improve the terms on which exter- allow them to use assets they control, in nal capital can be raised. This makes more addition to their own wealth, to pay investment projects viable for entrepreneurs. bribes. The political insiders’ utility is a The result is faster economic growth. weighted average of the side payments The level of private property rights is con- received and economic growth, which is trollable by the government—either by higher if shareholder protection is strong. altering rights on the books or by varying Define an equilibrium as a level of private the enthusiasm with which they are property rights such that enforced. This means the government can 1. The political insiders’ utility at the determine capital market functionality. By equilibrium is higher than at any other raising the level of private property rights, it level, given the behavior of corporate can generate faster economic growth. insiders and outsiders. Consider a simple stylized political econo- 2. Neither the corporate insiders my model with the following cast of players. nor the outsiders are willing and able to First, political insiders value both economic pay side-payments to the political growth and side payments. The side pay- insiders sufficient to induce a change in ments can be outright bribery, campaign con- the private property rights regime. tributions, or anything in between. Second, If the corporate insiders lobby for strong corporate insiders control on-going business private property rights, they can sell their cap- operations with considerable internal cash. ital assets to more able entrants, who can Third, outsiders include both public investors more readily raise the necessary funds from and potential innovative entrepreneurs. We public investors. This permits the old corpo- assume the political insiders and corporate rate insiders to fade away into a comfortable insiders each act as one. In contrast, there are retirement. As strong private property rights a very large number of outsiders, each with a are implemented, the value of private benefits small amount of cash, who cannot act collec- of control, including those due to pyramids, tively. Outsiders supply their savings to capi- falls. This limits corporate insiders’ ability to tal users competitively. Some outsiders wish make side-payments and thus limits their to become entrepreneurs and are possibly political influence, reinforcing their decision more able than the corporate insiders. to sell out, assuming side-payments are Clearly, this is a highly simplified model of financed out of current cash flow. All of this the political and economic determinants of combines to create pressure for less oligarchic institutions. It omits obviously important and more diffuse capitalism. considerations, like the roles of organized In contrast, if the corporate insiders lobby labor, political ideologies, and pressure from for weaker property rights, and the attendant sp05_Article 1 8/18/05 9:32 AM Page 701

Morck, Wolfenzon, and Yeung: Corporate Governance 701

higher private benefits of control, they set undertake investment projects. This is the stage for economic entrenchment. because stronger private property rights Policies that compromise investor rights lower their costs of external capital and raise oligarchs’ private benefits of control, make more projects viable. This creates providing them with more resources to lobby product market competition for the further. This creates pressure for more oli- firms controlled by corporate insiders. garchic and less diffuse capitalism. This what It also creates capital market competi- we call economic entrenchment—oligarchs tion, as today’s entrepreneurs have lobby successfully for weak property rights, more wealth to invest in the future. And which provide them resources for further it allows creative destruction to pro- lobbying to preserve weak property rights. ceed, which erodes the value of the cor- This example is highly simplified. Private porate insiders’ existing capital assets. property rights protection is a multidimen- This competition effect is negative for sional phenomenon that certainly can take corporate insiders if private property more than two values. However, two values rights grow stronger, and positive if they suffice to make our basic point: economic grow weaker. entrenchment is but one possible equilibri- These considerations belie any simple um. There can be others—perhaps many predictions about the relationship between others—in a more nuanced model. the level of private property rights and the Establishing the equilibrium is nontrivial. corporate insider’s welfare. Morck, However, some intuitive considerations must Stangeland, and Yeung (2000) and Rajan and apply. Any deviation from a preexisting level Zingales (2003) assume that weaker private of private property rights has several effects: property rights always benefit the corporate 1. Any improvement in private prop- insider. That is, they presume that the trans- erty rights transfers wealth from exist- fer, cost of capital, and competition effects ing users of capital to existing providers combine to generate a net increase in the of capital. The opposite occurs if private corporate insider’s welfare if the private property rights fall. We call this the property rights regime switches from strong transfer effect. to weak. This is likely a special case, albeit an 2. Any improvement in private prop- important one, for the sum of the effects can erty rights reduces the cost of external have the opposite sign. For example, if cor- capital, and thus raises capital invest- porate insiders’ future investment needs ment by both corporate insiders and greatly exceed their internal cash, the cost of outsider entrepreneurs. The opposite is capital effect might be sufficiently large to true when private property rights are overwhelm the transfer and competition downgraded. We call this the cost of effects. In that case, the corporate insiders capital effect. The cost of capital effect and the outsiders concur on the need for is positive for outsiders when private stronger private property rights. property rights are strengthened and Dynamic changes can be intriguing. Rajan negative when they are downgraded. and Zingales (2003) point out that many For corporate insiders, the cost of capi- countries that currently have small stock mar- tal effect from stronger (weaker) private kets had large and dynamic stock markets a property rights is positive (negative) if century ago—a phenomenon they call the they need outside capital and zero if great reversal. They propose that, in these they do not. countries, a first generation of corporate 3. A change in private property insiders raised money from public investors rights changes the number of outsiders and then lobbied for policies that caused who become entrepreneurs and financial market atrophy, or at least stood by sp05_Article 1 8/18/05 9:32 AM Page 702

702 Journal of Economic Literature, Vol. XLIII (September 2005)

while ideologues enacted them. This weaken- This may not be entirely realistic if organ- ing of private property rights was to their ized labor prefers job security for members advantage because they no longer needed out- over general economic growth, as Roe side capital. In our terminology, the “cost of (2003) and Högfeldt (2003) argue is the case capital” effect faded, and the “transfer” and in social democracies like Sweden. However, “competition” effects dominated. They call in such countries, we might think of organized their idea an interest group theory of financial labor as insiders too. development, which parallels our concept of This and all the other simplifications in this economic entrenchment. In essence, Rajan section highlight the need for rigorous theo- and Zingales (2003) propose that the prefer- retical development of this area and for care- ences of the first generation of corporate ful empirical work to guide theory. insiders regarding investor property rights Developing models that formalize different protection changed as the insiders aged—or flavors of capitalism is a fascinating uncharted that their heirs had markedly different prefer- territory for creative theorists. ences. (For recent work, see, e.g., Perotti and Ernst-Ludwig von Thadden forthcoming.) 7.4 Determinants of Economic This reasoning, while perhaps historically Entrenchment plausible, highlights a time inconsistency Economic entrenchment is a self sustain- problem. Rational public investors should ing, stable equilibrium that seems to charac- have anticipated this volte-face. Expecting a terize some, but not all oligarchic capitalist future erosion of shareholders’ property economies. The stark divergence between rights, they should have been loath to pro- high and low income economies appears to vide the first generation of entrepreneurs result from the latter becoming trapped in with capital on generous terms. This objec- weak property rights regimes characterized tion can be circumvented in several ways. by economic entrenchment. This begs the First, public shareholders make mistakes questions of what determines whether or not and learn from experience. The statutory any particular country falls into an economic property rights protection most great rever- entrenchment trap. We believe this to be a sal countries provided their first generation fundamental economic problem. of public investors was weak, and public investors may have learned from experience 7.4.1 Generic Considerations not to offer equity capital on generous terms. For economic entrenchment to occur, cor- Alternatively, the erosion may have been porate insiders must be willing and able to uncertain, for investors might have expected pay the political insider sufficient side-pay- the enactment of strong property rights. ments to refrain from establishing strong Finally, public investors might have expected property rights protection, and outsiders the erosion to occur many years down the must be unable or unwilling to offer large road. Perhaps the expected present value of enough side-payments to oppose this. insider expropriation at the time of the initial Equally important are political insiders’ pref- provision of finance was small. Such issues erences—the relative values they place on highlight the difficulty of constructing fully general economic prosperity versus personal rational expectations equilibrium models of side-payment income. These presumably institutional development. depend on how strongly property rights A simplification in our conceptual frame- affect economic growth as well as social con- work is that outsiders always prefer strong pri- straints on political insiders’ acceptance and vate property rights because the transfer and valuation of side-payments. cost of capital effects are both positive for Obviously, current economic condition them and the competition effect is irrelevant. and future economic outlooks affect the sp05_Article 1 8/18/05 9:32 AM Page 703

Morck, Wolfenzon, and Yeung: Corporate Governance 703

corporate insider’s stance—she is more able bribes. An absence of democratic accounta- and willing to pay large side-payments to bility, a high general level of corruption, a political insider for weak private property controlled press, and an ignorant, illiterate, rights if her existing operations generate and impoverished citizenry probably shift larger cash flows and her future investment this balance toward accepting bribes. Of needs are smaller. In our terminology, the course, all these factors are also outcomes, transfer effect is larger, for larger cash flows potentially caused by economic entrench- mean the corporate insider has more oppor- ment. Glaezer and Shleifer (2003) describe a tunities to appropriate wealth. Interested in “Curley Effect” in Boston politics, whereby sustaining these cash flows, the corporate its Irish–American Mayor James Michael insider views the competition effect associat- Curley used wasteful redistribution and ed with high private property rights as a neg- inflammatory rhetoric to drive more affluent ative. Also, the corporate insider has more and educated Anglo–Saxon voters out of internal cash to finance future investment, Boston, thereby cementing his political con- so the cost of capital effect is less important. trol. Nonetheless, many authors argue per- The corporate insiders’ current cash flows suasively that these sorts of factors are and future investment needs certainly largely historically rooted and therefore depend on exogenous “initial conditions,” at exogenous in fundamental ways. However, least to some extent. Corporate insiders with the partial endogeneity of such factors might more experience in rent-seeking than in contribute critically to stabilizing economic genuine entrepreneurship, as in some transi- entrenchment. tion economies, might tend to favor weak Further considerations, and our reading property rights because these let them play of the available literature, let us resolve to their advantage. Joel S. Hellman, Geraint these generic considerations into a set of Jones, and Daniel Kaufmann (2003) call this fundamental determinants of economic state capture—in our terminology, the entrenchment. These consist of preexisting acquisition of control over the organs of the endowments and industrial conditions; state by corporate oligarchs. If such initial political, legal, social, and cultural traditions; conditions induce weak institutions, these and economic openness. We consider each may well further limit future investment set in turn. opportunities and alter political leaders’ con- 7.4.2 Initial Endowment and Industrial cern for the public interest. Ensuing path Conditions dependence might well lock in economic entrenchment. William Easterly and Levine (2003) con- All else equal, oligarchs’ lobbying should sider three broad classes of hypotheses. The be more effective if political insiders value first they dub the “endowments hypothe- side payments more and economic growth sis”—variables like settler mortality, lati- less. This weighting is inherent in the param- tude, natural resources, and access to the eters of the political insiders’ utility func- sea matter most in explaining general pros- tions, though this is a vast oversimplification. perity. The second, their “institutions Enumerating the likely determinants of the hypothesis,” focuses on private property weights politicians assign to the public inter- rights, accountability, political stability, gov- est versus their personal gain requires fur- ernment effectiveness, regulatory burden, ther reflection. This relative valuation most rule of law, and an absence of corruption. plausibly depends on the country’s political, The third, their “public policy hypothesis,” legal, and cultural traditions. holds that more standard economic variables Political insider’s preferences presumably like inflation, trade openness, and exchange also depend on the costs and benefits of taking rate overvaluation. sp05_Article 1 8/18/05 9:32 AM Page 704

704 Journal of Economic Literature, Vol. XLIII (September 2005)

Using the logarithm of 1995 real per capi- the privileged, even after the abolition of ta GDP as their dependent variable in mul- slavery. tiple regressions, they report that A second category is Spanish America. endowments explain economic growth, with These economies had valuable mineral natural resources and mortality rate the most resources and abundant native labor with significant factors. They also find that low human capital. The Spanish colonial endowments help explain cross-country vari- master built on preconquest social institu- ation in institutional development, with nat- tions, whereby Indian elites extracted trib- ural resources and mortality rate again the ute from the general population to control most significant factors; and that endow- huge landholdings. This system locked in the ments do not explain growth except through status of the earliest settlers, who became their influence on institutions. Finally, they formal representatives of the Spanish report that policy variables do not explain Crown, while restricting further settlement. growth once the institutional effects are The result is a highly unequal distribution of accounted for. The Easterly and Levine political and economic power. (2003) study is valuable not only for its In both of these categories of economies, empirical findings, but for demonstrating institutional development was directed how researchers can test the importance of toward protecting the economic and political different theoretical and conceptual expla- power of entrenched oligarchs. This was nations for these issues. We anticipate a accomplished by curtailing democracy and large literature developing along these lines. otherwise limiting the economic and politi- Other important studies provide valuable cal power of the commoners, say, by spend- insights into why endowments are critically ing little on basic public education. This is important. consistent with the oligarchs’ established Such a view is embedded in Kenneth L. investments generating high cash flows, but Sokoloff and Stanley L. Engerman (2000), with broader investment having relatively who study the economic history of the low returns to the oligarchs. Moreover, the Americas from the sixteenth to the eigh- oligarchs, the first wave of European settle- teenth century. They point out (p. 221) that ment, by favoring weak institutions, might plantation economies like Barbados, Cuba, well have limited the returns to further and Jamaica, which “specialized in the pro- investment by later immigrants or natives. duction of sugar and other highly valued The resulting skewed wealth distribution crops associated with extensive use of slaves induced politicians to attach little weight to had the highest per capita income in the the interests of the common people, further New World” in the sixteenth century, reinforcing economic entrenchment. including slaves in the calculation. The Sokoloff and Engerman (2000) contrast greater efficiency of the very large planta- these two categories of economies with a tions and the overwhelming fraction of the third—British North America. These areas, populations that came to be black and slave especially north of Virginia, had climates made the distributions of wealth and human unfavorable to large-scale plantations. capital extremely unequal. The long-run Their development was based instead on success and stability of the members of yeoman farmers and townsfolk of European these elites (white plantation owners) were descent with relatively high and similar lev- also facilitated by their disproportionate els of human capital—especially after set- political influence. Together with the legally tlement displaced the fur trade. codified inequality intrinsic to slavery, the Consequently, the areas that became greater inequality in wealth contributed to Canada and the United States developed the evolution of institutions that protected more egalitarian distributions of political sp05_Article 1 8/18/05 9:32 AM Page 705

Morck, Wolfenzon, and Yeung: Corporate Governance 705

power, and their institutions developed to propose that ethnic fragmentation causes provide more equal treatment and opportu- groups in power to invest in competitive nities to their populations. As well, they short-term “extractive” policies for fear of spent far more on education for the popula- another ethnic group usurping power. They tion as a whole. These colonies presented further suggest that ethnic diversity, or few opportunities for the first wave of set- more general political fragmentation, tlers to appropriate cash flows, but offered a reduces the consensus for funding good broad range of investment opportunities public goods like infrastructure, education, that required outside capital. In this envi- and financial regulation. They provide ronment, the initial settlers favored proper- robust supportive empirical evidence. ty rights protection and sound institutions in Acemoglu, Johnson, and Robinson (2002, general. The resulting relatively even distri- 2003), using urbanization to proxy for eco- bution of wealth induced politicians to place nomic development, demonstrate a “rever- a greater weight on general prosperity. sal of fortune” effect. Many countries that Acemoglu, Johnson, and Robinson (2001) were relatively rich in 1500 are now rela- propose a different explanation as to how the tively poor. They argue that this is inconsis- initial conditions confronted by European tent with a simple geographic or climatic settlers in different countries generated dif- story. Instead, they argue that “[i]n prosper- ferent investment opportunities and hence ous and densely settled areas, Europeans different property rights regimes. They introduced or maintained already existing examine the economic growth of British extractive institutions to force the large pop- colonies to ascertain why some became rich ulation and the slaves imported from Africa and others did not. They argue that large to work in mines and plantations.” “In con- European settler populations preferred trast, in previously sparsely settled areas, institutions that enforced the rule of law and Europeans settled in large numbers and encouraged investment. In places where cli- created institutions of private property … matic conditions led to high settler mortality, [that] encourage commerce and industry.” large settler populations were unviable, and This institutional development made possi- small cadres of European colonial overlords ble the spillover benefits associated with took charge and established institutions industrialization in the eighteenth and nine- based on exploiting indigenous populations. teenth centuries that built their current high Taking European settlers’ early mortality standards of living. rate as an exogenous instrument, they find Thus, the work by Sokoloff and Engerman that low mortality encouraged European set- suggests that an initial round of investment tlement; inducing early institutions support- opportunities creates either a skewed or ive of the rule of law, public education, etc., egalitarian distribution of wealth, and that which form the basis of current institutions. this in turn provides politicians with either High settler mortality fostered “extractive” an apolitical peasantry or a politically active institutions, which also persist. They go on to middle class. The former allows politicians show that current per capita income is high- to place greater weight on their personal er the better the country’s current protection gains than does the latter. Acemoglu, against state expropriation; and that this, in Johnson, and Robinson suggest that large turn, is highly negatively correlated with populations of European settlers, who early settler mortality rates. expected to build a new homeland, trans- In a similar vein, Easterly and Levine planted institutions supportive of the rule of (1997) use African data from the 1960s law and property rights. The ensuing devel- through the 1980s to show that ethnic opment of social and political traditions diversity is related to slow growth. They allows these institutions to persist. These sp05_Article 1 8/18/05 9:32 AM Page 706

706 Journal of Economic Literature, Vol. XLIII (September 2005)

observations lead to the following additional Political systems are probably not exoge- considerations. nous in the long run, for corporate insider’s lobbying targets might include the political 7.4.3 Traditions—Political, Legal, Social, process itself, as well as private property and Cultural rights. Engerman and Sokoloff (1997) and Rajan and Zingales (2003) point out that Sokoloff and Engerman (2000) argue that the the financial atrophy, or “great reversals,” colonial elites of Latin American countries, they detect did not occur with equal intensi- intent on controlling the state to control natu- ty in all countries, and that many stock mar- ral resources, actively opposed democracy to kets flourished through the entire twentieth prevent the poor majority from gaining power century. They propose that different politi- and redistributing rents from those resources. cal, legal, and cultural traditions instill dif- Stephen Haber (2000) makes similar argu- ferent degrees of susceptibility to economic ments. Olson (2000) and others present gen- entrenchment, much as certain genes create eral models of how entrenched special dispositions toward specific infirmities given interests take control of the political process. the necessary environmental triggers. Rajan and Zingales (2003) propose that the A tradition of democracy is probably U.S. political process is largely immunized important, see Robert J. Barro (1991, 1997). against economic entrenchment (which they If political insiders are elected, and if out- call “the interest group theory of financial siders substantially outvote insiders, political development”) because of its multitudinous insider’s should assign a lower weight to branches of government, which provide checks side-payments than they would absent and balances, and because of the diversity and democracy. Democratic traditions are the activity of competing interest groups. work of decades and centuries, and so can be In addition to democracy, an efficient and regarded as exogenous—at least compared independent judiciary seems important. It to more ephemeral economic phenomena. raises the penalty for being caught accepting For instance, religious dissidents and other bribes, and so plausibly shifts political insid- antiestablishment groups from Europe set- ers’ cost–benefit analysis toward greater con- tled early in the thirteen colonies that cern for economic growth. La Porta, became the United States. Their popular Lopez-de-Silanes, Shleifer, and Vishny suspicion of excessively centralized political (1998) and others propose that ancient legal and economic power regularly resurges in traditions, established centuries or even mil- American political history; as when lennia ago through colonial occupation, inva- President Woodrow Wilson (1913–21) sion, or economic domination, still influence wrote: “No country can afford to have its levels of general corruption, judicial efficien- prosperity originated by a small controlling cy, and basic legal rights. La Porta, Lopez-de- class. The treasury of America does not lie in Silanes, Shleifer, and Vishny (1998) echo the brains of the small body of men now in Friedrich Hayek (1960) in proposing that, in control of the great enterprises, . . . It Civil Code countries, which subordinate depends upon the inventions of unknown courts to the State, elites use the judicial sys- men, upon the originations of unknown tem to become entrenched. In contrast, men, upon the ambitions of unknown men. Common Law traditions have more inde- Every country is renewed out of the ranks of pendent judiciaries that protect private prop- the unknown, not out of the ranks of the erty owners from the State.25 See Beck, already famous and powerful in control.” 24 25 Another difference is that Common Law systems, based on continually reinterpreted precedents, are better 24 See Louis Brandeis (1914) chapter 10, “The than Civil Code traditions at adjusting to changing social Inefficiency of the Oligarchs.” and economic circumstances. sp05_Article 1 8/18/05 9:32 AM Page 707

Morck, Wolfenzon, and Yeung: Corporate Governance 707

Levine, and Demirgüç-Kunt (2003) for the poor majority from gaining power and empirical evidence consistent with this view. redistributing rents from natural resources. Yet the evidence in Rajan and Zingales (2003) Even less tangible institutions, like culture that many civil law countries (such as Belgium and mores, may also be important. Landes or France) were much more financially devel- (1949) attributes the generally poor perform- oped than common countries (such as the ance of the French economy compared to United States and Australia) undermines the those of Germany, Great Britain, and the stronger version of this theory, and brings United States throughout the nineteenth cen- back a central role for political choice on leg- tury to a predominance of family control in islation which seems to have led to a reversal France. He argues that French family firms in the interwar period (see also Perotti and were highly risk averse and more interested in von Thadden forthcoming). survival and succession than in growth and Another important adjunct to democracy, innovation. Thus, family businesses lobbied that also probably raises the costs of taking for protectionism and bailouts, and regarded bribes, is a free and independent press. A the state as “a sort of father in whose arms free press exposes and criticizes political [they] could always find shelter and consola- insiders who value side-payments more tion” (p. 50). The cost was slow overall econo- than social objectives. A free press also my growth. Consistent with this, Murphy exposes controlling insiders’ extractions of (2004) argues that French culture places a private benefits. Finally, a free press huge weight on each generation’s sacred duty informs outsiders, making them less manip- to pass on a patrimony to the next, and that the ulable, and thus making insiders’ lobbying French Civil Code makes it very difficult for less effective. Dyck and Zingales (2002) the patriarch of a family business to disinherit show that an independent and free press his son. mitigates theft by corporate insiders. Max Weber (1958), La Porta, Lopez-de- However, Djankov, Clara McLeish, Silanes, Shleifer, and Vishny (1997a), and Nenova, and Shleifer (2003) show that the others argue that religion, perhaps the most media virtually everywhere are controlled intangible of institutions, also matters, with either by the State or by a few wealthy fam- some religions more amenable to commerce ilies. Where those families are part of oli- and finance than others. Weber (1958) pro- garchic elites, the watchdog role of the poses that the Protestant religion, more press may be seriously compromised specifically the Calvinist Reformation, by despite seeming free of State interference. emphasizing individual accountability, fos- Less quantifiable social institutions also tered a cultural shift that favored entrepre- probably determine a country’s propensity for neurs over old-money elites. La Porta, economic entrenchment. For example, Lopez-de-Silanes, Shleifer, and Vishny whether a country provides sound basic pub- (1997a) argue that hierarchical religions, lic education might matter. Better educated which they define to include Roman outsiders are better able to understand and Catholicism and Islam, are less conducive to participate in political debates. This presum- the growth of large publicly owned business- ably lowers outsiders’ lobbying costs relative es. Like Weber, they emphasize the impor- to insiders’. A better educated populace also tance of cultural institutions that encourage presumably generates more capable and individual initiative and cooperation among politically astute outsider entrepreneurs. unrelated individuals. These characteristics Engerman and Sokoloff (1997) and Sokoloff embolden outsiders and limit popular and Engerman (2000) argue that Latin acceptance of oligarchic rule. Stulz and American colonial elites actively opposed Rohan Williamson (2003) show that a coun- public investment in human capital to prevent try’s principal religion (a proxy for culture) sp05_Article 1 8/18/05 9:32 AM Page 708

708 Journal of Economic Literature, Vol. XLIII (September 2005)

helps predict cross-country variation in credi- capital flows—affect the cost of capital tor rights. For example, Catholic countries among other things. They find that liberaliza- provide creditors weaker rights than other tions decrease the cost of capital by signifi- countries do; and firms in Catholic countries cant margins: 5 to 75 basis points. Similarly, use less long-term debt. Religion actually pre- Peter Blair Henry (2000a, 2000b) reports a dicts a country’s creditor rights better than significant favorable impact of capital open- openness to international trade, language, ness on the domestic cost of capital, and thus income per capita, and even legal origin. on investment. John Coffee (2002), Andrew Religion and language, both as proxies for cul- Karolyi and Stulz (2003), Stulz (1999), and ture, are also important explanatory variables Jordan Siegel (2005) discuss how cross-listing of property rights enforcement. on foreign exchanges and submitting to for- Further research along the lines of all of eign regulators can credibly signal investors the studies discussed above is likely to gen- about firm quality. erate exciting new insights. We are only Second, autarky allows better control over beginning to understand these very basic the pace of creative destruction. Capital bar- issues, which are fundamental not only to riers can keep local innovators from devel- economics and finance, but to the whole oping their innovations, cutting into existing spectrum of social sciences. firms’ markets, and accumulating wealth. Trade barriers can keep innovative foreign 7.4.4 Globalization products out of domestic markets and thus One particularly rapidly changing aspect of preserve the value of existing capital assets. the economic environment of many countries Suppressing domestic innovation is of no is openness to the global economy. European value if foreign innovators can flood the countries are merging into an ever closer market with their products. economic and political union, regional free Third, autarchy gives governments much trade blocks and bilateral free trade treaties greater policy freedom. Confiscatory taxes, are much in vogue, and pressure continues skewed regulations, inefficient subsidies, for further multilateral action to reduce trade and other politically related private benefits and capital flow barriers. Globalization mer- of control are difficult to sustain in an open its special attention here, for the persistence economy. And bilateral and multilateral of economic entrenchment requires a degree treaties limit governments’ powers to enact of economic autarky for several reasons26. such policies. Moreover, footloose capital, First, autarchy preserves monopoly and skilled labor, and consumers can take their monopsony power in product and factor mar- business elsewhere if the local economy kets in general. If domestic entrants can bor- becomes too distorted. row abroad, list on foreign stock exchanges, Finally, locals may imitate foreigners. or otherwise obtain capital from foreign Richard E. Caves (1974) and a huge subse- investors, they can establish themselves quent literature argue that entry by foreign despite ill functioning local capital markets multinationals raises productivity because of and institutions. Geert Bekaert and Campbell technology spillovers and heightened compe- R. Harvey (2000) examine how emerging tition in both goods and factor markets. Locals market liberalizations—regulatory changes, may also come to demand the same financial the introduction of depositary receipts and and political rights they see foreigners country funds, and structural breaks in equity enjoying. That is, openness may create popular support for institutional reform. Local governments and regulators may be pressed 26 For a very recent discussion on the limits of globalization in the context of concentrated control corporations, to adopt international best practices in see Stulz (forthcoming). disclosure, governance, and regulation. sp05_Article 1 8/18/05 9:32 AM Page 709

Morck, Wolfenzon, and Yeung: Corporate Governance 709

For example, Stulz (1999) analyzes how Stangeland, and Yeung (2000) show that openness alters domestic investors demand for economies with more inherited billionaire better corporate governance. Locals can learn wealth as a fraction of GDP have more from foreign investors the skills to seek infor- restrictions on inward foreign direct invest- mation about domestic companies and to ment. Rajan and Zingales (2003) find greater monitor performance. Khanna and Palepu “great reversals,” or financial atavism, in (2001) examine Indian firms and show that countries with more closed economies. They foreign investors identify better performing argue that dual openness, that is trade open- India firms and propose that their monitoring ness and capital flow openness, is essential to improves performance. Active foreign counter the forces that undermine capital investors both train personnel for domestic market development. With only trade open- investors and serve as role models. Also, the ness, industrial incumbents press for govern- presence of active investors raises the demand ment subsidies as well as the maintenance of for information services, such as auditing and their privileged access to capital. The end financial analysis. And foreign investors can result is the sustained repression of domestic activate a dormant market for corporate con- capital market development. Moreover, they trol. All of this raises the net benefits to corpo- argue that capital openness requires private rate insiders of providing good governance and property rights, for otherwise information lowers the cost of sending a “good governance” asymmetry continues to mar domestic capi- signal to investors. tal markets, preventing small local firms to Thus, economic openness, the freedom of seek foreign capital and thus sustaining the locals to do business with foreigners, ought dominance of the local oligarchs (meaning also to be numbered among private proper- that the foreign capital will flow to the oli- ty rights; and economic autarchy is probably garch’s companies). Rajan and Zingales a lobbying goal of oligarchs seeking econom- (2003) use exports plus imports over GDP to ic entrenchment. At the same time, open- proxy for trade openness, international capi- ness, once installed, probably will make tal flows to proxy for capital flow openness, economic entrenchment more difficult to and measures such as equity issues, the attain for elite. number of listed companies, and stock mar- While direct evidence of this is limited, ket capitalization to proxy for financial devel- circumstantial evidence is supportive.27 opment. They report that trade openness Faccio (2003) shows that economies in heightens the positive impact of industrial- which the corporate and political elites have ization on financial development. They also stronger blood ties also have more restric- note that this impact is much more impor- tions on cross-border capital flows. Morck, tant early in the twentieth century than in its latter decades. The impact disappears in the midcentury period when international capi- 27 By direct evidence, we mean evidence that links tal mobility was limited, roughly from 1930 openness, or the lack thereof, to lobbying behavior. An to the 1970s. extensive literature links openness to development. This Some direct evidence pertaining to deserves another survey, but some related contributions merit mention here. Sachs and Warner (1995) report a changes in openness is available to support positive impact of globalization on reforming countries. these arguments. Morck, Stangeland, and Bekaert and Harvey (forthcoming) find that financial lib- Yeung (2000) show that the unexpected eralization has a greater positive impact on growth in countries with better institutional environments. These include enactment of free trade with the United countries with English legal origins, better investor pro- States caused the stock prices of Canadian tection, and speedier judicial processes. Li, Morck, Fan firms controlled by old money families to fall Yang, and Yeung (2004) find that when emerging markets grow more open to foreign investors, their stock markets relative to other Canadian stocks in the same grow more discriminating in valuing individual companies. industries. They also show that Canadian sp05_Article 1 8/18/05 9:32 AM Page 710

710 Journal of Economic Literature, Vol. XLIII (September 2005)

firms controlled by old money families had constituents, like organized labor and pro- abnormally high capital intensity prior to fessional managers certainly matter. Also, free trade, but that their capital intensity fell the framework assumes a unitary govern- to industry, firm size, and firm age corrected ment, sidestepping complicated political norms subsequent to free trade. Johnson processes and tensions associated with and Mitton (2003) show that, when the alternative constitutional arrangements, Malaysian government imposed capital con- e.g., systems with substantial separation of trols in 1998 in response to allegations that powers, federal systems, and alternative foreign “hot money” caused the East Asian electoral systems. These are important economic crisis, the stock prices of firms issues. with ties to the ruling party rose and the Expanding the list of players beyond “cor- stock prices of other firms fell. Their result porate insiders” and “outsiders” may lead to suggests that capital controls create a screen useful and competing explanations for con- for cronyism, in which the government for- centrated corporate control other than eco- wards resources to connected firms via their nomic entrenchment. Roe (2003) suggests policies. However, Siegel (2004) finds that that corporate control must be concentrated increasing openness in Korea primarily in economies with powerful organized labor expanded and strengthened that countries’ and other powerful interest groups who most politically connected firms. Further would otherwise exploit providers of capital. evidence is needed to clarify the issue. Perotti and von Thadden (forthcoming) The arguments of Morck, Stangeland, and examine political economy determinants of Yeung (2000), Johnson and Mitton (2003), corporate control and argue that corporate and Rajan and Zingales (2003) suggest that control arrangements reflect the median vot- openness to global capital flows, as well as to ers’ preference. Laborers and managers with international trade, ought to augment undiversified human capital dislike risky cor- national economic performance. However, porate strategies. If these constituencies pro- other students of international finance dis- vide the median voter, a democratic society agree out of concern that capital market would grant corporate control to banks or openness might heighten economic volatili- others averse to risky strategies, whose pref- ty. For example, Jagdish Bhagwati (1998) erences resemble those of voters more than argues that capital market openness can those of wealth maximizing diversified spread financial crises, and Kristin J. Forbes shareholders. and Roberto Rigobon (2002) study conta- Political market competition and constitu- gion as a factor in market fluctuations. Be tions, mechanisms that determine winners in this as it may, avoiding economic entrench- the political market, could lead to important ment clearly counts as a serious weight on additional insights. The intensity of competi- the side of the balance favoring capital open- tion in the political market might affect the ness. Further work exploring this issue is equilibrium path of institutional develop- likely to be of considerable value. ment. In the framework of Engerman and Sokoloff (1997), Sokoloff and Engerman 7.5 Critique (2000), and Acemoglu, Johnson, and Robinson (2001), institutional development The above political economy framework follows the preferences of the dominantly is deliberately simplified for clarity, and powerful. Easterly and Levine (2003) uses a limited number of players. The evo- describe a political market crowded with lution of institutions is certainly affected by unstable outcomes so that all ruling parties more than “corporate insiders” and “out- act myopically, impeding property rights siders” lobbying “political insiders.” Other development. Also, constitutions might sp05_Article 1 8/18/05 9:32 AM Page 711

Morck, Wolfenzon, and Yeung: Corporate Governance 711

affect the behavior of political actors and that might compromise economic growth. their weighing of constituents’ interests. This section proposes the concept of econom- Research addressing these issues is need- ic entrenchment as a feedback loop, whereby ed for a clearer understanding of institution- weak institutions place sweeping corporate al development. A useful example is Pagano governance powers in the hands of a tiny elite and Volpin (forthcoming), who model entre- group, who then lobby for weak institutions preneurs, rentiers (minority shareholders), to preserve their concentrated control over workers, and two competing political parties. the countries large corporations—oligarchic They show that the electoral system, capitalism. whether proportional (where a political But sometimes, highly concentrated cor- party dominates by winning a majority of porate control does not induce economic votes) or majoritarian (where a political entrenchment. Strong institutions develop party aims to win in a majority of districts), and diffuse capitalism takes hold. This affects the equilibrium level of investor pro- apparent indeterminacy in the coevolution tection. Under a proportional rule, parties of institutions and corporate control is an cater to social groups whose votes might add important puzzle. Economic entrenchment up to a dominant position, and an outcome explains why family business groups, an favorable to entrepreneurs or employees, institutional advantage early in economic but unfavorable to rentiers, ensues. Under a development, become oligarch capitalism in majoritarian rule, parties compete for votes some countries, but not others. Exploring in pivotal districts, which may not be ideo- the political economy basis of financial logically committed. If these districts are development provides answers, and is likely populated by rentiers, who focus on the to yield further insights soon.28 implications of policies for overall economic Some core concepts underlie economic growth, stronger shareholder protection and entrenchment as a possible political econo- weaker employment protection are more my outcome. First, entrusting the corporate likely. The paper vividly demonstrates the governance of the greater part of a country’s fruitfulness of injecting multiple con- corporate sector to a tiny elite, especially via stituents and political competition into control pyramids, gives that elite huge advan- the research on the political economy of tages in political lobbying. Second, such oli- institutional development. garchs lobby for weak property rights if they prefer increased private benefits of control 7.6 Summary over better access to outside investors’ capi- The core inquiry in this survey is the con- tal. Thus, the elite prefer strong private centrated control of corporations and its property rights only under some circum- micro and macro level economic implica- stances, while the rest of the population tions. Concentrated control of a country’s always prefers strong private property rights. large corporations, as in family control pyra- Third, economic entrenchment requires that mids, could serve important positive econom- political insiders see the benefits of cooper- ic functions. It might help overcome market ating with the oligarchs as outweighing the inefficiencies by facilitating monitoring, con- costs, despite its detrimental effects on the trol, and capital allocation. But the concen- populace in general. trated control of a country’s large This framework does not imply that eco- corporations might also induce microeco- nomic entrenchment is inevitable. The elite nomic problems associated with entrenched and the general citizenry might both prefer management wielding control rights vastly 28 This approach is gaining attention from, among oth- disproportionate to their ownership stakes ers, Acemoglu, Johnson, and Robinson (2004), Perotti and and macroeconomic resource misallocation Volpin (2004), and most recently Stulz (forthcoming). sp05_Article 1 8/18/05 9:32 AM Page 712

712 Journal of Economic Literature, Vol. XLIII (September 2005)

strong property rights, particularly if the speculation, and to clarify the trade-offs future economic outlook is bright and the involved. elite, not just the general citizenry, need The success of those entrusted with cor- external financing. However, the framework porate governance powers in influencing does suggest a high degree of path depend- institutional development doubtless depends ence. A historical shock that induced eco- on existing institutions—social, political, nomic entrenchment in the past could have judicial, and educational institutions. These a long-lasting effect on a country’s institu- institutions determine the effectiveness and tional development; for once in place, eco- nature of successful lobbying strategies, as nomic entrenchment can be self-sustaining well as the payoffs available to political insid- in the absence of further shocks. ers. Democracy, the rule of law, a free press, How corporate control and the political comprehensive public education, and indi- influence it engenders are path dependent is vidualistic cultures all probably raise political of great interest. For on this issue turns the insiders’ potential costs when accepting most important economic question of all— bribes from corporate insiders. why some countries are rich and others poor. This newly emerging literature suggests This section is deliberately speculative. Our conditions deep in each country’s economic central idea is that economists ought to history strengthened or weakened these explore how institutional development, as a sorts of institutions, either locking in oli- political economy issue, relates to the distri- garchic capitalism and economic entrench- bution of corporate control in an economy. ment or opening the way to rapid growth and Corporate control provides resources and more diffuse capitalism. But these historical, power to influence institutional develop- social, political, and judicial institutional ment. The whole body of economics sug- conditions are, in turn, endogenous, if we gests that such resources and power would push deeper into economic history. Our the- be used to further the self-interest of those oretical and empirical understanding about who wield them. how these interactions play out is severely The preliminary work surveyed above, limited. For example, we are only beginning notably Acemoglu et al. (2001, 2002, 2003), to understand the importance of legal tradi- Engerman and Sokoloff (1997), Easterly and tions in determining the property rights of Levine (2003), and Rajan and Zingales outside investors, and have little understand- (2003), highlights the importance of an ini- ing of other ways in which legal traditions tial distribution of economic power, the pref- might matter. We have barely scratched the erences of those wielding it, and the path surface in learning how different legal sys- dependence of institutional development. tems affect financial and economic institu- This research agenda is long and incom- tions such as the distribution of corporate plete. We know little about when or whether governance powers, let alone social, political, initial conditions matter more than the judicial, and educational institutions, not to elite’s preferences. We know little about how mention the complicated interactions that balance varies across countries and over between them all. Yet all of these almost cer- time. Much important work remains. tainly affect the political economy equilibri- We speculate that the preferences of um that determines property rights among entrenched elites regarding institutional other things. Further work—conceptual, development depend on their expected theoretical, and empirical—is needed. future needs for broader cooperation—for Overall, rigorous theoretical models of example external financing to undertake this sort of political economy game are likely future economic opportunities. But further to run up against difficult problems. Rajan work is needed to confirm or reject such and Zingales (2003) explicitly sidestep an sp05_Article 1 8/18/05 9:32 AM Page 713

Morck, Wolfenzon, and Yeung: Corporate Governance 713

apparent time inconsistency in institutional 8. Conclusion change. How should investors in a strong property rights institutional environment Economic growth seems related to the deal with the possibility that property rights distribution of control over an economy’s might be weakened in the future? Do large corporate sector. Outside of the investors buy stocks in emerging markets as United States and United Kingdom, most ways of betting on those countries develop- large corporations have controlling owners, ing good institutions? This might explain typically very wealthy families. Using pyram- some of the findings of Morck, Yeung, and idal structures, cross shareholding, and Yu (2000). But we have little real under- super voting rights, they typically control standing of how stable property rights many listed companies via control pyramids. regimes are and what contributes to their These drive wedges between such families’ stability. dominant control rights and their often While we suggest using formal political miniscule actual investments in companies economy models and hypothesis testing to they control. advance our understanding of institutional This separation permits the simultaneous development, this needs not be the only presence of two well known corporate gov- approach. At present, nontechnical and histor- ernance problems: a divergence of the inter- ically descriptive approaches, like Haber ests of the controlling owner from that of (2000), Roe (2003), and Morck (forthcoming), outside shareholders and the entrenchment can still provide critical insights. of the controlling owner. Tiers of fully con- Assembling either cross country and his- trolled, but only partially owned firms let the torical data for formal econometric analysis controlling owner utilize corporate can be labor intensive. One way of mitigat- resources she does not de jure own to fund ing this problem is for researchers to share private benefits with no fear of a successful their data. Some plausibly important vari- shareholder revolt or hostile takeover. By ables also measure concepts more com- orchestrating intercorporate transactions at monly associated with anthropology, artificial prices, the controlling owner can sociology, political science, or other social tunnel wealth between pyramid firms to sciences than with economics and finance. generate greater private benefits. The con- Borrowing data from other fields might sequence is inefficient resource allocation at thus help build a deeper understanding of the firm level. institutions, and of the extent to which they Many countries effectively entrust the really are creatures of economics. Surveys corporate governance of substantial parts of of the attitudes of different segments of their corporate sectors to the large pyrami- society toward different sorts of institution- dal groups of a few extremely wealthy fami- al changes might also be useful.29 Carefully lies. This can potentially magnify poor designed economic experiments might also governance by a few family patriarchs into help our understanding institutional devel- inefficient economywide capital allocation, opment while we await sufficient panel reduced investment in innovation, and data on real economic and institutional retarded economic growth. Moreover, such variables. elite families understandably value the status quo. They sometimes, but not always, appear to influence public policies to curtail private property rights development, capital 29 For example, Jorge Domínguez (1982) surveys atti- market development, economic openness, tudes of Latin American elites and outsiders to economic openness to detect what he calls national bourgeois and other threats to the status quo. Much coalitions—akin to our entrenched economic elites. existing work points to this economic sp05_Article 1 8/18/05 9:32 AM Page 714

714 Journal of Economic Literature, Vol. XLIII (September 2005)

entrenchment as a significant issue in many transition of the political economy from a countries. We propose a simple conceptual suboptimal to a better equilibrium. Further political economy framework to define eco- empirical and theoretical work that clarifies nomic entrenchment and the conditions these issues is likely to be of first order under which it takes hold. importance. While we develop our arguments in the REERENCES context of wealthy families with extensive Acemoglu, Daron, and James A. Robinson. 2002. control pyramids, the fundamental concept “Economic Backwardness in Political Perspective.” of economic entrenchment is more general. NBER Working Paper 8831. Tiny elites of professional managers running Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. “The Colonial Origins of keiretsu corporate groups in Japan, or of Comparative Development: An Empirical bureaucrats running state controlled pyra- Investigation.” American Economic Review, 91(5): mids in France, might also be entrenched. 1369–1401. Acemoglu, Daron, Simon Johnson, and James A. The essential issue is the negative conse- Robinson. 2002. “Reversal of Fortune: Geography quences for growth of entrusting an econo- and Institutions in the Making of the Modern World my’s capital allocation to a tiny elite that Income Distribution.” Quarterly Journal of Economics, 117(4): 1231–94. cannot be sacked. Acemoglu, Daron, Simon Johnson, and James A. Our literature survey suggests several key Robinson. 2003. “Understanding Prosperity and issues. Poverty: Geography, Institutions and the Reversal of Fortune.” Massachusetts Institute of Technology. First, the archetypal corporate gover- Mimeo. nance problem in the modern United States Acemoglu, Daron, Simon Johnson, and James A. economy, a conflict between atomistic share- Robinson. 2004. “Institutions as the Fundamental Cause of Long-Run Growth.” NBER Working Paper holders and professional managers, does not 10481. generalize to most other countries. Rather, Aghion, Philippe, Eve Caroli, and Cecilia Garcia- large firms in most countries are typically Penalosa. 1999. “Inequality and Economic Growth: The Perspective of the New Growth Theories.” organized into pyramidal groups controlled Journal of Economic Literature, 37(4): 1615–60. by a few wealthy families. The ensuing cor- Agnblad, Jonas, Erik Berglöf, Peter Högfeldt, and porate governance problem is conflict Helena Svancar. 2001. “Ownership and Control in Sweden: Strong Owners, Weak Minorities, and between the controlling shareholder of the Social Control,” in The Control of Corporate Europe. pyramidal group and public shareholders. Fabrizio Barca and Marco Becht, eds. Oxford: Second, the distribution of control over the Oxford University Press, 228–58. Aikawa, Yoshisuke. 1934. New Capitalism and Holding corporate sector affects economic develop- Companies. Tokyo: Tokyo Bankers Association. ment. Such highly concentrated control over Aitken, Brian J., and Ann E. Harrison. 1999. “Do corporate assets plausibly leads to a range of Domestic Firms Benefit from Direct Foreign Investment? Evidence from Venezuela.” American market power distortions, especially in capi- Economic Review, 89(3): 605–18. tal markets. It also may curtail investment in Aitken, Brian, Ann E. Harrison, and Robert E. Lipsey. innovation and augment rent-seeking. All of 1996. “Wages and Foreign Ownership: A Comparative Study of Mexico, Venezuela, and the these effects likely retard economic growth. United States.” Journal of International Economics, Third, public policy regarding important 40(3–4): 345–71. issues, like property rights, the development Allen, Jeffrey W., and Gordon M. Phillips. 2000. “Corporate Equity Ownership, Strategic Alliances, of financial markets and institutions, and and Product Market Relationships.” Journal of economic openness, is usefully thought of as Finance, 55(6): 2791–2815. a political economy outcome. From this per- Almazan, Andres, and Javier Suarez. 2003. “Entrenchment and Severance Pay in Optimal spective, “public policy” in many countries Governance Structures.” Journal of Finance, 58(2): cannot be considered a discretionary vari- 519–47. able that can be “adjusted” to cure econom- Almeida, Heitor, and Daniel Wolfenzon. Forthcoming-a. “A Theory of Pyramidal Ownership and Family ic ills. We need to identify those factors that Business Groups.” Journal of Finance. are “adjustable” and that might induce a Almeida, Heitor, and Daniel Wolfenzon. Forthcoming-b. sp05_Article 1 8/18/05 9:32 AM Page 715

Morck, Wolfenzon, and Yeung: Corporate Governance 715

“Should Business Groups be Dismantled? The Costs of Separating Control from Cash-Flow Rights,” Equilibrium Costs of Efficient Internal Capital in Concentrated Corporate Ownership. R. K. Morck, Markets.” Journal of Financial Economics. ed. Chicago: University of Chicago Press, 445–60. Anderson, Ronald C., and David M. Reeb. Beck, Thorsten, Asli Demirgüç-Kunt, and Ross Levine. 2003a. “Founding-Family Ownership, Corporate 2005. “Bank Supervision and Corruption in Diversification, and Firm Leverage.” Journal of Lending.” University of Minnesota. Mimeo. Law and Economics, 46(2): 653–84. Beck, Thorsten, Ross Levine, and Asli Demirguc-Kunt. Anderson, Ronald C., and David M Reeb. 2003. “Law and Finance: Why Does Legal Origin 2003b. “Founding-Family Ownership and Firm Matter?” Journal of Comparative Economics, 31(4): Performance: Evidence from the S&P 500.” Journal 653–76. of Finance, 58(3): 1301–28. Beck, Thorsten, Ross Levine, and Norman Loayza. Anderson, Ronald C., Sattar A. Mansi, and David M. 2000. “Finance and the Sources of Growth.” Journal Reeb. 2003. “Founding Family Ownership and the of Financial Economics, 58(1–2): 261–300. Agency Cost of Debt.” Journal of Financial Bekaert, Geert, and Campbell R. Harvey. 2000. Economics, 68(2): 263–85. “Foreign Speculators and Emerging Equity Attig, Najah, Klaus Fischer, and Yoser Gadhoum. 2003. Markets.” Journal of Finance, 55(2): 565–613. “On the Determinants, Costs, and Benefits of Bekaert, Geert, and Campbell R. Harvey. Pyramidal Ownership: Evidence on Expropriation of Forthcoming. “Does Financial Liberalization Spur Minority Interests.” Working Paper. Growth?” Journal of Financial Economics. Attig, Najah, Yoser Gadhoum, and Larry Lang. 2003. Bennedsen, Morten, and Daniel Wolfenzon. 2000. “Bid–Ask Spread, Asymmetric Information and “The Balance of Power in Close Corporations.” Ultimate Ownership.” Working Paper. Journal of Financial Economics, 58(1–2): 113–40. Bae, Kee-Hong, Jun-Koo Kang, and Jin-Mo Kim. 2002. Berglof, Erik, and Enrico Perotti. 1994. “The “Tunneling or Value Added? Evidence from Mergers Governance Structure of the Japanese Financial by Korean Business Groups.” Journal of Finance, Keiretsu.” Journal of Financial Economics, 36(2): 57(6): 2695–2740. 259–84. Baek, Jae-Seung, Jun-Koo Kang, and Kyung Suh Park. Bergstrom, Clas, and Kristian Rydqvist. 1990. 2004. “Corporate Governance and Firm Value: “Ownership of Equity in Dual-Class Firms.” Journal Evidence from the Korean Financial Crisis.” Journal of Banking and Finance, 14(2–3): 255–69. of Financial Economics, 71(2): 265–313. Bergstrom, Clas, and Kristian Rydqvist. 1992. Baker, George P. 1992. “Beatrice: A Study in the “Differentiated Bids for Voting and Restricted Creation and Destruction of Value.” Journal of Voting Shares in Public Tender Offers.” Journal of Finance, 47(3): 1081–1119. Banking and Finance, 16(1): 97–114. Barca, Fabrizio, and Marco Becht. 2001. The Control of Berle, Adolph A., and Gardiner C. Means. 1932. The Corporate Europe. Oxford: Oxford University Press. Modern Corporation and Private Property. New Barclay, Michael J., and Clifford G. Holderness. 1989. York: Macmillan Company. “Private Benefits from Control of Public Bertrand, Marianne, Paras Mehta, and Sendhil Corporations.” Journal of Financial Economics, Mullainathan. 2002. “Ferreting Out Tunneling: An 25(2): 371–95. Application to Indian Business Groups.” Quarterly Barclay, Michael J., Clifford G. Holderness, and Jeffrey Journal of Economics, 117(1): 121–48. Pontiff. 1993. “Private Benefits from Block Bhagwati, Jagdish. 1998. “The Capital Myth: The Ownership and Discounts on Closed-End Funds.” Difference between Trade in Widgets and Dollars.” Journal of Financial Economics, 33(3): 263–91. Foreign Affairs, 77(3): 7–12. Barro, Robert J. 1991. “Economic Growth in a Cross Bianchi, Marco, Magda Bianco, and Luca Enriques. Section of Countries.” Quarterly Journal of 2001. “The Separation Between Ownership and Economics, 106(2): 407–43. Control in Italy.” Bank of Italy. Mimeo. Barro, Robert J. 1997. Determinants of Economic Billett, Matthew T., and David C. Mauer. 2000. Growth: A Cross-Country Empirical Study. “Diversification and the Value of Internal Capital Cambridge: MIT Press. Markets: The Case of Tracking Stock.” Journal of Barth, James R., Gerard Caprio Jr., and Ross Levine. Banking and Finance, 24(9): 1457–90. 2004. “Bank Regulation and Supervision: What Black, Bernard, Reinier Kraakman, and Anna Works Best?” Journal of Financial Intermediation, Tarassova. 2000. “Russian Privatization and 13(2): 205–48. Corporate Governance: What Went Wrong?” Bebchuk, Lucian Arye. 1999. “A Rent-Protection Stanford Law Review, 52: 1731–1808. Theory of Corporate Ownership and Control. NBER Boyd, John H., and Edward C. Prescott. 1986. Working Paper 7203. “Financial Intermediary-Coalitions.” Journal of Bebchuk, Lucian Arye, and Christine Jolls. 1999. Economic Theory, 38(2): 211–32. “Managerial Value Diversion and Shareholder Brandeis, Louis. 1914. “Other People’s Money and Wealth.” Journal of Law, Economics, and How the Bankers Use It.” Brandeis School of Law Organization, 15(2): 487–502. Library, University of Louisville. Bebchuk, Lucian Arye, Reinier Kraakman, and George Brunello, Giorgio, Clara Graziano, and Bruno M. G. Triantis. 2000. “Stock Pyramids, Cross-Ownership, Parigi. 2003. “CEO Turnover in Insider-Dominated and Dual Class Equity: The Mechanisms and Agency Boards: The Italian Case.” Journal of Banking and sp05_Article 1 8/18/05 9:32 AM Page 716

716 Journal of Economic Literature, Vol. XLIII (September 2005)

Finance, 27(6): 1027–51. Studies, 51(3): 393–414. Burkart, Mike, Fausto Panunzi, and Andrei Shleifer. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de- 2003. “Family Firms.” Journal of Finance, 58(5): Silanes, and Andrei Shleifer. 2002. “The Regulation of 2167–2201. Entry.” Quarterly Journal of Economics, 117(1): 1–37. Caprio, Gerard, Luc Laeven, and Ross Levine. 2003. Djankov, Simeon, Rafael La Porta, Tatiana Nenova, “Governance and Bank Valuation.” NBER Working and Andrei Shleifer. 2003. “Who Owns the Media?” Paper 10158. Journal of Law and Economics, 46(2): 341–81. Carkovic, Maria, and Ross Levine. 2002. “Does Domínguez, Jorge. 1982. “Business Nationalism: Latin Foreign Direct Investment Accelerate Economic American National Business Attitudes and Behavior Growth?” Paper presented at the World Bank towards Multinational Enterprises,” in Economic Conference, May 30–31, 2002. Issues and Political Conflict: US–Latin American Caselli, Francesco, and Nicola Gennaioli. 2003. Relations. Jorge Dominguez, ed. Boston: “Dynastic Management.” NBER Working Paper Butterworth Scientific, 16–68. 9442. D’Souza, Julia, and John Jacob. 2000. “Why Firms Castro, Rui, Gian Luca Clementi, and Glenn Issue Targeted Stock.” Journal of Financial MacDonald. 2003. “Investor Protection, Optimal Economics, 56(3): 459–83. Incentives, and Economic Growth.” Stern School of Durnev, Art, Kan Li, Randall Morck, and Bernard Business Working Paper. Yeung. 2004. “Capital Markets and Capital Caves, Richard E. 1974. “Multinational Firms, Allocation: Implications for Economies in Competition, and Productivity in Host-Country Transition.” Economics of Transition, 12(4): 593–634. Markets.” Economica, 41(162): 176–93. Durnev, Art, Randall Morck, and Bernard Yeung. Choi, Dosoung. 1991. “Toehold Acquisitions, 2001. “Does Firm-Specific Information in Stock Shareholder Wealth, and the Market for Corporate Prices Guide Capital Allocation?” NBER Working Control.” Journal of Financial and Quantitative Paper 8093. Analysis, 26(3): 391–407. Durnev, Art, Randall Morck, and Bernard Yeung. 2004. Claessens, Stijn, Simeon Djankov, and Larry H. P. “Value-Enhancing Capital Budgeting and Firm- Lang. 2000. “The Separation of Ownership and Specific Stock Return Variation.” Journal of Finance, Control in East Asian Corporations.” Journal of 59(1): 65–105. Financial Economics, 58(1–2): 81–112. Dyck, Alexander, and Luigi Zingales. 2002. “The Claessens, Stijn, Simeon Djankov, Joseph Fan, and Corporate Governance Role of the Media,” in The Larry Lang. 2002. “Disentangling the Incentive and Right to Tell: The Role of Mass Media in Economic Entrenchment Effects of Large Shareholdings.” Development. Roumeen Islam, ed. Washington, Journal of Finance, 57(6): 2741–71. D.C.: World Bank, 107–40. Coase, Ronald. 1937. “The Nature of the Firm.” Dyck, Alexander, and Luigi Zingales. 2004. “Private Economica, 4: 386–405. Benefits of Control: An International Comparison.” Coffee, John. 2001. “Do Norms Matter? A Cross- Journal of Finance, 59(2): 537–600. Country Examination of Private Benefits of Control.” Easterly, William, and Ross Levine. 1997. “Africa’s Columbia University Law School Working Paper. Growth Tragedy: Policies and Ethnic Divisions.” Coffee, John. 2002. “Competition Among Securities Quarterly Journal of Economics, 112(4): 1203–50. Markets: A Path Dependent Perspective.” Columbia Easterly, William, and Ross Levine. 2003. “Tropics, Law and Economics Working Paper 192. Germs, and Crops: How Endowments Influence Cull, Robert, Jana Matesova, and Mary Shirley. 2002. Economic Development.” Journal of Monetary “Ownership and the Temptation to Loot: Evidence Economics, 50(1): 3–39. from Privatized Firms in the Czech Republic.” Easterly, William. 2002. “Inequality Does Cause Journal of Comparative Economics, 30(1): 1–24. Underdevelopment: New Evidence from Commodity De Jong, Abe, and Chris Veld. 2001. “An Empirical Endowments, Middle Class Share, and Other Analysis of Incremental Capital Structure Decisions Determinants of Per Capita Income.” New York under Managerial Entrenchment.” Journal of University Working Paper. Banking and Finance, 25(10): 1857–95. Edwards, Jeremy, and Klaus Fischer. 1994. Banks, De Long, J. Bradford. 1990. “Did J. P. Morgan’s Men Finance and Investment in Germany. Cambridge: Add Value? A Historical Perspective on Financial Cambridge University Press. Capitalism.” NBER Working Paper 3426. Engerman, Stanley L., and Kenneth L. Sokoloff. 1997. Denis, Diane K., and John J. McConnell. 2003. “Factor Endowments, Institutions, and Differential “International Corporate Governance.” Journal of Paths of Growth among New World Economies: A Financial and Quantitative Analysis, 38(1): 1–36. View from Economic Historians of the United De Soto, Hernando. 1989. The Other Path: The States,” in How Latin America Fell Behind: Essays Invisible Revolution in the Third Worlds. New York: on the Economic Histories of Brazil and Mexico, Harper and Row. 1800–1914. S. Haber, ed. Stanford: Stanford De Soto, Hernando. 2000. The Mystery of Capital: University Press, 260–304. Why Capitalism Triumphs in the West and Fails Faccio, Mara. 2003. “Politically Connected Firms.” Everywhere Else. New York: Basic Books. Vanderbilt University Working Paper. Diamond, Douglas W. 1984. “Financial Intermediation Faccio, Mara, and Larry H. P. Lang. 2002. “The and Delegated Monitoring.” Review of Economic Ultimate Ownership of Western European sp05_Article 1 8/18/05 9:32 AM Page 717

Morck, Wolfenzon, and Yeung: Corporate Governance 717

Corporations.” Journal of Financial Economics, Chicago: University of Chicago Press. 65(3): 365–95. Hayek, Friedrich. 1960. The Constitution of Liberty. Federov, Oleg. 2000. “Case Studies on Abusive Self- London: Routledge and Kegan Paul. Dealing.” Paper presented at the OECD/World Hellman, Joel S., Geraint Jones, and Daniel Kaufmann. Bank Corporate Governance Roundtable for Russia, 2003. “Seize the State, Seize the Day: State Capture February 24–25, 2000. and Influence in Transition Economies.” Journal of Fields, L. Paige, and Eric L. Mais. 1994. “Managerial Comparative Economics, 31(4): 751–73. Voting Rights and Seasoned Public Equity Issues.” Henry, Peter Blair. 2000a. “Stock Market Liberalization, Journal of Financial and Quantitative Analysis, Economic Reform, and Emerging Market Equity 29(3): 445–57. Prices.” Journal of Finance, 55(2): 529–64. Filatotchev, Igor, Mike Wright, and Michael Bleaney. Henry, Peter Blair. 2000b. “Do Stock Market 1999. “Privatization, Insider Control and Managerial Liberalizations Cause Investment Booms?” Journal Entrenchment in Russia.” Economics of Transition, of Financial Economics, 58(1–2): 301–34. 7(2): 481–504. Högfeldt, Peter. 2004. “The History and Politics of Forbes, Kristin J., and Roberto Rigobon. 2002. “No Corporate Ownership in Sweden.” Stockholm Contagion, Only Interdependence: Measuring Stock School of Economics Working Paper. Market Comovements.” Journal of Finance, 57(5): Holmén, Martin, and Peter Högfeldt. 2005. “Pyramidal 2223–61. Discounts: Tunneling or Agency Costs?” Stockholm Friedman, Eric, Simon Johnson, and Todd Mitton. School of Economics Working Paper. 2003. “Propping and Tunneling.” Journal of Holderness, Clifford G., and Dennis P. Sheehan. 1988. Comparative Economics, 31(4): 732–50. “The Role of Majority Shareholders in Publicly Held Gertner, Robert H., David S. Scharfstein, and Jeremy Corporations: An Exploratory Analysis.” Journal of C. Stein. 1994. “Internal versus External Capital Financial Economics, 20(1–2): 317–46. Markets.” Quarterly Journal of Economics, 109(4): Holderness, Clifford G., Randall S. Kroszner, and 1211–30. Dennis P. Sheehan. 1999. “Were the Good Old Days Gigler, Frank, and Thomas Hemmer. 2002. That Good? Changes in Managerial Stock “Informational Costs and Benefits of Creating Ownership since the Great Depression.” Journal of Separately Identifiable Operating Segments.” Finance, 54(2): 435–69. Journal of Accounting and Economics, 33(1): 69–90. Hoshi, Takeo, Anil Kashyap, and David Scharfstein. Glaezer, Edward, and Andrei Shleifer. 2003. “The 1991. “Corporate Structure, Liquidity, and Curley Effect.” Harvard University Working Paper. Investment: Evidence from Japanese Industrial Glaezer, Edward, Jose Sheinkman, and Andrei Shleifer. Groups.” Quarterly Journal of Economics, 106(1): 2003. “The Injustice of Inequality.” Journal of 33–60. Monetary Economics, 50(1): 199–222. Jensen, Michael C., and William H. Meckling. 1976. Gompers, Paul A., Joy L. Ishii, and Andrew Metrick. “Theory of the Firm: Managerial Behavior, Agency 2003. “Corporate Governance and Equity Prices.” Costs and Ownership Structure.” Journal of Quarterly Journal of Economics, 118(1): 107–55. Financial Economics, 3(4): 305–60. Gompers, Paul A., and Josh Lerner. 1999. The Venture Joh, Sung Wook. 2003. “Corporate Governance and Capital Cycle. Cambridge: MIT Press. Firm Profitability: Evidence from Korea before the Gompers, Paul A., and Andrew Metrick. 2001. Economic Crisis.” Journal of Financial Economics, “Institutional Investors and Equity Prices.” 68(2): 287–322. Quarterly Journal of Economics, 116(1): 229–59. Johnson, Simon, and Todd Mitton. 2003. “Cronyism Gorton, Gary, and Andrew Winton. 2003. “Financial and Capital Controls: Evidence from Malaysia.” Intermediation,” in Handbook of the Economics of Journal of Financial Economics, 67(2): 351–82. Finance. George Constantinides, Milton Harris, and Johnson, Simon, Peter Boone, Alasdair Breach, and René Stulz, eds. Amsterdam: North Holland, Eric Friedman. 2000. “Corporate Governance in the 431–552. Asian Financial Crisis.” Journal of Financial Graham, Edward M. 2003. Reforming Korea’s Economics, 58(1–2): 141–86. Industrial Conglomerates. Washington, D.C.: Johnson, Simon, Rafael La Porta, Florencio Lopez-de- Institute for International Economics. Silanes, and Andrei Shleifer. 2000. “Tunneling.” Grossman, Sanford J., and Joseph E. Stiglitz. 1980. “On American Economic Review, 90(2): 22–27. the Impossibility of Informationally Efficient Johnson, W. Bruce, Robert P. Magee, Nandu J. Markets.” American Economic Review 70(3): Nagarajan, and Henry A. Newman. 1985. “An 393–408. Analysis of the Stock Price Reaction to Sudden Haber, Stephen. 2000. Political Institutions and Executive Deaths: Implications for the Managerial Economic Growth in Latin America: Essays in Labor Market.” Journal of Accounting and Policy, History, and Political Economy. Stanford: Economics, 7(1–3): 151–74. Hoover Institution Press. Kang, David C. 2002. “Bad Loans to Good Friends: Hass, Jeffrey J. 1996. “Directorial Fiduciary Duties in a Money Politics and the Developmental State in South Tracking Stock Equity Structure: The Need for a Korea.” International Organization, 56(1): 177–207. Duty of Fairness.” Michigan Law Review, 94(7): Karolyi, Andrew, and René Stulz. 2003. “Are Financial 2089–2177. Assets Priced Locally or Globally?” in Handbook of Hayek, Friedrich. 1941. The Pure Theory of Capital. the Economics of Finance. George Constantinides, sp05_Article 1 8/18/05 9:32 AM Page 718

718 Journal of Economic Literature, Vol. XLIII (September 2005)

Milton Harris, and René Stulz, eds. Amsterdam: La Porta, Rafael, Florencio Lopez-de-Silanes, and North Holland, 975–1020. Andrei Shleifer. 1999. “Corporate Ownership around Khanna, Tarun. 2000. “Business Groups and Social the World.” Journal of Finance 54 (2): 471–517. Welfare in Emerging Markets: Existing Evidence La Porta, Rafael, Florencio Lopez-de-Silanes, and and Unanswered Questions.” European Economic Guillermo Zamarripa. 2003. “Related Lending.” Review, 44(4–6): 748–61. Quarterly Journal of Economics, 118(1): 231–68. Khanna, Tarun, and Krishna Palepu. 1999a. “Policy Lease, Ronald C., John J. McConnell, and Wayne H. Shocks, Market Intermediaries, and Corporate Mikkelson. 1983. “The Market Value of Control in Strategy: The Evolution of Business Groups in Chile Publicly-Traded Corporations.” Journal of Financial and India.” Journal of Economics and Management Economics, 11(1–4): 439–71. Strategy, 8(2): 271–310. Lease, Ronald C., John J. McConnell, and Wayne H. Khanna, Tarun, and Krishna Palepu. 1999b. “The Mikkelson. 1984. “The Market Value of Differential Right Way to Restructure Conglomerates in Voting Rights in Closely Held Corporations.” Journal Emerging Markets.” Harvard Business Review, of Business, 57(4): 443–67. July–August 1999. Leff, Nathaniel. 1976. “Capital Markets in the Less Khanna, Tarun, and Krishna Palepu. 2000a. “Is Group Developed Countries: The Group Principle,” in Affiliation Profitable in Emerging Markets? An Money and Finance in Economic Growth and Analysis of Diversified Indian Business Groups.” Development. R. McKinnon, ed. New York: Marcel Journal of Finance, 55(2): 867–93. Dekker, 97–122. Khanna, Tarun, and Krishna Palepu. 2000b. “The Leff, Nathaniel H. 1978. “Industrial Organization and Future of Business Groups in Emerging Markets: Entrepreneurship in the Developing Countries: The Long Run Evidence from Chile.” Academy of Economic Groups.” Economic Development and Management Journal, 43(3): 268–85. Cultural Change, 26(4): 661–75. Khanna, Tarun, and Krishna Palepu. 2001. “Emerging Lemmon, Michael L., and Karl V. Lins. 2003. Market Business Groups, Foreign Intermediaries, “Ownership Structure, Corporate Governance, and and Corporate Governance,” in Concentrated Firm Value: Evidence from the East Asian Financial Corporate Ownership. Randall Morck, ed. Chicago Crisis.” Journal of Finance, 58(4): 1445–68. and London: University of Chicago Press, 265–94. Lenway, Stefanie, Randall Morck, and Bernard Yeung. Khanna, Tarun, and Krishna Palepu. Forthcoming. 1996. “Rent Seeking, Protectionism and Innovation “The Evolution of Concentrated Ownership in India: in the American Steel Industry.” Economic Journal, Broad Patterns and a History of the Indian Software 106(435): 410–21. Industry,” in The History of Corporate Governance Leuz, Christian, and Felix Oberholzer-Gee. 2003. around the World: Family Business Groups to “Political Relationships, Global Financing and Professional Managers. Randall Morck, ed. Chicago: Corporate Transparency.” Working Paper. University of Chicago Press. Levine, Ross. 1997. “Financial Development and Khanna, Tarun, and Jan W. Rivkin. 2001. “Estimating Economic Growth: Views and Agenda.” Journal of the Performance Effects of Business Groups in Economic Literature, 35(2): 688–726. Emerging Markets.” Strategic Management Journal, Levy, Haim. 1983. “Economic Evaluation of Voting 22(1): 45–74. Power of Common Stock.” Journal of Finance, 38(1): Kim, Se-Jik. 2004. “Bailout and Conglomeration.” 79–93. Journal of Financial Economics, 71(2): 315–47. Lewellen, Wilbur G. 1971. “A Pure Financial Rationale King, Robert G., and Ross Levine. 1993. “Finance, for the Conglomerate Merger.” Journal of Finance, Entrepreneurship, and Growth: Theory and 26(2): 521–37. Evidence.” Journal of Monetary Economics, 32(3): Li, Kan, Randall Morck, Fan Yang, and Bernard Yeung. 513–42. 2004. “Firm-Specific Variation and Openness in Krueger, Anne O. 1993. “Virtuous and Vicious Circles Emerging Markets.” Review of Economics and in Economic Development.” American Economic Statistics, 86(3): 658–69. Review, 83(2): 351–55. Lins, Karl V. 2003. “Equity Ownership and Firm Value Landes, David. 1949. “French Entrepreneurship and in Emerging Markets.” Journal of Financial and Industrial Growth in the Nineteenth Century.” Quantitative Analysis, 38(1): 159–84. Journal of Economic History, 9: 45–61. Lombardo, Davide, and Marco Pagano. 2002. “Law La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei and Equity Markets: A Simple Model,” in Corporate Shleifer, and Robert Vishny. 1997a. “Legal Governance Regimes: Convergence and Diversity. Determinants of External Finance.” Journal of Joe McCahery, Pieter Moerland, Theo Raaijmakesr Finance, 52(3): 1131–50. and Luc Renneboog, eds. Oxford: Oxford University La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Press, 343–62. Shleifer, and Robert Vishny. 1997b. “Trust in Large Lopez-de-Silanes, Florencio. 1997. “Determinants of Organizations.” American Economic Review, 87(2): Privatization Prices.” Quarterly Journal of 333–38. Economics, 112(4): 965–1025. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Mankiw, N. Gregory. 1995. “The Growth of Nations.” Shleifer, and Robert Vishny. 1998. “Law and Brookings Papers on Economic Activity, 1: 275–310. Finance.” Journal of Political Economy, 106(6): Modigliani, Franco, and Enrico Perotti. 2000. 1113–55. “Security Markets versus Bank Finance: Legal sp05_Article 1 8/18/05 9:32 AM Page 719

Morck, Wolfenzon, and Yeung: Corporate Governance 719

Enforcement and Investor Protection.” June 21–22, 2004, Alberta, Canada. International Review of Finance, 1(2): 81–96. Murphy, Kevin M., Andrei Shleifer, and Robert W. Megginson, William L., and Jeffry M. Netter. 2001. Vishny. 1991. “The Allocation of Talent: Implications “From State to Market: A Survey of Empirical for Growth.” Quarterly Journal of Economics, Studies on Privatization.” Journal of Economic 106(2): 503–30. Literature, 39(2): 321–89. Murphy, Kevin M., Andrei Shleifer, and Robert W. Mikkelson, Wayne H., and Richard S. Ruback. 1985. Vishny. 1993. “Why Is Rent-Seeking So Costly to “An Empirical Analysis of the Interfirm Equity Growth?” American Economic Review, 83(2): Investment Process.” Journal of Financial 409–14. Economics, 14(4): 523–53. Nenova, Tatiana. 2003. “The Value of Corporate Voting Mitton, Todd. 2002. “A Cross-Firm Analysis of the Rights and Control: A Cross-Country Analysis.” Impact of Corporate Governance on the East Asian Journal of Financial Economics, 68(3): 325–51. Financial Crisis.” Journal of Financial Economics, North, Douglass C. 1991. “Institutions.” Journal of 64(2): 215–41. Economic Perspectives, 5(1): 97–112. Morck, Randall. 1995. “The Economics of Olson, Mancur. 1963. “Rapid Growth as a Destabilizing Concentrated Ownership.” Canadian Business Law Force.” Journal of Economic History, 23(4): 529–52. Journal, 26(1): 63–85. Olson, Mancur. 1982. The Rise and Decline of Nations. Morck, Randall. 1996. “Revamping C.E.O. Stock New Haven: Yale University Press. Options,” in Value for Money. Edward Iacobucci, ed. Olson, Mancur. 2000. Power and Prosperity: Toronto: C. D. Howe Institute. Outgrowing Communist and Capitalist Morck, Randall. 1997. “The Corporate Governance Dictatorships. New York: Basic Books. Consequences of the Intercorporate Dividend Tax Pagano, Marco, and Paolo Volpin. Forthcoming. “The Exemption.” Corporate Governance Review, 6(9): Political Economy of Corporate Governance.” 9–12. American Economic Review. Morck, Randall, ed. Forthcoming. History of Corporate Perez-Gonzalez, Francisco. 2002. “Does Inherit Ownership. Chicago: University of Chicago Press. Control Hurt Firm Performance?” Columbia Morck, Randall, and Masao Nakamura. 1999. “Banks University Working Paper. and Corporate Control in Japan.” Journal of Finance, Perotti, Enrico, and Paolo Volpin. 2004. “Lobbying on 54(1): 319–39. Entry.” Paper presented at the World Bank–Stern Morck, Randall, and Masao Nakamura. 2004. “Been Entry Workshop, January 2005. There, Done That: A History of Corporate Perotti, Enrico C., and Ernst-Ludwig von Thadden. Ownership in Japan.” Paper presented at the NBER Forthcoming. “The Political Economy of Corporate 2003 Corporate Governance History Conference, Control.” Journal of Political Economy. Banff, Canada. Petersen, Mitchell A., and Raghuram G. Rajan. 1994. Morck, Randall, Masao Nakamura, and Anil Shivdasani. “The Benefits of Lending Relationships: Evidence 2000. “Banks, Ownership Structure, and Firm Value from Small Business Data.” Journal of Finance, in Japan.” Journal of Business, 73(4): 539–67. 49(1): 3–37. Morck, Randall, Andrei Shleifer, and Robert W. Vishny. Prowse, Stephen D. 1992. “The Structure of Corporate 1988. “Management Ownership and Market Ownership in Japan.” Journal of Finance, 47(3): Valuation: An Empirical Analysis.” Journal of 1121–40. Financial Economics, 20(1–2): 293–315. Rajan, Raghuram G. 1992. “Insiders and Outsiders: Morck, Randall, Andrei Shleifer, and Robert W. Vishny. The Choice between Informed and Arm’s-Length 1989. “Alternative Mechanisms for Corporate Debt.” Journal of Finance, 47(4): 1367–400. Control.” American Economic Review, 79(4): 842–52. Rajan, Raghuram G., and Luigi Zingales. 1998. Morck, Randall, Kathy He, and Bernard Yeung. 2000. “Financial Dependence and Growth.” American “Stability.” Unpublished Manuscript. Economic Review, 88(3): 559–86. Morck, Randall, David Stangeland, and Bernard Yeung. Rajan, Raghuram G., and Luigi Zingales. 2003. “The 2000. “Inherited Wealth, Corporate Control, and Great Reversals: The Politics of Financial Economic Growth: The Canadian Disease,” in Development in the Twentieth Century.” Journal of Concentrated Corporate Ownership. Randall Morck, Financial Economics, 69(1): 5–50. ed. Chicago: University of Chicago Press, 319–69. Roe, Mark J. 2001. “Rents and their Corporate Law Morck, Randall, and Bernard Yeung. 2004. “Family Consequences.” Columbia Law School Discussion Control and the Rent-Seeking Society.” Paper. Entrepreneurship: Theory and Practice, 28(4): Roe, Mark J. 2003. Political Determinants of Corporate 391–409. Governance: Political Context, Corporate Impact. Morck, Randall, Bernard Yeung, and Wayne Yu. 2000. Oxford and New York: Oxford University Press. “The Information Content of Stock Markets: Why Roll, Richard. 1988. “R2.” Journal of Finance, 43(3): Do Emerging Markets Have Synchronous Stock 541–66. Price Movements?” Journal of Financial Economics, Rydqvist, Kristian. 1996. “Takeover Bids and the 58(1–2): 215–60. Relative Prices of Shares That Differ in Their Voting Murphy, Antoin. 2004. “The History of Corporate Rights.” Journal of Banking and Finance, 20(8): Ownership in France.” Paper presented at the 1407–25. NBER History of Corporate Ownership Conference, Sachs, Jeffrey, and Andrew Warner. 1995. “Economic sp05_Article 1 8/18/05 9:32 AM Page 720

720 Journal of Economic Literature, Vol. XLIII (September 2005)

Reform and the Process of Global Integration.” Hees-Edper Corporate Group,” in Corporate Brookings Papers on Economic Activity, 1: 1–118. Decision-Making in Canada. R. J. Daniels and R. Schumpeter, Joseph A. 1912. Theorie der Wirtschaftlichen Morck, eds. Calgary: University of Calgary Press, Entwichlung. Leipzig: Dunker and Humbolt. 223–40. Translated by R. Opie as The Theory of Economic Stein, Jeremy C. 1997. “Internal Capital Markets and Development: An Inquiry into Profits, Capital, the Competition for Corporate Resources.” Journal Credit, Interest, and the Business Cycle (Cambridge: of Finance, 52(1): 111–33. Harvard University Press, 1934). Stigler, George. 1971. “The Theory of Economic Schumpeter, Joseph A. 1942. Capitalism, Socialism, Regulation.” Bell Journal of Economics and and Democracy. New York: Harper and Brothers. Management Science, 2: 3–21. Shleifer, Andrei. 1997. “Government in Transition.” Stulz, René. 1988. “On Takeover Resistance, European Economic Review, 41(3–5): 385–410. Managerial Discretion and Shareholder Wealth.” Shleifer, Andrei, and Robert W. Vishny. 1986. “Large Journal of Financial Economics, 20(1–2): 25–54. Shareholders and Corporate Control.” Journal of Stulz, René. 1999. “Globalization, Corporate Finance, Political Economy, 94(3): 461–88. and the Cost of Capital.” Journal of Applied Shleifer, Andrei, and Robert W. Vishny. 1993. Corporate Finance, 12(3): 8–25. “Corruption.” Quarterly Journal of Economics, Stulz, René. Forthcoming. “The Limits of Financial 108(3): 599–617. Globalization.” Journal of Finance. Shleifer, Andrei, and Robert W. Vishny. 1994. Stulz, René, and Rohan Williamson. 2003. “Culture, “Politicians and Firms.” Quarterly Journal of Openness, and Finance.” Journal of Financial Economics, 109(4): 995–1025. Economics, 70(3): 313–49. Shleifer, Andrei, and Robert W. Vishny. 1997. “A Tobin, James. 1982. “On the Efficiency of the Financial Survey of Corporate Governance.” Journal of System.” Lloyd’s Banking Review, 153: 1–15. Finance, 52(2): 737–83. Ugurlu, Mine. 2000. “Agency Costs and Corporate Shleifer, Andrei, and Daniel Wolfenzon. 2002. Control Devices in the Turkish Manufacturing “Investor Protection and Equity Markets.” Journal of Industry.” Journal of Economic Studies, 27(6): Financial Economics, 66(1): 3–27. 566–99. Siegel, Jordan. 2004. “Is Political Connectedness a Veblen, Thorstein. 1899. The Theory of the Leisure Paramount Investment after Liberalization? The Class: An Economic Study in the Evolution of Successful Leveraging of Contingent Social Capital Institutions. London: Macmillan and Company. and the Formation of Cross-Border Strategic Veblen, Thorstein. 1904. The Theory of Business Alliances Involving Korean Firms and their Global Enterprise. New York: Charles Scribner’s Sons. Partners (1987–2000).” Harvard Business School Veblen, Thorstein. 1923. Absentee Ownership and Working Paper. Business Enterprise in Recent Times. New York: A. Siegel, Jordan. 2005. “Can Foreign Firms Bond M. Kelley. Themselves Effectively by Renting U.S. Securities Volpin, Paolo F. 2002. “Governance with Poor Investor Laws?” Journal of Financial Economics, 75(2): Protection: Evidence from Top Executive Turnover 319–59. in Italy.” Journal of Financial Economics, 64(1): Slovin, Myron B., and Marie E. Sushka. 1993. 61–90. “Ownership Concentration, Corporate Control Weber, Max. 1958. The Protestant Ethic and the Spirit Activity, and Firm Value: Evidence from the Death of Capitalism. New York: Scribner’s Press. of Inside Blockholders.” Journal of Finance, 48(4): Williamson, Oliver. 1973. “Markets and Hierarchies: 1293–1321. Some Elementary Considerations.” American Smith, Brian F., and Ben Amoako-Adu. 1999. Economic Review, 63(2): 316–25. “Management Succession and Financial Performance Wurgler, Jeffrey. 2000. “Financial Markets and the of Family Controlled Firms.” Journal of Corporate Allocation of Capital.” Journal of Financial Finance, 5(4): 341–68. Economics, 58(1–2): 187–214. Sokoloff, Kenneth L., and Stanley L. Engerman. 2000. Zingales, Luigi. 1994. “The Value of the Voting Right: “Institutions, Factor Endowments, and Paths of A Study of the Milan Stock Exchange Experience.” Development in the New World.” Journal of Review of Financial Studies, 7(1): 125–48. Economic Perspectives, 14(3): 217–32. Zingales, Luigi. 1995a. “Insider Ownership and the Sonin, Konstantin. 2003. “Why the Rich May Favor Decision to Go Public.” Review of Economic Studies, Poor Protection of Property Rights.” Journal of 62(3): 425–48. Comparative Economics, 31(4): 715–31. Zingales, Luigi. 1995b. “What Determines the Value of Stangeland, David, Ronald J. Daniels, and Randall Corporate Votes?” Quarterly Journal of Economics, Morck. 1995. “In High Gear: A Case Study of the 110 (4): 1047–73.