مخاطرات اخلاقی و کژ گزینی در مدل منشاء توزیع اعتبار بانکی

Bank credit has evolved from the traditional relationship banking model to an originate-to-distribute model. We show that the borrowers whose loans are sold in the secondary market underperform their peers by about 9% per year (risk-adjusted) over the three-year period following the initial sale of their loans. Therefore, either banks are originating and selling loans of lower quality borrowers based on unobservable private information (adverse selection), and/or loan sales lead to diminished bank monitoring that affects borrowers negatively (moral hazard). We propose regulatory restrictions on loan sales, increased disclosure, and a loan trading exchange/clearinghouse as mechanisms to alleviate these problems.

The historic credit crisis of 2007–2008 brought an important question sharply into focus—to what extent should bank credit be allowed to evolve from its traditional relationship banking model to the transaction-oriented model that has largely emerged over the last two decades? This fundamental shift in banking has been due to the explosive growth in the secondary syndicated loan market.1 The presence of this market transforms bank credit to an “originate-to-distribute” model, where banks can originate loans, earn their fees, and then distribute them to other investors in a largely opaque manner. This shift to the originate-to-distribute model of bank credit has important implications for all market participants, including the originating banks, the participating loan investors, the borrowing firms and the regulators. The banks’ superior information about their borrowers gives rise to concerns about adverse selection—are the banks selling off loans about which they have negative private (unobservable) information? In a perfect market, this should lead to a breakdown of the secondary loan market due to the classic “lemons” problem. The issue of adverse selection is important from the perspective of the participating loan investors as well—can they trust that the bank selling the loan is doing so due to legitimate motives (like capital relief and risk management) rather than due to negative private information? Alternatively, does it lead to moral hazard in terms of an impairment in the monitoring function of banks, thereby having a negative effect on the borrowers? There are several policy questions that arise from this debate. Should the regulatory authorities restrict the originate-to-distribute activities of banks? Should they enforce enhanced disclosure of the banks’ activities in the loan sales market? How are the borrowing firms being affected, in the long run, by banks moving from relationship banking to the originate-to-distribute model of credit? Does this shift lead to value creation or value reduction in the corporate sector? These questions are, ultimately, empirical ones. Using extensive data from the syndicated loan market, this paper is the first empirical investigation of these important but as yet unanswered questions.2 Banks could sell loans in the secondary market due to negative private information about the borrower, or for legitimate reasons such as capital relief, risk diversification, improving balance-sheet liquidity, and reducing financing frictions and their cost of capital. The positive effects of loan sales on banks have led to a point of view that the originate-to-distribute model of bank credit is “socially desirable”.3 There is also a vast literature on banks being “special”, since they generate proprietary information about the borrowers in the course of lending to them.4 The loan buyers who do not have a lending relationship with the borrowers are then likely to be at an information disadvantage when buying a loan originated by a relationship bank. This could lead to moral hazard and adverse selection problems (Gorton and Pennacchi, 1988; Pennacchi, 1988). Banks that sell loans would have a reduced incentive to engage in costly screening and monitoring of the borrowers. In addition, they would have an incentive to sell the loans of the borrowers about whom they have negative private information. Duffee and Zhou (2001) examined these issues in a theoretical setting with bank loans and the presence of credit risk mitigation via the default swap market or the loan sales market. From a borrower's perspective, there are potentially positive as well as negative consequences of their loans being sold in the secondary market. The positive effects include a lower cost of capital (Gupta et al., 2008), increased access to debt capital (Drucker and Puri, 2008), and information effects (Gande and Saunders, 2008). The negative effects include a breakdown of lending relationships, reduced monitoring which could lead to suboptimal investment and operating decisions, harsher covenants (Drucker and Puri, 2008), and difficulties in renegotiation (Carey et al., 1993).5Parlour and Plantin (2008) presented a theoretical model which embeds some of the bank and borrower incentives and effects outlined above. However, from an empirical standpoint, it is not clear which of these effects dominate. Furthermore, if the originate-to-distribute model of credit creates incentives for banks to originate bad loans and then sell them off in the secondary market, such borrowers should underperform their peers in the long run. Since theoretical arguments on this issue can go either way, it needs to be resolved empirically. Our paper is the first one in the literature to empirically examine the long-run performance of borrowers with and without an active secondary market for their loans. The existing empirical literature has largely focused on the impact of bank loan announcements on the borrowers’ stock returns. Most studies have shown that loans are “special”—their announcements elicit positive short-term abnormal returns for borrowers, in contrast to the announcement effect of most other forms of corporate financing such as common stock, preferred stock, straight debt and convertible debt. 6 This result has been somewhat reversed by Billett et al. (2006), who show that firms announcing bank loans suffer negative abnormal returns in the long run. The literature on the effects of loan sales on the borrower's stock price is rather sparse. While Dahiya et al. (2003) documented a negative announcement effect of the sale of a borrower's loans by its lending bank, Gande and Saunders (2008) documented the opposite (positive) announcement effect. However, none of these studies has measured the long-run performance of the borrowers whose loans trade in the secondary loan market. We study a large sample of 1054 borrowers, the largest sample analyzed in this literature thus far. Our results show that borrowers with an active secondary market for loans significantly underperform their peers by about 9% per year on a risk-adjusted basis over the three-year period subsequent to their loans first being traded in the secondary market. This result is robust to most techniques of measuring long-run abnormal returns. The underperformance is stronger for small, high-leverage, speculative-grade (SG) borrowers, which is intuitive since these are precisely the firms where moral hazard and adverse selection problems may be more severe. For the borrowers that have an active secondary loan market, using Tobin's q we find a significant reduction in value (as a percentage of total assets) of about 14% over three years when compared to their peers. The significant long-run underperformance and value reduction of borrowers with an active secondary loan market is a striking result, for which we offer two possible explanations. First, banks may be cherry picking by preferentially selling loans of the borrowers about whom they have negative private information that is unobservable to outsiders. Alternatively, banks may be knowingly originating some lemons, primarily to expand their origination fee based income, since they are able to sell these loans in the secondary market to outside investors (mostly non-bank financial institutions and hedge funds).7 In a perfect market, reputation concerns should prevent a bank from cherry picking and/or selling lemons on a systematic basis. If it is still happening, it is perhaps an indication of a market failure, where the investors have not (yet) recognized the adverse selection that they are facing in the secondary syndicated loan market. Our second explanation is based on the moral hazard argument. When borrowers lose the discipline of bank monitoring, they may be more prone to making suboptimal investment and operating decisions, which may lead to a negative long-run performance and value reduction.8 Based on our tests and results, it is not possible to clearly distinguish which one of the two explanations dominates. It is likely that both these mechanisms play some role in explaining our results. In addition, despite our extensive robustness tests, there is always a possibility that some of the abnormal returns that we observe may be partly due to inadequate risk adjustments. While the borrowers with an active loan market underperform their peers, those without an active loan market do not show any significant long-run underperformance. Our findings refine the results of Billett et al. (2006), who claim that bank loans are not special. This is especially interesting in light of the results of Gande and Saunders (2008) who claim that banks are “special” even in the presence of a secondary market for loans.9Our paper shows that bank loans are still “special”, but only if the bankers do not sell them. Our results have important policy implications for regulators. Whether the underperformance and value reduction of borrowers with an active secondary loan market is due to banks originating and selling lemons, or due to diminished monitoring, it raises serious questions about the extent to which the originate-to-distribute model of bank credit is “socially desirable”. While there are clear benefits of enhancing the liquidity of the secondary syndicated loan market, we demonstrate some of its long-term undesirable consequences. It is likely that one of the major reasons for the latter is the highly deregulated nature of the secondary syndicated loan market. Should the regulators impose restrictions on the sales of bank loans by originating banks? Perhaps. At the minimum, they could require the originating banks to retain a certain proportion of the loans on their balance sheet to limit the moral hazard and adverse selection problems. Also, there must be additional disclosure requirements about the loans being traded in the secondary market, along with disclosure about the market participants that are trading them. A loan trading exchange with a clearinghouse could be a possible solution. It is certainly clear that the originate-to-distribute model of bank credit needs to be modified, and the transactions made more transparent. The rest of our paper is organized as follows. In Section 2, we provide information about our data along with some descriptive statistics. In Section 3, we explain the different methods used in this paper for examining the long-run performance of the borrowing firms. We describe and interpret our results in Section 4. Section 5 concludes.

We investigate the effects of the transition in bank credit from the relationship banking model to the “originate-to-distribute” model, on a large sample of borrowers in the syndicated loan market. This shift has mainly been due to the growth in the secondary market for syndicated loans, which has allowed banks to sell loans to participating investors in a largely opaque manner. While the prior literature has documented several benefits of the loan sales market for the banks as well as the borrowers, the long-run effects of the existence of this market on the borrowing firms have never been examined. This is precisely what we study in this paper. When banks sell syndicated loans in the secondary market, it raises moral hazard and adverse selection questions. Are the banks selling lemons, i.e., the loans of borrowers about whom they have negative private information that is unobservable to outside investors? Are they deliberately originating bad loans to enhance their fee income, just because there is an active secondary market where they can sell these loans? How does this affect the incentives of the bank to monitor their borrowers? Is the severance of their lending relationship harmful for borrowers? What is the consequent impact on the long-run valuation of the borrowers? Theory alone is insufficient to answer these questions, as there are both positive and negative effects of the secondary market for syndicated loans. Ultimately, these issues must be resolved empirically, which is the focus of our paper. We find that borrowers with an active secondary market for their loans underperform their peers by about 9% per year in terms of annual risk-adjusted abnormal returns, over the three-year period subsequent to the initial sale of their loans. These abnormal returns are largely concentrated among small, high-leverage, speculative-grade borrowers. In addition, the borrowers with an active loan market suffer a valuation loss of about 11–14% of the value of their total assets over a three-year period when compared to their peers. We offer two explanations for our results. First, banks may indeed be selling lemons based on their unobservable private information about the borrower. This would be an indication of a market failure, since, in an efficient market, reputation concerns should inhibit such actions on the part of banks. It bears a remarkable resemblance to the events that have unfolded in the monumental subprime mortgage crisis that began in 2007. Second, borrowers might suffer due to their diminished relationship with banks, since selling the loans removes the discipline of bank monitoring. This might diminish firm value in the long run. Our tests cannot conclusively confirm one or the other explanation. In addition, there is always the possibility that the abnormal returns that we find in our paper are, at least partly, due to inadequate risk adjustment in our factor models. We try and alleviate these concerns by estimating an array of alternative factor models and abnormal return calculations. Our results are robust across the different model specifications and approaches that we use in this paper. We also find that the borrowers without an active secondary loan market do not suffer any negative long-run effects after obtaining syndicated bank loans. This reaffirms the inference that, for some borrowers, bank loans are indeed “special” when compared to other forms of corporate financing such as equity or public debt, all of which result in a negative long-run performance of the firms raising capital. Our results refine the findings of Billett et al. (2006) who documented a negative long-run performance for all firms that borrow in the syndicated loan market. We show that this negative long-run performance is limited to the borrowers that have an active secondary market for their loans. This is especially interesting in light of the results in Gande and Saunders (2008), who inferred that banks continue to be special even in the presence of a secondary market for loans. Gande and Saunders also suggested that secondary market loan trading is valuable to equity investors. Their inferences are based just on the announcement effect of loan sales on the borrowers’ stock price, whereas we examine the long-run performance of the borrowing firms. Our results suggest that bank loans are “special”, but only for the subset of borrowers that do not have secondary market trading in their loans. We show that the originate-to-distribute model of bank credit may not entirely be “socially desirable”, since we document some of the negative effects of this model on the long-run performance and valuation of borrowers. Our results have important policy implications for regulators. The highly deregulated nature of the secondary loan market is perhaps one of the reasons for the moral hazard and adverse selection problems that we detect. One solution could be to impose restrictions on the sale of the loans that the banks originate, in terms of requiring them to hold at least a certain percentage of those loans on their books. This would hinder banks from originating bad loans as well as preserve some of the benefits of bank monitoring of borrowers. There should of course be additional disclosure requirements on all participants in the loan sales market, in order to reduce the occurrence of adverse selection. Lastly, the establishment of a loan trading exchange with a clearinghouse could benefit all market participants by way of greater transparency and regulatory oversight.