1、3200 英文單詞， 英文單詞，1.7 萬英文字符，中文 萬英文字符，中文 5250 字文獻(xiàn)出處： 文獻(xiàn)出處：Nadauld T D, Weisbach M S. Did securitization affect the cost of corporate debt? [J]. Journal of Financial Economics, 2012, 105(2):332-352.Did securitization affect
2、the cost of corporate debt?Taylor D. Nadauld, Michael S. WeisbachAbstractThis paper investigates whether the securitization of corporate bank loan facilities had an impact on the price of corporate debt. Our results sugg
3、est that loan facilities that are subsequently securitized are associated with a 17 basis point lower spread than that of facilities that are not subsequently securitized. We consider facility characteristics that are as
4、sociated with the likelihood of securitization and estimate the extent to which these characteristics are related to spreads. We document that Term Loan B facilities, facilities of B-rated firms, and facilities originate
5、d by banks that originate CLOs are securitized more frequently than other facilities. Spreads on facilities estimated to be more likely to be subsequently securitized have lower spreads than otherwise similar facilities.
6、 The results are consistent with the view that securitization caused a reduction in the cost of capital.Keywords: Securitization；Collateralized loan obligations (CLO)；Corporate bank loans1. IntroductionWhile much attenti
7、on has focused on the role of securitization in the mortgage market, relatively little has been paid to the securitization of corporate bank loans in the form of collateralized loan obligations (CLOs).1 This lack of atte
8、ntion is surprising given that both the volume of non-investment-grade bank loans and the number of newly originated CLOs spiked dramatically between the years 2002 and 2007, from over $125 billion of loans and 43 CLOs
9、in 2002 to over $540 billion of loans and 213 CLOs in 2007.2 One potential explanation for this pattern was that the popularity of CLOs created a demand for collateral that could be used to construct them (see, for examp
10、le, Gennaioli, Shleifer, and Vishny, in press). Consequently, banks active in the securitization business had incentives to adjust their lending behavior to increase the quantity of loans that could be used for collater
11、al, affecting the cost of capital for firms in the primary lending market.In this paper we estimate the extent to which the practice of securitizing bank debt influenced the cost of such debt for borrowers in the primary
12、 corporate debt market. To do so, we rely on a sample of over 4,000 loan facilities from the Dealscan database. We match these loan facilities to a database of CLOs provided by Moody’s. The Moody’s sample provides data o
13、n the characteristics and identity of all the securities serving as collateral in any Moody’s-rated CLO as of the first quarter 2009, which can be used to predict the attributes of ‘‘securitizable’’ loan facilities. We f
14、irst address the question of whether a loan facility that was ultimately securitized had a lower spread than an otherwise similar loan facility that was not securitized. Our estimates indicate that the answer to this que
15、stion is a clear ‘‘yes’’; they suggest that of the most common type of securitized facility, a B-rated, Term Loan B, a facility that is ultimately securitized has about a 17 basis point lower yield than an otherwise iden
16、tical loan that was not securitized.There are a number of possible interpretations to this result. It is possible that CLOs, for some unidentified reason, were more likely to add relatively low-yielding loans in their po
17、rtfolio, or that an omitted variable is driving both spreads and securitization demand in our specification. in spreads to be larger in B-rated debt than in Ba-rated debt because B-rated debt is securitized more frequent
18、ly. For this reason, in a sample of Term Loan B facilities only, we estimate the difference in spreads between B-rated facilities originated by securitization-active banks and compare those estimates to spreads on Ba-rat
19、ed facilities originated by securitization-active banks. Consistent with the securitization demand-driven hypothesis, our estimates indicate that the difference in the spread on loans issued by securitization-active bank
20、s to B-rated borrowers was almost 21 basis points lower than the spread on loans made by non-securitization-active banks to B-rated borrowers. In contrast, there is virtually no difference in spreads between Ba-rated loa
21、ns from securitization-active banks and non-securitization-active banks. In addition, if the demand to securitize Term Loan B facilities issued to B-rated borrowers was highest among securitization-active banks, we would
22、 expect to find differences in the spreads of Term Loan B facilities with these securitization-friendly attributes relative to Term Loan A facilities or revolving facilities. Our estimates indicate that the incremental e
23、ffect of being a securitization-friendly Term Loan B facility is 30.6 basis points relative to Term Loan As and is 20.8 basis points relative to revolvers. These equations imply that the drop in spreads for securitizatio
24、n-friendly facilities does not reflect a bank-specific characteristic, such as an increased risk tolerance, since this characteristic would likely be reflected in the spreads of securitization-friendly Term Loan A facili
25、ties and revolvers as well.Another way to do the comparison is by estimating the factors that affect the differences across facilities within a particular loan. This approach has the advantage of having common underlying
26、 risks within the loan, although the data are limited to loans that contain multiple types of facilities. In this case, the differences in spreads between Term Loan B and either Term Loan A or revolvers decline with char
27、acteristics associated with securitization, again consistent with the view that demand for collateral for CLOs affects the pricing of the facilities in the primary market.Overall, the results are consistent with the view
28、 that CLOs’ demand for collateral affected pricing in the corporate debt market. The effect of the securitization-driven debt market on the housing market during the 2002–2007 period is well-known. Shivdasani and Wang (2
29、011) and Axelson et al. (2011) show that the debt market was an important driver of the leveraged buyout (LBO) boom during that period. Our results suggest that securitization had a pricing impact on the corporate debt m
30、arket more broadly.Perhaps the most closely related work to ours is Ivashina and Sun (2011), who provide evidence that demand for loans from institutional investors in the secondary market, measured by the time a loan re
31、mains in syndication, is negatively related to spreads on these loans. The authors present evidence that some but not all of their effect is due to demand from CLOs. Our analysis extends their work in a number of directi
32、ons. First, we focus exclusively on the role of securitization in loan pricing. Second, we provide direct evidence on the types of loans most associated with securitization, in particular, Term Loan B facilities, loans i
33、ssued by banks that also issue CLOs, and loans that are made to B-rated borrowers. Third, we use these characteristics as a way to identify the causal nature of the relation between pricing and securitization. To the ext
34、ent that Ivashina and Sun’s (2011) and our identification strategies are different from one another and our results are nonetheless similar, each paper’s analysis complements the other’s and strengthens the conclusion th
評(píng)論
0/150
提交評(píng)論