گروه های کسب و کار بازار سرمایه داخلی : شواهدی از وام درون گروهی

We study business groups’ internal capital markets using a unique data set on intra-group lending in Chile (1990–2009). In line with groups’ financing advantage, firms that borrow internally have higher investment, leverage, and return on equity (ROE) than other firms. At the margin, controlling shareholders have higher cash-flow rights in borrowing firms than in lending firms. However, there is no robust evidence of minority shareholders losing out from intra-group loans as tunneling predicts. Our evidence is consistent with the idea that strict regulation and disclosure requirements for intra-group loans, which are features of the Chilean market, reduce the risk of expropriation in pyramids.

Business groups are prevalent around the world. They are typically controlled by families and organized as pyramidal ownership structures. The literature describes bright and dark sides to business groups (Khanna and Yafeh, 2007). The bright side of business groups is their ability to overcome market frictions. For instance, firms with limited access to intermediated funds can benefit from the support of the rest of the group when they suffer negative cash-flow shocks (Gopalan et al., 2007 and Khanna and Yafeh, 2005). The dark side of business groups refers to the potential expropriation of minority shareholders. This behavior is called tunneling (Johnson, La Porta, López-de-Silanes, and Shleifer, 2000). For example, Bertrand, Mehta, and Mullainathan (2002) show that Indian groups channel resources away from firms in which the controlling shareholder has low cash-flow rights toward firms in which he has high cash-flow rights, although Siegel and Choudhury (2012) question the robustness of their result using the same data set of Indian firms. More generally, recent research suggests that the financing advantage of business groups outweighs tunneling problems (Masulis, Pham, and Zein, 2011), and thus, the debate about the advantages and disadvantages of business groups remains open. Despite their ubiquity, the inner workings of business groups remain relatively unexplored because of data limitations. In this paper we have direct access to internal capital markets, which represent a key dimension of business groups. We assemble a unique data set that covers all intra-group lending of Chilean firms. This information must be disclosed as a separate line in the balance sheet of every firm according to Chilean regulation.1 The fact that we have almost two decades of data (1990–2009) is an advantage over other studies with more limited samples.2 Furthermore, we can extract crucial details about intra-group lending from the notes to financial statements, which are available in electronic form for the years 2001–2009. In the notes each firm reports a firm-by-firm loan balance with other affiliated firms. This means that we can identify specific lending relationships, i.e., who is borrowing and from whom, and not only the net borrowing or lending position of each firm within the group. As far as we know, there is no other public data set with such a comprehensive coverage of intra-group lending in other countries. As with business groups in general, intra-group lending can be motivated by tunneling or by a financing advantage. We develop the main predictions of these two hypotheses and then test them with our data. Our results suggest that, although tunneling can be a concern at the margin, the activity of internal capital markets is better explained by the financing advantage. The regulatory oversight and disclosure requirements of the Chilean market, which are stricter than in other emerging markets such as China or India, probably stack the cards against the tunneling hypothesis in this sample. Our results are consistent with the idea that appropriate regulation can reduce the risk of expropriation in the internal capital markets of pyramids without canceling their financing benefits. However, we cannot fully discard other causes (e.g., culture), because, at the end of the day, we observe only one type of regulation (with minor adjustments) throughout the entire sample period. More generally, our results show that the link between pyramidal groups and the expropriation of minority shareholders is not an unquestionable axiom (see also Almeida and Wolfenzon, 2006a and Khanna and Yafeh, 2007). As Siegel and Choudhury (2012) put it, our evidence points towards a role for business groups beyond being mere “expropriation devices.” Pyramids not only allow large shareholders to achieve control while investing relatively little capital; they also give large shareholders access to more cash flows than the ones directly produced by a firm, which in a world of imperfect capital markets can be an important advantage. In terms of the tunneling hypothesis, we first study the impact of the controlling shareholder's cash-flow rights on the direction of intra-group loans. Cash-flow rights (CFRs) represent the proportional claim of the controlling shareholder to the dividends of the firm (Adams and Ferreira, 2008). The tunneling hypothesis predicts that loans go from firms in which the controlling shareholder has low CFRs toward firms in which he has high CFRs. However, the key ingredient for tunneling is that minority shareholders of the lending firm are harmed by this opportunistic behavior of the controlling shareholder. In particular, we expect to find lower-than-normal return on equity (ROE) and lower-than-normal dividends in connection to intra-group loans. Without this evidence the abusive behavior of the controlling shareholder cannot be asserted. In the data we find that firms that are net receivers of intra-group loans have, at the margin, higher CFRs than firms that are net providers of intra-group loans. On average, there is a ten percentage-point difference in CFRs between receivers and providers. However, there are receivers and providers of intra-group loans in all layers of control pyramids. Neither are all receivers at the top of the pyramid nor all providers at the bottom, which would happen if loans were only intended to siphon resources away from minority shareholders. On the contrary, lending relationships are generally created between firms that are close to each other in the control pyramid. Furthermore, product market relationships (e.g., firms in the same industry or firms in closely integrated industries) predict the creation of lending relations. In other words, considerations beyond the direct benefit of the controlling shareholder also play a role. When we look at ex post outcomes we do not find robust evidence of lower ROE in providers of intra-group loans when compared to other firms in pyramids and to firms that do not belong to pyramids. There is also no evidence that providers pay lower dividends. Based on our results, it is hard to make the case for tunneling as the primary reason for internal capital markets in pyramids. The financing advantage hypothesis has two key predictions. First, investment and ROE have to increase in the borrowing firm, because alleviating financial constraints is the main reason for the intra-group loan. Second, internal loans affect the capital structure of the firm, because financial contracting within the group is easier, and consequently less expensive, than with outside intermediaries. Since intra-group lending is only available to affiliated firms, firms that receive loans should be over-levered when compared to similar firms outside groups. At the same time, these firms should not only increase leverage, but also replace expensive external debt (i.e., bank debt, bonds, etc.) with cheap loans from related firms. We find in the data that receivers of intra-group loans are typically small, capital-intensive firms with higher investment rates than providers. This difference in investment rates is statistically significant and robust to multiple regression specifications. There is also evidence that ROE in receivers goes up, although the effects are marginally significant. We find that receivers have leverage ratios that are 7%–10% higher than leverage ratios of other firms, while there is no discernible effect on the leverage ratio of providers. External leverage (external debt over total assets) is about 6% lower in receivers, which speaks of a strong substitution between external debt and intra-group loans in these firms. Overall, the evidence is consistent with the financing advantage hypothesis. There is a vast literature on business groups, although less so on intra-group loans. In one of the recent studies, Jiang, Lee, and Yue (2010) argue that the controlling shareholders of Chinese groups use these loans mostly for tunneling. On the other hand, Gopalan, Nanda, and Seru (2007) argue that intra-group loans in India obey the financing advantage logic, as these loans reduce the likelihood of bankruptcy in affiliated firms. We contribute to these papers by showing that, while tunneling can be a concern at the margin, groups also use intra-group lending to alleviate financial constraints and increase investment in some firms. Other papers in the literature study the financing advantage and tunneling aspects of business groups in a variety of settings. Tunneling has been shown in private placements of equity (Baek, Kang, and Lee, 2006), mergers and acquisitions (Bae et al., 2002 and Cheung et al., 2006), dilutive equity offerings (Atanasov, Black, Ciccotello, and Gyoshev, 2010), privatizations (Atanasov, 2005), dividend payments (Faccio, Lang, and Young, 2001), and board compensation (Urzúa I., 2009). In terms of the financing advantage, Almeida, Park, Subrahmanyam, and Wolfenzon (2011) show that groups use internal revenues to set up or acquire capital-intensive firms, which are more likely to be constrained in financial markets (see also Belenzon, Berkovitz, and Rios, 2013). Similarly, Gopalan, Nanda, and Seru (2013) find that firm investment is partly financed by the dividends of other firms in the group. Many papers explore the investment-cash flow sensitivities of group and non-group firms (Hoshi et al., 1991, Lee et al., 2009 and Shin and Park, 1999). Recently, Almeida and Kim (2012) study the investment of Korean chaebol and non-chaebol firms after the Asian financial crisis. They argue that the differences in investment can be attributed to the active use of internal capital markets. Because of data availability, most of these papers focus on the outcome of internal capital markets (investment), rather than the underlying flows of capital. Testing the mechanism with data like ours is important for understanding the differences in behavior between group and non-group firms. We also make a contribution to the literature on internal capital markets by focusing on the interplay between internal and external capital markets. While the theoretical literature on internal capital markets is focused on the case of conglomerates with fully owned subsidiaries [see Stein (2003) for a survey], firms in business groups can be separately listed in the stock market and issue debt independently. As a consequence, the internal capital markets of business groups can have capital structure effects that are absent from conglomerates. The presence of debtholders and minority shareholders in different layers of the control pyramid also implies that the welfare and efficiency implications of intra-group loans are more complicated than in conglomerates. For example, it could be optimal, from the standpoint of investment efficiency, to engage in winner picking in conglomerates (Stein, 1997), but the redistribution of funds between firms in a business group can benefit a group of debtholders and minority shareholders at the expense of others. Even if firms that provide loans have poor investment opportunities, and therefore relatively low ROE, it is not clear that making intra-group loans is justified. The minority shareholders of those firms can prefer an increase in dividends rather than being lenders to other firms.3 The rest of the paper is organized as follows. In Section 2 we describe the main hypotheses and empirical predictions. In Section 3 we describe our data in detail. Section 4 presents the main results for the provider-receiver status, and the regressions using pairwise balances of intra-group loans. In Section 5 we look at the effect of the provider-receiver status on firm outcomes. Section 6 describes briefly the regulation of the Chilean market in comparison to that of other emerging markets. Section 7 presents our conclusions.

We study business groups’ internal capital markets using a unique data set on intra-group lending in Chile between 1990 and 2009. Lending relationships are formed by firms that are close to each other in the control pyramid. Also, firms that belong to the same industry, and firms in more integrated industries, tend to form lending relationships. At the margin, firms in which controlling shareholders have high cash-flow rights receive more loans than other firms. However, loans do not typically go from the bottom of the control pyramid straight to the top of the pyramid, as tunneling suggests. Capital-intensive and small firms tend to receive more intra-group loans than other firms. The receivers of intra-group loans increase investment strongly, and to a lesser extent ROE and extraordinary dividends. Receivers are more levered than other firms, but have lower external leverage. On the other hand, there is little evidence that providers of intra-group loans behave differently than other firms except for the fact that they invest less and they have higher external leverage. Overall, this evidence is consistent with the financing advantage hypothesis, but less so with tunneling. The key missing link in favor of tunneling is a clear sign of underperformance in firms that provide internal loans. In other words, there is no conclusive evidence that the controlling shareholder hurts minority shareholders of the lending firm by directing intra-group loans to firms in which he has high cash-flow rights. The interpretation under the financing advantage hypothesis is that business groups allow firms to increase their debt levels beyond the levels permitted by financial markets. Subject to less stringent financial constraints, these firms can invest more and increase ROE and dividends. The Chilean regulation of intra-group loans is stronger than in other emerging markets such as China or India. Therefore, our results are consistent with the idea that better regulation can reduce the risk of expropriation in the internal capital markets of pyramids without canceling their financing benefits.