Journal of International Business Studies (2010) 41, 88–104

& 2010 Academy of International Business All rights reserved 0047-2506 www.jibs.net

Pyramiding vs leverage in corporate groups: International evidence Mara Faccio1, Larry HP Lang2 and Leslie Young3 1

Krannert School of Management, Purdue University, West Lafayette, Indiana, USA; 2The Chinese University of Hong Kong, Hong Kong, China; 3The Asia Pacific Institute of Business, The Chinese University of Hong Kong, Hong Kong, China Correspondence: M Faccio, Krannert School of Management, Purdue University, 403 W. State Street, West Lafayette, IN 47907-2076, USA. Tel: þ 1 765 496 1951; Fax: þ 1 765 494 9658; E-mail: [email protected]

Received: 26 November 2005 Revised: 28 January 2009 Accepted: 21 February 2009 Online publication date: 20 August 2009

Abstract Among firms listed in Western Europe and East Asia, when creditor protection is strong, the controlling shareholder trades off retaining control in bad states through pyramiding against retaining the upside in good states via leverage. This result might arise because the controlling shareholder uses both leverage and pyramiding to expand control of resources, but in different circumstances since they have different outcomes under downside shocks. When creditor protection is weak, the controlling shareholder no longer prefers pyramiding in bad states, because creditors will not be able to seize the firm in bad states. Therefore pyramiding and leverage are used together. Journal of International Business Studies (2010) 41, 88–104. doi:10.1057/jibs.2009.33 Keywords: finance; corporate governance; capital structure

INTRODUCTION Research on corporate capital structure often focuses on the leverage ratio. Research on the structure of corporate ownership and control often focuses on the controlling shareholder’s ratio of ownership to control rights in a firm that s/he controls indirectly through a corporate pyramid. In this paper we link these two research areas by testing the connection between the two key ratios. The general reason to hypothesize a connection is that the corporate governance literature ties each ratio to corporate governance. Denis and McConnell (2003), John and Senbet (1998) and Shleifer and Vishny (1997) survey this literature, and the connection between leverage and governance indicators. The corporate governance literature suggests two contrasting predictions, depending on the postulated role of debt in corporate governance. Jensen and Meckling (1976) and Jensen (1986, 1989) argue that debt constrains managerial expropriation by imposing fixed obligations on corporate cash flow in situations in which default on debt would deprive managers of control and related benefits. In the same vein, Easterbrook (1984) argues that debt forces managers to be accountable to the external capital market. Lang, Ofek, and Stulz (1996) show that debt curtails investment by firms with poor-quality opportunities, and that leverage increases when growth opportunities are smaller (see also Kim & Sorensen, 1986; Titman & Wessels, 1988). Maloney, McCormick, and Mitchell (1993) find that leverage improves managerial decisionmaking on key issues such as acquisitions. If a greater separation of

Pyramiding vs leverage in corporate groups

Mara Faccio et al

89

ownership from control encourages expropriation by the controlling shareholder of a firm in a pyramid, then, to relax the constraint that debt imposes on his expropriation, the manager would choose lower leverage. Thus the Jensen–Meckling analysis of the role of debt in corporate governance predicts that a lower ratio of ownership to control rights would be associated with lower leverage. Stulz (1988) and Ellul (2008) suggest a contrasting prediction. These authors argue that higher leverage allows the controlling shareholder to control more resources without diluting his or her control over the corporation. A greater separation of ownership from control through pyramiding also achieves this result, but the two methods share risk differently. A controlling shareholder who uses debt will not share upside shocks with creditors, but risks losing control of the corporation under severe downside shocks. If s/he uses pyramiding, the controlling shareholder will share both upside and downside shocks with the minority shareholders. Given this difference, the two methods should be substitutes: that is, greater leverage should be associated with less pyramiding. We test this hypothesis in the five largest European economies and nine East Asian economies, using all listed corporations with credible accounting data. We find supporting evidence when creditor protection is good. When creditor protection is poor, a controlling shareholder might be able to shift downside risk onto the debt-holders by delaying or repudiating payments. Furthermore, as Harvey, Lins, and Roper (2004) note: ‘‘In emerging markets, a domestically issued short-maturity debt contract is less likely to discourage overinvestment, because family groups or governments typically control the banks and can use them for their own purposes.’’ Under these circumstances, debt and pyramiding would have similar consequences for the controlling shareholder under downside shocks. Given this similarity, the two methods should be complements that are favored in similar circumstances: that is, higher leverage should be associated with more pyramiding. We find evidence consistent with this hypothesis for firms that can access loans from financial institutions in the same corporate group, but not for firms that lack such access. Presumably, poorly protected creditors avoid lending that would expose them to the downside risk of nonrelated parties. Our analysis leads us to ask whether the controlling shareholder seeks to control more resources to

expropriate those resources from debtors and minority shareholders or to exploit new investment opportunities.1 We provide evidence that the first motivation drives leverage in our sample, but we do not claim to have resolved the matter. Instead, we recognize that each motivation might operate at different times; indeed, both might operate at the same time.

HYPOTHESIS DEVELOPMENT The introduction argued that the relationship between leverage and pyramiding depends on the role postulated for debt in corporate governance, and also on the quality of creditor protection. To discriminate among these postulates, while testing for the impact of creditor protection, we regress leverage vi on (O/C)i, the ratio of ownership to control rights, Creditor Rightsi, and the product of the latter two terms (to capture any interaction between them). Thus our regression equation is

 O O þa2
Creditor Rightsi y i ¼ a0 þ a1
C i C i ð1Þ N X þ a3
Creditor Rightsi þ an
xni þ ei n¼4

In Eq. (1) the variables xni control for other determinants of leverage that might otherwise induce spurious correlations. When creditors have few rights, they will surely refrain from lending that would expose them to their debtors’ downside risk2 – unless creditor and debtor belong to the same corporate group. To mitigate expropriation via related party loans, a number of countries impose severe formal restrictions on such loans. Indonesia, for example, prohibits a bank from lending more than 10% of its capital to an affiliated company. However, the restrictions are hardly enforced: for example, in 1995 Indonesian regulators found that Bank Anrico was lending 1925% of its net capital to affiliated companies. In 1998, 76% of the loans of the Bank Dagang Nasional Indonesia were made to affiliated companies. Unsurprisingly, 96% of this bank’s loans were found to be non-performing shortly thereafter. Also in 1998, after the collapse of Orient Bank (Philippines), regulators found that 75% of the bank’s loans were to directors, the controlling shareholder, and its associates; furthermore, 60% of loans were unsecured. Akerlof and Romer (1993) and Laeven (2001) show that related party loans within Russian and Chilean business

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

90

groups facilitated the looting of banks. In contrast, Krozner and Strahan (2001) argue that board linkages between banks and non-financial firms did not systematically lead to conflicts of interests or to a misuse of funds in the US, which has strong creditor protection. Since access to related party loans can affect the outcome, we perform separate regressions for firms with and without such access. In each case we use two alternative measures of leverage, and we correct standard errors for heteroskedasticity and for the clustering of firms within industries in each country (see Table 2). We then repeat the regression for firms with access to related party lending, but divide them according to whether creditor rights are strong or weak in their jurisdiction (see Table 3). Finally, we reaffirm our conclusions using a generalized method of moments (GMM) instrumental variables estimator for the O/C ratio to neutralize any endogeneity bias that might arise because controlling shareholders choose this ratio jointly with leverage (see Tables 4 and 5).

Other Determinants of Leverage Previous papers identify important determinants of leverage. These determinants, which we use as control variables in our regressions, include growth opportunities, firm size, asset tangibility, diversification, operating risk, bankruptcy risk, and taxes. Growth opportunities. The empirical evidence in the US shows that corporations with a high Q tend to have low leverage. Rajan and Zingales (1995) report a negative relationship between leverage and the market-to-book ratio for a sample of large corporations in the US, Germany, France, the United Kingdom, and Canada.3 Tobin’s Q is often interpreted as a proxy for a corporation’s growth opportunities.4 Titman (1984), Bradley, Jarrell, and Kim (1984), Titman and Wessels (1988), Maksimovic and Titman (1991), and Frank and Goyal (2003), among others, find a negative relationship between leverage and other proxies for growth opportunities, such as the human capital of its employees, the brand image of its products, or other intangible assets that cannot be accepted as collateral by prudent lenders. This negative relationship is consistent with Myers’ (1977) analysis of debt overhang as a constraint on a corporation’s willingness to undertake positive NPV projects financed by stockholders because this

Journal of International Business Studies

would benefit bondholders. Higher-growth corporations might exhibit lower leverage because they face higher costs of financial distress (Fama & French, 1992).5 We control separately for this latter risk.

Firm size. Rajan and Zingales (1995) argue that size could proxy for the probability of default, which is higher for smaller firms. On the other hand, larger, more visible firms suffer less from informational asymmetry, have easier access to equity markets, and therefore should be less levered. Studies by Frank and Goyal (2003), Hoshi, Scharfstein, and Kashyap (1990), Kester (1986), Kim and Sorensen (1986), and Rajan and Zingales (1995) provide mixed evidence on the relation between leverage and size. Tangibility. Rajan and Zingales (1995) argue that fixed assets, which are easier to collateralize, reduce the agency costs of debt. However, Berger and Udell (1995) argue that this relationship would be weaker in relationship-oriented economies. Myers (1977) suggests that the debt-overhang problem would be less for corporations with tangible assets, which could imply a positive association between leverage and tangible assets. Harris and Raviv (1990) argue that corporations with more tangible assets have a higher liquidation value, which increases the usefulness of information to stockholders. Since debt provides them with information (e.g., on the corporation’s ability to service debt), they require higher leverage in corporations with more tangible assets. Diversification. As we know, diversification can affect leverage in at least two ways. First, through diversification, a corporation can reduce its firmspecific risk, indicating higher leverage. Second, diversified corporations might be able to access internal capital markets, indicating lower consolidated leverage. In line with the first prediction, most of the empirical evidence (e.g., Ahn, Denis, & Denis, 2006; Comment & Jarrell, 1995; Lewellen, 1971) shows that diversified firms tend to have higher leverage. Operating and bankruptcy risks. In line with Myers (1977), leverage has been found to decrease with operating risk (Bancel & Mittoo, 2004; Kim & Sorensen, 1986) and return volatility (Bradley et al., 1984). Additionally, in line with the predictions of the trade-off theory of capital

Pyramiding vs leverage in corporate groups

Mara Faccio et al

91

structure, Harris and Raviv (1990) find that leverage is negatively correlated with the interest coverage ratio and the probability of reorganization following default. Ross (1977) and Harris and Raviv (1990), among others, find that leverage is positively related to the probability of default. Friedman, Johnson, and Mitton (2003) argue that entrepreneurs can use debt to prop up their corporations when they get into financial distress, thereby retaining the option for future expropriation (‘‘tunneling’’). Our analysis will control for financial distress in the form of bankruptcy risk. Doing so allows us to focus on other ways in which debt might facilitate expropriation, which might be captured through the impact of the O/C ratio on leverage.

Taxes. Most countries’ corporate tax systems favor debt financing, in that interest on corporate debt is tax deductible, so debt financing should be more pronounced in countries with higher marginal corporate tax rates. However, if a corporation has a low or negative income before interest and taxes, then the tax benefit of debt also depends on whether the tax code allows carrybacks and carryforwards (Graham, 2003). The predicted correlation between taxes and leverage might also be undermined if a firm can exploit investment-related tax shields in place of interest deductions, or if investment and capital structure decisions are endogenous (Dammon & Senbet, 1988).

DATA AND DEFINITIONS Corporate Data We obtain our data from the Worldscope database. We start from the 1996 accounting data of all corporations listed for the five largest West European economies (France, Germany, Italy, Spain, and the UK) and nine East Asian economies (Hong Kong, Indonesia, Japan, Malaysia, Philippines, Singapore, South Korea, Taiwan, and Thailand). We eliminate corporations that report data that are not credible (i.e., negative debt or negative sales), and corporations that miss data on short-term debt, long-term debt, book or market value of equity, total assets, sales, earnings, or income taxes. We also eliminate corporations whose main or secondary two-digit SIC is in the finance industry (SIC: 60–69), because their leverage ratios do not bear on agency issues. Consolidation forces the assets and liabilities of

each subsidiary to be recognized in the accounts of the parent corporation. This consolidation can significantly affect our measures of leverage in some countries. To ensure consistency in the reporting of debt we eliminate all 435 corporations that report unconsolidated accounts, as well as 81 corporations that provided no information about whether or not their accounts were consolidated. We note that by the end of our sample period most companies had moved to consolidated accounts, but this was true for only 63.6% of companies in Korea, 78% in Germany, and 80.9% in Japan. Thus we have screened out a significant number of firms from these countries, thereby understating the role of some large groups, such as the Korean chaebols. The consolidated accounts of a parent corporation recognize the assets and liabilities of the subsidiaries that they ‘‘control’’, as defined in the accounting rules of their host country. This accounting definition is typically much more restrictive than that of the United States. For example, the European Union Directive 7/83 requires a parent corporation to produce consolidated accounts if it holds a majority of the subsidiary’s voting rights, or controls the majority of its board. Therefore corporation A might control corporation B, in our sense of holding at least 20% of B’s voting rights, yet A would not control B in the accounting sense. Thus corporations A and B would not consolidate their accounts. Conversely, corporation A could control an unlisted corporation B in the accounting sense: therefore they would consolidate their accounts. However, on that basis, A and B would not be affiliated with a group according to our definition, which requires that a group include at least two listed firms. Thus affiliation to the same group in our sense is neither necessary nor sufficient for two firms to consolidate their accounts. Our analysis incorporates the debt between two listed corporations affiliated with the same group, provided that neither is controlled by the other in the accounting sense: such debt is relevant to the agency issues addressed in this paper. Our analysis ignores debt between a listed corporation and the unlisted subsidiaries that it controls in the accounting sense, which is eliminated by consolidation. Such debt is not relevant to agency issues, since in view of its transparency in the consolidated accounts it is hardly likely to constrain the management of the parent corporation, or to facilitate expropriation.

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

92

Our analysis excludes the unlisted subsidiaries of corporations that report consolidated accounts: these subsidiaries usually have a few block shareholders, and thus are not exposed to the agency problems that are our focus. Non-financial companies do not consolidate accounts with financial firms, so our leverage measures include loans from group banks and financial companies. Our analysis is based on the 3448 non-financial corporations that have consolidated accounts and have ownership data available. We obtain the 1996–1997 ownership and group affiliation data on these corporations from Worldscope, national stock exchanges, national company handbooks, and the other sources listed in Appendices A and B. To identify all the ultimate owners of each corporation that own at least 5% of its shares we trace back the network of indirect ownership through other corporations. We obtain these ownership and control data from Claessens, Djankov, and Lang (2000) for East Asia, and from Faccio and Lang (2002) for Western Europe.

Definitions of Leverage and Creditor Protection We define debt as the sum of long-term and shortterm financial debt. This excludes non-financial liabilities, such as accounts payable, provisions for pensions, deferred taxes, and other provisions for future liabilities. We use book values rather than market values, which already reflect market expectations. We use two alternative measures of leverage: the debt/total asset ratio (D/TA), where total assets include debt, non-financial liabilities, and shareholder equity; and the debt/(debt þ equity) ratio (D/(D þ E)), in which the denominator includes debt and shareholder equity, but excludes all non-financial liabilities. To measure creditor protection we use the index developed in La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1998), which aggregates four categories of creditor rights. Definitions of Ownership and Control Dispersed shareholders have difficulty concerting their actions, so the largest shareholder can control a corporation if it holds enough voting shares. For each corporation in our sample, we identify the ‘‘controlling shareholder’’, if present, as the largest shareholder holding at least 10% of control rights. This standard of control is used in earlier studies, such as La Porta, Lopez-de-Silanes, and Shleifer (1999) and Claessens et al., (2000). If the controlling

Journal of International Business Studies

shareholder is a corporation or financial institution, then we identify its owners, its owners’ owners, etc. If the controlling shareholder is an unlisted company, then we consider the corporation to be family controlled. The controlling shareholder of a corporate group can gain control rights in a corporation Z in excess of its ownership rights by pyramiding, that is, owning Z indirectly through other corporations. If the controlling shareholder owns a fraction x of the shares of corporation X, which owns a fraction y of the shares in corporation Y, which owns a fraction z of the shares in Z, then through this ownership chain the controlling shareholder owns a fraction xyz of the shares of Z. However, the controlling shareholder’s share of the control rights of Z through this control chain can be measured by its weakest link, that is, the minimum of x, y, and z. Let O be the controlling shareholder’s share of the ownership rights in a corporation, and let C be its share of the control rights, aggregated over all control chains. The O/C ratio indicates the extent to which the controlling shareholder has pyramided his ownership stake to secure more control rights over the resources contributed by minority shareholders. This O/C ratio will generally be lower for corporations further down the pyramid. Thus a lower O/C ratio indicates greater pyramiding. A controlling shareholder could expropriate minority shareholders by intragroup transactions, for example, by directing a firm in which s/he has small ownership rights to buy an overpriced asset from a firm in which s/he has large ownership rights. We define a corporation to be ‘‘group-affiliated’’ if it meets one of the following four criteria: it is controlled by a shareholder through pyramiding, that is, indirectly through a chain of other corporations in the sample; it controls another corporation in the sample; it has the same controlling shareholder as at least one other corporation in the sample; or its controlling shareholder is a corporation or financial institution that is ‘‘widely held’’ in that no shareholder holds 10% or more of the voting rights. Our proxy for access to related party loans is affiliation to a group that also controls some financial institution. Financial institutions include banks, insurance companies, mutual funds, private pension funds, merchant banks, and venture capitalists. (We note that we obtain similar results if we confine financial institutions to banks.)

Pyramiding vs leverage in corporate groups

Mara Faccio et al

93

Definitions of Control Variables We control for growth opportunities using Tobin’s Q, that is, the ratio of market value of equity plus book value of debt to the sum of book value of equity plus book value of debt. We measure firm size by the logarithm of the corporation’s total assets, ln(TA), and asset tangibility by the ratio of fixed to total assets, Tangib. We follow Lang and Stulz (1994) in measuring diversification by the number of different two-digit SIC industries in which the firm operates, NoSic. We control for volatility using asset betas, bA. We first define a corporation’s equity beta as sI rI;M ð2Þ bS ¼ sM in which sI is the standard deviation of its stock return, rI,M is the correlation coefficient between its stock return and the return on the market index (see below), and sM is the standard deviation of the market return. We compute standard deviations and correlation coefficients using the monthly stock returns over the period from January 1994 to December 1996; for corporations that went public through 1994, the period is from January 1995 to December 1996. We assume that the beta of debt equals zero, and compute the asset beta from the relation bA ¼

SbS Bð 1  t c Þ þ S

ð3Þ

in which S is the market value of equity, B is the book value of debt, and tc is the corporation’s tax rate. We compute the tax rate by dividing the corporation’s taxes by pretax income. To control for bankruptcy risk we rank corporations in ascending order of the ratio of earnings before interest, taxes, and depreciation (EBITDA) to interest expense. BankrDec assigns corporations to their decile in this ranking. Corporations in the first decile, which have the lowest ratio of EBITDA to interest expense, face the most difficulty in meeting interest payments. This variable also indirectly accounts for profitability, since higher values of the EBITDA to interest expense ratio imply higher profitability. Thus we do not control further for profitability. Finally, we control for the impact of taxes on leverage by including each firm’s effective tax rate (Tax), measured by dividing each firm’s income taxes by its pretax income. We constrain this variable to range between zero (for firms with a negative pretax income) and the highest marginal

corporate tax rate in the country of each given firm. Given the complexities of tax systems discussed by Dammon and Senbet (1988) and Graham (2003), this variable might fail to control fully for the cross-country impact of taxation of leverage. However, a more detailed examination of taxes is beyond the scope of this paper.

RESULTS AND INTERPRETATIONS Results Table 1 summarizes the regression variables, broken down by country. These include the number of corporations in our sample, creditor rights, the percentage of corporations that can access related party loans, leverage, and Tobin’s Q. Table 1 also reports the results of tests for differences between the mean leverage for various subsets of corporations. Corporations that can borrow from related parties have significantly higher leverage than those that cannot; corporations in economies with weak creditor rights (index o 3) have significantly higher leverage than those with strong creditor rights (index X 3); corporations in which the ownership and control rights of the controlling shareholder are identical (O/C¼1) have significantly lower leverage than corporations where O/Co 1. Because there is a significant difference between the leverage of corporations that can or cannot access related party loans, we perform separate regressions for corporations in these two categories. The observations for a given industry or country may be affected by factors special to that country, such as accounting conventions, taxes, and political risk (Burgman, 1996; Graham, 2003), so we adjust the standard errors for clustering at the country/industry level and for heteroskedasticity. We report adjusted p-values in parentheses below our coefficient estimates. Table 2, Panel A, presents our results for corporations that can access related party loans. Leverage is significantly higher for corporations with a lower O/C ratio, and for those that are domiciled where creditor protection is weaker. The significant positive coefficient on the product of the O/C ratio and the index of creditor protection indicates that, in economies with better creditor protection, a higher O/C ratio has a more positive (or less negative) impact on leverage. We find that, among the other control variables in the regression, leverage is positively related to size and negatively

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

94

Table 1

Country

Access to related party lending, ownership/control and mean leverage ratios by country

Number of corporations

Panel A: Summary statistics France 372 Germany 309 HK 212 Indonesia 81 Italy 96 Japan 832 Malaysia 149 Philippines 36 Singapore 145 South Korea 138 Spain 82 Taiwan 83 Thailand 70 The UK 843 All 3448

Creditor rights

% corps in a group that controls some financial institution

O/C

D/TA

D/(D+E)

Q

0 3 4 4 2 2 4 0 4 3 2 2 3 4 2.74

11.56 13.59 21.70 40.74 28.13 64.42 23.49 36.11 3.45 38.41 25.61 0.00 12.86 43.42 35.64

0.941 0.835 0.882 0.789 0.720 0.596 0.844 0.873 0.794 0.908 0.920 0.851 0.939 0.833 0.794

22.23 23.00 24.49 35.32 22.82 33.12 24.55 23.64 22.52 52.30 18.98 25.06 40.58 17.57 25.95

36.68 41.75 30.63 41.45 37.68 47.58 32.23 27.91 27.79 69.43 28.77 29.17 47.47 28.17 37.99

1.61 1.90 1.43 1.40 1.19 1.38 2.91 1.73 1.77 0.99 1.67 2.05 1.35 2.22 1.74

6.68*** 2.03*** 2.07**

0.46*** 0.17*** 4.10***

Panel B: t-statistics for differences between means Strong (43) vs weak (o3) creditor rights Corporations affiliated vs non-affiliated to a group that controls some financial institution O/C¼1 vs O/Co1 corporations

4.20*** 3.46*** 1.05

Notes: Creditor Rights is the index of creditor protection developed in La Porta et al. (1998). % corps in a group that controls some financial institution is the percentage of companies that are affiliated to a group that includes at least one financial institution. O/C is the ratio of ownership rights to voting rights. D/TA is the ratio of book value of short- and long-term financial debt to total assets (%). D/(D+E) is the ratio of book value of short- and long-term financial debt to the sum of book value of debt plus book value of equity (%). Q is the ratio of market value of equity plus the book value of debt divided by the book value of equity plus the book value of debt. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively.

related to the corporation’s riskiness and exposure to bankruptcy. In contrast, Panel B shows that neither the O/C ratio nor the quality of creditor protection has a significant impact on the leverage of corporations that cannot access related party loans, either because they do not belong to a corporate group or because their group does not include a financial institution. As mentioned earlier, it is harder for these corporations to access the debt financing that would allow the controlling shareholder to control more resources, since creditors could refuse a loan that they perceive as not creditworthy. Table 2 reports a significant positive interaction between the O/C ratio and the index of creditor protection on leverage. Therefore we partition our sample of corporations that can access related party loans between those in economies with good creditor protection (index X 3) and those in economies with poor creditor protection (index o 3). Table 3 displays the results of separate regressions for these two subsamples. Panel A shows

Journal of International Business Studies

that in countries with strong creditor protection, corporations that have a lower O/C ratio have significantly lower leverage, as measured by the ratio of debt to debt plus equity: that is, a wider separation of ownership from control is associated with less leverage. In contrast, Panel B shows that for countries with weak creditor protection, corporations that have a lower O/C ratio have significantly higher leverage (as measured by the ratio of debt to total assets): that is, a wider separation of ownership from control is associated with higher leverage.

Interpretation To interpret our findings in terms of how risk is shared between the controlling and the minority shareholders we propose the following argument. As noted earlier, if the controlling shareholder increases control over resources through pyramiding then s/he shares both upside and downside risk with the minority shareholders. If the controlling shareholder increases control over resources

Pyramiding vs leverage in corporate groups

Mara Faccio et al

95

Table 2

OLS regressions of leverage on the ownership/control ratio by access to related party lending

Dependent variable

O/C O/C  Creditor rights Creditor rights Q ln(TA) Tangib NoSic bA BankrDec Tax Intercept

Adj. R2 Prob 4F

Panel A: Corporations affiliated to a group that controls some financial institution (N¼1229)

Panel B: Corporations not affiliated to a group that controls a financial institution (N¼2219)

D/TA

D/(D+E)

D/TA

D/(D+E)

0.161 (0.00) 0.050 (0.00) 0.063 (0.01) 0.006 (0.50) 0.032 (0.00) 0.023 (0.65) 0.007 (0.37) 0.075 (0.00) 0.057 (0.00) 0.052 (0.64) 0.522 (0.00)

0.121 (0.01) 0.044 (0.01) 0.051 (0.00) 0.005 (0.35) 0.018 (0.00) 0.121 (0.01) 0.005 (0.33) 0.051 (0.00) 0.040 (0.00) 0.018 (0.81) 0.370 (0.00)

0.119 (0.21) 0.033 (0.21) 0.038 (0.18) 0.000 (0.97) 0.027 (0.00) 0.011 (0.80) 0.002 (0.79) 0.043 (0.16) 0.054 (0.00) 0.045 (0.33) 0.449 (0.00)

0.081 (0.22) 0.024 (0.19) 0.022 (0.30) 0.005 (0.15) 0.015 (0.05) 0.138 (0.00) 0.001 (0.84) 0.012 (0.55) 0.040 (0.00) 0.080 (0.05) 0.278 (0.01)

46.16% 0.00

47.39% 0.00

41.64% 0.00

41.89% 0.00

Notes: D/TA is the ratio of book value of short- and long-term financial debt to total assets (%). D/(D+E) is the ratio of book value of short- and long-term financial debt to the sum of book value of debt plus book value of equity (%). O/C is the ratio of ownership rights to voting rights. Creditor Rights is the index of creditor protection developed in La Porta et al. (1998). Q is the ratio of market value of equity plus the book value of debt divided by the book value of equity plus the book value of debt. ln(TA) is the natural logarithm of the book value of total assets. Tangib is the ratio of fixed to total assets. NoSic is the number of different two-digit SIC code sectors in which the firm reports at least 10% of sales. bA is the asset beta. BankrDec is the company’s bankruptcy decile, based on the ratio of earnings before interest, taxes and depreciation to interest expenses. Tax is the ratio of a firm’s income taxes divided by its pretax income. This variable is constrained between 0 and the highest marginal corporate tax rate in the country of each given firm. All regressions include industry fixed-effects. p-values, adjusted for heteroskedasticity and clustering at the country/industry level, are reported in parentheses below the coefficients estimates.

through leverage then s/he does not share any upside gains with the debt-holders who provide the capital, but s/he does increase the risk that, if severe downside shocks leave him/her unable to honor his debts, then s/he will lose control of the corporation. Such radical differences between the consequences of using leverage and using pyramiding to increase control over resources suggest that these two strategies will be used in different circumstances, that is, that a high separation between ownership and control should be associated with low leverage. This association is what we find when creditor protection is good. However, the differences between the two forms of control are muted if creditor protection is poor. In this case the controlling shareholder could shift

downside risk onto the debt-holders by delaying or repudiating his/her interest and principal payments. Moreover, if the controlling shareholder has only a low ratio of ownership to control rights, then s/he also enjoys only a little of the upside shocks to the corporation. The control of more resources is then the principal benefit of greater leverage. Because the two ways to increase control over resources now have similar consequences for the controlling shareholder, we expect them be used in similar circumstances. This association is what we find when creditor protection is weak. These interpretations of our results presume that the controlling shareholder determines a corporation’s leverage. This presumption seems reasonable when the controlling shareholder can access loans

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

96

Table 3

OLS regressions of leverage by creditor protection

Dependent variable

O/C Q ln(TA) Tangib NoSic bA BankrDec Tax Intercept

Adj. R2 Prob 4F

Panel A: Corporations based in countries with strong (43) creditor rights (N¼589)

Panel B: Corporations based in countries with weak (o3) creditor rights (N¼640)

D/TA

D/(D+E)

D/TA

D/(D+E)

0.045 (0.07) 0.004 (0.47) 0.039 (0.00) 0.087 (0.04) 0.011 (0.07) 0.061 (0.00) 0.052 (0.00) 0.062 (0.40) 0.200 (0.00)

0.057 (0.00) 0.011 (0.01) 0.020 (0.00) 0.065 (0.03) 0.006 (0.14) 0.036 (0.01) 0.037 (0.00) 0.023 (0.65) 0.162 (0.00)

0.058 (0.01) 0.048 (0.00) 0.026 (0.00) 0.110 (0.06) 0.018 (0.00) 0.086 (0.00) 0.064 (0.00) 0.181 (0.00) 0.440 (0.00)

0.031 (0.05) 0.021 (0.06) 0.016 (0.00) 0.226 (0.00) 0.012 (0.01) 0.063 (0.00) 0.045 (0.00) 0.104 (0.03) 0.279 (0.00)

44.78% 0.00

48.75% 0.00

47.89% 0.00

46.76% 0.00

Notes: This sample includes only corporations affiliated to a group that controls some financial institution. D/TA is the ratio of book value of shortand long-term financial debt to total assets (%). D/(D+E) is the ratio of book value of short- and long-term financial debt to the sum of book value of debt plus book value of equity (%). O/C is the ratio of ownership rights to voting rights. Creditor Rights is the index of creditor protection developed in La Porta et al. (1998). Q is the ratio of market value of equity plus the book value of debt divided by the book value of equity plus the book value of debt. ln(TA) is the natural logarithm of the book value of total assets. Tangib is the ratio of fixed to total assets. NoSic is the number of different two-digit SIC code sectors in which the firm reports at least 10% of sales. bA is the asset beta. BankrDec is the company’s bankruptcy decile, based on the ratio of earnings before interest, taxes and depreciation to interest expenses. Tax is the ratio of a firm’s income taxes divided by its pretax income. This variable is constrained between 0 and the highest marginal corporate tax rate in the country of each given firm. All regressions include industry fixed-effects. p-values, adjusted for heteroskedasticity and clustering at the country/industry level, are reported in parentheses below the coefficients estimates.

from a financial institution in which s/he owns a controlling block of shares. However, some jurisdictions restrict the percentage of loans that can be made to related parties or they may require the approval of related party loans by an independent vote of the minority shareholders.

WHY DO CONTROLLING SHAREHOLDERS SEEK TO CONTROL MORE RESOURCES? In the introduction, we noted two motivations for controlling shareholders to control more resources: for subsequent expropriation, or to exploit growth opportunities. How can we distinguish between these two motivations? Bebchuk, Kraakman, and Triantis (2000), Claessens, Djankov, Fan, and Lang (2002), and Faccio, Lang, and Young (2001) assert that a low ratio of ownership to control gives the controlling shareholder a greater incentive to

Journal of International Business Studies

expropriate minority shareholders relative to his ability to do so. Thus the O/C ratio might indicate the controlling shareholder’s tendency to expropriate minority shareholders. In that case, our finding that a lower O/C ratio is associated with greater leverage would indicate that, to control more resources that they can expropriate, controlling shareholders foist debt on corporations near the bottom of their corporate pyramids. This interpretation is consistent with many anecdotes that emerged from the Asian financial crisis. Before accepting this interpretation, we acknowledge that the O/C ratio might be endogenously determined, for example because the controlling shareholder chooses leverage and the O/C ratio jointly. To uphold our interpretation we need an instrumental variable for O/C, that is, a variable that is correlated with O/C, but is independent

Pyramiding vs leverage in corporate groups

Mara Faccio et al

97

of leverage. Gompers, Ishii, and Metrick (2007) observe that ‘‘one possible signal that insiders place a high value on the private benefits of control is if the company is named for one of the insiders.’’ This observation suggests that the O/C ratio will be negatively correlated with the dummy variable that equals 1 if the firm is named after an insider, and 0 otherwise. If this dummy variable were actually a proxy for reputation, then it would be positively correlated with the O/C ratio, but it would still be a valid instrumental variable. However, the decision to name a firm after an insider might still be made jointly with the choice of leverage: that is, the dummy variable just proposed might still be endogenous. To get around this issue we compute the dummy variable from the data of other firms that match our target firm in terms of nationality and size decile. We define an instrumental variable Average Name, whose value for firm j is the proportion of other firms in j’s country and size decile (as measured by Total Assets) that are named after one of their insiders (CEO, chairman of the board, president, a vice president, or secretary of the board as recorded in Worldscope). We define another instrument for the O/C ratio as follows. Let Average ROA for firm j be the average return on assets of other firms in j’s country and size decile. A high Average ROA could indicate a strong incentive for firm j’s controlling shareholder to increase investments in j, and hence to choose a high O/C ratio. Alternatively, it could indicate a strong incentive for the controlling shareholder to expropriate debtholders and minority shareholders, and hence to choose a low O/C ratio. We compute the Average ROA for firm j without regard to firm j’s industry. To capture industryspecific factors that might affect the choice of the O/C ratio we define Average Size for firm j as the average book value of firms in j’s industry, taking the average over all countries. We note that we compute Average Name, Average ROA, and Average Size for firm j by using only data from other firms. Since firm j’s controlling shareholder undoubtedly has little influence on these other firms, we can assume that these variables are determined independently of firm j’s leverage. We address the endogeneity of the O/C ratio using the GMM instrumental variables estimator, which is robust to heteroskedasticity (Baum, Schaffer, & Stillman, 2003; Wooldridge, 2002). Like any instrumental variables technique, GMM

assumes that the instruments are significantly correlated with the possibly endogenous variable(s), after controlling for the other exogenous variables, and that the instruments are independent of the error process. To test the first assumption we consider the goodness of fit of the first-stage regressions. Tables 4 and 5 report the first-stage coefficients, their p-values (in parentheses), the partial R2 between the excluded instruments and each possibly endogenous regressor, and the F-test corresponding to the R2. Across all first-stage regressions the p-value of the F-test is 0.02 or less. This result confirms that our instruments are significantly correlated with the possibly endogenous variable(s). We can also test the second assumption, as we have more instruments than endogenous variables. The appropriate test for our GMM estimator is the Hansen J test (Hansen, 1982). Here, the joint null hypothesis is that the excluded instruments are uncorrelated with the error term, and that they are correctly excluded from the estimated equation. Under this joint hypothesis the test statistic J has a w2 distribution. A high J would undercut the hypothesis and the validity of our instruments. Tables 4 and 5 report the estimated coefficients and their associated p-values (in parentheses) of GMM regressions with the instrumental variables discussed above. The results are similar to those reported in Tables 2 and 3: that is, corporations in which the controlling shareholder enjoys private benefits from control tend to have lower leverage when creditor protection is strong, but higher leverage when creditor protection is weak. These results are not consistent with the controlling shareholder increasing leverage to extract private benefits when creditor protection is strong. They are consistent with his/her doing so when creditor protection is weak. When D/(D þ E) is the dependent variable, the J statistic is high enough to raise doubts as to whether our instruments are valid in some of the specifications. When D/TA is the dependent variable, the J statistic is acceptable, which suggests that we have valid instrumental variables. However, they may not fully reflect the private benefits of control. Even if they did, we could not exclude the other motivation for a corporation with a low O/C ratio to seek high leverage: to exploit opportunities for growth. For example, both motivations could have operated in East Asia at different times, or even at the same time.

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

98

Table 4

IV regressions of leverage on the ownership/control ratio by access to related party lending

Dependent variable

O/C

O/C  Creditor rights

D/TA

First stage (1)

Second stage (2)

Panel A: Corporations affiliated to a group that controls some financial institution (N¼1229) O/C O/C  Creditor Rights Creditor rights Q ln(TA) Tangib NoSic bA BankrDec Tax Average name Average ROA Average size Intercept Partial R2 of excluded instruments F-test of excluded instruments Prob4F Hansen J statistic w2 p-value Dependent variable:

0.006 (0.66) 0.020 (0.04) 0.017 (0.01) 0.161 (0.00) 0.011 (0.15) 0.010 (0.72) 0.004 (0.44) 0.330 (0.00) 0.341 (0.10) 0.024 (0.00) 0.011 (0.59) 0.434 (0.15) 2.38% 9.88 0.00

O/C

D/(D+E)

0.787 (0.00) 0.075 (0.01) 0.056 (0.00) 0.417 (0.00) 0.030 (0.15) 0.034 (0.67) 0.009 (0.52) 0.850 (0.00) 0.061 (0.92) 0.043 (0.00) 0.076 (0.18) 0.254 (0.76) 0.83% 3.41 0.02

O/C  Creditor Rights

(3)

(4)

1.330 (0.00) 0.344 (0.08) 0.264 (0.09) 0.013 (0.31) 0.033 (0.00) 0.089 (0.13) 0.003 (0.58) 0.102 (0.00) 0.054 (0.00) 0.238 (0.05)

1.266 (0.00) 0.427 (0.01) 0.329 (0.01) 0.006 (0.57) 0.014 (0.04) 0.183 (0.00) 0.003 (0.48) 0.081 (0.00) 0.039 (0.00) 0.105 (0.28)

1.278 (0.00)

1.240 (0.00)

0.14 0.71

2.50 0.11

D/TA

D/(D+E)

First stage (5)

Second stage (6)

(7)

(8)

2.723 (0.12) 1.145 (0.23) 1.060 (0.23) 0.009 (0.54) 0.015 (0.03)

1.127 (0.48) 0.892 (0.30) 0.808 (0.31) 0.013 (0.31) 0.012 (0.07)

Panel B: Corporations not affiliated to a group that controls a financial institution (N¼2219) O/C O/C  Creditor rights Creditor rights Q ln(TA)

Journal of International Business Studies

0.017 (0.00) 0.010 (0.00) 0.017 (0.00)

0.885 (0.00) 0.032 (0.00) 0.026 (0.00)

Pyramiding vs leverage in corporate groups

Mara Faccio et al

99

Table 4 Continued Dependent variable:

O/C  Creditor Rights

O/C

D/TA

D/(D+E)

First stage

Tangib NoSic bA BankrDec Tax Average name Average ROA Average size Intercept Partial R2 of excluded instruments F-test of excluded instruments Prob4F Hansen J statistic w2 P-value

Second stage

(5)

(6)

(7)

(8)

0.001 (0.95) 0.003 (0.38) 0.040 (0.00) 0.001 (0.78) 0.048 (0.23) 0.236 (0.00) 0.014 (0.00) 0.002 (0.82) 1.106 (0.00) 2.76% 20.88 0.00

0.033 (0.61) 0.023 (0.02) 0.094 (0.01) 0.001 (0.80) 0.084 (0.46) 0.516 (0.01) 0.023 (0.00) 0.003 (0.91) 0.205 (0.54) 1.11% 8.22 0.00

0.035 (0.49) 0.019 (0.26) 0.038 (0.27) 0.053 (0.00) 0.011 (0.87)

0.103 (0.03) 0.018 (0.24) 0.053 (0.07) 0.040 (0.00) 0.048 (0.42)

3.001 (0.07)

0.763 (0.61)

0.13 0.72

1.68 0.19

Notes: This table reports the estimated coefficients and associated p-values (in parentheses) of instrumental variables (IV) regressions in which O/C and O/C  Creditor Rights are treated as endogenous. The instrumental variables used are Average Name, Average ROA, and Average Size.

Table 5

IV regressions of leverage by creditor protection

Panel A: Corporations based in countries with strong (43) creditor rights (N¼589) Dependent variable

O/C

D/TA

First stage

ln(TA) NoSic bA Tangib

Second stage

(1)

(2)

(3)

0.022 (0.03) 0.012 (0.21) 0.010 (0.35) 0.033 (0.34) 0.035 (0.56)

0.626 (0.01) 0.016 (0.14) 0.028 (0.00) 0.005 (0.50) 0.074 (0.01) 0.077 (0.10)

0.717 (0.00) 0.024 (0.00) 0.008 (0.26) 0.001 (0.84) 0.054 (0.06) 0.058 (0.20)

O/C Q

D/(D+E)

Panel B: Corporations based in countries with weak (o 3) creditor rights (N¼640) O/C

D/TA

First stage

D/(D+E)

Second stage

(4)

(5)

(6)

0.001 (0.97) 0.021 (0.03) 0.009 (0.36) 0.010 (0.83) 0.276 (0.00)

1.261 (0.00) 0.050 (0.22) 0.053 (0.00) 0.005 (0.72) 0.089 (0.11) 0.484 (0.00)

0.993 (0.00) 0.028 (0.38) 0.037 (0.00) 0.002 (0.89) 0.063 (0.16) 0.550 (0.00)

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

100

Table 5 Continued Panel A: Corporations based in countries with strong (43) creditor rights (N¼589) Dependent variable

O/C

D/TA

First stage

BankrDec Tax Average name Average ROA Average size Intercept Partial R2 of excluded instruments F-test of excluded instruments Prob4F Hansen J statistic w2 P-value

D/(D+E)

Second stage

Panel B: Corporations based in countries with weak (o 3) creditor rights (N¼640) O/C

D/TA

First stage

D/(D+E)

Second stage

(1)

(2)

(3)

(4)

(5)

(6)

0.000 (1.00) 0.239 (0.06) 0.591 (0.10) 0.039 (0.17) 0.062 (0.02) 1.596 (0.00) 1.89% 3.71 0.01

0.050 (0.00) 0.279 (0.01)

0.035 (0.00) 0.238 (0.02)

0.046 (0.00) 0.609 (0.00)

0.030 (0.00) 0.450 (0.01)

0.140 (0.33)

0.201 (0.14)

0.014 (0.06) 0.276 (0.02) 0.645 (0.02) 0.029 (0.01) 0.012 (0.70) 0.103 (0.82) 2.46% 5.29 0.00

0.624 (0.00)

0.419 (0.01)

3.55 0.17

6.76 0.03

1.02 0.60

0.19 0.91

Notes: This sample includes only corporations affiliated to a group that controls some financial institution. This table reports the estimated coefficients and associated p-values (in parentheses) of instrumental variables (IV) regressions in which O/C is treated as endogenous. The instrumental variables used are Average Name, Average ROA, and Average Size.

CONCLUSION A central theme of corporate finance is that debt and equity share risk differently between those who provide capital and those who control its use. Both pyramiding and leverage permit the controlling shareholder of a corporation to control more resources. Our analysis indicates that when creditor protection is strong, wider separation of ownership from control is associated with low leverage: that is, controlling shareholders who use one method tend to avoid the other. This conclusion suggests that controlling shareholders regard the two methods as having different implications for risk sharing. However, when creditor protection is weak, controlling shareholders who use one method tend to use the other also, suggesting that controlling shareholders regard the two methods as having similar implications for risk sharing. It follows that improved creditor protection can help to allocate risk more efficiently, not only by limiting the use of debt to secure private benefits, but also, more fundamentally, by sharpening the distinction between debt and equity as risk-sharing instru-

Journal of International Business Studies

ments. Thus improved creditor protection can force entrepreneurs to confront their risks and opportunities with capital structures that are more precisely tailored to their individual circumstances. This conclusion illustrates how improving capital market institutions can promote economic growth and stabilize financial markets, as emphasized by ¨ c¸-Kunt and Maksimovic authors such as Demirgu (1998), Greenwood and Jovanovic (1990), Levine (1997), and Rajan and Zingales (1998). Direct evidence that the separation of ownership from control and access to related party lending leads to expropriation through debt would require that we document actual harm to banks, minority shareholders, or countries/governments. This task would require further research, a cross-corporation counterpart to the cross-country comparison of Johnson, Boone, Breach, and Friedman (2000), who show that, after controlling for a number of macroeconomic variables, poor investor protection had a significant effect on the extent of currency depreciation and stock market performance during the Asian financial crisis.6

Pyramiding vs leverage in corporate groups

Mara Faccio et al

101

ACKNOWLEDGEMENTS We thank Stijn Claessens, Simeon Djankov, and Joseph Fan for providing their data for East Asia. We acknowledge helpful comments from the referees of this journal, Lemma Senbet (the Editor), as well as Amber Anand, Mark Huson, Rafael La Porta, Maria-Teresa Marchica, Ron Masulis, Ike Mathur, John McConnell, Randall Morck, Roberto Mura, seminar participants at the University of Alberta, Harvard Business School, Notre Dame, Tulane, Vanderbilt, Virginia Tech, York, and participants at the meetings of the American Economic Association, the European Financial Management Association, the Mitsui Life Symposium on Global Financial Markets, and the Western Finance Association. We also thank Sandra Sizer for editorial help. Larry Lang gratefully acknowledges the financial support of the China Guang Hua Science and Technology Foundation and the Hong Kong Government UGC Earmarked Grant. NOTES We are grateful to a referee for pointing out this second possible motivation. 2 Desai, Foley, and Hines (2004) show that foreign affiliates of multinational firms rely less on external debt in countries with weak creditor protection. 3 However, when researchers measure leverage at book value, the relationship is not significant in Italy 1

and Japan. It becomes significant when leverage is measured at market value. McConnell and Servaes (1995) find that for high-growth firms Q is negatively affected by leverage; for low-growth firms they find a positive correlation between Q and leverage. Lang, Ofek, and Stulz (1996) find that higher levels of leverage have a negative impact on the growth of the firm when Qo1, but a positive (though nonsignificant) impact when Q41. They argue that debt disciplines management when Qo 1, preventing them from investing in negative NPV projects. For firms with Q41 (i.e., good investment opportunities) high leverage does not constrain management. 4 Another widely used proxy for growth opportunities is the historical sales growth rate, which we used as a robustness check; the results did not differ substantially. We use the Q-ratio because lenders should be more concerned about future growth (hence the ability to repay debt) than historical growth. 5 The preceding studies do not control for this possibility in analyzing the relationship between leverage and growth, so it is not possible to distinguish between the two hypotheses. 6 We are grateful to referees for the points made in this paragraph.

REFERENCES Ahn, S., Denis, D. J., & Denis, D. K. 2006. Leverage and investment in diversified firms. Journal of Financial Economics, 79(2): 317–337. Akerlof, G. A., & Romer, P. M. 1993. Looting: The economic underworld of bankruptcy for profit. Brookings Papers on Economic Activity, 2: 1–73. Bancel, F., & Mittoo, U. R. 2004. Cross-country determinants of capital structure choice: A survey of European firms. Financial Management, 33(4): 103–132. Baum, J. 1994. The money machine. Far Eastern Economic Review, 11 (August); 62–66. Baum, C. F., Schaffer, M. E., & Stillman, S. 2003. Instrumental variables and GMM: Estimation and testing. Stata Journal, 3(1): 1–31. Bebchuk, L. A., Kraakman, R. H., & Triantis, G. G. 2000. Stock pyramids, cross-ownership, and dual class equity: The creation of agency costs of separating control from cash flow rights. In R. Morck (Ed.) Concentrated ownership, : 295–315 Chicago: University of Chicago Press. Berger, A. N., & Udell, G. F. 1995. Relationship lending and lines of credit in small firm finance. Journal of Business, 68(3): 351–381. Bradley, M., Jarrell, G., & Kim, E. H. 1984. On the existence of an optimal capital structure: Theory and evidence. Journal of Finance, 39(3): 857–878. Burgman, T. A. 1996. An empirical examination of multinational corporate capital structure. Journal of International Business Studies, 27(3): 553–570.

China Credit Information Service. 1997. Business groups in Taiwan, 1996–1997. Taipei, Republic of China: China Credit Information Service. Chu, Y.-W., & Chu and Hamilton, 1993. Business networks in Hong Kong, mimeo, University of California, Davis. Claessens, S., Djankov, S., & Lang, L. H. P. 2000. The separation of ownership and control in East Asian corporations. Journal of Financial Economics, 58(1–2): 81–112. Claessens, S., Djankov, S., Fan, J. P. H., & Lang, L. H. P. 2002. Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance, 57(6): 2741–2771. Comment, R., & Jarrell, G. A. 1995. Corporate focus and stock returns. Journal of Financial Economics, 37(1): 67–87. Commerzbank. 1997. Wer geho¨rt zu wem [Who belongs to whom]. http://www.commerzbank.com/navigate/date_frm.htm. Comision Nacional del Mercado de Valores. 1998. Participaciones significativas en sociedades cotizadas. http://www .cnmv.es/english/cnmve.htm. CONSOB. 1997. Bollettino – edizione speciale n. 4/97 – Compagine azionaria delle societa` quotate in borsa o ammesse alle negoziazioni nel mercato ristretto al 31 dicembre 1996. http://www.consob.it. Dammon, R. M., & Senbet, L. W. 1988. The effect of taxes and depreciation on corporate investment and financial leverage. Journal of Finance, 43(2): 357–373. Demirgu¨c¸-Kunt, A., & Maksimovic, V. 1998. Law, finance, and firm growth. Journal of Finance, 53(6): 2107–2137.

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

102

Denis, D. K., & McConnell, J. J. 2003. International corporate governance. Journal of Financial and Quantitative Analysis, 38(1): 1–36. Desai, M. A., Foley, C. F., & Hines Jr., J. R. 2004. A multinational perspective on capital structure choice and internal capital markets. Journal of Finance, 59(6): 2451–2487. Dodwell Marketing Consultants. 1997. Industrial groupings in Japan: The anatomy of the Keiretsu, (12th ed.), 1996/1997 Tokyo, Japan: Dodwell Marketing Consultants. Easterbrook, F. H. 1984. Two agency-cost explanations of dividends. American Economic Review, 74(4): 650–659. Ellul, A. 2008. Control motivations and capital structure decisions, Working paper, Indiana University. Faccio, M., & Lang, L. H. P. 2002. The ultimate ownership of western European corporations. Journal of Financial Economics, 65(3): 365–395. Faccio, M., Lang, L. H. P., & Young, L. 2001. Dividends and expropriation. American Economic Review, 91(1): 54–78. Fama, E., & French, K. 1992. The cross-section of expected stock returns. Journal of Finance, 47(2): 427–466. Financial Times. 1997. Extel financial. Fisman, R. 1998. Announcement Effects of Suharto’s illnesses on Related Companies, mimeo, Harvard Business School, September. Frank, M. Z., & Goyal, V. K. 2003. Testing the pecking order theory of capital structure. Journal of Financial Economics, 67(2): 217–248. Friedman, E., Johnson, S., & Mitton, T. 2003. Propping and tunneling. Journal of Comparative Economics, 31(4): 732–750. Gompers, P. A., Ishii, J. L., & Metrick, A. 2007. Extreme governance: An analysis of dual-class companies in the United States, Working paper, University of Pennsylvania. Graham, J. R. 2003. Taxes and corporate finance: A review. Review of Financial Studies, 16(4): 1075–1129. Greenwood, J., & Jovanovic, B. 1990. Financial development, growth, and the distribution of income. Journal of Political Economy, 98(5): 1076–1107. Hansen, L. P. 1982. Large sample properties of generalized method of moments estimators. Econometrica, 50(4): 1029–1054. Harris, M., & Raviv, A. 1990. Capital structure and the informational role of debt. Journal of Finance, 45(2): 321–349. Harvey, C. R., Lins, K. V., & Roper, A. H. 2004. The effect of capital structure when expected agency costs are extreme. Journal of Financial Economics, 74(1): 3–30. The Herald Tribune. 1997. French company handbook 1997. SFB-Paris Bourse. Hiscock, G. 1998. Asia’s wealth club, London: Nicholas Brealey. Hong Kong Stock Exchange. 1997. Hong Kong Company Handbook. Hoshi, T., Scharfstein, D., & Kashyap, A. 1990. The role of banks in reducing the costs of financial distress in Japan. Journal of Financial Economics, 27(1): 125–160. Il Sole 24 ore. 1997. Il taccuino dell’azionista. Indobusiness. 1998. 1995 ranking of the largest Indonesian conglomerates. http://indobiz.com/company/warta/conglo/ htm. Jensen, M. C. 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review, 76(2): 323–329. Jensen, M. C. 1989. Eclipse of the public corporation. Harvard Business Review, 67(5): 61–75. Jensen, M. C., & Meckling, W. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3(4): 305–360. John, K., & Senbet, L. W. 1998. Corporate governance and board effectiveness. Journal of Banking and Finance, 22(4): 371–403. Johnson, S., Boone, P., Breach, A., & Friedman, E. 2000. Corporate governance in the Asian financial crisis. Journal of Financial Economics, 58(1–2): 141–186.

Journal of International Business Studies

Kester, W. C. 1986. Capital and ownership structure: A comparison of United States and Japanese manufacturing corporations. Financial Management, 15(1): 5–16. Kim, W. S., & Sorensen, E. H. 1986. Evidence on the impact of agency costs of debt on corporate debt policy. Journal of Financial and Quantitative Analysis, 21(2): 131–144. Korean Fair Trade Commission. 1997. 1996 List of the largest 30 chaebols. Seoul, Korea: Korean Fair Trade Administration. Krozner, R. S., & Strahan, P. E. 2001. Bankers on boards of directors: Monitoring, conflicts of interest, and lender liability. Journal of Financial Economics, 62(3): 415–452. Laeven, L. 2001. Insider lending and bank ownership: The case of Russia. Journal of Comparative Economics, 29(2): 207–229. Lang, L. H. P., & Stulz, R. M. 1994. Tobin’s q, corporate diversification, and firm performance. Journal of Political Economy, 102(6): 1248–1280. Lang, L. H. P., Ofek, E., & Stulz, R. M. 1996. Leverage, investment, and firm growth. Journal of Financial Economics, 40(1): 3–29. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. W. 1998. Law and finance. Journal of Political Economy, 106(6): 1113–1155. La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 1999. Corporate ownership around the world. Journal of Finance, 54(2): 471–518. Levine, R. 1997. Financial development and economic growth: Views and agenda. Journal of Economic Literature, 35(2): 688–726. Lewellen, W. G. 1971. A pure financial rationale for the conglomerate merger. Journal of Finance, 26(2): 527–537. Lim, U. 1998. Ownership structure and family control in Korean conglomerates: With cases of the 30 largest chaebols. Korea: Seoul University. London Stock Exchange. 1997. The London Stock Exchange yearbook. London: The London Stock Exchange. Maksimovic, V., & Titman, S. 1991. Financial policy and reputation for product quality. Review of Financial Studies, 4(1): 175–200. Maloney, M. T., McCormick, R. E., & Mitchell, M. 1993. Managerial decision making and capital structure. Journal of Business, 66(2): 189–217. McConnell, J. J., & Servaes, H. 1995. Equity ownership and the two faces of debt. Journal of Financial Economics, 39(1): 131–157. Myers, S. 1977. Determinants of corporate borrowing. Journal of Financial Economics, 5(2): 147–175. The Nation. 1998. Thai tycoons: Winners and losers in the economic crisis. Special Issue, July. Philippine Stock Exchange. 1997. Investment guide 1996. Manila: The Philippine Stock Exchange. Rajan, R. G., & Zingales, L. 1995. What do we know about capital structure? Some evidence from international data. Journal of Finance, 50(5): 1421–1460. Rajan, R. G., & Zingales, L. 1998. Financial dependence and growth. American Economic Review, 88(3): 559–586. Ross, S. 1977. The determination of financial structure: The incentive-signalling approach. Bell Journal of Economics, 8(1): 23–40. Sato, K. 1984. The anatomy of Japanese businesses, Chapter 4. Armonk, NY: M.E. Sharpe Publishing. Shleifer, A., & Vishny, R. W. 1997. A survey of corporate governance. Journal of Finance, 52(2): 737–783. Singapore Stock Exchange. 1997. Singapore company handbook. Singapore: Singapore Stock Exchange. Stulz, R. M. 1988. Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics, 20(1–2): 25–54. Tan, E. 1993. Interlocking directorates, commercial banks, other financial institutions, and non-bank corporations. Philippine Review of Economics and Business, 30: 1–50.

Pyramiding vs leverage in corporate groups

Mara Faccio et al

103

Tara Siam Company 1997. Thai business groups 1996–1997: A unique guide to who owns what. Bangkok, Thailand: Tara Siam Company. Taylor, M. 1998. Have cash, will travel. Far Eastern Economic Review, Special Section on the Li Ka-Shing Conglomerate, 5 March. Titman, S. 1984. The effect of capital structure on a firm’s liquidation decision. Journal of Financial Economics, 13(1): 137–151.

Titman, S., & Wessels, R. 1988. The determinants of capital structure choice. Journal of Finance, 43(1): 1–19. Vatikiotis, 1997. From chickens to microchips: The story of Thai conglomerates. Far Eastern Economic Review, 23 (January): 38–43. W. I. Carr Banque Indosuez Group. 1997. Indonesian group connections. Jakarta, Indonesia: Banque Indosuez. Wooldridge, J. M. 2002. Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press.

APPENDIX A See Table A1. Table A1

Data sources for East Asian corporations

Country

Immediate ownership data

Dual-class shares

Business groups: Pyramids and cross-holdings

Hong Kong

Worldscope, (1998) Asian Company Handbook (1998) Hong Kong Stock Exchange, 1997 Worldscope, (1998) Asian company handbook, 1998 Institute for Economic and Financial Research, 1996 Worldscope, (1998) Japan company handbook, 1998 Worldscope, (1998) Asian company handbook, 1998 Worldscope, (1998) Asian company handbook, 1998

Datastream International, 1998

Chu, Y.-W. and Hamilton, G., 1993, Taylor, M., 1998, Hong Kong Stock Exchange, 1997.

Datastream International, 1998 Institute for Economic and Financial Research, 1996

Fisman, R., 1998, W. I. Carr Banque Indosuez Group, 1997, Indobusiness, 1998.

Datastream International, 1998

Dodwell Marketing Consultants, 1997, Sato, K., 1984. Korean Fair Trade Commission, 1997, Lim, U., 1998.

Worldscope, (1998) Asian company handbook, 1998 Philippine Stock Exchange, 1997 Worldscope, (1998) Asian company handbook, 1998 Worldscope, (1998) Asian company handbook, 1998 Worldscope, (1998) Asian company handbook, 1998 Securities Exchange of Thailand, 1997

Datastream International, 1998 Philippine Stock Exchange, 1997

Indonesia

Japan South Korea

Malaysia

Philippines

Singapore

Taiwan

Thailand

Datastream International, 1998

Datastream International, 1998 Kuala Lumpur Stock Exchange (1997)

Datastream International, 1998 Singapore Stock Exchange, 1997 Datastream International, 1998

Datastream International, 1998 Securities Exchange of Thailand, 1997

Hiscock, G., 1998, http://www. ambg.com.my for A-M Banking Group, http://www.berjaya.com. my for Berjaya Group, http:// ww.simenet.com for Sime Darby Group, http://www.lion.com.my for Lion Group, http://www. hongleong-group.com.sg for Hong Leong Group. Philippine Stock Exchange, 1997, Tan, E., 1993.

Singapore Stock Exchange, 1997, Hiscock, G., 1998. China Credit Information Service, 1997, Baum, J., 1994. Tara Siam, 1997, The Nation, 1998, Vatikiotis, M., 1997.

Source: Claessens et al. (2000).

Journal of International Business Studies

Pyramiding vs leverage in corporate groups

Mara Faccio et al

104

APPENDIX B See Table B1. Table B1

Sources of ownership and control data for West European corporations

Country

Immediate ownership data

Dual-class shares

Business groups

France

Herald Tribune, 1997, Financial Times, 1997, Worldscope (1998), http:// www.bourse-de-paris.fr/fr/market8/ fsg830.htm Commerzbank, 1997 (http:// www.commerzbank.com/navigate/ date_frm.htm), Financial Times, 1997, Worldscope (1998) CONSOB, 1997, Il Sole 24 ore, 1997.

Datastream, 1999, Financial Times, 1997, Les Echos, 1996, Muus, 1998

Herald Tribune, 1997, Financial Times, 1997.

Datastream, 1999, Financial Times, 1997, Die Welt, 1996, Becht and Boehmer, 1998

Commerzbank, 1997, Extel financial

Datastream, 1999, Il Sole 24 ore, 1997.

Il Sole 24 ore, 1997, http:// www.olivetti.it/group/, http:// www.pirelli.com/company/ index.htm Comision Nacional del Mercado de Valores, 1998., Extel financial

Germany

Italy

Spain

Comision Nacional del Mercado de Valores, 1998.

United Kingdom

Financial Times, 1997, London Stock Exchange, 1997, Financial Times, Worldscope, (1998), http:// www.hemscott.com/equities/company/

ABOUT THE AUTHORS Mara Faccio (PhD, Dottorato, Finance, Universita` Cattolica (Milan)) is Hanna Chair in Entrepreneurship & Associate Professor of Finance at the Krannert School of Management, Purdue University. Her research interests include international corporate governance, family firms, mergers and acquisitions, and corporate political connections. She was born in Italy, and can be reached at [email protected]. Larry Lang (PhD, University of Pennsylvania) is a professor of Finance at The Chinese University of

Datastream, 1999, Financial Times, 1997, ABC, 1996, Crespi-Cladera and Garcia-Cestona, 1998 Datastream, 1999, Financial Times, 1997, Financial Times, 1996

Extel financial

Hong Kong. His interests are especially in the area of Business/Merger & Acquisitions. He was born in China, and can be reached at [email protected] .cuhk.edu.hk. Leslie Young (DPhil, Mathematics, Oxford University) is a professor of Finance and Executive Director of the Asia Pacific Institute of Business at The Chinese University of Hong Kong. His research interests include corporate governance, international finance, and political economy. He was born in China and is a citizen of New Zealand. E-mail: [email protected].

Accepted by Lemma Senbet, Area Editor, 21 February 2009. This paper has been with the authors for four revisions.

Journal of International Business Studies