Testing Financial Constraints against Expropriation: Explaining the Use of Intra-Group Financing in China
Joseph P. H. Fan The Chinese University of Hong Kong [email protected]
Li Jin Harvard Business School [email protected]
Guojian Zheng School of Business, and Contemporary Accountancy and Finance Research Center, Sun Yat-sen University [email protected]
March, 2009
Abstract
Intra-group financing in business groups in emerging market is studied using Chinese data. The focus is on two motivations of intra-group financing: cross-financing to relieve the severe financing constraints that are often prevalent in emerging market economies, and expropriation from minority shareholders in environments with weak corporate governance. We document the existence of both, and discuss their implications on the efficiency and magnitude of intra-group financing in emerging markets. We find that, from the business group perspective, the internal capital market is most efficient when firms are well governed and have a pressing need to mitigate financing constraints.
Keywords: financial constraint, tunneling, business group, emerging market finance, corporate governance.
JEL Classification: G31 G32 G34 G15
We thank Joshua Coval, Robin Greenwood, Wei Jiang, Tarun Khanna,Josh Lerner, Randall Morck, Stewart Myers, Enrico Perotti, David Scharfestein, Shangjin Wei, and participants at seminars and conferences at the Chinese University of Hong Kong, Harvard Business School, Binghamton University, George Mason University, Singapore Management University, Queensland University of Technology, University of Queensland, the 2008 NBER China Workshop, the 2008 CICF Conference, and the 2009 American Finance Association Annual Meeting. We are grateful for financial support from the Institute of Economics and Finance of CUHK, Division of Research of Harvard Business School, and the National Natural Science Foundation of China project number 70532003 (Zheng).
1 1. Introduction
Complex group-like organizations are commonplace. We are familiar with a conglomerate structure in which business activities are organized into different units. These units are interdependent, as resources are frequently transferred among them under the direction of headquarters. Another common structure is the business group. As with conglomerates, there are often intensive transfers of resources among the units of the business group. However, distinct from conglomerate units, which are typically fully owned by the same legal corporation, many business group units are legally separate but effectively controlled by a group parent company, either directly or indirectly through intermediate companies in a control pyramid. Moreover, unlike the conglomerate units that depend solely on headquarters for financing, the units of a business group can raise external capital. In fact, their legal separation from their parent firm is an outcome of such an equity raising (or carve-out) transaction.
A body of literature focuses on the conflicts of interest arising from the legal separation of business group units, in particular between the controlling parent and the minority shareholders of business group subsidiaries, as resource allocation within the group sometimes benefits the controlling parent at the expense of minority shareholders of the subsidiaries. The internal transfer of financial resources between the units of a conglomerate is less of an issue because of the common ownership of all the units within the conglomerate. However, tunneling between the units of a business group (Johnson et al., 2000), and in particular between units with a different ownership structure, is controversial not only economically but also legally, because the interests of various shareholders can be affected. Numerous studies report that the cost of such conflicts of interest is reflected in lower security prices and therefore ultimately borne by the controlling parent (owner) (Claessens et al., 2002; La Porta et al., 2002; and many others).
Given that a group structure is costly, the question arises as to the countervailing benefits that explain its persistent existence and wide prevalence around the world. Although the literature provides abundant conjecture (Morck, Wolfenzon, Yeung, 2004; Almedia and Wolfenzon, 2006; Khanna and Yafeh, 2007), empirical evidence has only recently emerged (Gopalan et al., 2007; Khanna and Thomas, forthcoming). In addition to data unavailability, a key challenge to such research is that it is difficult
2 to disentangle the resources diverted to facilitate group efficiency from those diverted to satisfy the private benefits of the controlling shareholder. From the perspective of minority shareholders both constitute tunneling, the cost of which is borne by the group owner. However, in under-developed financial markets, tunneling may still be welfare enhancing from the overall group perspective if it helps to mitigate severe financing constraints faced by member firms of the group, even though not the first- best outcome.
In this study, we focus on transfers of financial resources within business groups and test the hypothesis that financial tunneling may be motivated by both group capital allocation efficiency and pure expropriation of minority shareholders. Using data on Chinese business groups, we provide evidence that both mitigating financing constraints and expropriation of minority shareholders are operative in our sample, and that both occasion a greater magnitude of intra-group financing activity. We also provide evidence on the efficiency of internally allocated capital from the perspective of the business group as a whole. We show that intra-group financing that is aimed at mitigating financing constraints promotes the more efficient use of internally generated funds, whereas internal financing activities aimed at expropriation results in a low efficiency in the use of the group’s capital. Which effect will dominate depends on the business group’s corporate governance and financing constraints: to look at only the positive or the negative side of the internal capital market is an oversimplification.
The China dataset has several advantages for our tests. First, the sample includes pairs of non-publicly listed parents and publicly listed subsidiaries. These subsidiaries have been carved out from their parents and listed on China’s two stock exchanges. We have separate (non-consolidated) financial data for both the non-publicly listed parent and the listed subsidiary of each business group in our sample, and also detailed ownership data and information on the fund flow between the two. The detailed pair- wise data structure facilitates our tests and allows us to disentangle the different incentives and effects of tunneling. Second, most listed subsidiaries were carved out from their parent companies in response to China’s enterprise reforms, and hence the pyramidal group structure is not primarily determined by managers or controlling owners. This makes our empirical analysis less subject to firm-level endogeneity issues than studies based on different institutional settings.
3 To our knowledge, this study is the first to disentangle the internal financing of business groups from both sides and to document both the magnitude and relative efficiency of such financing activities. The study also documents an asymmetry in the fund flow pattern. We often see funds flow from subsidiary to parent firm. This may be because funds tend to flow to places with more authority, but may also be due to the search for better investment opportunities by the (more constrained) parent firm.
Both expropriation of minority shareholders and financial constraints are likely to be severe in emerging markets (Shleifer and Vishny, 1997; Stein, 2002), and may even be of concern in developed markets such as those of the United States. The findings of this paper from the largest emerging economy in the world clearly have important implications for both developing and developed economies.
The remainder of the paper is organized as follows. Section 2 describes a conceptual model that gives rise to several empirical predictions. In Section 3, we detail the institutional background of our study and define our measures. The empirical results are presented and some interpretations are offered in Section 4. Section 5 discusses the robustness of our results. Section 6 concludes.
2. A Model of Financial Tunneling
Previous research has discussed separately the negative and positive roles of the business group. On the negative side, some of the intra-group activities might reflect expropriation of minority shareholders by controlling shareholders (Bae, Kang, and Kim, 2002; Bertrand, Mehta, and Mullainathan, 2002; Baek, Kang, and Park, 2004; Baek, Kang, and Lee, 2006), and deter creative destruction and slow down economic growth (Morck, Strangeland, Yeung, 2000; Morck, Wolfenzon, and Yeung, 2004). On the positive side, the structure may afford member firms easier access to capital (Hoshi, Kashyap, and Scharfstein, 1991; Almeida and Wolfenzon, 2006), facilitate the coordination of internal transfers of resources and improve risk management (Khanna and Palepu, 2000; Khanna and Yafeh, 2005), and provide support to financially distressed member firms (Gopalan, Nanda, and Seru, 2007).
4 However, it is often difficult to empirically disentangle the positive and negative effects of business groups. Taking financial tunneling as an example, transfers of cash among the units within a business group can be motivated both by the consumption of private benefits and the bypassing of financial constraints. The latter role of internal capital markets has been extensively discussed in the conglomerate literature (Williamson, 1985; Stein, 1997; Rajan, Servaes, and Zingales, 2000; Scharfstein and Stein, 2000; and many others).
Before we test this dual role of the intra-group capital market, we use a simple model to show that the pyramidal business group structure can facilitate both capital allocation efficiency and expropriation. We conceptually describe the model and its predictions in the following, and provide details of the model development in the Appendix.
We start with an owner of a business carving out a subset of the assets of the firm and listing the carve-out on a stock exchange. Through the parent company, the owner maintains majority ownership and therefore control of the listed subsidiary. The public listing allows the owner to raise external capital and creates a class of minority shareholders in the subsidiary. Assuming that the parent company and the subsidiary are not related operationally, therefore any transfers between the two companies will be financial, and will either serve to fund investment projects in the parent or subsidiary or be consumed by the controlling owner as private benefits. We further assume that the legal environment in which the business group is located is unable to fully prevent such tunneling activities (Johnson et al., 2000), in that the probability of being discovered is not very high and the expected punishment not prohibitively severe. We also assume that the parent company cannot effectively commit to refraining from tunneling because such commitment would be costly in terms of the opportunity losses from private benefits and investment opportunities, or the costs of self-imposed corporate governance constraints. However, minority shareholders can price protect themselves by buying the stocks at a discount (Claessens et al., 2002; La Porta et al., 2002; Lemmon and Lins, 2003). Therefore, the cost of tunneling, including the expected cash expropriated from minority shareholders and the resources required to materialize such expropriation, is entirely borne by the controlling shareholder.
5 Assume that the parent company, because it is not publicly listed, is much more financially constrained than the listed subsidiary, and that cash is therefore almost always transferred from the subsidiary to the parent company. We can now consider the controlling owner’s tunneling incentive. The benefit of tunneling for private benefit consumption decreases with the parent company’s percentage stake in the subsidiary. At the extreme, if the subsidiary is completely owned by the parent (a typical multi-division conglomerate structure), there is no private benefit for tunneling.
In addition to private benefits, tunneling can create value for the parent company by relieving financial constraints. The parent company by itself may not have sufficient capital to take on all of the available positive net present value (NPV) projects. Assuming (without loss of generality) that lost investment opportunities are permanently lost, then, tunneling enables the parent company to undertake more value-creating investment. For simplicity, we assume that the first extra dollar invested in the parent firm generates a high marginal value (a positive NPV) for the parent company, but that every incremental dollar put into the parent firm generates a decreasing amount of NPV, thus reflecting a depletion of valuable investment opportunities available to the parent firm.
Intuitively, we should observe more tunneling if the parent company and the listed subsidiary are more severely misaligned in incentives (concomitant with a lower ownership stake of parent in the subsidiary), or face more severe financial constraints. For given parameter values of legal enforcement (the probability of tunneling being discovered and of being penalized when it is discovered) and tunneling costs, we can plot the “actions” of tunneling against a two-dimensional diagram of parent company ownership stake in the listed subsidiary and financing constraints (see the Appendix).
In this setting, when the ownership stake of the parent firm in the subsidiary is low – indicating less alignment of interests between the parent and the subsidiary – tunneling occurs even if the financing constraints are not severe. In fact, it is possible that tunneling will occur even if the parent firm has worse investment opportunities than the subsidiary, suggesting that parent firms sometimes engage in tunneling to expropriate minority investors in the subsidiary despite the fact that so doing hurts the overall capital allocation efficiency of the group.
6 In cases in which the parent ownership stake in the subsidiary is high but financing constraints are severe, a firm may optimally engage in tunneling. In such cases, mitigating financing constraints serves as the main motivation for tunneling, and expropriation is less of a concern. Almost by definition, the efficiency of such tunneling is the highest.
In summary, the model shows that tunneling may be motivated by both private benefit consumption and investment efficiency, and that the extent to which one or the other is true depends on the controlling owner’s ownership stake in the subsidiary and the extent of the external financing constraints faced by the parent firm.
3. Institutional Background and Methodology
3.1 Institutional Background
China began to experiment with shareholder reforms in the early 1990s, when firms selected by the government were given quotas of equity that they were permitted to float on the newly established Chinese stock exchanges in Shanghai and Shenzhen. To limit the total float and stabilize stock prices, the government imposed a size quota system, and mandated that large firms carve out subsidiaries of an appropriate size and list them. The size of the listed company and the proportion to be floated were determined by careful political deliberation aimed at balancing regions and inducing regional competition (Pistor and Xu, 2005), and firm managers did not have a major influence in the decision. The quota system was scrapped in 2001, and since then securities firms have been allowed to nominate enterprises for public listing subject to screening by an independent listing committee of the Chinese Securities Regulatory Commission. Most of the companies in our sample (96%) went public before 2001 and their public listing was therefore subject to the quota regulation.
Under the quota system, controlling firms were typically not allowed to sell their stake, again out of concern that floating too much equity could flood China’s fledging market and dampen stock prices. By 2004, roughly two thirds of the equity of publicly listed Chinese companies was in the form of “non-tradable” shares held by the parent
7 firm, and in some cases directly by the State. At the start of our sample period in 2000, we observe a number of firm pairs consisting of a publicly listed subsidiary and a non-listed, controlling parent firm. This is a typical pyramid structure in which control is one-directional, with parents controlling listed subsidiaries. The controlling parent companies (hereafter parents) typically retain a controlling stake in the listed subsidiaries (hereafter termed subs), and the rest of their shares are distributed among outside equity investors. In our sample, the mean (median) percentage ownership of the listed subsidiary by the parent firm is 51.5% (54.2%). This structure closely resembles what is found in business groups elsewhere in the world, particularly in emerging markets, that is, legally independent entities with close financial and operational relationships. As an emerging market economy, China lacks a smooth, well-functioning external financial market. The tradable public equity market in early 2004 (roughly the middle of our sample period) was equivalent to only 17% of China’s total GDP (McKinsey report, 2006). The public corporate debt market in China in 2004 was virtually non-existent and the state-owned banking system quite rigid in its lending practices, resulting in many firms with good investment projects being starved of cash (Allen, Qian, and Qian, 2005; Cull and Xu, 2000, 2003; Maskin and Xu, 2001).
Contrary to practice in many developed markets, private equity and venture capital (PEVC) investor operations were in a legal grey area in China and not well protected. The scale of such operations was typically quite small, rendering them less relevant for the funding of large-scale investment projects. Not surprisingly, for such PEVC funds that are available, the cost of capital is on par with, if not higher than, that in more developed markets, and much higher than that in the public equity market. As a consequence, Chinese firms often operate in rigid financial environments, and external financing is frequently not readily available. IPOs are heavily regulated in China, and publicly listed firms are viewed as a precious resource and command prestige, which means that they are more likely to be able to obtain bank loans and trade credits (see Allen, Qian, and Qian, 2005; McKinsey report, 2006). Listed firms can, in theory, conduct seasoned equity offerings to raise new equity finance, whereas non-listed parent companies face more severe financing constraints. At the same time, expropriation of minority shareholders is a serious concern in China, and the fact that many managers and directors of publicly traded companies are bureaucrats or appointed by parent companies can give rise to incentive problems, as it is unclear
8 whether they will put the interests of the minority shareholders of listed companies ahead of those of parent companies (Fan, Wong, and Zhang, 2007).
The weak financial and legal environment gives parent firms at least two potential reasons to rely on intra-group capital flow from listed subs: to expropriate the minority shareholders of the sub and to mitigate the constraints on securing external financing. It is the interaction between the two rationales that yields the interesting characteristic that Chinese firm pairs may have an asymmetric internal capital market: parents have control over listed subs, and not vice versa, and financing constraints are also more severe for parents than for listed subs. We can thus expect to observe more cash flowing from listed subs to parents.
3.2. Methodology
We measure internal capital market activity using a conventional method in the literature, namely, the sensitivity of the investment of one firm to the available cash flow of another firm in the group (Lamont, 1997, Shin and Stulz, 1998). We also modify the standard approach of measuring the efficiency of investment of internal capital by the sensitivity of investments to a measured investment opportunity (Gertner, Powers, and Scharfstein, 2002). To cope with the context of the business group, we adjust the methodology used in the conglomerate literature by amending the cash flow measurement and the measure of the relative investment opportunity between parent and subsidiary.
3.2.1. Measurement of cash flow
Tunneling comes in many guises. Theoretically, tunneling can take the form of unfair transfer pricing in related-party transactions involving raw materials or intermediate or final products. However, we do not have direct information on the terms of trade between group members, and thus cannot rule out the possibility that the prices on these related-party transactions are fair. In addition, a large amount of related-party transactions arise naturally due to the operational inter-dependency of the parents and the listed subs. As most of the listed subs in our sample are subsidiaries that have been artificially carved out from the parents to meet institutional IPO requirements, a high degree of operational interdependence can be expected. For example, a parent might
9 carve out its manufacturing assets to create a listed sub but retain the functions of finished product sales and raw material procurement. In this case, neither the parent nor the listed sub can be independent from the other, and related-party transactions between the two will naturally be extensive, even if there is no intentional transfer of financial resources between the parties involved.
Because transfer pricing through related-party transactions is reflected in many accounting and information disclosures and can therefore be readily detected and monitored by auditors and market regulators, any irregularities are likely to result in strong share price reactions or even disciplinary action by the regulators (Cheung, Rau, and Stouraitis, 2006). Indeed, the first universal accounting standard for industrial firms published in China in 1997 details accounting treatment and information disclosure requirements for related-party transactions, and the Chinese regulators have similarly provided guidelines on such matters. Also, as has already been argued, transfer pricing does not necessarily mean tunneling, as the transaction price may be fair. For these reasons, we filter out related-party transactions related to raw materials and intermediate goods to reveal any remaining inter-company cash flow activities such as equity co-investments, purchases or sales of assets, loans, debt guarantees, unreported cryptic transactions, and so forth. These other channels of inter-company cash flow are more obscure, and often take place in legal grey areas or are outright illegal.
We filter out the impact of related-party transactions by focusing on a revised measure of the investment sensitivity of one firm to another’s cash flow. In the literature, the sensitivity of investment to cash flow (typically measured as earnings before interest and tax (EBIT) plus depreciation and amortization) is used to test for internal capital markets among the divisions of a conglomerate (Kaplan and Zingales, 1997).1 In our data, however, a large fraction of EBIT may take the form of trade credits between parents and listed subs (i.e., accounts receivable and payable). To determine the true cash flow available for internal use, and thus for internal capital market activities, we further subtract from the EBIT the amount of increase in trade credits extended since
1 Kaplan and Zingales (1997) also defined cash flow as earnings before extraordinary items plus depreciation, and other researchers have used similar definitions. For example, Shin and Stulz (1998) defined cash flow as operating profit plus depreciation, and Hoshi, Kashyap, and Scharfstein (1991) defined it as income after tax plus depreciation minus dividends. 10 the last period.2,3 By this definition, we essentially treat all related-party transactions as if there were no price distortion and thus no tunneling, and look beyond such transfer pricing for evidence of internal capital markets. Note that by this conservative definition, if any internal capital market activities were indeed hidden in related-party transaction activities, then we would clearly be under-estimating the true magnitude of the internal capital market.
The traditional EBIT measure of cash flow also neglects the effect of income tax. This is less significant in studies of the business segments of conglomerates, as income tax is not levied at the segment level. However, as listed subsidiaries and their non-listed parent firms are legally separate entities, a further adjustment needs to be made to the cash flow measurement. The new definition of cash flow is therefore EBIT plus depreciation and the net increase in accounts payable, minus income tax.
A distinctive difference between the segments in a conglomerate and the subsidiary firms in a business group is the ability of subsidiary firms to secure external financing through bank debt and seasoned equity offerings. Because the segments of a conglomerate typically cannot obtain financing independently, cash flow from financing is not considered in the literature on internal capital markets within conglomerates. Funds secured by listed subs through external financing do, however, affect the available cash flow, thus necessitating a further adjustment to the traditional cash flow measure.
In summary, we use three cash flow measures in our tests of internal capital market activity. Measure 1 is the traditional cash flow measure adjusted for related-party transactions, that is, EBIT + depreciation + the net increase in accounts payable minus receivable attributable to related-party transactions. Measure 2 is the after-tax version of Measure 1. Measure 3 is further adjusted to include external financing, that is, EBIT + depreciation + the net increase in accounts payable minus receivables
2 Suppose that firm A has $100 in annual sales and no depreciation and amortization. If no trade credit is extended to customers, then its cash flow would, by definition, be $100, but if $30 of its sales is in the form of accounts receivable, then the actual cash flow for the current period would be only $70. Alternatively, suppose that firm A has $100 in annual sales and has not extended any trade credits to customers, but receives as accounts payable $30 in trade credits from a supplier. In this case, the current period actual cash flow is $130. Here, the existence of trade credits affects the amount of actual cash flow available for investment activities. 3 Due to a lack of available data on amortization, we do not include it in our calculation of cash flow. We obtain similar results when we substitute EBIT with definitions similar to those referenced in note 2. 11 attributable to related-party transactions - income tax + the net increase in bank debt and equity.
3.2.2. Measurement of relative investment opportunity
The literature measures investment opportunity by using Tobin’s q. Following Gertner, Powers, and Scharfstein (2002), the interaction term between the q of the receiving company and other firm’s cash flows should measure the efficiency of the cash flow through the internal capital market. A positive and significant coefficient on this interaction term indicates that when the investment opportunity of the receiving company is good (as denoted by a higher q), then more capital will flow into the company.
However, because we focus on the efficiency of capital flow from the perspective of the business group, we modify the methodology by replacing Tobin’s q with the relative q, which is defined as the difference between the q of the parent and the q of the listed sub. Intuitively, more capital flowing from the listed sub to the parent when the parent’s investment opportunity is better than that of the listed sub (denoted by a positive relative q) is an indication of the efficient use of internally generated funds. Conversely, more capital flowing from the listed sub to the parent when the parent’s investment opportunity is worse than that of the listed sub is an indication of the inefficient use of internal capital. To implement this approach with our data, when the parent and listed sub are not in the same industry defined at the three-digit level of the Chinese Standard Industry Classification System, the relative q is defined as the parent’s industry q minus the listed sub’s industry q, where the industry q is the mean Tobin’s q of all of the listed firms in that three-digit SIC industry category at the beginning of the year. We use the industry q rather than the individual firm q for two reasons. First, the individual q is not available for the non-listed parents, and second, the industry average q avoids measurement errors that might be specific to a particular firm. If the parent and listed sub are in the same industry, then the relative q is defined as the parent’s industry q minus the listed sub’s individual q. Almost 39% of the observations in our sample have parents and listed subs in the same three-digit industry. As an alternative, we exclude these observations from our sample, but the exclusion does not significantly change the empirical results that are based on the full sample.
12 3.3.3. Regression model
We estimate the following model for both the listed sub and parent samples.
CapitalExpenditure 0 1OwnCashFlow 2OtherCashFlow
3 RelativeQ 4OtherCashFlow*RelativeQ YearDummies .
Capital Expenditure is measured as fixed asset investment. In the listed sub regressions, Own Cash Flow is the cash flow of the listed sub and Other Cash Flow is the cash flow of the parent. Conversely, in the parent regressions, Own Cash Flow is the cash flow of the parent and Other Cash Flow is the cash flow of the listed sub. All of these variables are adjusted by the total assets of the listed sub (parent) at the beginning of the fiscal year. Year dummy variables and firm fixed effects are included in the regression.
If internal capital markets exist, then, controlling for Own Cash Flow, a firm’s capital expenditure should be sensitive to the cash flow of the other firm, which means that β2 should be significantly positive. The higher the value of β2, the more sensitive one company’s capital expenditure is to the other’s cash flow, that is, the greater the magnitude of the internal capital market. Similar to Gertner, Power, and Scharfstein
(2002), we use β4 to measure the investment efficiency of internal capital market activities. A significantly negative β4 means that a firm tends to invest more using the other firm’s cash flow when its relative investment opportunity (as measured by the difference in the q of the parent and the q of the listed sub) is worse, thus reflecting an inefficient use of the internal capital market. A significantly positive β4 means that a firm tends to invest more of the other firm’s cash flow when its relative investment opportunity is better, thus reflecting an efficient use of the internal capital market.
We are fully aware of the recent debate in the literature over whether investment sensitivity to cash flow is truly a reflection of financing constraints, particularly given the critiques by Kaplan and Zingales (1997, 2000), Erickson and Whited (2000), Gomes (2001), and Moyen (2004), among others. However, our empirical analysis uses a measure of financing constraints that is unaffected by this recent debate.
13 4. Empirical Analysis
4.1. Data and Source
Our sample consists of 1,324 pair-year observations. Each pair consists of a listed firm and its non-listed parent firm. The study period is 2000 to 2005. We have income statement and balance sheet data for both the parents and the listed subs, which enables us to directly test the direction of internal capital markets by running regressions on both the non-listed parent firms and their listed subsidiaries. Data on the trade credits between parents and listed subs, which is hand collected from the annual reports of the listed subs, is used to adjust cash flow and construct the investment-cash flow model. To further test our hypothesis regarding internal capital markets, we also collect data on the cash flow rights of parents in listed subs and bank ownership.
Data other than that hand-collected from public channels comes from two main sources: the annual reports of the listed firms, which are publicly available from the Chinese Securities Regulatory Commission Web site, and the National Bureau of
Statistics’ (NBS) Annual Industrial Survey Database, which surveys and publishes in the official China Statistics Yearbooks financial and operating information on industrial firms.4 The NBS Database includes all state-owned industrial enterprises and private industrial firms with total annual sales of more than RMB5 million (about US$600,000 using the exchange rate of December 31, 2005). The Database is representative of the national economy and has been increasingly used in academic research (Chow, 1993; Chuang and Hsu, 2004; and Li, Yue, and Zhao, 2006; Fan, Huang, and Zhu, 2007).
We collect data on all of the firms listed in the aforementioned sources. To ensure that we are not including “shell” or holding companies that have no real operations, we include only observations in which the parent is an industrial firm as surveyed by
4 As of the beginning of 1993, all industrial firms in China, regardless of ownership type and size, must report their financial statements according to the same “Accounting Standards for Enterprises.” As almost our entire sample is composed of industrial firms, the financial and operating information of the listed firms and the parent firms is thus consistent. Note also that the NBS data is non-consolidated, and therefore the financial data reported for an entity does not incorporate the numbers for its legally separate affiliates. 14 NBS. We further remove instances for which we have missing data, or for which the parent firm of the listed sub cannot be identified in the NBS Database.5 Finally, to ensure that the parent firms have at least a relative controlling position in the associated listed firms, we drop observations in which the parent company’s stake in a listed subsidiary is 20% or less.6 This yields a final sample of 1,324 firm-year observations.
4.2 Summary Statistics
Table 1 reports the summary statistics for the main variables. The average (median) capital expenditure adjusted by total assets at the beginning of the period is 10.9% (7.01%) for the listed subs and 7.21% (3.83%) for the parents, representing a 2.88% (3.18%) difference in fixed asset investment level. The net trade credit between the parents and listed subs normalized by total assets at the beginning of the period is -3.71% (-1.47%) for the listed subs and 2.86% (1.37%) for the parents, which means that, on average, the listed subs tend to be net trade credit providers. For the listed subs, the average (median) of (EBIT + Depreciation) is 9.87% (9.49%) and of Cash Flow Measure 1 is 6.20% (8.10%). For parents, the average (median) of (EBIT + Depreciation) is 7.02% (5.85%) and of Cash Flow Measure 1 is 10.21% (8.03%). As expected, the trade credit adjustments significantly affect the cash flow levels of both the listed subs and the parents, but in different directions. For the listed subs, the average (median) of Cash Flow Measures 1, 2 and 3 is 6.20% (8.1%), 4.61 % (5.97%), and 7.83% (9.40%), respectively.
Table 1 also reports the industry q and industry growth for the listed sub and parent samples. The average relative q of the sample is -0.06, reflecting an average lower investment opportunity of the parents relative to the listed subs. Parents control an
5 This can occur, for example, when organizational changes (e.g., restructuring, merger, or acquisition) result in a firm legally changing its name and identification code such that what is recorded in a listed firm’s annual report might not be precisely the same entity that is recorded in the NBS database. Data entry errors made during the NBS survey and data collection process may also account for some mismatches. 6 Although not all of the parent firms have at least a 50% of stake in the associated listed firms, our descriptive statistics (not reported) show the cash flow rights of the parent firms to be about six times larger by mean, or ten times larger by median, than the sum of the cash flow rights of the second and third largest block holders. The non-listed parents in our sample are thus dominant shareholders because the other large shareholders are not sufficiently large to have a comparable impact on the listed firms. 15 average of 52.92% of the shares of the listed subs, and 22.66% of the listed subs have an ownership stake in local banks. Insert Table 1 here
4.3. Empirical Tests
4.3.1. On the Existence of Internal Capital Markets
In Table 2, which reports the results of the total sample regression, we can see that in both the listed sub and parent regressions capital expenditure is significantly related to Own Cash Flow, but the coefficients are about four times larger in the parent than in the listed sub regressions.7 More importantly, the coefficients of Other Cash Flow are significantly positive in the parent regressions but not significant in the listed sub regressions, demonstrating the existence and asymmetric nature of the internal capital market between parents and listed subs, with parent investments relying on the cash flow of the listed subs but not vice versa. These findings are consistent with the hypothesis that financing constraints are greater for parents (because they are non- listed), but also consistent with the hypothesis that parents have a greater incentive to expropriate from the minority shareholders of listed subs. The economic significance is obvious: as the cash flow of listed sub increases by 10%, the parent investment increases by 3.7% (for Cash Flow Measure 1), 2.21% (for Cash Flow Measure 2), and 1.32% (for Cash Flow Measure 3), respectively. We also see that the coefficients of the interaction between other cash flow and relative q are positive in the regressions with Measures 1 and 2, but slightly negative in the regressions with Measure 3. Theoretically, there is no ex ante prediction of the sign of this coefficient. In general, the efficiency of the use of the internal capital market may be either positive or negative depending on whether the intra-group capital flow is used productively (to mitigate financing constraints) or unproductively (to expropriate from minority shareholders).
As our focus is on the flow of cash out of the listed sub and into the parent, the following analysis emphasizes the empirical results of the parent regressions. Further,
7 We do not wish to over-interpret this. We are aware of the argument that a high coefficient on β2 (Own Cash Flow) could be caused by a change in the unobserved investment opportunity, but believe that the systematic difference in the magnitude of the coefficient for the listed sub and parent regressions is striking, and not readily reconciled by the explanation of changing investment opportunity alone. 16 as we obtain very similar results using the three cash flow measures, we only report the results using the Cash Flow Measure 1.
Table 2 suggests that there are, besides the trade credits associated with related-party transactions, many other less observable channels through which parents can respond to their incentive to expropriate resources from listed subs. Because we have documented their existence after controlling for trade credit, the internal capital market should be even stronger and more pervasive.
Insert Table 2 here
4.3.2. On the Role of Corporate Governance
According to the literature, ownership is fundamental to – and cash flow rights theoretically at the core of – corporate governance (Jensen and Meckling, 1976; Shleifer and Vishny, 1997) and empirically strongly related to the incentives of large shareholders to tunnel the listed firms that they control (Bertrand, Mehta, and Mullainathan, 2002; Claessens, Djankov, Fan, and Lang, 2002), especially when legal protection for outside investors is weak (La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 1997,1998, 1999). We therefore use the cash flow rights of the parents as our measure of corporate governance.
We divide the sample into two equal halves according to the cash flow rights of the parents relative to the listed subs. We conjecture that the incentive to tunnel will be weaker for controlling parents with a higher percentage ownership of their listed subs because the interests of the shareholders of the parents and listed subs will be more closely aligned. Everything else being equal, we expect to observe stronger internal capital market activity due to tunneling when the percentage ownership of parents in their listed subs is lower. The main purpose of tunneling in this case being to expropriate minority shareholders, there is less consideration of the efficiency of the activity, and the parents may aim to steal as much as possible. The efficiency of internal capital market activities resulting from this tunneling motivation is thus lower.
17 As can be seen in Table 3, the investments of parents with lower cash flow rights in their listed subs are more dependent on the cash flow of the listed subs. The coefficients are, on average, about four times larger when the cash flow rights are below the median level, and the F tests for the difference in the coefficients between the high and low cash flow rights samples are significant at the 1% level. Investment efficiency is lower (the coefficients of the interaction term are more negative, or less positive) for parents with lower cash flow rights. The F test for the difference in the coefficients on the interaction term between the high and low cash flow rights samples is significant at the 5% level. The asymmetry of the existence and efficiency of internal capital markets between firms with different cash flow rights in emerging markets is consistent with the expropriation hypothesis. In contrast, unreported results on the investment of listed subs from the parent cash flow show no significant difference in investment-cash flow sensitivity between the high cash flow rights and low cash flow rights samples.
Insert Table 3 here
4.3.3. On the Role of Mitigating Severe External Financing Constraints
Bank ownership has been argued to be important to firms for raising external finance. For example, Hoshi, Kashyap, and Scharfstein (1991) examined two sets of Japanese firms, one with close financial ties to large Japanese banks that serve as the primary source of external finance and one without such ties. They found investment to be more sensitive to liquidity for the latter set of firms, demonstrating bank ownership to be an important mechanism in mitigating financing constraints. In China, firms often have ties to local banks through an ownership stake. We hypothesize that parent-listed sub pairs in which the listed sub owns shares in a local bank will have a good relationship with a bank and thus easier access to bank credit.
We partition our data set using the existence of bank ownership as a proxy for good external financing availability. We hypothesize that firms with no bank ownership will have more severe financial constraints, and thus need to use intra-group financing as a mitigating mechanism. If this holds, then we would expect more internal capital market activity in the sub-sample with no bank ownership, controlling for everything else. We would also expect the intra-group fund flow so occasioned to be value-
18 enhancing, as the need to mitigate financing constraints would be greater when investment opportunities are greater for the parent than for the listed sub.
We also report the regression results on the role of mitigating severe external financing constraints in Table 3. We can see that the investments of parents with no bank ownership are more dependent on the cash flows of their listed subs, with the coefficients of Other Cash Flow being on average more than three times larger. It is also clear that investment efficiency is better for this sub-sample, as the interaction term is more positive in column (4), which contains the parents with no bank ownership. In contrast, in unreported tests in which we break down the sample in the same way, we find no pattern in the listed sub regressions. Overall, the empirical results are consistent with the hypothesis that the intra-group cash flow from listed subs to parents is stronger when the parents face more severe financing constraints, and that the efficiency of such cash flow is higher than when the parents that do not face severe financing constraints.
4.3.4. The Interaction of Corporate Governance and Financing Constraints
Table 3 reports separately the effects of expropriation and the mitigation of financing constraints. Table 4 reports the results of the joint effects of corporate governance and financing constraints on the existence and efficiency of internal capital markets. We control for governance (financing constraints) effects when considering the effects of financing constraints (governance). Specifically, we perform the same regression on four sub-samples: “strong governance and financially constrained,” “strong governance and not financially constrained,” “weak governance and financially constrained,” and “weak governance and not financially constrained.” If both incentives (to expropriate minority shareholders and to mitigate financing constraints) are important, then we would expect to see both leading to more internal capital market activity but with an offsetting effect on the efficiency of such activity. Specifically, we would expect the efficiency of internal capital market activity to be highest for the sub-sample in which the parent ownership of the listed subs is highest but there is no bank ownership, because the most likely reason for internal capital market activity in this context is to mitigate financing constraints. In contrast, we would expect the efficiency of internal markets to be the lowest for the sub-sample in which the parent ownership of the listed subs is lowest and there is bank ownership,
19 because the most likely reason for internal capital market activity in this context is expropriation of minority shareholders of the listed subs.
We report the results in Table 4. In most of the regressions, the effects of financing constraints (corporate governance) on internal capital markets are still obvious after controlling for corporate governance (financing constraints). When low cash flow rights are accompanied by greater financing constraints, the magnitude of the internal capital market is largest (see the coefficient of Other Cash Flow in column (4) of Table 4). It is also interesting that in most of the regressions the coefficients of the interaction terms are most negative when the cash flow rights of the parents are lower and the parents are less financially constrained. For example, in column (3) of Table 4, the coefficient of Other Cash Flow*Relative Q is -0.2353, the only negative coefficient among the four sub-samples. In contrast, in column (2), when the cash flow rights of parents are higher and the parents are more financially constrained, the interaction term is 0.3230, the most positive coefficient among the four sub-samples.
Our basic results for the existence and efficiency of internal capital markets are robust to controlling for the effects of expropriation and the mitigation of financing constraints. We further find that the effects of financing constraints on internal capital markets are more pronounced when governance is bad, and that the effects of governance on internal capital markets are more pronounced when firms face less severe financing constraints. In other words, we document that expropriation decreases the efficiency of internal capital markets, and that this is most prominent when firms do not face severe financing constraints (i.e., when there is no “good” reason to choose to transfer funds between companies rather than through external markets).
Insert Table 4 here
5. Robustness Checks
The results of a variety of robustness tests are reported in Tables 5-10. The main results reported in this paper are robust to these additional tests. The focus of the tests is the interaction of parent cash flow rights and parent financing constraints. Note that because the other cash flow measures yield similar results, we report only the results using Cash Flow Measure 1.
20 5.1. Absolute Instead of Relative Investment Opportunity
Similar results are obtained using the industry q of the parent firm (Table 5) and the industry average growth of the parent (Table 6) to proxy for the investment opportunity of the parent. The industry average q of the parent (rather than the difference between the industry average q of the parent and the q of the listed sub) provides the absolute measure of the investment efficiency of the parent, whereas industry growth is the mean of the firm-level lagged sales growth in the same industry as calculated from the NBS Annual Industrial Survey Database. Whereas the former measure is more consistent with the original measure of investment efficiency proposed by Gertner, Powers, and Scharfstein (2002), the latter is more likely to be exogenous and potentially more robust to noise in the observations, as it is calculated from the previous year’s sales growth data drawn from a large sample of non-listed firms (approximately 200,000 observations per year) and applied to our relatively small sample. The results are quite similar to those reported in the previous tables. For example, the coefficient of Other Cash Flow*Industry Growth is most negative and significant when the parent cash flow rights are lower and the parent financing constraints more severe.
5.2. Effects of Change in the Regulatory Environment
Table 7 reports robust test results when capital market regulation is used to determine how changes in corporate governance influence internal capital markets. In response to the increasing numbers of severe expropriation cases reported in the Chinese capital markets over the past ten years and the ensuing huge cost to minority shareholders, the Chinese regulators developed a series of capital market regulations in the middle of 2003 that cover the investigation and punishment of expropriation of minority shareholders by large shareholders (parents) of publicly listed companies. Of all of the regulations against insider expropriation, two documents have been widely discussed by the Chinese media and market participants. The first was issued on August 28th, 2003, and mainly deals with the very detailed regulations of the many kinds of fund transfers and debt guarantees between large shareholders and listed firms that can be conducted through various types of normal (or reported) and abnormal (or unreported) related-party transactions. The second, issued on December
21 4th, 2004,8 is more important, and deals with the new regulations and governance mechanisms that protect minority shareholders, including giving more power to independent outside directors and demanding more compelling information disclosures. Assuming that these regulations have taken effect to some extent, we expect to observe changes in the internal market activity after the publication of the regulations. We therefore divide the sample into a 2000-2003 sub-sample and a 2004- 2005 sub-sample. In Table 7, we observe a relatively smaller coefficient of Other Cash Flow and a more positive coefficient of the interaction term in the 2004-2005 sub-sample, which is consistent with our hypothesis that when the outside governance environment improves, the expropriation effect seems less severe. We further interact market regulation with bank ownership and find similar but weaker evidence.
Insert Table 7 here
5.3. Size as an Alternative Measure of Financial Constraints
In accordance with the large body of literature on financing constraints (for example, Almeida and Campello, 2007; Erickson and Whited, 2000), we also use size to proxy for financing constraints. In Table 8, we rank the parent size at the beginning of the year and form two sub-samples: the bottom 33% and the top 33% of the parent firms. We expect the bottom 33% of the parent firms to be more financially constrained than the top 33% firms. Table 8 reports very similar results when bank relationship is replaced by size in the joint tests of corporate governance and financing constraints. That is, when the size of the parent is larger and the cash flow rights of the parent lower, the coefficient of the interaction term is most negative, which is consistent with our hypothesis (column (4)).
Insert Table 8 here
5.4. Excluding Observations with Negative Cash Flow
Our main results include listed sub firms with negative adjusted cash flows, which warrant a different interpretation from the other firms. Specifically, the regression specification implicitly assumes that when a listed sub experiences a negative cash
8 The draft for open comments on this document was issued at the beginning of 2004. Details of the regulations can be found at http://www.csrc.gov.cn/n575458/n4239016/n6634558/n9768098/n9768450/index_1.html. 22 flow, the parent correspondingly invests less. In other words, capital may flow in the reverse direction from parent to listed sub to support a sub that might be in financial trouble. The literature documents evidence that even a controlling shareholder or corporate insider of a listed firm may prop up a firm if this would preserve future opportunities for more expropriation (Friedman, Johnson, and Mitton, 2003; Jin and Myers, 2006). To check whether our results are mainly driven by these propping up activities, we rerun our main analysis excluding the sub-sample of listed subs with negative cash flow. The results, reported in Table 9, reveal the effect of negative cash flow on the main results to be minor.
Insert Table 9 here
5.5. Other Tests
We perform additional robust tests but do not report them in the tables. To check the robustness of the relative q measure, we run a regression for a sub-sample that includes only listed subs and parents that are not in same industry. As another robustness check, we include the natural logarithm of firm assets as the basic model, and examine the effects of cash flow on investment change due to the addition. These additional tests generally yield results similar to those of our main tests.
6. Conclusion
We study the internal capital market phenomenon in emerging market business groups using 1,324 pair-years of Chinese listed firms and their non-listed parents. We document the existence of two aspects of intra-group financing, namely, cross- financing to mitigate severe financing constraints and the exploitation of weak corporate governance to expropriate minority shareholders. Our results suggest that both can account for the rise of internal capital markets in emerging market economies, but that the implications for the efficiency of such intra-group financing are opposite: the efficiency is higher when the context is the mitigation of financial constraints and lower when it is expropriation from outside investors.
23 References
Almeida, H., and M. Campello, 2007, Financial Constraints, Asset Tangibility, and Corporate Investment, Review of Financial Studies 20 (5), 1429–1460.
Almeida, H., M. Campello, and M. Weisbach, 2004, The Cash Flow Sensitivity of Cash, Journal of Finance 59, 1777–1804.
Almeida, H., M. Campello, and M. S. Weisbach, 2006, Corporate Financial and Investment Policies When Future Financing Is Not Frictionless, NBER Working Paper 12773.
Almeida, H. V., and D. Wolfenzon, 2006, A Theory of Pyramidal Ownership and Family Business Groups, Journal of Finance 61, 2637–2679.
Allen, F., J. Qian, and M. J. Qian, 2005, Law, Finance, and Economic Growth in China, Journal of Financial Economics 77, 57–116.
Bae, K., J. Kang, and J. Kim, 2002, Tunneling or Value Added? Evidence from Mergers by Korean Business Groups, Journal of Finance 57, 2695–2740.
Baek, J., J. Kang, and I. Lee, 2006, Business Groups and Tunneling: Evidence from Private Securities Offerings by Korean Chaebols, Journal of Finance 61 (5), 2415– 2449.
Baek, J., J. Kang, and K. S. Park, 2004, Corporate Governance and Firm Value: Evidence from the Korean Financial Crisis, Journal of Financial Economics 71, 265– 313.
Bertrand, M., P. Mehta, and S. Mullainathan, 2002, Ferreting Out Tunneling: An Application to Indian Business Groups, Quarterly Journal of Economics 117 (1), 121–148.
Bertrand M., S. Johnson, K. Samphantharak, and A. Schoar, 2008, Mixing Family with Business: A Study of Thai Business Groups and the Families Behind Them, Journal of Financial Economics 88, 466–498.
Cheung, Y., P. R. Rau, and A. Stouraitis, 2006, Tunneling, Propping, and Expropriation: Evidence from Connected Party Transactions in Hong Kong, Journal of Financial Economics 82 (2), 343–386.
Chow, G. C., 1993, Capital Formation and Economic Growth in China, Quarterly Journal of Economics 108, 809–842. Chuang, Y. C., and F. Hsu, 2004, FDI, Trade and Spillover Efficiency: Evidence from China’s Manufacturing Sector, Applied Economics 36, 1103–1115.
Claessens, S., S. Djankov, and L. H. P. Lang, 2000, The Separation of Ownership and Control in East Asian Corporations, Journal of Financial Economics 58 (1-2), 81– 112.
24 Claessens, S., S. Djankov, J. P. H. Fan, and L. H. P. Lang, 2002, Disentangling the Incentive and Entrenchment Effects of Large Shareholdings, Journal of Finance 57 (6), 2741–2771.
Claessens, S., S. Djankov, J. P. H. Fan, and L. H. P. Lang, 2003, When Does Corporate Diversification Matter to Productivity and Performance? Evidence from East Asia, Pacific-Basin Finance Journal 11 (3), 365–392.
Claessens, S., J. P. H. Fan, and L. H. P. Lang, 2006, The Benefits and Costs of Group Affiliation: Evidence from East Asia, Emerging Markets Review 7 (1), 1–26.
Cull, R., and L. C. Xu, 2000, Bureaucrats, State Banks, and the Efficiency of Credit Allocation: The Experience of Chinese State-Owned Enterprises, Journal of Comparative Economics 28, 1–31.
Cull, R., and L. C. Xu, 2003, Who Gets Credit? The Behavior of Bureaucrats and State Banks in Allocating Credit to Chinese State-Owned Enterprises, Journal of Development Economics 71, 533–559.
Djankov S., R. La Porta, F. Lopez-de-Silanes, and A. Shleifer, 2008, The Law and Economics of Self-Dealing, Journal of Financial Economics 88, 430–465.
Erickson T, Whited TM, 2000, Measurement Error and the Relationship between Investment and q, Journal of Political Economy 108, 1027-1057.
Fan, J. P. H., J. Huang, R. Morck, and B. Yeung, 2006, Institutional Determinants of Vertical Integration: Evidence from China, Unpublished Manuscript.
Fan, J. P.H., T. J. Wong, and T. Zhang, 2007, Politically Connected CEOs, Corporate Governance and Post-IPO Performance of China’s Partially Privatized Firms, Journal of Financial Economics 84, 330–357.
Fan, J. P. H., J. Huang, and N. Zhu, 2007, Distress without Bankruptcy: An Emerging Market Perspective, Unpublished Manuscript.
Fazzari, S. M., R. G. Hubbard, and B. P. Petersen, 1988, Financing Constraints and Investment, Brookings Papers on Economic Activity 1, 141–195.
Fazzari, S. M., R. G. Hubbard, and B. P. Petersen, 2000, Investment-Cash Flow Sensitivities Are Useful: A Comment on Kaplan and Zingales, Quarterly Journal of Economics 115, 695–705.
Fee, C. E., C. J. Hadlock, and J. R. Pierce, 2008, Investment, Financing Constraints, and Internal Capital Markets: Evidence from the Advertising Expenditures of Multinational Firms, Review of Financial Studies, forthcoming.
Friedman, E., S. Johnson, and T. Mitton, 2003, Tunneling and Propping, Journal of Comparative Economics 31, 732–750.
Gertner, R., E. Powers, and D. Scharfstein, 2002, Learning about Internal Capital Markets from Corporate Spin-offs, Journal of Finance 57, 2479–2506.
25 Gomes J.F., 2001, Financing Investment, American Economic Review 91, 1263–1285.
Gopalan R., V. Nanda V, and A. Seru, 2007, Affiliated Firms and Financial Support: Evidence from Indian Business Groups, Journal of Financial Economics 86, 759–795.
Hadlock, C., and J. Pierce, 2008, Does the KZ Index Provide a Useful Measure of Financial Constraints? Working Paper, Michigan State University.
Hoshi, T., A. Kashyap, and D. Scharfstein, 1991, Corporate Structure, Liquidity, and Investment: Evidence from Japanese Industrial Groups, Quarterly Journal of Economics CVI, 3360.
Jensen, M., and W. Meckling, 1976, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics 3, 305–360.
Jin, L., and S. Myers, 2006, R2 Around the World: New Theory and New Tests, Journal of Financial Economics 79, 257–292.
Johnson, S., R. La Porta, F. Lopez-de-Silanes, and A. Shleifer, 2000, Tunneling, American Economic Review 90, 22–27.
Kaplan, S. N., and L. Zingales, 1997, Do Investment-Cash Flow Sensitivities Provide Useful Measures of Financial Constraints, Quarterly Journal of Economics 112, 159– 216.
Kaplan, S. N., and L. Zingales, 2000, Investment-Cash Flow Sensitivities Are Not Valid Measures of Financial Constraints, Quarterly Journal of Economics 115, 707– 712.
Khanna, T., and K. Palepu, 2000, Is Group Affiliation Profitable in Emerging Markets? An Analysis of Diversified Indian Business Groups, Journal of Finance 55 (2), 867–891.
Khanna, T., and C. Thomas, 2008, Synchronicity and Firm Interlocks in an Emerging Market, Journal of Financial Economics, forthcoming.
Khanna, T., and Y. Yafeh, 2005, Business Groups and Risk Sharing around the World, Journal of Business 78 (1), 301–340.
Khanna, T., and Y. Yafeh, 2007, Business Groups in Emerging Markets: Paragons or Parasites? Journal of Economic Literature 45, 2 (June), 331–372.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1997, Legal Determinants of External Finance, Journal of Finance 52 (3), 1131–1150.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1998, Law and Finance, Journal of Political Economy 106 (6), 1113–1155.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, 1999, Corporate Ownership around the World, Journal of Finance 54, 471–517.
26 La Porta R, F. Lopez-De-Silanes, A. Shleifer, and R. Vishny, 2002, Investor Protection and Corporate Valuation, Journal of Finance 57, 1147–1170.
La Porta, R., F. Lopez-de-Silanes, and G. Zamarripa, 2003, Related Lending, Quarterly Journal of Economics 118 (1), 231–268.
Lamont, O., 1997, Cash Flow and Investment: Evidence from Internal Capital Markets, Journal of Finance 52, 83–109.
Lemmon, M. L., and K. V. Lins, 2003, Ownership Structure, Corporate Governance, and Firm Value: Evidence from the East Asian Financial Crisis, Journal of Finance 58, 1445–1468.
Li, K., H. Yue, and L. Zhao, 2006, Ownership, Institutions, and Capital Structure: Evidence from Chinese Firms, Unpublished Manuscript.
Lins, K. V., and H. Servaes, 2002, Is Corporate Diversification Beneficial in Emerging Markets? Financial Management 31, 2 (Summer), 5–31.
Maskin, E., and C. Xu, 2001, Soft Budget Constraint Theories from Centralization to the Market, Economics of Transition 9, 1–27.
McKinsey, 2006, Putting China’s Capital to Work: The Value of Financial System Reform (May), McKinsey Global Institute.
Mitton, T., 2002, A Cross-Firm Analysis of the Impact of Corporate Governance on the East Asian Financial Crisis, Journal of Financial Economics 64, 215–241.
Morck, R., Strangeland, D., Yeung, B., 2000. Inherited Wealth, Corporate Control, and Economic Growth: The Canadian Disease. In Morck, R., ed., Concentrated Corporate Ownership. National Bureau of Economic Research Conference Volume. University of Chicago Press.
Morck, R., Wolfenzon, D., and Yeung B., 2005. Corporate Governance, Economic Entrenchment and Growth. Journal of Economic Literature 43, 657–722.
Moyen, N., 2004, Investment-Cash Flow Sensitivities: Constrained Versus Unconstrained Firms, Journal of Finance 59 (4), 2061–2092.
Pistor, K., and Xu, C., 2005, Governing Stock Markets in Transition Economies: Lessons from China, American Review of Law and Economics 7 (1), 184–210.
Rajan R. G., and L. Zingales, 1998, Financial Dependence and Growth, American Economic Review 88, 3 (June), 559–586.
Rajan R., H. Servaes, and L. Zingales, 2000, The Cost of Diversity: The Diversification Discount and Inefficient Investment, Journal of Finance 55 (1), 35– 80.
Rauh, J. D., 2006, Investment and Financing Constraints: Evidence from the Funding of Corporate Pension Plans, Journal of Finance 61 (1), 33–72.
27 Scharfstein, D. S., and J. C. Stein, 2000, The Dark Side of Internal Capital Markets: Divisional Rent-Seeking and Inefficient Investment, Journal of Finance 55 (6), 2537– 2564.
Shin, H., and R. M. Stulz, 1998, Are Internal Capital Markets Efficient? Quarterly Journal of Economics 113 (2), 531–552.
Shleifer, A., and R. Vishny, 1997, A Survey of Corporate Governance, Journal of Finance 52, 737–783.
Stein J., 1997, Internal Capital Markets and the Competition for Corporate Resources, Journal of Finance, 52 (1), 111-133.
Stein, J., 2002, Agency, Information, and Corporate Investment, in G. Constantinides, M. Harris, R. Stulz, eds., Handbook of the Economics of Finance, Amsterdam: North- Holland.
Whited, T., and G. Wu, 2006, Financial Constraints Risk, Review of Financial Studies 19 (2), 531–559.
Williamson, O. E., 1985, The Economic Institutions of Capitalism, New York: Free Press.
28 Appendix: Sketch of a Theoretical Model
In the following, we sketch a simple setup to illustrate the intuition behind the empirical tests. It is a parsimonious model, with just enough features to show all of the intuition.
We have a listed subsidiary. At the time of listing, the parent company sold shares of the subsidiary on the stock exchange to outsiders. These outside investors factored in the expectation of tunneling activities. Thus, the investors were not fooled and were compensated for the tunneling through a lower IPO price given the future earnings. From the perspective of the controlling shareholder, expropriation of minority investors is ex ante not a first best choice because of potential ex ante resource allocation inefficiency and ex post deadweight loss if additional resources must be used to conceal the tunneling. Nevertheless, due to the lack of cost-effective governance mechanisms, the controlling shareholder cannot commit to refraining from expropriating ex post, and must thus suffer from a lower welfare due to the expropriation.
Tunneling here means taking the cash flow of the subsidiary away and using it in the parent firm.9 From the perspective of the business group, tunneling may be conducted for good and bad reasons. A good reason is to mitigate the financing constraints of the parent to support positive NPV investment opportunities. A bad reason is to expropriate from minority shareholders of the listed firm.
From the perspective of the listed firm, once the shares have been purchased at the IPO, outside equity investors will not want to see any subsequent tunneling, regardless of the motivation, because tunneling strictly reduces the cash flow of the listed firm and consequently the value of the equity stake held by the outside equity investors.
The legal system protects the shareholders of the listed firm, and therefore any tunneling will be penalized if exposed, regardless of the ultimate motivation.
Expropriation Motivation and Punishment as a Deterrent to Tunneling
The probability of being discovered increases with the scale of tunneling (X0): the more tunneling by the parent, the more likely it is to be discovered.
Prob (discovery) = min(δX0, 1).
As a numerical example, δ=0.01% and the maximum probability of being discovered is 1. If tunneling is discovered, then the punishment is proportional to the crime committed: punishment=ξX.
For example, ξ=5 means paying a fine that is 500% of the amount tunneled.
9 The other direction of cash flow – from parent firm to listed subsidiary – is also possible. Theoretical and empirical studies have discussed the possibility of “propping” by controlling shareholders of listed firms to preserve future opportunities to expropriate from minority shareholders. For simplicity, such propping is not modeled here. 29 There may also be a cost associated with tunneling and covering up tunneling. For example, a special investment vehicle might have to be set up to enable tunneling and bribes might have to be paid to auditors and other individuals involved. The cost also increases with the tunneling activity. For simplicity, assume that it is also proportional to the size of tunneling, where cost = cX.
The benefit from expropriation decreases with the parent firm’s percentage stake in the subsidiary. The higher the percentage stake, the less meaningful it is to tunnel money out of the subsidiary. At the extreme, if the subsidiary is 100% owned by the parent, then there is no benefit at all to tunneling, because that constitutes taking money out of the right pocket and putting it into the left.10 Mathematically, before adding other costs, the benefit of transferring $X0 from the subsidiary to the parent is equal to X0 (1 – θ), where θ is the percentage ownership of the parent in the subsidiary.
Mitigating Financing Constraints as Another Motivation for Tunneling
Tunneling transfers financial resources from the subsidiary to the parent. In this setting, the receipt of this transfer may create value for the parent, because the parent may be initially financially constrained and have insufficient capital to take on all positive NPV projects. Assuming that lost investment opportunities are lost forever, there is an added benefit of transferring resources to the parent in that it enables more value-creating investment.11
Due to financing constraints, each additional $1 that flows into the parent firm will bring an additional amount of value due to the realization of positive NPV investment opportunities. As an increasing amount of capital flows in and the financing constraints are relaxed, the marginal benefit of an incremental dollar of capital flow will decrease, as positive NPV projects are depleted. Eventually, the additional benefit of a dollar investment could go to zero, as illustrated by the downward sloping curve in the figure below.
For simplicity, assume that the first dollar of investment brings in an additional value of $Y. The incremental contribution of the additional investment decreases with the total amount of tunneling. Assume that the decreasing relation between tunneling and the incremental value of another dollar can be approximated by a straight line, and that the slope of the line is k. k is intuitively a measure of the scale diseconomy of the resource transfer from the subsidiary to the parent firm to mitigate the financing constraints faced by the parent. Thus, if the amount of tunneling is $X0, then the total additional benefit from relaxing the financing constraints will be
½ * [Y + (Y-kX0)]*X0, which is the area of the trapezoid (shaded area in Figure 1).
Incremental contribution to firm NPV per $1 additional investment
10 Note that we neglect the potential benefit of using the cash flow of the listed sub to better mitigate the financing constraints faced by the parent. We discuss that possibility in great detail later. 11 Note that we assume that listed subs have much less financing constraints than parent firms. This is likely to be the case in China, given the institutional details that we discuss in the paper. 30 Y
Total amount of tunneling (X0)
Figure 1: The relation between the incremental value of relaxing financing constraints and the total amount of tunneling.
The Maximization Problem for the Parent
At any time, controlling shareholders solve the following problem to determine the amount of tunneling (X0*).
Max X0 (1-θ – c) – δX0 ξX0 + ½ * [Y + (Y-kX0)]*X0 X0.
We can see that from the parent’s perspective, there are two components of value creation. The first is expropriation from minority shareholders, which creates the value X0 (1-θ – c) – δX0 ξX0 . Within a certain range of the value of X0, this is an increasing and concave function of X0. The second is the benefit of mitigating financing constraints, which is ½ * [Y + (Y-kX0)]*X0. Within a certain range of the value of X0, this is also increasing and concave in X0. These two components are additive rather than multiplicative, meaning that they can each separately influence the total value creation but there is no interaction effect between them.12
This is a simple quadratic form of the objective function. Solving for the maximum, we obtain X0* = (1-θ – c + kY) / (2δξ+1).
Thus, for given parameter values of the cost of tunneling, the scale diseconomy of mitigating financing constraints, and the probability and penalty of the discovery of tunneling (c, K, δ, and ξ), X0* increases with Y (the value of the first dollar of additional investment in the parent, or a measure of the severity of financing constraints) and decreases with θ (the percentage ownership of the parent in the subsidiary). Intuitively, we would expect to see more tunneling if there is greater misalignment between the incentives of the parent firm and those of the subsidiary (a lower ownership stake θ), or more severe financing constraints (Y).
We can plot the “actions” of tunneling with given parameter values of c, K, δ, and ξ against a two-dimensional diagram of θ and Y as in Figure 2.
12 We could add higher order terms, for example making the δX0 ξX0 cubic rather than quadratic, but so doing would only add complexity to the problem without adding much intuition. 31 Y 1-θ-c+KY=0
1-θ-c+KY>0, do tunneling Tunnel to invest
1-θ-c+KY<0
θ
Tunnel to expropriate
Figure 2: The Tunneling with Given Parameter Values of c, K, δ, and ξ
At values above and to the left of the line 1-θ-c+KY=0, the firm will optimally engage in tunneling. Furthermore, the further the (Y, θ) combination is above the line, the more tunneling will be optimally performed. As this illustrative graph shows, for a small value of θ (less alignment between the interests of parent and subsidiary), tunneling happens even if Y is negative. These are cases of firms tunneling to expropriate from minority investors in the subsidiary rather than to mitigate financing constraints, despite the fact that doing so hurts the overall capital allocation efficiency of the group. This is marked on the plot by the area “tunnel to expropriate.”
There are also cases in which θ is high but Y is also high, and the firm might optimally engage in tunneling. In these cases, mitigating financing constraints serves as the main motivation for tunneling, as expropriation is less of a concern. Almost by definition, the efficiency of such tunneling, as measured by the sensitivity of parent investment to the interaction term between subsidiary cash flow and relative Q, will be the highest. This is marked on the plot by the area “tunnel to invest.”
Discussion of the Welfare Implications from the Perspective of the Business Group
Tunneling to expropriate is value destroying, despite the fact that everyone rationally expects it and outsiders are compensated for it by a lower IPO share price. In a rational equilibrium model, outside investors will not be systematically fooled, and thus get what they paid for and break even. Instead, the insiders bear all of the dead weight loss in welfare from tunneling, as they will have to sell the equity at a lower price to outsiders if some tunneling is expected. This is not a first best outcome for insiders, but as they cannot commit to refraining from tunneling after the IPO, they cannot achieve a first best outcome.
32 In some circumstances, tunneling to mitigate financing constraints may be value creating, even from the perspective of the business group. And, given that outsiders break even ex ante, this also means that tunneling could be value-creating from the perspective of the insider (here we treat the insider and the non-listed parent firm as the same).
The gain or loss from tunneling, from the perspective of the business group, will be as follows.
Gain: ½ * [Y + (Y-kX0)]*X0, which is the value created by reducing the parent financing constraints.
Loss: X0 c – δξX0, which is the value absorbed by the cost of tunneling and covering it up, or by the legal system in preventing tunneling. From the perspective of the business group, this value disappears rather than being transferred from one group member to another.
It is conceivable that, given different parameter values, the gain could be larger than, equal to, or smaller than, the loss. Thus, from the perspective of the business group, tunneling could, at least in theory, be value enhancing when the parent firm faces severe financing constraints.
33 Table 1 Summary Statistics
This table reports summary statistics for the listed sub and parent samples. Capital Expenditure is the change in the net value of fixed assets from year t-1 to year t, plus depreciation. We have three measures for cash flow: Cash Flow Measure 1 (for both the listed sub and parent firms), which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation – the listed sub (parent) net increase in trade credits with the parent (listed sub), where net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable; Cash Flow Measure 2 (only for the listed sub firms), which is defined as Cash Flow Measure 1 - income tax; and Cash Flow Measure 3 (only for the listed sub firms), which is defined as Cash Flow Measure 2 + (the net increase in bank debt + the net increase in equity offerings). The capital expenditure and cash flow measures are all adjusted by the total assets of the listed sub (parent) at the beginning of the year in the listed sub (parent) regressions. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = parent industry q- listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = parent industry q- listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. Industry Growth is the mean of the firm-level lagged sales growth in the same three-digit industry as calculated from the National Bureau of Statistics’ Annual Industrial Survey Database. Cash Flow Right is the percentage of total listed sub shares owned by the parent. Bank Ownership is equal to 1 when the listed sub owns shares in a local bank and 0 otherwise. Size is the total assets of the listed sub (parent) at the beginning of the year in thousand Yuan.
Variable Obs. Mean Median Mean Median Listed Sub Parent Capital Expenditure 1324 0.1090 0.0701 0.0721 0.0383 Net Trade Credits 1324 -0.0371 -0.0147 0.0286 0.0137 EBIT+ Depreciation 1324 0.0987 0.0949 0.0702 0.0585 Cash Flow Measure 1 1324 0.0620 0.0810 0.1021 0.0803 Cash Flow Measure 2 1324 0.0461 0.0597 —— —— Cash Flow Measure 3 1324 0.0783 0.0940 —— —— Industry Q 1324 2.8955 2.8223 2.8814 2.8223 Industry Growth 1324 0.1133 0.0911 0.1185 0.0948 Cash Flow Right of Parent 1324 51.5292 54.2300 —— —— Bank Ownership Dummy 1324 0.2266 0.0000 —— —— Size (thousand Yuan) 1324 220993 129575 533833 241418 Common Variables Relative Q 1324 -0.0627 0
34 Table 2 The Magnitude and Efficiency of Internal Capital Markets This table reports regression results on the magnitude and efficiency of internal capital markets for the total sample. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. In the listed sub (parent) regressions, Own Cash Flow is the cash flow of the listed sub (parent) and Other Cash Flow is the cash flow of the parent (listed sub). We have three measures for cash flow: Cash Flow Measure 1 (for both the listed sub and parent firms), which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation – the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable – the listed sub (parent) accounts payable; Cash Flow Measure 2 (only for the listed sub firms), which is defined as Cash Flow Measure 1 - income tax; and Cash Flow Measure 3 (only for the listed sub firms), which is defined as Cash Flow Measure 2 + (the net increase in bank debt + the net increase in equity offerings). Own Cash Flow in all of the regressions and Other Cash Flow in the listed sub regressions is Cash Flow Measure 1. In the parent regressions, the three cash flow measures are used to proxy for Other Cash Flow. The capital expenditure and cash flow measures are all adjusted by the total assets of the listed sub (parent) at the beginning of the year in the listed sub (parent) regressions. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = parent industry q – the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively. Other Cash Flow Measure 1 Measure 2 Measure 3 Measure Listed Sub Regression Parent Regression Parent Regression Parent Regression (1) (2) (3) (4) (5) (6) (7) (8) Own Cash Flow 0.2992 0.3064 1.1027 1.0959 1.1149 1.1226 1.0682 1.0714 (0.0374)*** (0.0376)** (0.0677)*** (0.0677)* (0.0689)*** (0.0687)*** (0.0684)* (0.0687)*** * ** ** Other Cash Flow 0.0034 0.0098 0.3700 0.3633 0.2210 0.2314 0.1316 0.1303 (0.0201) (0.0219) (0.0598)*** (0.0599)* (0.0445)*** (0.0445)*** (0.0380)* (0.0381)*** ** ** Relative Q 0.0198 0.0262 0.0171 0.0063 0.0130 0.0040 0.0156 0.0199
35 (0.0145) (0.0151)* (0.0262) (0.0271) (0.0264) (0.0265) (0.0267) (0.0281) Other Cash Flow * 0.0310 0.2026 0.2562 -0.0394 Relative Q (0.0204) (0.1278) (0.0964)*** (0.0822) Obs. 1324 1324 1324 1324 1324 1324 1324 1324 Adj_R2 0.09 0.10 0.31 0.32 0.30 0.31 0.29 0.29
36 Table 3 Cash Flow Rights, Bank Ownership, and Internal Capital Markets This table reports the parent firm regression results based on cash flow rights and bank ownership. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. Bank Ownership is Yes when the listed sub owns shares in a local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation – the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable – the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = the parent industry q - the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm- level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Variables Cash Flow Right Bank Ownership High Low Yes No (1) (2) (3) (4) Own Cash Flow 0.8758 1.0952 1.1038 1.0968 (0.1388)*** (0.0826)*** (0.1889)*** (0.0734)*** Other Cash Flow 0.0715 0.4417 0.1451 0.4432 (0.0355)** (0.0753)*** (0.0616)** (0.1431)*** Relative Q 0.0248 0.0008 -0.0748 0.0212 (0.0459) (0.0357) (0.0744) (0.0291) Other Cash Flow * Relative Q 0.2617 -0.1540 0.0723 0.2038 (0.1228)** (0.3139) (0.4889) (0.0828)** Obs. 662 662 300 1024 Adj_R2 0.24 0.36 0.25 0.33
37 Table 4 The Interaction Effect of Cash Flow Rights and Bank Ownership on Internal Capital Markets This table reports the parent firm regression results based on the interaction of cash flow rights and bank ownership. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. Bank Ownership is Yes when the listed sub owns shares in a local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = the parent industry q - the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Cash Flow Right High Low Bank Ownership Yes No Yes No (1) (2) (3) (4) Own Cash Flow 1.2263 0.8765 0.9183 1.3680 (0.3366)*** (0.1544)** (0.2559)*** (0.0869)*** Other Cash Flow -0.0234 0.1044 0.2214 0.5300 (0.1225) (0.0515)** (0.1041)** (0.0750)*** Relative Q -0.0041 0.0342 -0.0882 0.0055 (0.1164) (0.0505) (0.1182) (0.0367) Other Cash Flow * Relative Q 0.1264 0.3230 -0.2353 0.0708 (0.3896) (0.1578)** (0.1665) (0.0656) Obs. 160 502 140 522 Adj_R2 0.27 0.23 0.24 0.35
38 39 Table 5 Use of the Industry Q to Proxy for Investment Opportunity This table reports the parent firm regression results using the industry Q based on the interaction of cash flow rights and bank ownership. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. Bank Ownership is Yes when the listed sub owns shares in a local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Cash Flow Right High Low Bank Ownership Yes No Yes No (1) (2) (3) (4) Own Cash Flow 1.0074 0.9280 0.9646 1.3798 (0.2986)*** (0.1310)*** (0.1986)*** (0.0765)*** Other Cash Flow 0.1432 0.3859 0.2659 0.7451 (0.2357) (0.1798)** (0.1557)* (0.2150)*** Industry Q 0.0215 0.0374 0.0118 -0.0027 (0.0615) (0.0316) (0.0368) (0.0206) Other Cash Flow * Industry Q -0.0991 0.0188 -0.3598 -0.0112 (0.0845) (0.0484) (0.1570)** (0.0381) Obs. 160 502 140 522 Adj_R2 0.28 0.32 0.35 0.49
40 Table 6 The Use of Industry Growth to Proxy for Investment Opportunity This table reports the parent firm regression results using industry growth based on the interaction of cash flow rights and bank ownership. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. Bank Ownership is Yes when the listed sub owns shares of the local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Industry Growth is the mean firm-level lagged sales growth in the same three-digit industry as calculated from the National Bureau of Statistics’ (NBS’s) Annual Industrial Survey Database. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Cash Flow Right High Low Bank Ownership Yes No Yes No (1) (2) (3) (4) Own Cash Flow 1.4959 0.8382 1.0616 1.7709 (0.3084)*** (0.1469)*** (0.2405)*** (0.1030)*** Other Cash Flow 0.1394 0.2996 0.2187 0.6712 (0.2854) (0.1152)*** (0.2382) (0.1005)*** Industry Growth 0.2836 -0.0608 0.5212 0.1473 (0.4298) (0.2236) (0.5783) (0.2288) Other Cash Flow * Industry -0.0318 0.0915 -2.9912 -1.4927 Growth (1.7540) (0.8222) (1.2601)** (0.6951)** Obs. 160 502 140 522 Adj_R2 0.29 0.16 0.35 0.50
41 Table 7 The Effect of Capital Market Regulation on Internal Capital Markets This table reports the parent firm regression results based on capital market regulation. To capture the effect of capital market regulation against insider expropriation, we divide our sample year into two sub-samples: pre- regulation (2000-2003) and post-regulation (2004-2005). Bank Ownership is Yes when the listed sub owns shares in a local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = the parent industry q - the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year for the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The Whited (1980) adjusted standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Year 2000-2003 2004-2005 2000-2003 2004-2005 Bank Yes No Yes No Ownership (1) (2) (3) (4) (5) (6) Own Cash 0.5357 1.1486 0.5147 0.5458 1.2585 1.2452 Flow (0.1758)** (0.2163)** (0.2656) (0.2032)** (0.4114)* (0.2449)* * * * * ** ** Other Cash 0.3583 0.2111 0.2184 0.3749 0.1505 0.2348 Flow (0.1311)** (0.1257)* (0.1877) (0.1453)** (0.1669) (0.1254)* * Relative Q -0.0014 -0.0276 -0.0213 0.0007 -0.0700 -0.0180 (0.0130) (0.0236) (0.0376) (0.0142) (0.0585) (0.0255) Other Cash 0.1314 0.5711 0.4332 0.1166 0.2382 0.3929 Flow * Relative (0.1505) (0.2234)** (0.6608) (0.1608) (0.2368) (0.1981)* Q * Obs. 829 495 184 645 116 379 Adj_R2 0.18 0.33 0.21 0.21 0.24 0.39
42 43 Table 8 The Use of Firm Size to Proxy for Financing Constraints This table reports the parent firm regression results based on the interaction of cash flow rights and size. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. We rank the parent size at the beginning of the year and form two sub-samples for the bottom 33% and top 33% of the parent firms. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = the parent industry q - the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Cash Flow Right High Low Parent Size Bottom Top 33% Bottom Top 33% 33% 33% (1) (2) (3) (4) Own Cash Flow 1.1128 0.8414 1.2853 0.8940 (0.2580)*** (0.2240)*** (0.1515)*** (0.1385)*** Other Cash Flow 0.1655 -0.0517 0.4056 0.1817 (0.0806)** (0.1970) (0.1861)** (0.1161)* Relative q 0.1447 0.0717 -0.0717 0.0095 (0.0872) (0.1062) (0.0477) (0.0566) Other Cash Flow * Relative q 0.2709 0.1084 -0.3162 -0.9022 (0.1320)** (0.1248) (0.4957) (0.6642) Obs. 185 245 230 227 Adj_R2 0.35 0.17 0.39 0.31
44 Table 9 Excluding Firms with Negative Cash Flow This table reports the parent firm regression results using the relative Q based on the interaction of cash flow rights and bank ownership, but excluding firms with negative cash flow. Cash Flow Right is the percentage of listed sub shares owned by the parent. We define “High Cash Flow Right” as instances when the cash flow rights are above the median for the total sample and “Low Cash Flow Right” as instances when the cash flow rights are below the median for the total sample. Bank Ownership is Yes when the listed sub owns shares in a local bank and No otherwise. The dependent variable is Capital Expenditure, which is defined as the change in the net value of fixed assets from year t-1 to year t, plus depreciation. Own Cash Flow is the cash flow of the parent and Other Cash Flow the cash flow of the listed sub. In this table, we use only one cash flow measure to proxy for both own cash flow and other cash flow: Listed Sub (Parent) Cash Flow, which is defined as the listed sub (parent) EBIT + the listed sub (parent) depreciation - the listed sub (parent) net increase in trade credits with the parent (listed sub), where the net trade credits are equal to the listed sub (parent) accounts receivable - the listed sub (parent) accounts payable. The capital expenditure and cash flow measures are all adjusted by the total assets of the parent at the beginning of the year. Relative Q is the difference in investment opportunity between the parent and the listed sub. If the parent and listed sub are not in the same industry, then Relative Q = the parent industry q - the listed sub industry q. If the parent and listed sub are in the same industry, then Relative Q = the parent industry q - the listed sub individual q, where Listed Sub Industry Q is the mean firm-level Tobin’s q at the beginning of the year in the same three-digit industry as calculated from the total listed firms in China’s capital market and Parent Industry Q is matched from the listed sub industry q for the same three-digit industry. All of the regressions include year dummies (not reported) and all of the variables are winsorized at the 1% and 99% points. The regressions employ the firm-level fixed effects method. The standard errors are in parentheses. ***, **, and * denote significance at the 1%, 5%, and 10% level, respectively.
Cash Flow Right High Low Bank Ownership Yes No Yes No (1) (2) (3) (4) Own Cash Flow 1.1630 0.7555 1.2159 1.4346 (0.3986)** (0.2136)* (0.4745)* (0.1215)*** * ** ** Other Cash Flow -0.2933 0.2478 0.1234 0.4421 (0.5721) (0.1224)* (0.1421) (0.1552)*** * Relative q 0.0021 -0.0302 -0.1919 0.0191 (0.1416) (0.0745) (0.1236) (0.0530) Other Cash Flow * Relative q 0.0211 0.8220 -0.7419 0.2253 (0.2376) (0.4699)* (0.3471)* (0.6327) * Obs. 122 400 93 429 Adj_R2 0.24 0.25 0.38 0.43
45 46