آزاد سازی مالی، ساختار بازار و نفوذ اعتباری

This paper shows that the effects of financial liberalization on the credit market of a small and capital constrained economy depend on the market structure of domestic banks prior to liberalization. Specifically, under perfect competition in the domestic credit market prior to liberalization, liberalization leads to lower domestic interest rates, in turn leading to increased credit penetration. However, when the initial market structure is one of imperfect competition, liberalization can lead to the exclusion of less wealthy entrepreneurs from the credit market. This provides a rationale for the mixed empirical evidence concerning the effects of liberalization on access to credit in developing markets. Moreover, the analysis provides new insights into the consequences of foreign lenders’ entry into developing economies.

The last three decades have witnessed a wave of financial liberalization that has increased less developed countries’ access to international capital markets and to foreign financial banks. Financial liberalization was regarded as a means to foster competition in the financial sector, increasing access to credit for firms. As a result, an increased growth rate was expected. In some countries this was observed (Claessens et al., 2001 and Micco et al., 2007). In others, while the access and conditions faced by larger firms improved, the effects on smaller firms have been mixed or negative. For example, focusing on a specific type of liberalization, i.e., the entry of foreign banks, the study of Gormley (2010) for India showed a reduction in lending to small and medium enterprises after financial liberalization. What could explain this wide divergence in the results of financial liberalization? We suggest that one possible explanation for this contradictory evidence is that the impact of financial liberalization depends on the market structure of the domestic banking system prior to liberalization. Assume that financial liberalization is defined as the entry of foreign banks coupled to improved access to international capital markets. We show that if the domestic banking market is competitive, financial liberalization leads to the expected positive results. On the other hand, when markets are initially non-competitive, liberalization can lead to the exclusion of smaller (or weaker) businesses from the credit market, while the stronger companies are served by international banks. Thus, we show that the ex ante competitive situation in the financial market can explain the diverging observations on the effects of financial liberalization. The proponents of financial liberalization have argued that the entry of foreign banks would reduce costs by the use of improved financial technology; because they would demand improved financial regulation; and because they would be less corrupt and would not lend to related firms in banking conglomerates. Opponents of liberalization pointed out that foreign entrants are at a disadvantage because they have less information about clients and the domestic legal system, and because their lobbying capacity is lower than that of domestic banks. One of the implications of the informational disadvantages is that for foreign banks, the entrepreneurial rent received by entrepreneurs should be higher, ceteris paribus, than for domestic banks. Since increased competition by foreign banks will weaken domestic banks and reduce their lending, these two effects could combine to reduce total lending in the economy post-liberalization. This combination of advantages and weaknesses suggests that foreign banks tend to cherry pick firms with good internal accounting systems, solid financial positions and therefore a low probability of default. Given their better technology and lower loan origination costs, they can outcompete the domestic banks for these clients, leaving the riskier and more information-demanding clients for domestic banks, which can apply their relatively better monitoring ability. In these conditions, each type of bank specializes in the type of firm in which it has a comparative advantage, and this should lead to an increase in efficient lending. However, this conclusion rests on the assumption that the entry of foreign banks reduces or does not alter the cost of funds for domestic banks. If the entry of foreign banks were to raise the cost of funds for domestic banks, the effects of entry would be different. In particular, if the domestic banking system is imperfectly competitive, and the cost of funds is initially kept artificially low due to the lack of alternatives for savers, post-liberalization, the increased access to the international financial markets and the increased competition for domestic funds by foreign banks leads to an increase in the cost of funds for local banks. This may lead to a reduction in lending to the riskiest clients. There is empirical evidence for these effects. Rashid (2011), using data from 81 developing and emerging countries for the period 1995–2009, shows that foreign banks compete for deposits with the domestic banking system, and that this leads to an increase in the reliance on costlier (and more volatile) non-deposit funds by domestic banks. Moreover, the author presents evidence that increases in the share of deposits by foreign banks leads to a decline in credit to the private sector. Another paper, by Detriagache et al. (2008), uses data from the banking sector of 62 low income countries. They show that increased foreign bank penetration is associated with lower access to credit by the private sector and that foreign banks have a safer (less risky) loan portfolio.1 Why would foreign banks lend to a less risky loan portfolio than domestic banks in emerging economies? One possibility is that there are costs associated to operating in countries in environments that are different, with legal systems that are often inefficient or corrupt. In this regard, Mian (2006) has shown, using a sample of 80.000 loans in Pakistan, that the greater the physical and cultural distance between the headquarters of a foreign bank and a local subsidiary, the less willing they are to grant loans that are informationally complex, even though they are willing to lend to sound firms requiring relational contracting. Greater distance also makes banks less likely to bilaterally renegotiate loan contracts, and less successful at recovering from defaults. Several models have used information problems as the basis of their explanations for the observed phenomena. In their paper, Detriagache et al. (2008) propose a model in which, prior to the entry of foreign banks, domestic banks may find it convenient to reduce monitoring costs by pooling creditors, therefore implicitly providing a subsidy to weaker borrowers. When foreign banks arrive, with improved ability to screen hard numerical information, they compete in advantageous conditions for the financial needs of larger firms. This produces a segmentation of the credit market, reducing the average quality of firms being pooled by domestic banks and, under certain conditions on the monitoring cost, a pooling equilibrium is no longer feasible. In a separating equilibrium, loans become too expensive for weak borrowers, who drop out of the market. Foreign bank entry only benefits more transparent firms, while other firms are either indifferent or worse off. Rueda Maurer (2008), using the arguments of Detriagache et al. (2008), notes that different levels of credit protection may lead to credit constraints for opaque small and medium sized enterprises (SMEs), and tests this conclusion in a cross-section of 22 transition countries. Similarly, Gormley (2011) uses a model where information acquisition is relatively more expensive for foreign banks but these banks have a lower cost of funds. Initially, if screening costs are high, there is a pooling equilibrium. This equilibrium is broken by the entry of foreign firms, which cream-skim the firms that can use large amounts of capital (so that the lower financing cost of foreign banks offsets their higher screening costs), lowering the average quality of firms that remain for domestic banks. Now pooling may not be feasible and if screening costs are high, some banks exit the market, reducing lending to weaker firms. In this paper we present a complementary explanation for the observation that foreign entry may either increase or reduce firm’s access to loans. In our model it is the initial intensity of competition in the domestic financial market that drives the results. Consider a continuum of potential entrepreneurs with different wealth levels, who can invest in risky projects but whose individual assets are insufficient to carry out their projects without a loan from the credit market. Moreover, the ownership of assets by agents is private information and there is ex-post moral hazard, modeled as an incomplete verifiability of returns. The fact that a share of the returns is non-verifiable ex-post gives rise to entrepreneurial rent, as defined in Holstrom and Tirole (2011). The extent of entrepreneurial rent depends on factors such as the quality of the legal institutions in a country. The moral hazard problem reduces the surplus that can be contracted with a lender to an amount smaller than the surplus of the project. This means that entrepreneurs with small amounts of wealth will not receive loans and will be unable to set up their projects, even though it would be efficient to do so, if the capital were available. Moreover, we assume that domestic banks have the advantage of facing a lower rate of entrepreneurial rents, relative to foreign lenders. This is explained by better knowledge of the idiosyncracies of the legal system, or better monitoring ability due to immediacy. However, foreign banks have the advantage of a lower cost of funds, either because they have better access to less expensive international funds, or because of better financial and administrative technology. Consider a small, closed and capital constrained economy, so that under a competitive domestic financial market, the internal deposit rate is higher than the world rate. When the economy liberalizes, the entry of foreign banks does not change the competitive situation of the internal financial system. However, the supply of loanable funds ceases to be restricted, as the domestic financial system has access to lower cost foreign capital. Since foreign banks have a lower cost of funds, but face higher entrepreneurial rents, they lend to select wealthier entrepreneurs, whose contractible surplus relative to the loan size is higher.2 Therefore, these agents face lower lending costs after entry through two channels: the lower financing costs after liberalization and by the fact that they operate with more efficient foreign banks. The remaining entrepreneurs who gain access to credit operate with domestic banks, which face relatively lower entrepreneurial rents than foreign banks. These entrepreneurs benefit from the lower costs of funds after liberalization because competition transfers the lower rates from banks to borrowers. Moreover, the lower costs of funds leads to an increase in the number of potential entrepreneurs with access to loans for their projects. With a supply of capital that is no longer constrained by domestic supply, there is an increase in the number of entrepreneurs that receive funding. However, when the initial market structure is one of imperfect financial markets, the results are different. The reason is that both the credit and the deposit market are non-competitive. This means that banks pay a lower rate to savers than the rate paid in the case of perfect competition. When the economy liberalizes, the deposit rate for domestic banks rises because the foreign banks compete for deposits with the domestic monopoly bank. Again, foreign banks lend to the wealthier entrepreneurs, leaving the weaker entrepreneurs for the domestic banks. However, the increase in the cost of funds for the domestic bank may lead it to stop lending to marginal entrepreneurs, who are no longer profitable. Hence, the model is capable of explaining the conflicting evidence on the effects of liberalization on access to credit, via a mechanism that is complementary to those based on the observation that foreign entry leads to cream-skimming, making pooling equilibria inviable. While in our formal model we have imposed interest rate equalization in order to simplify the presentation, this is not needed for our results, which depend only on the fact that under imperfect competition, liberalization could result in the costs of funds for domestic banks going up. If, on the other hand, international interest rates are lower than those paid on deposits under an oligopsonistic banking system, the negative effects of liberalization disappear, i.e., the results are analogous to those obtained under a competitive financial system. Similarly, if there are competitive alternatives for savers in the closed economy, so there is no financial repression even under imperfect competition in the banking system, we recover the qualitative results obtained under a competitive financial system. We conduct a comparative statics analysis to investigate how changes in the various parameters of the economy alter the effects of financial liberalization. An important parameter is the entrepreneurial rent rate, which reflects, among other factors, the efficiency of the domestic legal system. When there is competition in the closed economy, a fall in the rate of entrepreneurial rents facing domestic and foreign banks increases welfare through two channels. Lower levels of entrepreneurial rents means that foreign banks can cut into the segment served by domestic banks and with their lower cost of funds, they increase the welfare of those entrepreneurs. Similarly, when domestic banks face a lower rate of entrepreneurial rents, firms that were excluded from the loan market now have access to it, which improves the welfare of their owners. When domestic banks are not competitive, lowering the rate of entrepreneurial rents facing the domestic bank increases access to credit, raising welfare as before. However, when it is the foreign banks that face lower rates of entrepreneurial rents, there is an additional effect. There is a direct effect because foreign banks can lend to more firms, increasing the welfare of the owners due to their lower costs. But there is also an indirect effect, because when agents switch to foreign banks, the domestic bank begins to lend to entrepreneurs that were not being served before. Our model explains the observation of Rueda Maurer (2008), who found that foreign entry interacts favorably with increases in creditor protection (i.e., reductions in the rate of entrepreneurial rents), leading to increases in lending to the smallest SMEs. Regarding the wealth distribution, the most interesting results are those that relate to ex ante competition. In that case, we show that more inequality in the distribution of wealth leads to higher benefits from liberalization in economies that are wealth constrained i.e., relatively capital poor. The opposite holds when the economies are relatively capital rich. The intuition is that in capital poor countries, concentrating wealth leads to a reduction in the number of agents that are excluded from the loan market because of the level of entrepreneurial rents, and in the open economy, these additional agents are not constrained in obtaining credit by the limited amounts of domestic capital. Finally, observe that the model has testable implications: add a variable that describes the competitive status of a country’s banking system (previous to the entry of foreign banks) to the standard regression of credit penetration on financial liberalization. This variable should be positively correlated with ex post credit penetration. Moreover, one can test for how the interaction between creditor rights and foreign entry described by Rueda Maurer (2008) is affected by the ex ante degree of competition among domestic banks.

6. Conclusions One of the puzzles of the recent experience with financial liberalization in developing countries has been the observation that in some countries access to credit falls rather than increases, as would have been expected from the increased competition and better lending technologies of multinational banks. This paper presents an explanation for this observation based on the notion that the effects of liberalization depend on the ex ante financial market structure in the country. In particular, we show that when the country has an initially competitive banking structure, financial liberalization is unambiguously beneficial, in terms of access to lower international rates, increased availability of capital, lower loan costs for some firms and increased access to credit. In the case of initially imperfect competition, access among weaker firms may decrease, because the cost of funds for domestic banks rises, as a result of increased competition for funds by foreign banks (or because some savers now have access to international markets). The increased cost of funds coupled to a reduction in lending to weaker firms are consistent with the observations of the effects of financial deregulation in the US (Berger et al., 1995), and there is recent evidence of these effects in other countries after financial liberalization (Rashid, 2011). Finally, observe that it is possible to test the model by including, in the models that examine the effects of financial liberalization on market access, a variable representing the competitiveness of the financial system previous to liberalization. The logic of our results is driven by the fact that under imperfect competition, the cost of funds for banks is lower than under competition. Therefore, once the financial system in liberalized, there is a rise in the cost of funds for banks, and increased competition for wealthier entrepreneurs, because of the entry of more efficient foreign banks. The end result is a possible reduction in lending to weaker entrepreneurs and lower credit penetration. The reader may question the relevance of our results, since it is unusual for the banking industry in a country to be controlled by a single firm. However, it is straightforward to show that the results extend without any changes to the case of a banking cartel, which is not uncommon in developing countries with closed financial markets. Moreover, the main results should continue to hold in more general models of imperfect competition, but it would make the algebra much more complicated without further gain in intuition beyond that obtained in the monopoly case. Nonetheless, it is possible to speculate on possible models of imperfect competition to tackle the issues studied here. One possibility would be to extend our model to the double Bertrand Competition model; i.e., having competition on both loans and deposits, in the absence of an infinitely elastic supply of funds as in Freixas and Rochet (1997, p. 63–67). In these models, there is an equilibrium in which banks make zero profits, there is one active bank and the profit margin (net of fixed costs) is positive. A second approach is to have a differentiated products Salop-type model with n banks equidistantly distributed on the unit circle and where, at each location there is a mass 1 of entrepreneurs and the wealth distribution is the same at each location. Again, this model would predict that the profit margin is positive and falls as the number of banks increases, suggesting that our results would continue to hold when there are a small number of banks prior to liberalization.