Leuz - Gee 2006 JFEC.pdf

ARTI AR TICL CLE E IN PR PRESS ESS

Journal of Financial Economics 81 (2006) 411–439 www.elsevier.com/locate/jfec

Political relationships, global financing, and corporate transparency: Evidence from Indonesia$ Christian Leuza,, Felix Oberholzer-Geeb a

The Wharton School, University of Pennsylvania, Philadelphia, PA 19104, USA b Harvard Business School, Harvard University, Boston, MA 02163, USA

Received 8 September 2003; received in revised form 4 May 2005; accepted 22 June 2005 Available online 20 February 2006

Abstract

This study examines the role of political connections in firms’ financing strategies and their longrun performance. We view political connections as an example for domestic arrangements which can reduce the benefits of global financing. Using data from Indonesia, we find that firms with strong political political connections connections are less likely to have publicly traded foreign securities. securities. As a result, result, estimates estimates of the performance consequences of foreign financing are severely biased if value-creating domestic arrangements such as political relationships are ignored. Connections not only alter firms’ financing strategies, they also influence long-run performance. Tracking returns across several regimes, we show that firms have difficulty re-establishing connections with a new government when their patron

$

We thank two anonymous reviewers, Michael Backman, Phil Berger, Jack Coffee, Alexander Dyck, Mara Faccio, Ray Fisman, Tarun Khanna, Greg Miller, Todd Mitton, DJ Nanda, Jordan Siegel, Benny Tabalujan, Ross Watts, Greg Waymire, Peter Wysocki, and Jerry Zimmerman, as well as seminar participants at the AEA Annual Meeting, Chinese University of Hong Kong, EAA Annual Meeting, Harvard Business School, Norwegian School of Economics and Business, Singapore Management University, Stanford Summer Camp, University of Rochester, Stockholm University, and the Wharton School for helpful comments. We are grateful to Simeon Djankow and Ray Fisman for sharing data. Arief Budiman, Shanshan Cao, Bryan Chao, Christina Hsiung Chen, Robert Irwan, Randy Jusuf, Ian Lin, Svetlana Ni, and Julie Wong provided excellent research assistance. The generous financial support of the Wharton–Singapore Management University Research Center and the Baker Foundation is gratefully acknowledged.  Correspon Corresponding ding author. author. Tel.: Tel.: +1 215 898 2610; 2610; fax: +1 +1 215 573 2054. 2054. E-mail address: [email protected] (C. Leuz). 0304-405 0304-405X/$ X/$ - see front matter matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jfineco.2005.06.006

ARTI AR TICL CLE E IN PR PRESS ESS 412

C. Leuz, F. Oberholzer-Gee / Journal of Financial Economics 81 (2006) 411–439

falls falls from from power, power, leading leading closely closely connec connected ted firms firms to underp underperf erform orm under under the new regime regime and subsequently to increase their foreign financing. r 2006 Elsevier B.V. All rights reserved. JEL Classifications: P16; G32; G38; K22; K42; M41; G18 Keywords: Cross listing; Financing choices; Emerging markets; Asian financial crisis; Indonesia; Disclosure

1. Introduction Introduction

Political connections are believed to be a valuable resource for many firms (Fisman, (Fisman, 2001), 2001 ), but there are only a handful of studies that document how such connections impact firms’ strategic choices and their long-run financial performance (Faccio, (Faccio, 2002; 2002; Johnson and Mitton, 2003). 2003). Studying the performance consequences of political ties is of particular inte intere rest st beca becaus usee thes thesee ties ties are are ofte often n inco incons nsist isten entt with with othe otherr va valu luee-cr crea eatin ting g busi busine ness ss strategies. Moreover, the performance of closely connected firms might vary dramatically over time depending on the political fortunes of their backers, suggesting that connectionbase based d stra strate tegi gies es coul could d be part partic icul ular arly ly risk risky. y. In this this stud study, y, we anal analyz yzee how how clos closel ely y connected firms make strategic decisions and how these decisions affect their long-run financial performance. In the first part of the paper, we examine the link between political ties and firms’ global financing strategies. We are interested in this link because foreign capital has become an increasingly increasingly important important source of finance for firms in emerging emerging markets (e.g., Karolyi, (e.g., Karolyi, 1998; 1998; Beka Be kaer ertt et al al., ., 20 2002 02). ). Empiric Empirical al studies studies find that that compan companies ies that that access access global global capita capitall markets markets experience experience substantial substantial benefits. benefits. For instance, instance, firms with U.S. cross-listing cross-listingss enjoy a lower cost of capital, improved visibility and reputation, and better growth opportunities (e.g., Errunza (e.g., Errunza and Miller, 2000; 2000; Doidge et al., 2004; 2004; Hail and Leuz, 2004). 2004). Taking these benefits at face value, it is difficult to understand why only a minority of comp compan anie iess acce access ss fore foreign ign capi capital tal mark markets ets.. A core core idea idea in this this pape paperr is that that domes domesti ticc opportunities significantly reduce the net benefits of foreign securities for some firms. For inst instan ance ce,, firms firms with with poli politi tica call ties ties ofte often n rece receive ive cheap cheap loan loanss from from stat statee-own owned ed banks banks (Faccio, 2002; 2002; Wiwattanakantang et al., 2006), 2006), so they do not need to tap into foreign capi capita tall mark market ets. s. It is also also poss possib ible le that that glob global al finan financi cing ng impo impose sess extr extra a cost costss on closel closely y connect connected ed firms firms because because the decisio decision n to cross-l cross-list ist shares shares on foreig foreign n exchan exchanges ges often forces firms to adapt to the regulations that govern these markets (Coffee, (Coffee, 2002; 2002; Reese and Weisbach, 2002; 2002; Siegel, 2005). 2005). If minority shareholders are better protected abroad, issuing foreign securities becomes expensive for controlling owners accustomed to expl exploit oitin ing g domes domestic tic inve invest stors ors.. Simil Similar arly ly,, firms firms with with fore foreign ign secu securit ritie iess attr attrac actt the the attent attention ion of foreig foreign n analys analysts ts and the interna internatio tional nal busine business ss press press (Ba Baker ker et al al., ., 200 2002 2; Lang et al., 2003). 2003). However, high levels of public scrutiny can be difficult to reconcile with political favors of often dubious legality. These hidden costs of foreign financing can perhaps explain why few companies finance themselves globally despite the apparent bene benefit fitss of doing doing so. so. Whil Whilee this this stud study y is focus focused ed on poli politi tica call conne connect ctio ions ns,, we belie believe ve the hidden costs of global financing to be significant in many contexts. Political patronage is just just one one exam example ple of domes domesti ticc arra arrang ngeme ement ntss that that redu reduce ce the the net net bene benefits fits of havi having ng foreign securities.

ARTICLE IN PRESS C. Leuz, F. Oberholzer-Gee / Journal of Financial Economics 81 (2006) 411–439

413

We examine the link between political connections and global financing using Indonesian data. Indonesia’s crony capitalism under former President Suharto provides a particularly suitable setting for examining the role of connections for firms’ foreign financing choices. There is ample evidence that political connections were particularly valuable under the Suharto regime and often involved access to finance (Borsuk, 1993; McBeth, 1994; Fisman, 2001). Moreover, Indonesia has low levels of mandatory disclosure, suggesting that publicly traded foreign securities have substantial informational consequences. Finally, the Asian financial crisis and subsequent regime changes provide shocks that we can exploit in assessing the performance consequences of foreign financing and political connections. Our results provide strong support for the view that foreign securities and close political connections are substitutes. Firms that were close to the Suharto regime were significantly less likely to have publicly traded securities abroad. These findings hold after controlling for firm size, financial leverage, profitability, and industry characteristics. We carefully test for the possibility that our findings are due to unobserved heterogeneity among the sample firms. Using an instrumental-variable strategy, our closeness measure does not appear to be endogenous to the choice of foreign securities. We also demonstrate that our results are robust to alternative proxies for political connections and are not driven by company differences in volatility, susceptibility to bad news, dependence on external financing, or exposure to foreign product markets. Our finding that political relationships influence the likelihood of global financing has important implications for studies estimating the performance effects of foreign securities. Recent papers show that Asian firms with foreign securities, higher-quality disclosures, and less problematic ownership structures exhibit significantly higher returns during the Asian financial crisis (Johnson et al., 2000; Mitton, 2002; Lemmon and Lins, 2003; Baek et al., 2004). However, if domestic sources of firm value are omitted from these analyses, the resulting estimates are likely to be biased. In a re-analysis of returns during the crisis, we do indeed find significant bias in simple performance regressions that fail to control for political connections. Conceptually, this bias provides empirical evidence that, during the crisis, President Suharto was propping up firms that were close to his regime.1 The analysis also illustrates why it is important to consider the endogeneity of foreign financing decisions and hence firms’ domestic opportunities when estimating performance benefits. The second part of the paper is concerned with the long-run consequences of investments in political relationships. Unlike many other resources at the firm’s disposal, political connections can lose their value overnight when the government fails to win an election, suggesting that investments in political relationships might be particularly risky. On the other hand, many formal models of political economy assume that politicians auction off favors to firms with the highest willingness to pay for political patronage (Bernheim and Whinston, 1986). Under this view, well-connected managers can easily rebuild their relationships once a regime has changed, and regime changes have little effect on long-run firm performance. 1

While empirical support for propping is still rare, there is ample anecdotal evidence that Suharto tried to protect well-connected firms. For example, the Texmaco group received loans in excess of US$ 1 billion from Bank Negara Indonesia, one of Indonesia’s largest state banks. The loans far exceeded the bank’s legal lending limit, but were approved by Suharto ‘‘as a means to prop up the conglomerate after the Asian financial crisis.’’ Texmaco’s founder, Marimutu Sinivasan, is said to be a long-time friend of President Suharto ( Solomon, 1999).

ARTICLE IN PRESS 414


Indonesia provides ample opportunity to study the performance consequences of regime changes. After the long-lived Suharto regime and its sudden demise on 21 May 1998, President Habibie, a long-time Suharto ally, took office. In October 1999, Habibie was replaced by President Wahid, a cleric critical of the Suharto regime (Symonds, 1998). These regime changes allow us to study both a friendly transition, when Suharto’s prote ´ge ´ Habibie came to power, and the inauguration of the Wahid government which treated the Suharto cronies less favorably. Studying the financial consequences of such changes, we find that firms closely connected to Suharto underperformed during the Asian financial crisis as long as Suharto was in power; recovered under Habibie; but significantly underperformed once Wahid took office. This pattern of changes in long-run financial performance suggests that it was not easy for companies connected to the Suharto regime to establish close connections with the new and less friendly Wahid government. If the empirical regularity documented in the first part of this study—that closely connected firms shy away from global financing—holds generally, Suharto firms should have found it more attractive to finance their operations globally once Wahid came to power. In an empirical analysis of the Wahid period, we document that former Suharto cronies became more inclined to issue foreign securities. This set of results has important implications for the consequences of financial liberalization because it suggests that there are important complementarities between capital market liberalization and political reform (Chui et al., 2000; Stulz, 2005). Firms closely connected with the Suharto regime chose not to access global capital markets because the benefits of global financing remained small for these firms as long as they benefited from political connections. Once the connections had lost their value, raising capital abroad became more attractive. Changes in financing strategies reflect changes in the domestic opportunities that firms must forgo when they issue foreign securities. This paper contributes to the literature on global financing and the benefits of cross-listing by developing a more complete model of financing choices (e.g., Saudagaran, 1988; Saudagaran and Biddle, 1995; Karolyi, 1998; Doidge et al., 2004). An appealing feature of this model is that it is easier to understand why the vast majority of large firms do not access global capital markets even though the group of globally financed companies experiences substantial benefits. Our study also complements the literature on the importance of political connections by documenting that close political ties have important consequences for firm strategy and performance (Stigler, 1971; Kroszner and Strahan, 1999; Baron, 2001; Faccio, 2002; Johnson and Mitton, 2003; Friedman et al., 2003). Political ties affect firm performance in two ways. In the short run, firms with close connections underperformed during the Asian financial crisis. In the long run, investments in political relationships stopped paying off because a new, less friendly government came to power. Investments in political ties, we conclude, are less likely to be a source of long-term value in environments with macroeconomic and political instability. The paper is organized as follows. Section 2 describes the institutional setting of this study. Section 3 explains the sample and the data. In Section 4, we present the main results for firms’ foreign financing decisions. Section 5 explores whether our results are specific to the type of foreign security that we study. In Section 6, we analyze the performance effects of foreign securities. In Section 7, we explore the dynamics of political connections, and Section 8 concludes the paper.


415

2. Institutional setting

A key premise of our approach is the idea that political connections constitute a source of firm value. There is empirical evidence supporting this view, both for Indonesia (Fisman, 2001) and for a larger set of economies. For instance, connected firms pay lower taxes and have larger market shares (Faccio, 2002). In Indonesia, the Suharto regime often arranged preferential financing for well-connected firms (so-called ‘‘memo-lending’’). An example from the early 1990s is Golden Key, a little-known chemical and manufacturing group, which received an unsecured loan of $430 million from the state-owned Bank Pembangunan Indonesia. Court proceedings subsequently revealed that Hutomo Mandala Putra, the youngest son of President Suharto, was an early investor in Golden Key and had introduced the firm to bank officials who approved the loan at ‘‘neck-breaking speed’’ (McBeth, 1994). Similarly, the Barito Pacific group received huge loans from state banks prior to the Asian financial crisis. Political connections were widely cited as the reason behind the state banks’ generosity (Borsuk, 1993). The benefits of political connections are not confined to debt financing. Barito Pacific’s 1993 Indonesian stock offering, for instance, was greatly helped by the state civil service pension fund acquiring a 20% stake. Barito denied allegations that it needed the pension fund’s entry to ‘‘shore up the company before it could go public,’’ but analysts noted that the fund’s investment substantially boosted the company’s capital (Borsuk, 1993). A further source of value for politically well-connected firms is the granting of important licenses. The Salim Group, one of the largest Indonesian conglomerates, had very close ties to President Suharto and was awarded lucrative franchises in banking, flour milling, and telecommunications (Shari, 1998). These anecdotes illustrate that political connections are a way to obtain low-cost financing and other economic advantages. An alternative strategy to increase firm value is to access foreign capital markets. The issuance of foreign securities can lower the cost of capital, help overcome the obstacles of segmented markets (Stulz, 1981, 1999; Errunza and Miller, 2000), and increase the firm’s value by fostering its recognition among analysts and investors (Merton, 1987; Lang et al., 2003). Some authors have also argued that U.S. cross-listings improve corporate transparency, investor protection, and hence the value of the firm to outsiders. This claim is the subject of an ongoing debate.2 Coffee (1999, 2002), Mitton (2002), Reese and Weisbach (2002), and Doidge et al. (2004) argue in favor of the hypothesis. Fanto (1996), La Porta et al. (2000), Licht (2001), and Siegel (2005) are more skeptical. In assessing the performance and governance effects of global financing strategies, it is important to understand why firms choose to issue foreign securities. The incentives to do so depend in part on the relation between the value of access to foreign capital markets and firms’ political relationships. A priori, it is not obvious whether firms with strong political connections are more or less likely to access global capital markets. As close political ties afford more attractive business opportunities, closely connected, fast-growing firms with a high demand for capital might find it particularly attractive to tap into foreign markets. Well-connected firms might also be the most interesting investment opportunities for 2

Our analysis highlights that domestic opportunities, such as political connections, are important when estimating the benefits of legal bonding, but it does not contradict or support the bonding hypothesis. We also do not analyze whether politically connected firms are more or less likely to expropriate outside shareholders, which is a separate issue (see Cheung et al., 2005, for the link between connections and expropriation).



foreign investors. These arguments suggest that close political ties and access to foreign capital markets are complements. On the other hand, there appear to be equally valid reasons to believe that firms with stronger political relationships are less likely to raise capital globally. Low-cost financing from state-owned banks makes it less attractive to tap into foreign markets, and the additional scrutiny that comes with publicly traded foreign securities could be particularly costly to firms with close political ties. Moreover, high levels of transparency and public attention might expose political favors of questionable legality. For instance, firms in weakly regulated markets are often free to engage in undisclosed related-party transactions benefiting controlling insiders and political backers. Transactions of this type would have to be reported if the firm’s securities trade on a major U.S. exchange. These arguments suggest that political connections and global financing are substitutes. In view of the conflicting arguments, the relation between firms’ financing strategies and their political connections is ultimately an empirical question. To examine this question, we analyze the likelihood of Indonesian firms having publicly traded foreign debt or equity securities. We focus on publicly traded securities because they are likely to subject firms to additional scrutiny by foreign investors, financial analysts, and the business press. We also study which firms have debt or equity securities traded on major U.S. exchanges, such as NYSE or NASDAQ. Firms with exchange-traded securities have to file Form 20-F with the SEC, which requires extensive disclosures (e.g., on relatedparty transactions) and a reconciliation of foreign financial statements to U.S. GAAP. By virtue of filing with the SEC, firms are also subject to SEC enforcement, legal threats from shareholders, and the provisions of the Foreign Corrupt Practices Act, under which most SEC enforcement actions are brought (Coffee, 2002; Siegel, 2005; Karpoff et al., 2005). Having gained a better understanding of the tradeoff between political relationships and foreign financing, we turn to the performance consequences of these strategic choices. We first ask whether firms with foreign securities outperformed other companies during the Asian financial crisis. The 1997/1998 crisis is widely seen as a crisis of confidence, and it is interesting to investigate whether firms with foreign securities outperformed other companies during this period (Mitton, 2002). Next, we look beyond the immediate crisis to study the long-run consequences of investments in political relationships. We are particularly interested in the performance of the Suharto cronies under President Abdurrahman Wahid. Wahid ran on an anti-establishment platform, promising to end cronyism in Indonesia. He ordered the arrest of Tommy Suharto, the son of the former president, and refused to pardon Tommy when he pleaded for clemency to avoid an 18month prison sentence for corruption (BBC, 2000). The Wahid presidency thus affords the opportunity to study the performance consequences of political ties when a less friendly government comes to power. 3. Sample and data

We begin our sample construction with a worldwide search for foreign securities issued by Indonesian firms. To compile our data, we use the databases of Datastream, Bloomberg, Global Access, and SDC as well as the SEC filings on Edgar, lists of foreign listings from major international stock exchanges, and the ADR lists provided by the Bank of New York and Citibank. Based on this extensive search, we identify 22 firms with


417

publicly traded debt and equity securities as of June 30, 1997, shortly before the beginning of the Asian crisis. This count does not include foreign securities that are private debt agreements or private equity placements because these arrangements allow investors to be informed via private channels rather than through public disclosure. We separately analyze these securities in Section 5. Our tests require financial statement and share price data. We obtain financial data from the Worldscope database for the 22 firms with foreign securities as well as all remaining Indonesian firms covered by Worldscope in 1997. This search results in a sample of 151 firms. We hand-collect financial information from the Indonesian Capital Market Directory (1997) and firms’ annual reports if the necessary financial data are missing in Worldscope. We lose 13 firms because we are unable to find share price data on Datastream. In addition, we drop eight firms that are not traded over our sample period. Thus, the final sample consists of 130 firms, representing over 80% of the Indonesian market capitalization in December 1996. Our measure of political connections is based on Fisman (2001). His study shows that firms that were close to Suharto suffered negative returns when bad news about the President’s health hit the stock market. Based on this result, we compute for each firm a cumulative stock return over the six health-related news events identified by Fisman. The event days are January 30–February 1, 1995; April 27, 1995; April 29, 1996; July 4–9, 1996; July 26, 1996; and April 1–3, 1997.3 The cumulative return is on average 4.6% and exhibits considerable cross-sectional variation. Some firms lost more than 20% of their value over these six events. We multiply the cumulative returns by 1 so that larger realizations indicate greater closeness to Suharto. This variable is our main proxy for political connections. Table 1 reports descriptive statistics for our sample firms. All financial statement data are measured as of the fiscal year-end in 1996. As expected, firms with publicly traded foreign securities are significantly larger than those without such securities. They are also more capital intensive and have more long-term debt. Both groups exhibit similar accounting returns on assets for fiscal 1996.

4. The choice of foreign securities

4.1. Main results We begin our analysis by studying firms’ decisions to have publicly traded foreign securities. In our empirical model, the net benefit of foreign securities yi depends on a vector of firm characteristics, X i, the closeness to the Suharto regime, C i, and industry fixed effects, ms : yi ¼ X i b þ gC i þ ms þ ei .

(1)

If firms with closer connections to Suharto are less likely to have foreign securities, we observe go0. Prior studies identify firm size and the export level of a firm’s industry as 3

For further details on the events see Fisman (2001). There are seven firms for which we do not have return data for all six events. Dropping these firms does not materially alter our results or inferences. The results are also very similar using the average rather than the cumulative return over the six events.



Table 1 Summary statistics for the variables in the analysis The table reports means and standard deviations (in parentheses) for a sample of 130 Indonesian firms and a subsample of 22 firms with publicly traded foreign equity securities as of June 1997 and foreign debt securities that mature in or after 1996. ‘‘Closeness to Suharto’’ is the log stock return over six news events indicating that President Suharto is in bad health, multiplied by –1. ‘‘Closeness to Suharto (resignation)’’ is the log stock return prior to and at Suharto’s resignation (5/12/1998–5/21/1998), multiplied by –1. ‘‘Suharto family-owned or SOE’’ is a binary variable that is equal to one if Suharto, his wife, or any of their children were important owners, or if the firm is state-owned. The variable is also coded as one if another company in our dataset, which was itself owned by the Suharto family, is listed as an important owner. Information on the seven most important owners comes from the Indonesian Capital Market Directory and from Worldscope. Firm characteristics are measured at the end of fiscal year 1996. ‘‘ROA’’ is the ratio of operating income to total assets. ‘‘Capital intensity’’ is the ratio of fixed assets to total assets. ‘‘Financial leverage’’ is the ratio of long-term debt to total assets. ‘‘Age’’ is the number of years since the firm’s incorporation. ‘‘External financing needs’’ is the firm’s actual asset growth rate minus the maximum growth rate that can be financed if a firm relies only on its internal resources and maintains its dividend, as suggested by Demirgu ¨c-Kunt and Maksimovic (2002). ‘‘President Director’’ indicates whether the head (CEO) of the managing board of directors is Chinese ( ¼ 1). ‘‘Volatility’’ is the standard deviation of the weekly stock returns during 1996. ‘‘Pre-crisis returns’’ and ‘‘Crisis returns’’ are annualized log stock returns for the periods indicated in the table. ‘‘Days with bad news from Hong Kong (Singapore)’’ is the cumulative returns on the worst five trading days for the Hang Seng Index (Strait Times Index) between January 1995 and April 1997. ‘‘Days with large exchange rate fluctuations’’ is the sum of the absolute returns over the three days with the most positive and the three days with the most negative changes in the rupiah–dollar exchange rate. ‘‘Days with large positive exchange rate fluctuations’’ is the cumulative returns over the five days with the most positive changes in the rupiah–dollar exchange rate. We denote statistically significant differences between the two subgroups as follows: y significant at 10%, *significant at 5%, **significant at 1% (using a nonparametric Wilcoxon test).

Closeness to Suharto Closeness to Suharto (resignation) Suharto family owned or SOE Total assets (millions of Rupiah) ROA Capital intensity Financial leverage External financing needs Age President Director is Chinese Volatility Pre-crisis log returns 7/1/96–6/30/97 Crisis log returns 7/1/97–8/31/98

Full sample (N ¼ 130)

Firms with foreign securities (N ¼ 22)

Firms w/o foreign securities (N ¼ 108)

0.072 (0.107) 0.109 (0.208) 0.092 (0.291) 2390 (4990) 0.068 (0.069) 0.340 (0.237) 0.190 (0.168) 0.224 (0.425) 24.346 (13.806) 0.585 (0.495) 0.062 (0.026) 0.267 (0.420) 1.353 (1.153)

0.058 (0.040) 0.009 (0.125) 0.182 (0.395) 6430 (8730) 0.070 (0.055) 0.422 (0.244) 0.292 (0.145) 0.368 (0.493) 26.182 (14.090) 0.591 (0.503) 0.060 (0.036) 0.135 (0.346) 1.077 (0.849)

0.075 0.116) 0.129 (0.216)** 0.074 (0.263) 1570 (3320)** 0.068 (0.072) 0.324 (0.233)y 0.169 (0.165)** 0.193 (0.405) 23.972 (13.784) 0.583 (0.495) 0.062 (0.024) 0.294 (0.430) 1.410 (1.200)


419

Table 1 (continued )

Cumulative returns on days With bad news from Hong Kong With bad news from Singapore With large exchange rate fluctuations With large positive exchange rate fluctuations Industry classification Agriculture Mining Manufacturing Transport Trade Finance Services

Full sample (N ¼ 130)

Firms with foreign securities (N ¼ 22)

Firms w/o foreign securities (N ¼ 108)

0.034 (0.124) 0.063 (0.100) 0.138 (0.112) 0.011 (0.075)

0.046 (0.057) 0.086 (0.083) 0.172 (0.125) 0.002 (0.066)

0.031 (0.134) 0.059 (0.103)y 0.131 (0.109) 0.013 (0.076)

0.038 (0.193) 0.015 (0.124) 0.508 (0.502) 0.062 (0.241) 0.092 (0.291) 0.238 (0.428) 0.046 (0.211)

0.045 (0.213) 0

0.037 (0.190) 0.019 (0.135) 0.500 (0.502) 0.056 (0.230) 0.093 (0.291) 0.241 (0.430) 0.056 (0.230)

0.545 (0.510) 0.091 (0.294) 0.091 (0.294) 0.227 (0.429) 0

important factors influencing the decision to cross-list shares abroad (Saudagaran, 1988; Saudagaran and Biddle, 1995; Karolyi, 1998). A basic specification therefore controls for industry effects and firm size, measured by total assets (model 1).4 Firms’ financing needs and profitability can also influence their inclination to tap into global capital markets. We add to the base model capital intensity and return on assets as proxies for financing needs and profitability, respectively (model 2). The former is computed as the ratio of fixed assets to total assets, and the latter is measured as the ratio of operating income to total assets. Another control variable, which is frequently used in the literature, is financial leverage (Healy and Palepu, 2001; Johnson and Mitton, 2003). We compute leverage as the ratio of long-term debt to total assets (model 3). The net benefit of foreign securities yi is unobserved, but we know which firms have such securities: yi ¼ 4

(

1

if yi 40

0

if yi p0

;

(2)

:

Using the market value of equity as a proxy for firm size yields even stronger results, but we use total assets because market capitalization is likely affected by firms’ financing and capital structure choices.



Given the binary nature of our dependent variable, we present probit estimates in Table 2. Standard errors in parentheses are clustered on business group affiliation to account for the possibility that within-group financing strategies might be correlated (Khanna and Palepu, 2000; we use affiliation data from Claessens et al., 2000; Fisman, 2001). The models explain a substantial fraction of the cross-sectional variation in firms’ foreign financing choices. The key result is that firms with strong political connections are less likely to have foreign securities. The estimated coefficient in the first specification of Table 2 implies that increasing closeness to Suharto by one standard deviation reduces the likelihood of a firm having foreign securities by about five percentage points. The result that political connections and foreign securities are negatively associated continues to hold in the extended models 2 and 3 where we control for firm profitability (ROA), financing needs (capital intensity), and financial leverage. The results remain similar if we use other proxies for financing needs and profitability, namely average sales growth and EBITDA over total assets (not tabulated). We also control for a firm’s average trading volume over the event days to address the concern that infrequent trading of some stocks could affect our results by biasing the Suharto measure towards zero; but including trading volume leaves our results virtually unchanged. Eleven firms in our sample are subsidiaries and affiliates of foreign firms. Because the tradeoff between domestic political benefits and foreign financing could be different for these firms, we drop them to test the robustness of our results (model 4). As before, we find that firms with better political connections are less likely to have foreign securities. A valid measure of political connections is critical for this study. To see if our results hinge on a particular definition of political connections, we examine two alternative proxies for the closeness to Suharto. The first is firms’ stock returns during the days leading up to Suharto’s resignation in 1998. The idea is that corporations close to Suharto were likely to have experienced negative returns when he resigned. It is not clear at which point the stock market expected Suharto to step down.5 To compute resignation returns, we chose to accumulate returns over the period from May 12 through May 21, 1998. We use this cumulative return, multiplied by 1, along with all our controls in model 5. We continue to find that more closely connected firms are less likely to have foreign securities. We also code a direct, family based measure for close connections. We use information on the identity of the seven most important owners of the firms in our data set to create a binary variable that is one if Suharto, his wife, or any of their children were important owners. The variable is also coded as one if another company in our data set, which was itself owned by the Suharto family, is listed as an important owner. In addition, we include state-owned enterprises because they were directly controlled by the president. As expected, companies that were owned by members of the Suharto family are less likely to have foreign securities (model 6 in Table 2). That said, there is good reason to believe that the family parameter is biased downward. The Suharto family maintained a complex web of connections and hid many of its assets. This is one of the reasons why subsequent Indonesian governments found it difficult to prove charges of cronyism and corruption (Center for Public Integrity, 2005). However, in the specification with the family indicator, 5

On May 12, student protests calling for Suharto’s resignation gained momentum and widespread support. On May 15, a wing of the ruling Golkar party called for his resignation. The upper house of the Parliament joined these calls on May 18 (Cohen, 1998; DJ Newswire, 5/18/1998). Suharto finally resigned on May 21.



Table 2 Foreign securities and political connections The table reports probit estimates of the likelihood that the 130 Indonesian firms in our sample have publicly traded foreign securities. The dependent variable takes on a value of one if the firm has foreign securities and zero otherwise. ‘‘Closeness to Suharto’’ is the log stock return over six news events indicating that President Suharto is in bad health, multiplied by –1. ‘‘Closeness to Suharto (resignation)’’ is the stock return prior to and including Suharto’s resignation (5/12/1998–5/21/1998), multiplied by –1. ‘‘Suharto family owned or state-owned enterprise’’ is a binary indicator, which is 1 if Suharto, his wife, any of their children, or another company that itself is owned by the Suharto family is one of the firm’s largest owners. The indicator is also 1 if the company is a state-owned enterprise. Firm characteristics are measured at the end of fiscal year 1996. ‘‘Firm size’’ is the log of total assets in millions of rupiah. ‘‘ROA’’ is the ratio of operating income to total assets. ‘‘Capital intensity’’ is the ratio of fixed assets to total assets. ‘‘Financial leverage’’ is the ratio of long-term debt to total assets. ‘‘Industry’’ indicators are included for agriculture, manufacturing, transport, trade, and finance. Standard errors (in parentheses) are clustered based on group affiliations reported by Fisman (2001) and Claessens et al. (2000). Models 4, 5, and 6 drop firms that are affiliates or subsidiaries of foreign firms. We denote (two-sided) levels of statistical significance as follows: y significant at 10%, * significant at 5%, ** significant at 1%. At the bottom of the table, we report two types of tests: an F -test for the hypothesis that the coefficient on the first-stage instrument is zero ( H 0: b1 (Instrument) ¼ 0), and a Smith and Blundell (1986) test with the null hypothesis that our closeness measures are exogenous. ^

(1) Closeness to Suharto Closeness to Suharto (resignation returns) Suharto-family owned or State-owned enterprise Firm size ROA Capital intensity Financial leverage Industry indicators Constant Observations Pseudo R2

4.018 (1.654)*

(2) 4.021 (1.640)*

(3) 4.618 (1.752)**

(4)

(5)

(6)

5.042 (1.742)** 1.920 (0.938)*

0.940 (0.477)* 0.852 0.837 0.806 0.793 0.635 0.829 (0.163)** (0.177)** (0.189)** (0.190)** (0.154)** (0.173)** 0.615 1.287 2.648 0.411 3.416 (2.750) (3.077) (3.793) (3.785) (3.991) 0.959 0.435 0.640 1.034 0.691 (1.031) (0.925) (0.923) (0.915) (0.948) 1.777 1.364 0.992 1.160 (0.941)y (0.911) (0.923) (0.934) Included Included Included Included Included Included 18.986 19.063 18.749 18.250 15.060 19.241 (3.425)** (3.821)** (4.152)** (4.146)** (3.471)** (3.807)** 130 0.38

130 0.39

130 0.41

119 0.41

119 0.40

119 0.38

H 0: b1 (age) ¼ 0 (Prob4F ) (Smith–Blundell test (Prob4w2))

0.009 (0.500)

0.003 (0.643)

0.015 (0.868)

0.120 (0.862)

0.095 (0.965)

N/A

H 0: b1 (Chinese) ¼ 0 (Prob4F ) (Smith–Blundell test (Prob4w2))

0.037 (0.455)

0.048 (0.415)

0.018 (0.502)

0.011 (0.706)

0.107 (0.518)

N/A

H 0: b1 (Chinese+age) ¼ 0 (Smith–Blundell test (Prob4w2))

0.015 (0.927)

0.021 (0.848)

0.010 (0.691)

0.010 (0.688)

0.125 (0.640)

N/A

^

^

^



we effectively assume that companies not owned by a member of the Suharto family are not connected. This approach erroneously classifies non-family connections as nonconnections, thereby reducing differences between connected and unconnected firms and biasing the estimated coefficient downward. This issue is less likely to arise with our returnbased proxy for political connections.

4.2. Are the results driven by susceptibility to bad news or differences in foreign exposure? One concern with our results is that the return-based closeness measure could pick up unobserved cross-sectional variation that is unrelated to firms’ political connections. One possibility is that closeness captures how strongly a firm reacts to negative market news, irrespective of whether these news items concern President Suharto or other economic events. Another issue might be that health-based returns reflect firms’ differential exposure to domestic and foreign product markets. Days on which health problems of Indonesia’s leader are reported could simply be bad days for the Indonesian economy as a whole, which have a disproportionately larger negative effect on firms with a strong domestic orientation. We address these concerns by controlling for a firm’s susceptibility to bad news and its exposure to foreign markets. First, we compute the standard deviation of weekly returns in 1996 and add this measure of historical volatility to the probit model. Firms with more volatile stocks are likely to react more strongly to any news event. Table 3 reports this model in the first column. The coefficient on volatility is positive but not significant ( p-value ¼ 0.168). More importantly, closeness to Suharto continues to be highly significant and negatively associated with foreign securities. Second, we compute cumulative returns over alternative event days with extreme negative market news. To obtain events that are unrelated to the Suharto regime, we identify the worst five non-adjacent trading days for the Hang Seng Index (Hong Kong) as well as for the Strait Times Index (Singapore) between January 1995 and April 1997.6 This is the same time period over which the health events occurred. We construct three alternative proxies that capture how strongly a firm’s stock reacts to bad news. For each sample firm, we compute its cumulative return over (a) the Hong Kong events, (b) the Singapore events, and (c) the days that register as the worst days in both Hong Kong and Singapore. The latter days appear to be Asia-wide bad news. The average cumulative returns for our sample firms are 3.8%, 7.4%, and 8.0%, respectively. Table 3 reports probit models controlling for the returns during the Hong Kong and Singapore events (Columns 2 and 3).7 The results using returns on the combined event days are similar and not reported for brevity. In all cases, we find that the coefficient on political closeness slightly increases in magnitude and significance. Thus, these tests provide no evidence that differences in firms’ responsiveness to negative news in general drive our main result. 6

Adjacent trading days are combined into one event. We make sure that the event windows do not overlap with the Suharto health events. We also examine news reports for these days. The negative returns on these days primarily reflect worldwide equity market movements or interest rate or dollar exchange rate changes. 7 There are two firms for which we cannot compute returns for the Singapore events because the worst days in Singapore were in 1995 and early 1996 before these two firms started trading on the Jakarta exchange.



Table 3 Foreign securities and political connections—robustness tests The table reports probit estimates of the likelihood that the 130 Indonesian firms in our sample have publicly traded foreign securities. The dependent variable takes on a value of one if the firm has foreign securities and zero otherwise. ‘‘Volatility’’ is computed as the standard deviation of weekly returns in 1996. The ‘‘susceptibility to bad news’’ is based on returns measured on extreme days with negative market news that are unrelated to the Suharto regime. For each sample firm, we compute the cumulative return on the worst five non-adjacent trading days for the Hang Seng Index (Hong Kong) and, separately, for the Strait Times Index (Singapore) between January 1995 and April 1997. Column 2 (3) reports the results using returns on the Hong Kong (Singapore) events. ‘‘Exposure’’ is based on currency shocks, i.e., changes in the rupiah–dollar exchange rate of more than one percent. In Column 4, we use the sum of the absolute returns over the three worst and three best days of the exchange rate. In Column 5, we compute cumulative returns over the five worst days, i.e., days when the rupiah fell against the dollar. ‘‘External financing needs’’ is the firm’s actual asset growth rate minus the maximum growth rate that can be financed if a firm relies only on its internal resources and maintains its dividend, as suggested by Demirgu ¨c-Kunt and Maksimovic (2002). The other control variables are as defined before (see Table 2 for details). We denote (two-sided) levels of statistical significance as follows: y significant at 10%, * significant at 5%, ** significant at 1%.

Closeness to Suharto Volatility

(1)

(2)

(3)

(4)

(5)

(6)

(Volatility)

(HK returns)

(SG returns)

(D Rupiah)

(D Rupiah)

(Ext. Fin.)

4.371 (1.624)** 6.269 (4.751)

4.866 5.186 (1.806)** (1.842)**

0.810 (1.541)

Susceptibility to bad news

5.822 (1.895)**

5.623 (2.219)**

3.390 (1.751)*

6.114 (2.490)*

2.756 (1.839)

Exposure External financing needs Firm size ROA Capital intensity Financial leverage Industry Constant Observations Pseudo R2

0.826 (0.183)** 1.623 (2.980) 0.376 (0.915) 1.692 (0.919)y Included 19.474 (4.031)** 130 0.41

0.801 0.777 (0.189)** (0.191)** 1.351 1.723 (2.992) (2.990) 0.466 0.543 (0.932) (0.963) 1.747 1.892 y (0.951) (1.023)y Included Included 18.664 18.332 (4.136) (4.182)** 130 0.41

5.424 (1.851)**

128 0.41

0.882 (0.202)** 1.940 (3.102) 0.011 (0.929) 1.557 (0.946)y Included 20.724 (4.393)** 130 0.44

0.809 (0.341)* 0.957 0.802 (0.212)** (0.196)** 2.016 2.559 (3.214) (3.372) 0.165 0.433 (0.970) (0.979) 2.066 2.017 (1.004)* (0.989)* Included Included 21.797 19.040 (4.575)** (4.324)** 130 0.45

126 0.43

Next, we control for differences in firms’ exposure to foreign markets and the Indonesian economy. We compute stock returns on days with currency shocks, i.e., extreme changes in the rupiah–dollar exchange rate. Returns on these days are likely to reflect cross-sectional differences in the degree to which firms are exposed to foreign and domestic product markets. We accumulate returns in two ways. First, to



capture positive and negative rate fluctuations, we sum the absolute returns on the three ‘‘worst’’ and three ‘‘best’’ days of the rupiah–dollar exchange rate in the time period during which the health events occurred. Second, we compute cumulative returns over the five days during which the rupiah weakened most significantly against the dollar. On all these days, the rupiah–dollar exchange rate changed by more than one percent. Table 3 reports probit estimates using both exposure measures as additional controls (Columns 4 and 5). In both cases, the coefficient on exposure is significant and positive. That is, firms that are more affected by changes in the rupiah–dollar exchange rate are more likely to have foreign securities. The positive coefficient on returns for days on which the rupiah fell against the dollar suggests that firms with foreign securities generally benefit from a weak rupiah, perhaps because they are more export-oriented and more engaged in foreign markets. This finding is interesting in light of the recent debate about the performance effects of debt denominated in foreign currency. The macroeconomic literature on financial crises distinguishes two effects of currency devaluations: a drop in ‘‘net worth,’’ which weakens firms’ balance sheet positions, and a competitiveness effect which improves financial performance (Krugman, 1999; Aghion et al., 2001). As in Bleakley and Cowan (2002), our results seem to imply that the competitiveness effect dominates the ‘‘net worth’’ effect. In models 4 and 5 (Table 3), controlling for the exposure to foreign markets slightly increases the negative relation between closeness to Suharto and the likelihood of having foreign securities. We also check that controlling simultaneously for historical volatility, susceptibility to bad news, and exposure produces similar results to those reported in Table 3. But even then, the closeness measure remains highly significant. A final concern is that firms with foreign securities are financially less dependent on Suharto and hence less likely to experience negative returns on days with bad news about Suharto’s health. This logic suggests that the negative association between our return-based closeness measure and foreign financing might reflect financial dependence on Suharto. Our models control for leverage, capital intensity, and industry, all of which are likely to be associated with firms’ external financing needs, thus alleviating the concern to some extent. To directly address this issue, we compute a proxy for a firm’s dependence on external capital suggested by Demirgu ¨c-Kunt and Maksimovic (2002). If connected firms with higher external financing needs are more financially dependent on Suharto, introducing a proxy for these needs will attenuate the measured effect of closeness. As expected, we find that firms with larger external financing needs are more likely to have foreign securities (Table 3, model 6). However, the coefficient on closeness remains negatively associated with foreign securities and even increases relative to Table 2. In addition, our family-based connection measure, which is not subject to this final concern, produces similar results. Thus, our result that political connections and foreign financing are substitutes does not seem to be driven by financial dependence on Suharto. 4.3. Is closeness to Suharto endogenously determined? While a firm’s exposure to foreign markets and its susceptibility to bad news do not explain the relation between political connections and global financing, there is, more generally, a concern that the closeness measure could be endogenously determined. The


425

models in Table 2 assume that firms’ political connections are predetermined. This assumption is not unreasonable because many important political connections in Indonesia appear to be family related (see model 6 in Table 2 and Backman, 2001). Similarly, for Malaysia, another country with a centralized political power structure, Johnson and Mitton (2003) argue that political connections are based on chance and have long personal histories. Nevertheless, we develop an instrumental-variable strategy to address the endogeneity concern. Smith and Blundell (1986) suggest a simple exogeneity test for models with limited dependent variables. The test involves the estimation of a first stage with closeness as the dependent variable. The residuals from the first stage are then included as an additional covariate in the models in Table 2. Under the null hypothesis of exogeneity, the first-stage residuals have no explanatory power at the second stage. The standard order condition for identification applies, so we need at least one instrument for a firm’s closeness to the regime. We focus on two instruments that have the advantage of not being choice variables. The first instrument is the firm’s age. Controlling for closeness, age has no independent effect on the firm’s likelihood of having foreign securities, but younger firms are more likely to have close political connections, possibly because they are in greater need of ‘‘political help’’ early in their lives when they establish themselves in the fairly concentrated Indonesian business environment.8 Our second instrument is the ethnicity of a firm’s president director.9 Given the delicate state of race relations in Indonesia, it is likely that Chinese managers viewed close political connections with former President Suharto in a different light than indigenous Indonesians (Pribumis). In Indonesia, political favors of questionable legality typically needed to be repaid by kickbacks and side payments of an equally dubious nature. This practice was risky for any manager, but particularly perilous for Chinese executives because their ethnicity could be used against them. The trial of Golden Key owner Tan Tjoe Hong provides an example. Accused of having fraudulently secured a $430 million letter of credit, Hong was subject to a vocal anti-Chinese campaign throughout his trial. The Far Eastern Economic Review reports that Indonesians holding anti-Chinese views were paid to attend the court hearings (McBeth, 1994). Perhaps due to such fears and pressures, Chinese managers in our dataset are on average not as close to the Suharto regime as Pribumis. To be valid instruments, firm age and the ethnicity of the president director must be correlated with political closeness but uncorrelated with the choice of foreign securities. Consistent with this requirement, we do not find evidence that firm age or the ethnicity of the president director influence firms’ choices of foreign securities other than through the channel of political connections. Whether included separately or jointly in the models of Table 2, the coefficients on the instruments remain economically

8

We thank Benny Tabalujan for suggesting this instrument and providing anecdotal evidence. Indonesian firms have a two-tiered board structure. The president director heads the managing board of directors. Hence, the role of the president director broadly corresponds to the role of the CEO. Information on the ethnicity of the president director and the dominant owner, which we use as an alternative instrument, comes from a large number of publicly available sources such as press reports and company websites. We crosschecked information with an Indonesian accounting firm, an Indonesian stockbroker, and Indonesian students at the Wharton School. Michael Backman also kindly shared his expertise in these matters. A complete list of all sources is available upon request. 9



small and statistically insignificant, whereas the closeness variable remains largely unchanged.10 At the first stage, firm age and the president director’s ethnicity are significant predictors of our closeness measure (see F -tests for the null that the coefficients on the instruments are zero, reported at the bottom of Table 2.) Our instruments are weaker when we use resignation returns as the measure for political connections. Table 2 also reports p-values for the Smith–Blundell exogeneity test. As shown, we cannot reject exogeneity using either firm age, the ethnicity of the president director, or both instruments. In all cases, the p-values are far from conventional significance levels. Assuming that either firm age or the president director’s ethnicity are valid instruments, we find no evidence that our closeness measure is endogenous to the choice of foreign securities. Overall, the results presented in Tables 2 and 3 lend reasonable support to our hypothesis that domestic opportunities influence firms’ foreign financing choices. All three measures of political connections—stock returns during Suharto’s health crises, stock returns during Suharto’s resignation, and a direct family-based measure—indicate that firms with good political connections are less likely to have publicly traded foreign securities. 5. Public and private foreign securities

In this section, we investigate whether our results are specific to the type of foreign securities we have analyzed so far. We explore this question in two ways. First, we apply a narrower definition of foreign security using only a subset of issues that are publicly traded on major U.S. exchanges. These securities require a 20-F filing with the SEC and hence come with more stringent disclosure requirements. Second, we analyze foreign securities that are private debt agreements or private equity placements. These arrangements allow investors to be informed via private channels rather than public disclosure. Contrasting the results for private and public foreign securities can shed some light on the role that informational considerations play in firms’ foreign financing decisions. We identify eight firms with U.S. debt or equity securities that require a 20-F filing with the SEC. Table 4 reports the probit estimates using the full set of controls (Column 1). Firms that are closer to Suharto are significantly less likely to have U.S. securities that require a 20-F filing. The result is essentially the same as the one for our broader classification. To determine whether the earlier estimates for the more inclusive classification are solely driven by securities that are publicly traded in the U.S., we reestimate the model in Table 2 excluding 20-F firms (Column 2).11 Closeness to Suharto is still a significant predictor of global financing choices. We also check whether the distinction between debt and equity securities is relevant to our analysis. We re-estimate our models using either foreign debt or foreign equity securities only. In these analyses, closeness to Suharto is negatively associated with both types of foreign securities. For this reason, we do not think that our findings have capital structure implications.12 10

When included in specification (3) of Table 2, the coefficient on age is 0.002 with a standard error of 0.013. The coefficient on ethnicity is 0.285 with a standard error 0.490. 11 Eliminating 20-F firms still leaves one firm in the sample that trades in the U.S. but does not file with the SEC. Dropping this firm does not alter the results reported in Column 2 of Table 4. 12 Our analysis focuses on the question where firms obtain finance and not whether it is debt or equity. Capital structure theories should therefore apply in the usual way.



Table 4 U.S. securities with 20-F filing, private foreign securities, and political connections The table reports probit estimates of the likelihood that the 130 Indonesian firms in our sample have a certain type of foreign security. In the first column, the dependent variable takes on a value of one if the firm has securities that are publicly traded in the U.S. and require a 20-F filing with the SEC, and 0 otherwise. In model (2), firms with publicly traded foreign securities that do not require a 20-F filing are analyzed. In Column 3, the binary dependent variable indicates with a value of one that the firm has private securities (e.g., loans or private equity placements) that were arranged by at least one foreign investment bank as a lead manager. In Column 4, the dependent variable is based on information about the security’s (target) marketplace indicated in the SDC database. If the security is privately placed outside of Asia or specifically classified as a ‘‘foreign private placement,’’ we set the binary variable equal to one. In the fifth column, the dependent variable takes on a value of one if the firm has private placements in the U.S. All private securities are classified based on information in the SDC database. All other variables are as defined before (see Table 2 for details). We denote (two-sided) levels of statistical significance as follows: y significant at 10% * significant at 5% ** significant at 1%.

Closeness to Suharto Firm size ROA Capital intensity Financial leverage Industry Constant Observations Pseudo R2

(1)

(2)

(3)

(4)

(5)

Public U.S. securities (20-F filing)

Foreign securities w/o 20-F

Private securities (Foreign bank)

Private securities (Foreign market)

Private U.S. securities (144a placement)

6.791 (2.734)* 0.590 (0.168)** 2.070 (4.203) 3.251 (1.036)** 2.439 (1.393)y Included 16.213 (4.128)**

3.713 (1.797)* 0.699 (0.197)** 1.356 (3.216) 0.092 (1.047) 1.581 (0.933)y Included 16.448 (4.275)**

0.259 (1.345) 0.442 (0.104)** 0.327 (1.900) 0.855 (0.661) 1.610 (0.856)y Included 8.726 (2.179)**

0.578 (1.779) 0.691 (0.110)** 0.785 (3.101) 0.212 (0.595) 0.125 (0.954) Included 15.524 (2.279)**

1.341 (1.213) 0.336 (0.097)** 2.765 (2.957) 0.147 (0.526) 1.047 (0.837) Included 9.340 (2.000)**

130 0.47

122 0.33

130 0.23

130 0.34

130 0.15

Next, we analyze if firms that were close to the Suharto regime are also less likely to have private foreign securities such as loans from foreign banks and private placements in the U.S. under Rule 144a. We obtain data from the SDC database on private securities such as term loans, revolving credit facilities, syndicated loans, and private equity placements. As private securities are not traded, it is more difficult to determine what precisely constitutes a foreign security. We create three variables. The first indicates that at least one of the lead managers arranging the private security is a foreign investment bank. We identify 64 private securities of this type. The second variable is based on the ‘‘market place’’ indicated in SDC, i.e., the region that a private placement targets (e.g., Asia, Europe, or the U.S.). Our indicator equals one if the firm’s securities are privately placed outside Asia or if the placement is specifically classified as ‘‘foreign.’’ We identify 23 securities of this type. Finally, the third variable is even more specific indicating private placements in the U.S. There are eight such securities.



Table 4 reports probit estimates for the three definitions of private securities. All specifications include the full set of controls (Columns 3–5). While the standard errors are similar in magnitude to the errors in our previous analyses, the coefficients on the private securities indicators are insignificant in all cases. Firms that were close to Suharto are as likely as firms without political ties to have private foreign securities. The contrast between the findings for the U.S. securities requiring 20-F filings (Column 1) and U.S. private placements (Column 5) is particularly interesting because the comparison holds the foreign target market constant. To confirm that the coefficients on closeness are in fact statistically different across the models for public and private foreign securities, we estimated a series of bivariate probit models (not tabulated). Relative to their probability of having publicly traded securities, politically connected firms are significantly more likely to have private securities. The differences documented in Table 4 are not the result of weak statistical power. The results in this section indicate that it matters whether a foreign security is publicly traded or privately issued, suggesting that transparency considerations do play some role in the documented tradeoff between political connections and foreign financing. While the evidence in Table 4 points to one possible interpretation of our findings, we acknowledge that the evidence is merely suggestive. The mechanism question is not central to this paper, and we leave it to future work to determine why closely connected firms are reluctant to issue publicly traded foreign securities. 6. Returns to foreign securities before and during the Asian financial crisis

Another approach to testing our hypothesis that political relationships and foreign securities are alternative means of increasing firm value is to explicitly study the performance consequences of the two strategies. This analysis also demonstrates how important it is for empirical studies to account for the endogeneity of listing decisions. We analyze the stock returns of our sample firms one year prior to and during the financial crisis of 1997 and 1998. In a financial market equilibrium, it would be surprising if firms with foreign securities consistently outperformed firms with strong political relationships. In contrast, unexpected shocks such as the financial crisis in Asia are more likely to result in significant differences in performance. The Asian crisis, which many believe was due in part to weak corporate governance and low levels of transparency (Stiglitz, 1998; Harvey and Roper, 1999), might have created a premium for more transparent firms. Johnson et al. (2000), Mitton (2002), Lemmon and Lins (2003), and Baek et al. (2004) provide evidence to this effect. However, as the results in the previous section suggest, global financing and corporate transparency are only half the story. In measuring the performance effects of foreign securities, it is important to take into account firms’ political connections and consider how the regime responded to the economic turmoil. If President Suharto lost some of his ability to support politically well-connected firms during the crisis, return regressions that ignore political connections overestimate the value of foreign listings. In contrast, if Suharto supported ‘‘his’’ firms during the crisis, the benefits of foreign securities during the crisis might be larger than previously estimated. There is ample evidence that Suharto was personally involved in propping up connected companies during the 1997–1998 turmoil. For instance, Texmaco, a large Indonesian conglomerate with excellent connections, received loans in excess of $1 billion from Bank Negara Indonesia (BNI) during the crisis.


429

Suharto personally authorized these loans that violated Indonesian lending caps (Tesoro, 1999; see also Solomon, 1999).13 To investigate the importance of having foreign securities, we estimate a series of models explaining the stock price performance of our sample firms prior to and during the Asian financial crisis. In particular, we compare models that treat the presence of foreign securities as exogenous with models that explicitly take into account that foreign securities are chosen with a view to prior political investments. We investigate the year prior to the crisis (7/1/96–6/30/97) and the crisis period itself (6/30/97–8/31/98). The latter study period is chosen to make our results directly comparable to the analysis in Mitton (2002). We use annualized log returns r i as our dependent variable so that we can compare the magnitude of the estimated coefficients across time periods. We control for firm size (measured as the log of total assets), financial leverage (ratio of long-term debt to total assets), and the historical volatility of the stock (standard deviation of weekly returns in 1996): ri ¼ X i b þ f yi þ ms þ ei .

(3)

The results in Mitton suggest that firms with foreign securities outperform other firms during the crisis, i.e., f40 The performance effects of foreign financing are reported in Table 5. In a simple OLS regression, we find a positive and significant effect during the crisis (Column 4), consistent with Mitton. In contrast, firms with foreign securities do not outperform other firms in the year prior to the crisis (Column 1), which is in line with our expectations for returns in a financial market equilibrium. To the extent that these estimates are biased, the bias is not the result of an omitted variable problem. Adding our measure of closeness to these regressions, political connections bear no significant relation to stock returns and the coefficient on foreign securities changes little. Next, we estimate treatment effects models. These models explicitly account for the substitutive relation between political connections and global financing, thereby isolating the marginal effect of foreign securities on performance. The first stage of these models is the corresponding probit model from Table 2. At the second stage, we estimate Eq. (3). For all models, we clearly reject independence of the two stages (see the Wald tests reported at the bottom of the table). Thus, it is inappropriate to run simple performance regressions to measure the impact of foreign securities on stock returns. The first set of treatment effects results are presented in Columns 2 and 5 of Table 5. The performance effects of foreign securities are considerably larger than in the simple OLS regressions. Conceptually, the difference is an estimate of the benefits that Suharto provided to well-connected firms during the crisis. Controlling for political relationships, we now find a positive performance effect of foreign securities for the year prior to the Asian crisis, underscoring that both political connections and foreign securities contribute to firm value even outside the crisis. To address the concern that our results simply reflect long-run differences in performance across firms, a second set of treatment effects models controls for firm profitability prior to the crisis (ROA) and firms’ financing needs (capital intensity) (see Columns 3 and 6, Table 5). These controls reduce the magnitude of the estimated performance effects only minimally. :

13

Connected families can also prop up their ailing conglomerate using private funds, but such behavior while interesting is a separate issue (see Tabalujan, 2002; Friedman et al., 2003; Cheung et al., 2004).



Table 5 Returns to foreign securities The table reports regression results with annualized log returns for 130 Indonesian firms as the dependent variable. ‘‘Foreign Securities’’ is an indicator equal to 1 if a firm has publicly traded foreign securities and 0 otherwise. Firm characteristics are measured at the end of fiscal year 1996. ‘‘Firm size’’ is computed as the log of total assets. ‘‘Financial leverage’’ is the ratio of long-term debt to total assets. ‘‘Volatility’’ is the standard deviation of the weekly stock returns during 1996. ‘‘ROA’’ is the ratio of operating income to total assets. ‘‘Capital intensity’’ is the ratio of fixed assets to total assets. ‘‘Industry’’ indicators are included for agriculture, mining, manufacturing, transport, trade, finance, and services. Standard errors (in parentheses) are clustered based on group affiliations reported by Fisman (2001) and Claessens et al. (2000). In the two-stage treatment effects models, the first stage is a probit model (Models 2 and 3 in Table 2, respectively). We report the result of a Wald test for the null hypothesis that the first-stage and the second-stage equations are independent ( r ¼ 0). At the bottom of the table, we test whether the effect of foreign securities on returns is different before and during the crisis ( fpre-crisis ¼ f crisis ) using pooled models. The standard errors for the pooled models are bootstrapped using 1,000 replications. The coefficient on the interaction in Column 4 compares the estimated performance effects of foreign securities in Row 1 of Columns 1 and 4 (and analogously for Columns 5 and 6). We denote (two-sided) levels of statistical significance as follows: y significant at 10%, * significant at 5%, ** significant at 1%. 7/1/96–6/30/97 Pre-crisis

7/1/97–8/31/98 Mitton (2002)

(1)

(2)

(3)

(4)

(5)

(6)

OLS

2-stage estimates

2-stage estimates

OLS

2-stage estimates

2-stage estimates

0.029 (0.100) 0.034 (0.028) 0.291 (0.284) 2.473 (1.479)y

0.452 (0.202)* 0.104 (0.037)** 0.271 (0.263) 1.711 (1.586)

0.682 (0.271)* 0.079 (0.104) 1.293 (0.659) y 1.446 (3.430)

2.350 (0.470)** 0.342 (0.122)** 0.697 (0.808) 0.559 (2.446)

Industry Constant

Included 0.608 (0.593)

Included 2.095 (0.778)**

0.440 (0.205)* 0.094 (0.037)** 0.354 (0.291) 1.607 (1.585) 0.517 (0.448) 0.197 (0.169) Included 2.008 (0.773)**

Observations R-squared H 0: r ¼ 0 ðProb 4w2 Þ

130 0.21

130

130

Foreign securities Firm size Financial leverage Volatility ROA Capital intensity

H 0: f pre-crisis ¼ fcrisis Foreign Sec.  crisis Foreign securities Crisis

0.0059

Included 2.029 (2.190)

Included 4.382 (2.290) y

130 0.22

130

0.0060 0.507 (0.336) 0.119 (0.154) 1.706 (0.135)**

2.211 (0.203)** 0.290 (0.103)** 1.720 (0.691)* 0.943 (1.968) 2.295 (1.410)y 0.820 (0.501)y Included 3.112 (2.066) 130

0.0085

0.0000

0.415 (0.269) 0.971 (0.420)* 1.691 (0.114)**

0.420 (0.279) 0.942 (0.427)* 1.691 (0.118)**


431

Finally, it is interesting to know whether the benefits of having foreign securities increased during the Asian financial crisis. A casual glance at Table 5 suggests that the performance effects of foreign securities are larger during the crisis. To formally test this hypothesis, we pool pre-crisis and crisis returns to form a panel. We control for the period by introducing a time indicator (Crisis) that equals 1 for returns during the crisis. An interaction term Foreign Securities  Crisis allows the performance effect of foreign securities to vary by period. While we include all other covariates shown in the upper half of Table 5, we only report the coefficient estimates for foreign securities, the crisis, and the interaction term at the bottom of Columns 4, 5, and 6. Bootstrapped standard errors based on 1,000 replications are given in parentheses. As expected, returns are significantly lower during the crisis. Moreover, the coefficients on the interaction term are positive in all three models. However, they are not statistically significant at conventional levels. Thus, there is at best weak evidence that the Asian crisis increased the benefits of having publicly traded foreign securities. As we have panel data, we can also estimate these models using firm fixed effects. The resulting coefficient on the interaction remains virtually unchanged, indicating that time-invariant unobserved firm characteristics do not bias our estimates. Overall, the results presented in Table 5 suggest that President Suharto lent considerable financial support to politically well-connected firms before and during the financial crisis. As a consequence, conventionally measured performance effects of cross-listings or other forms of foreign financing are considerably downward biased if political relationships are ignored. 7. Dynamics of political relationships and foreign financing

Politically well-connected firms pursue different financing strategies than less connected companies, and these strategies help explain some of the variation in performance during the Asian financial crisis. In this section, we look beyond the immediate crisis to document how politically connected firms fare over longer periods of time. We are particularly interested in the performance of connected firms after their patrons leave office. Do investments in political relationships continue to pay off even when a new government is in charge? Indonesia provides an ideal context to study this question. After Suharto was forced to step down on 21 May 1998, B.J. Habibie, a long-time Suharto ally, was immediately sworn in as Indonesia’s new president. Habibie had personally benefited from Suharto’s patronage and was widely expected to continue the Suharto policies.14 The transition from Suharto to Habibie thus affords the opportunity to study the performance consequences of political relationships when a friendly government succeeds a regime. In contrast, President Abdurrahman Wahid, who was elected on October 21, 1999, ran on an anti-establishment platform and was much more critical of the Suharto regime. 14

While small in comparison to the Suharto empire, which Time Magazine estimated to be $73 billion prior to the financial crisis and $15 billion in 1998 ( Colmey and Liebhold, 1999), the Habibie family controlled interests in chemicals, construction, real estate, transport, and communications. Many of these businesses had profited from government contracts and state-granted monopolies ( Symonds, 1998). The BBC concluded that ‘‘President Habibie owes his rise to power entirely to his close friendship with former President Suharto’’ ( Head, 1998). Suharto met Habibie in the 1950s when he was stationed on the island of Sulawesi. In 1974, Suharto asked Habibie, who was then working in Germany, to return to Indonesia and placed him in charge of the state-owned oil company. In 1978, Habibie was appointed Minister of Research and Technology, a post he held until he was endorsed as vice-president ( Symonds, 1998).



However, even if the opposition takes over, it is not obvious that firms connected to the old regime will perform worse. First, some of the benefits associated with political connections are not easy to revoke in a legal manner, such as banking and mineral extraction licenses as well as long-term loans from state-owned banks. Second, companies that decided to invest in close links with the old regime might also find it beneficial to build close relations with the new government. Ultimately, it is the nature of political relationships that critically influences how connected firms perform when the regime changes. In models of political economy, political favors are often traded like other resources: politicians with influence offer more of their services to firms that offer higher payments (Bernheim and Whinston, 1986; Grossman and Helpman, 1994; Persson, 1998). In these models the same groups will be influential independent of the government that is in office. However, in richer models of the policy process, such as in specifications with partisan preferences (Alesina, 1988), politicians are no longer perfect substitutes, interest groups might favor particular candidates, and these candidates will have incentives to repay ‘‘their’’ groups when they are in power (for surveys, see Austen-Smith, 1997; Rodrik, 1995). The price of buying political favors is then path-dependent: supporting a particular regime in one period influences the future cost of favors. The study of the performance consequences of regime changes thus implicitly tests the plasticity of political relationships. An interesting null is that regime changes do not matter because firms can invest in new regimes just as easily as they were able to build connections with the old guard. To study the performance consequences of regime changes, we collected firmlevel performance data and controls for four periods: Suharto prior to the financial crisis (7/1/96–6/30/97), Suharto during the crisis (7/1/97–5/21/98), the Habibie period (5/22/98–10/19/99), and the Wahid government (10/21/99–6/30/01). We estimate models of the form rit ¼ X it b þ gC i þ lC i Rt þ yRt þ eit ,

(4)

where the r it are annualized log returns in period t, X it is a vector of firm characteristics, C i denotes the closeness to the Suharto regime, and R it is an indicator variable for a particular Indonesian government. The omitted time period is the Suharto regime before the crisis. The specification is a difference-in-difference model. The coefficient of interest, l , measures how the difference in annualized returns that is due to connections changes with changes in the regime. A potential concern with estimates from (4) is that politically well-connected firms differ in unobserved ways from less-connected companies and that these differences are correlated with financial performance. We can difference out such time-invariant heterogeneity by estimating rit ¼ X it b þ lC i Rt þ yRt þ ni þ eit ,

(4a)

where n i is a firm fixed effect. Note that n i subsumes the main effect of closeness on stock performance. The results for model (4) are reported in Table 6. There is clear evidence of path dependence. Compared to the pre-crisis period, firms with close connections to Suharto fared worse during the crisis when Suharto was in power and during the Wahid government. Under Habibie, connected firms performed as well as they did prior to the crisis. Model 4 of Table 6, where we include all three regimes, confirms that, relative to the pre-crisis period, the Habibie government was the only period during which



Table 6 Returns to political connections under different governments The table reports regression results with annualized log returns for 130 Indonesian firms as the dependent variable. ‘‘Closeness to Suharto’’ is the log stock return over six news events indicating that President Suharto is in bad health, multiplied by –1. There are three indicators for different periods. ‘‘Suharto Crisis’’ is the period from 7/1/97 to 5/21/98. Habibie is an indicator for the time period during which President Habibie was in power, 5/22/98–10/19/99. Another indicator denotes the government of President Wahid, 10/21/99–6/30/01. The omitted period is 7/1/96–6/30/97, a pre-crisis period during which President Suharto ruled Indonesia. ‘‘Firm size’’ is computed as the log of total assets. ‘‘Financial leverage’’ is the ratio of long-term debt to total assets. ‘‘Volatility’’ is the standard deviation of the weekly stock returns. ‘‘Industry’’ indicators are included for agriculture, mining, manufacturing, transport, trade, finance, and services. Model (5) includes firm fixed effects. Standard errors (in parentheses) are clustered based on group affiliations reported by Fisman (2001) and Claessens et al. (2000). At the bottom of the table, we report p-values from F -tests of the null hypothesis that the sum of the ‘‘Closeness to Suharto’’ variable and its interaction with a particular time indicator is zero. We denote (two-sided) levels of statistical significance as follows: y significant at 10%, * significant at 5%, ** significant at 1%. Pre-crisis vs. Suharto crisis

Pre-crisis vs. Habibie

Pre-crisis vs. Wahid

All periods

7/1/96–6/30/97 & 7/1/96–6/30/97 & 7/1/96–6/30/97 & 7/1/96–6/30/2001 7/1/97–5/21/98 5/22/98–10/19/99 10/21/99–6/30/01 & 7/1/96–6/30/97 (omitted) (1)

Closeness to Suharto Closeness  Suharto crisis Closeness  Habibie

0.952 (0.530) 2.278 (0.774)**

(2)

0.654 (0.570)

1.809 (0.762)* 0.766 (0.102)** 0.057 (0.112)

Habibie

0.046 (0.031) 0.330 (0.265) 4.169 (1.477)** 1.358 (0.583)* 0.363 (0.217) Included 1.648 (0.734)*

0.066 (0.028)* 0.222 (0.144) 4.433 (0.782)** 0.513 (0.204)* 0.316 (0.186) Included 0.908 (0.630)

0.473 (0.109)** 0.117 (0.028)** 0.278 (0.168) 1.402 (0.902) 1.293 (0.373)** 0.027 (0.181) Included 2.320 (0.653)**

251 0.50 0.0256

233 0.25 0.5274

225 0.48 0.1170

Wahid Firm size Financial leverage Volatility ROA Capital intensity Industry Constant

Observations R-squared p-values for H 0: close+close  period ¼ 0

0.956 (0.550)

0.315 (0.766)

Closeness  Wahid Suharto crisis

(3)

(4)

0.809 (0.578) 2.241 (0.855)** 0.248 (0.843) 2.063 (0.866)* 0.905 (0.108)** 0.184 (0.119) 0.666 (0.116)** 0.076 (0.022)** 0.081 (0.128) 0.834 (0.677) 0.480 (0.201)* 0.023 (0.157) Included 1.792 (0.524)** 488 0.43 0.0265 0.0925 0.0574

(5) Firm-fixed effects

2.258 (0.949)* 0.093 (0.930) 2.226 (0.965)* 0.921 (0.128)** 0.080 (0.152) 0.716 (0.145)** 0.090 (0.116) 0.198 (0.209) 2.310 (0.960)* 0.268 (0.278) 0.030 (0.363) – 1.899 (2.364)

488 0.44



Suharto-connected firms did as well as before the crisis. Controlling for unobserved timeinvariant firm characteristics as in (4a), we find the same pattern of results as in our previous models (model 5). Were political connections a good long-term investment in Indonesia? The bottom row of Table 6 reports p-values from F -tests of the null hypothesis that connections had no influence on financial performance (g+l ¼ 0). We find that connected firms underperformed during the crisis when Suharto was still in power, but they outperformed less connected firms during the Habibie period. Once Wahid came to office, however, it was not easy for the Suharto cronies to re-establish profitable political ties and the performance of their companies suffered correspondingly. Anecdotal evidence is consistent with the patterns in the data. For example, Texmaco’s chairman, Marimutu Sinivasan, who had received huge loans during the Suharto period, maintained close connections with Habibie. He was named Special Business Envoy for Asia, became vice treasurer for the ruling party, and received lucrative orders from the Indonesian army. Wahid, in contrast, ordered an investigation of Texmaco’s finances and insisted the case be settled ‘‘under the due process of law’’ (Tesoro, 1999). Having lost their political support, did the Suharto cronies find it more attractive to issue foreign securities? Such behavior would be consistent with our results in Table 2. We begin analyzing this conjecture by counting the firms with foreign securities under Suharto and under Wahid. In Table 7 (upper panel), we document that there is considerable variation over time. Of the 21 companies that had publicly traded securities in 1997, five gave up their securities, and 12 companies added foreign securities to their financing strategy. The middle panel in Table 7 looks at changes in financing for closely connected and less-connected firms. A firm is classified as being close to Suharto if its closeness index exceeds the median value of the sample. Nine closely connected but only three other firms issued new foreign securities under the Wahid government. This is our first evidence that Suharto firms responded to the change in government by adopting a more outwardoriented financing strategy. In the bottom panel of Table 7, we investigate the link between past political connections and current financing strategies more formally. The first specification is a probit model that is estimated on the sample of firms that did not have foreign securities during the Suharto period and were still active in 2001 ðN ¼ 101Þ. The dependent variable takes on a value of one if a company has publicly traded foreign securities in 2001 and zero otherwise. Controlling for firm-specific characteristics, we find that firms that were closer to Suharto have a greater tendency to issue foreign securities under Wahid. A one-standard-deviation increase in closeness raises the likelihood of having new securities by 2.1 percentage points, an effect that is statistically significant at the 6% level. The second model in Table 7 reports results for all firms that are included in the original sample and are still active in 2001 ðN ¼ 122Þ. This model uses information on firms that adopt new securities and information on firms that give up foreign securities during the Wahid government. The dependent variable is ordered. Adopting new foreign securities increases exposure to foreign investors and giving up securities has the opposite effect. Thus, we estimate an ordered probit model. The dependent variable is 1 if the firm no longer has a foreign security, 0 if there is no change, and 1 if the company issued a new security. We find that firms that were close to Suharto are more likely to adopt an outward-oriented financing strategy under Wahid. Similar to our previous result, the estimated coefficient on the closeness variable implies that a one-standard-deviation


435

Table 7 New issues of foreign securities after regime changes The first panel shows the number of firms with foreign securities in 1997 under Suharto and in 2001 under Wahid. In the second panel, we report changes in securities for closely connected and less connected firms, respectively. A firm is considered close to Suharto if its ‘‘Closeness Index’’ exceeds the median value for our sample. The bottom panel reports two specifications. The first is a probit model for the 101 firms that did not have foreign securities in 1997. The dependent variable takes on a value of 1 if the firm had a publicly traded foreign security in 2001 and zero otherwise. ‘‘Closeness to Suharto’’ is the log stock return over six news events indicating that President Suharto is in bad health, multiplied by –1. ‘‘Firm size’’ is the log of total assets in millions of rupiah. ‘‘ROA’’ is the ratio of operating income to total assets. ‘‘Capital intensity’’ is the ratio of fixed assets to total assets. ‘‘Financial leverage’’ is the ratio of long-term debt to total assets. ‘‘Industry’’ indicators are included for agriculture, manufacturing, trade, and finance. Standard errors (in parentheses) are clustered based on group affiliations reported by Fisman (2001) and Claessens et al. (2000). The second specification is an ordered probit model. The dependent variable is 1 (firm no longer has a foreign security), 0 (no change in foreign security), or 1 (issued new security), respectively. The sample consists of all firms from our original sample that survived. The covariates are as described above. We denote (two-sided) levels of statistical significance as follows: y significant at 10%, * significant at 5%, ** significant at 1%. Securities under Suharto and Wahid

Firms with foreign securities in 1997 Firms without foreign securities in 1997

Firms with foreign securities in 2001 16

Firms without foreign securities in 2001 5

12

89

Changes in securities, by strength of connections

Issued new foreign security No change Gave up foreign security

Firms close to Suharto 9 50 2

Other Firms 3 55 3

Securities and closeness

Closeness to Suharto Firm size ROA Capital intensity Financial leverage Industry Constant Observations Pseudo R2

Foreign securities in 2001 (but not in 1997) 3.299 (1.756)y 0.726 (0.250)** 5.070 (3.522) 0.466 (0.734) 2.698 (1.426)y Included 17.128 (5.660)**

Changes in new securities 1.950 (0.825)* 0.170 (0.103)y 1.271 (2.173) 0.221) (0.774 2.420 (0.959)** Included

101 0.432

122 0.139



increase in closeness increases the likelihood of having new foreign securities by 2.5 percentage points. This effect is statistically significant at the 1.8% level. The results in Tables 6 and 7 provide support for the view that political regime changes have significant consequences for the performance of companies that invested in political relationships and for the strategies that these companies adopt once a new government is in power. They also corroborate our earlier findings on the link between global financing and political connections. 8. Conclusions

In this study, we examine the role of political connections for firms’ financing strategies and their long-run financial performance. We view political connections as an example of domestic arrangements that reduce the benefits of global financing. Consistent with this argument, our first set of results shows that well-connected firms are less likely to have publicly traded debt or equity securities abroad, suggesting that connections and global financing are substitutes. We also provide return-based evidence that firms derive significant benefits from both foreign securities and political connections before and during the Asian crisis. For this reason, it is important to consider domestic arrangements, such as political connections, when estimating the performance benefits of global financing. There are at least three explanations for our findings. First, well-connected firms have access to preferential financing at home and therefore do not need to access foreign capital markets. Second, firms with political ties dislike the transparency and scrutiny that come with publicly traded securities. Third, foreign securities make it more difficult for insiders to extract private control benefits. Although the exact mechanism is not central to our study and conclusions, we make one attempt to shed some light on the question of whether the transparency consequences of foreign securities play a role in our findings. We show that Indonesian firms with close ties to the Suharto regime were as likely as non-connected firms to have privately arranged foreign securities, which allow access to global capital markets but do not come with the same outside scrutiny as publicly traded securities. Thus, the distinction between private and public securities appears to matter. While this finding is not a full explanation of why connected firms are less likely to have foreign securities, it is consistent with the view that the informational consequences of publicly traded securities play some role in the documented tradeoff between political connections and foreign financing. Our second set of results documents the financial consequences of regime changes for firms with strong political ties to Suharto. We find that closely connected firms underperformed during the Asian financial crisis as long as Suharto was in power, recovered under Suharto’s long-time ally Habibie, but significantly underperformed under Wahid, a cleric critical of the Suharto regime. This return pattern suggests that firms have difficulty re-establishing connections with a new government when their patrons fall from power. Given these difficulties, we document that Suharto firms are more likely to issue publicly traded foreign securities after Wahid’s surprise election. This evidence further supports the link between global financing and political connections. Two broader conclusions emerge from our findings. First, the results shed light on the difficulties of institutional reform and capital market liberalization in emerging market economies like Indonesia. Well-connected firms do not find global financing very attractive. As a result, the opening up of capital markets is likely to remain limited in


437

economies where political connections remain important (Stulz, 2005). To the extent that foreign financing strengthens the competitive position of less connected firms, firms with strong political connections can be expected to resist changes in domestic institutions that facilitate global financing, such as increases in corporate transparency (Rajan and Zingales, 2003). Institutional reform in this environment promises to be particularly difficult because it is the firms with political clout that prefer less financial liberalization (Chui et al., 2000). A second conclusion relates to research on firms operating in relationship-based economic systems. A growing literature investigates the performance effects of adopting corporate strategies that are consistent with the Anglo-Saxon model of arm’s-length finance. In many of these analyses, firms that pursue market-based strategies are compared to firms that do not. To make valid empirical inferences, however, it is important to recognize that these decisions are likely to be endogenously determined. In a relationshipbased economy, firms with weak connections (e.g., to political regimes or banks) have the strongest incentives to rely on market-based transactions. Unless this is taken into account, the debate about the performance and valuation effects of greater corporate transparency and improved governance is likely to be misinformed.

References Aghion, P., Bacchetta, P., Banerjee, A., 2001. Currency crises and monetary policy in an economy with credit constraints. European Economic Review 45 (7), 1121–1150. Alesina, A., 1988. Credibility and political convergences in a two-party system with rational voters. American Economic Review 78, 796–805. Austen-Smith, D., 1997. Interest groups: money, information, and influence. In: Mueller, D. (Ed.), Perspectives on Public Choice: A Handbook. Cambridge, pp. 296–321. Backman, M., 2001. Asian Eclipse: Exposing the Dark Side of Business in Asia. Wiley, Singapore. Baek, J., Kang, J., Park, K., 2004. Corporate governance and firm value: evidence from the Korean financial crisis. Journal of Financial Economics 71 (2), 265–314. Baker, K., Nofsinger, J., Weaver, D., 2002. International cross-listing and visibility. Journal of Financial and Quantitative Analysis 37, 495–521. Baron, D., 2001. Theory of strategic nonmarket participation: majority-rule and executive institutions. Journal of Economics & Management Strategy 10 (1), 47–89. BBC News, 2000. No pardon for Tommy Suharto. At: /http://news.bbc.co.uk/1/hi/world/asia-pacific/ 955578.stm S, accessed on 5 March 2005. Bekaert, G., Harvey, C., Lumsdaine, R., 2002. Dating the integration of world equity markets. Journal of Financial Economics 65 (2), 203–247. Bernheim, B., Whinston, M., 1986. Menu auctions, resource allocation, and economic influence. The Quarterly Journal of Economics 101 (1), 1–31. Bleakley, H., Cowan, K., 2002. Corporate dollar debt and depreciations: much ado about nothing? Working paper 02-5. Federal Reserve Bank of Boston. Borsuk, R., 1993. Outcry unlikely to hinder Barito Pacific’s stock issue. The Asian Wall Street Journal 07/16/ 1993, 1. Center for Public Integrity, 2005. Indonesia, Corruption Timeline. At: /http://www.public-i.org/ga/country. aspx?cc=id&act=timelineS, accessed on 1 May 2005. Cheung, S., Rau, R., Stouraitis, A., 2004. Tunneling, propping and expropriation: evidence from connected party transactions in Hong Kong. Unpublished working paper. City University of Hong Kong and Purdue University. Cheung, S., Jing, L., Rau, R., Stouraitis, A., 2005. Guanxi, political connections, and expropriation: the dark side of state ownership in Chinese listed companies. Unpublished working paper. City University of Hong Kong and Purdue University.



Chui, A., Titman, S., Wei, K., 2000. Corporate groups, financial liberalization and growth: the case of Indonesia. Working paper. Hong Kong Polytechnic University, Hong Kong. Claessens, S., Djankov, S., Lang, L., 2000. The separation of ownership and control in East Asian corporations. Journal of Financial Economics 58 (1–2), 81–112. Coffee, J., 1999. The future as history: the prospects for global convergence in corporate governance and its implications. Northwestern University Law Review 641–708. Coffee, J., 2002. Racing towards the top? the impact of cross-listings and stock market competition on international corporate governance. Columbia Law Review 102 (7), 1757–1832. Cohen, M., 1998. To the barricades: as economic hardship mounts, student protests are gaining support from other segments of society—and starting to look like a nationwide movement. Far Eastern Economic Review 5/14/1998, 23. Colmey, J., Liebhold, D., 1999. The family firm—Suharto Inc. Time Asia, May 24, 12. Demirgu ¨ c-Kunt, A., Maksimovic, V., 2002. Funding growth in bank-based and market-based financial systems: evidence from firm-level data. Journal of Financial Economics 65, 337–363. Doidge, C., Karolyi, G., Stulz, R., 2004. Why are foreign firms listed in the U.S. worth more? Journal of Financial Economics 71 (2), 205–238. Errunza, V., Miller, D., 2000. Market segmentation and the cost of capital in international equity markets. Journal of Financial and Quantitative Analysis 35 (4), 577–600. Faccio, M., 2002. Politically connected firms: can they squeeze the state? Unpublished working paper. University of Notre Dame, Indiana. Fanto, A., 1996. The absence of cross-cultural communication: SEC mandatory disclosure and foreign corporate governance. Journal of International Law and Business 17, 119–207. Fisman, R., 2001. Estimating the value of political connections. American Economic Review 91 (4), 1095–1102. Friedman, E., Johnson, S., Mitton, T., 2003. Propping and tunneling. Journal of Comparative Economics 31 (4), 732–751. Grossman, G., Helpman, E., 1994. Protection for sale. American Economic Review 84 (4), 833–850. Hail, L., Leuz, C., 2004. Cost of capital and cash flow effects of U.S. cross listings. Unpublished working paper. University of Pennsylvania. Harvey, C., Roper, A., 1999. The Asian bet. In: Harwood, A., Litan, R., Pomerleano, M. (Eds.), Financial Markets and Development: The Crisis in Emerging Markets. Washington, DC, pp. 29–115. Head, J., 1998. Profile: President BJ Habibie. BBC News. At /http://news.bbc.co.uk/1/hi/events/indonesia/ profiles/98092.stm S, accessed on 1 August 2005. Healy, P., Palepu, K., 2001. Information asymmetry, corporate disclosure, and the capital markets: a review of the empirical disclosure literature. Journal of Accounting and Economics 31, 405–440. Johnson, S., Mitton, T., 2003. Cronyism and capital controls: evidence from Malaysia. Journal of Financial Economics 67 (2), 351–382. Johnson, S., Boone, P., Breach, A., Friedman, E., 2000. Corporate governance in the Asian financial crisis. Journal of Financial Economics 58, 141–186. Karolyi, G., 1998. Why do companies list their shares abroad? a survey of the evidence and its managerial implications. Salomon Brothers Monograph Series 7(1), New York University, New York. Karpoff, J., Lee, D., Martin, G., 2005. The cost to firms of cooking the books. Unpublished working paper. University of Washington and Texas A&M University. Khanna, T., Palepu, K., 2000. Is group affiliation profitable in emerging markets? an analysis of diversified Indian business groups. Journal of Finance 55 (2), 867–891. Kroszner, R., Strahan, P., 1999. What drives deregulation? economics and politics of the relaxation of bank branching restrictions. Quarterly Journal of Economics 114 (4), 1437–1467. Krugman, P., 1999. Balance sheets, the transfer problem, and financial crises. International Tax & Public Finance 6 (4), 459–472. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2000. Investor protection and corporate governance. Journal of Financial Economics 58, 3–27. Lang, M., Lins, K., Miller, D., 2003. ADRs, analysts, and accuracy: does cross listing in the U.S. improve a firm’s information environment and increase market value? Journal of Accounting Research 41, 317–362. Lemmon, M., Lins, K., 2003. Ownership structure, corporate governance, and firm value: evidence from the East Asian financial crisis. Journal of Finance 58, 1445–1468.


439

Licht, A., 2001. Managerial opportunism and foreign listing: some direct evidence. University of Pennsylvania Journal of International Economic Law 22, 325–348. McBeth, J., 1994. Banking on friends: business and politics mix in Bapindo case. Far Eastern Economic Review 25. Merton, R., 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance 42, 483–510. Mitton, T., 2002. A cross-firm analysis of the impact of corporate governance on the East Asian financial crisis. Journal of Financial Economics 64 (2), 215–242. Persson, T., 1998. Economic policy and special interest politics. Economic Journal 108 (447), 310–327. Rajan, R., Zingales, L., 2003. The great reversals: the politics of financial development in the twentieth century. Journal of Financial Economics 69 (1), 5–51. Reese, W., Weisbach, M., 2002. Protection of minority shareholder interests, cross-listings in the United States, and subsequent equity offerings. Journal of Financial Economics 66, 65–104. Rodrik, D., 1995. Political economy of trade policy. In: Grossman, G., Rogoff, K. (Eds.), Handbook of International Economics, vol. 3. Elsevier, Amsterdam, pp. 1457–1494. Saudagaran, S., 1988. An empirical study of selected factors influencing the decision to list on foreign stock exchanges. Journal of International Business Studies, 101–127. Saudagaran, S., Biddle, G., 1995. Foreign listing location: a study of MNCs and stock exchanges in eight countries. Journal of International Business Studies, 319–341. Shari, M., 1998. A tycoon under siege. Business Week. 28/09/1998, p. 26. Siegel, J., 2005. Can foreign firms bond themselves effectively by renting U.S. securities laws? Journal of Financial Economics 75, 319–359. Smith, R., Blundell, R., 1986. An exogeneity test for a simultaneous equation Tobit model with an application to labor supply. Econometrica 54 (4), 679–686. Solomon, J., 1999. Indonesia to investigate Suharto over state-bank loan to firm. The Wall Street Journal 12/01/ 1999, A23. Stigler, G., 1971. The theory of economic regulation. Bell Journal of Economics and Management Science 2, 3–21. Stiglitz, J., 1998. The role of international financial institutions in the current global economy. Address to the Chicago Council on Foreign Relations, February 27. Stulz, R., 1981. On the effects of barriers to international investment. Journal of Finance 36, 923–934. Stulz, R., 1999. Globalization, corporate finance, and the cost of capital. Journal of Applied Corporate Finance 26, 3–28. Stulz, R., 2005. The limits of financial globalization, presidential address at the AFA meetings. Working paper. Ohio State University. Symonds, P., 1998. Who is B.J. Habibie? At: /http://www.wsws.org/news/1998/may1998/hab-m22.shtml S, accessed on 12 March 2005. Tabalujan, B., 2002. Why indonesian corporate governance failed—conjectures concerning legal culture. Columbia Journal of Asia Law, Spring 2002. Tesoro, J.M., 1999. Business scandals: past versus future. AsiaWeek, 24 December. Wiwattanakantang, Y., Kali, R., Charumlind, C., 2006. Connected lending: Thailand before the financial crisis. Journal of Business 79 (1), 181–217.

Leuz - Gee 2006 JFEC.pdf

Recommend Documents