فعالیت های خارج از ترازنامه تحت کژ گزینی : تجربه اروپا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|20495||2013||15 صفحه PDF||سفارش دهید||11630 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Economic Behavior & Organization, Volume 85, January 2013, Pages 176–190
In the crisis that started in 2007, banks’ off-balance sheet activity has been blamed for flooding the market with low-quality assets and contributing to spreading risk throughout the economic system. Nevertheless, this view is hardly sustainable within the context of sophisticated markets. This paper puts forward an alternative interpretation of the off-balance sheet market, the so-called adverse selection hypothesis. According to this hypothesis, an adverse selection problem characterizes the relation between banks and their counterparties in off-balance sheet deals. Empirically, this implies that off-balance sheet activity is expected to be negatively related to failure risk, whereas it is expected to be positively related to the quality of the assets used for off-balance sheet operations. We test the adverse selection hypothesis for a sample of banks in the 27 member countries of the European Union during the pre-crisis period 1996–2006 and the crisis period 2007–2009. In addition, we check for possible differences between banks in the first 15 members of the European Union and those in the 12 new members.
The financial turmoil that started in 2007 has changed the view on off-balance sheet (OBS) activity. Before the crisis, paraphrasing Alan Greenspan, OBS items were thought to have reinforced the flexibility, efficiency, and resiliency of the financial system. After the crisis, securitized assets, guarantees, committed credit lines, credit default swaps, and so forth were identified as part of the financial “trap” at the very heart of the downturn. OBS activity was considered responsible for having flooded financial markets with low-quality assets in a multilayered agency problem (Ashcraft and Schuermann, 2008), with final investors suffering the corresponding loss.2 However, for this new view to hold, investors operating in the OBS market would have to be willing to take disproportionate risks by investing in assets with a default probability close to one. This widespread lack of risk aversion is hardly acceptable as regards sophisticated investors in sophisticated markets. By contrast, if agents are assumed to be risk averse, the rapid expansion of the OBS market in the pre-crisis period cannot be explained in terms of low-quality assets systematically provided by “greedy” issuers and acquired by “innocent” investors. An alternative explanation is to look at OBS activity in terms of an adverse selection problem between banks and risk-averse investors. Since Akerlof's (1970) pioneering work, we know that, with risk-averse investors, information asymmetries between issuers and investors can lead to an adverse selection problem that can result in a no-trade equilibrium, that is, disappearance of the market. Under the assumption that investors are risk averse, the no-trade result associated with adverse selection can be avoided in the OBS market if high-quality assets offered by high-quality issuers are predominant. The reason is that this option for quality would be a positive signal sent by the issuers of OBS items and would give investors a guarantee that their investment is relatively safe. Indeed, as Shin (2009) points out, the severity of the credit crisis lies precisely in the fact that low-quality OBS items were not passed on to final investors. The assets in the hands of investors were mostly high quality, whereas low-quality assets were sitting on the balance sheet of financial intermediaries. This paper aims to provide an explanation of the operation of the European OBS market in terms of an adverse selection problem. Since this explanation provides an alternative view of the role of OBS activity in the current crisis, our analysis focuses on the pre-crisis period. However, the crisis period is also studied to gain an overall view of bank OBS activity. In a more specific way, our hypothesis is that an adverse selection problem characterizes the relation between a bank and its counterparties in OBS deals. According to standard economic theory (Freixas and Rochet, 2008), we assume that issuer banks are the risk-neutral, fully informed side of that relation, whereas their counterparties are the risk-averse, uninformed side. The latter's aim is to avoid selecting “lemons,” that is, low-quality assets from low-quality banks. Note that the fact that investors pursue this goal defines the degree to which banks are able to engage in OBS activity. Therefore, if investors receive signals that the issuer is a high-quality bank providing high-quality items, this bank's chances to participate and make profits in the OBS market are enhanced. From this perspective, banks’ risk indicators define the capacity to engage in OBS activity, and we expect to find that safe banks issuing safe OBS items should engage in this activity to a larger extent. More specifically, our hypothesis entails that banks’ OBS activity is expected to be negatively related to risk measures signaling a bank's probability of failure (such as measures of insolvency or portfolio risk), but it is expected to be positively related to risk measures signaling that banks are using high-quality assets for OBS activity (such as measures of credit or liquidity risk). Accordingly, the adverse selection hypothesis rejects associating OBS activity in the banking industry with “junk” assets. The relation between OBS activity and banks’ risk exposure has been thoroughly analyzed, and several hypotheses (briefly discussed in Section 2) are proposed in the literature on OBS activity to explain the relation. Nevertheless, this paper extends earlier analysis in several manners. First, it puts forward and tests the adverse selection hypothesis. This hypothesis enables us to explain the behavior of banks’ counterparties in OBS activity in terms of a standard adverse selection problem. As a result, the main focus of the analysis of the OBS market is not on the decision of issuer banks, but on their counterparties in OBS deals. Even more specifically, the key of the analysis is these investors’ risk aversion. Indeed, besides the contractual means that help to overcome the lemon problem in the OBS market, risk aversion would provide an additional self-regulating mechanism that controls for the quality standards acceptable in this market. Second, our analysis palliates the bias toward the US banking system in the research about the relation between OBS activity and risk. To our knowledge, this is the first study that specifically examines this relation in the European Union (EU) banking system. In addition, we separately analyze the OBS market in the first 15 members of the EU and in the countries that recently joined. Using a sample comprised of individual banks in the 27 EU member countries, we separately study the pre-crisis period 1996–2006 and the crisis years 2007–2009. This allows us to determine whether the financial downturn that started in 2007 modified the way in which banks operate in the OBS market. Specifically, we test the adverse selection hypothesis for the entire sample for both periods. Then, to check for any possible differences between mature and in-transition economies, we test this hypothesis for two subsamples. The first subsample comprises banks in the first 15 members of the EU (EU15), and the second banks in the 12 new EU joiners (EU12).3 Our testing process examines the nature of the relation between OBS deals and risk measures that reflect the probability of banks becoming insolvent. To capture this probability, we use the Z-score (ZS) ( Altman, 1968), as well as its two additive components, and the capital-to-assets ratio (EA). In addition, we analyze the relation of OBS activity with credit and liquidity risk. For this analysis, we use cross-sectional estimation based on average values and panel data techniques. For the entire sample, in the two periods studied, and regardless of the techniques used, our results support the adverse selection hypothesis; that is, OBS activity is negatively related to proxies of failure risk and positively related to credit and liquidity risk. Nevertheless, when first and new members are separately analyzed, a significant difference arises. The adverse selection hypothesis is consistent with bank behavior in EU15 for the periods 1996–2006 and 2007–2009, but not consistent with evidence obtained for the EU12 subsample in both periods. This difference suggests that the description of banking institutions using low-quality assets for OBS deals applies to the EU12 banking sector.
نتیجه گیری انگلیسی
For the adverse selection hypothesis OBS deals are not a means used by banks to get rid of that portion of risk they do not want to hold in their books. These deals are mainly a profit-seeking activity whose target consists of investors who place a high value on relatively safe assets issued by relatively safe institutions. Accordingly, the capacity to engage in OBS activity to a large extent requires, on the one hand, controlling indicators of failure risk. On the other hand, it entails that high-quality assets be used for OBS activity. Thus, banks’ OBS activity is expected to be negatively related to some types of risk (insolvency and portfolio risk) and positively related to other types (liquidity and credit risk). This paper tests these empirical predictions in two separate periods, 1996–2006 and 2007–2009, for the whole EU banking system and two subsamples, EU15 and EU12. Thus, we can check for any differences concerning the adverse selection hypothesis before and during the crisis, and between mature and in-transition economies. For EU and EU15,results support the conclusion that the adverse selection hypothesis is consistent with bank behavior in the OBS market during the pre-crisis and crisis periods. This result is obtained under cross-sectional analysis, using panel data techniques, and subjecting the results to different robustness tests. In an adverse selection problem, the option for quality is a means of avoiding the no-trade result. For the adverse selection hypothesis proposed in this paper, investors’ risk aversion is a self-regulating mechanism that provides incentives for banks to choose that option for quality in the OBS market. Our results suggest that in the two periods studied this option has been predominant in the OBS market of the EU15 banking system. The OBS activity in this banking system seems to have been a cherry-picking practice, with relatively high-quality assets used for the OBS deals. This may have helped reduce the risk f facing a chain of bank failures linked to (intra-European) OBS activity. Moreover, the contagion risk associated with OBS activity may also have been reduced by the fact that the EU15 banks that have been more active in the OBS market score better in insolvency and portfolio risk. Nevertheless, our results for the EU15 subsample also have a negative side. Consistent with Shin (2009), the test of the adverse selection hypothesis suggests that, when the crisis started, EU15 banks that had engaged in OBS activity to a large degree had low-quality assets sitting on their balance sheet. Indeed, those assets remained there during the crisis period. Regarding OBS activity in the EU12 subsample, the market discipline hypothesis—well known in the literature on OBS activity since the beginning of the 1990s—provides a better explanation of bank behavior than the adverse selection hypothesis for the two analyzed periods. This result implies that the banking institutions in new EU members used low-quality assets for OBS deals, thus supporting the association of these deals with “junk” assets.