Printer Friendly

An investigation of underwriting fees for asset-backed securities.

I. Introduction

One of the largest financial markets is the market for asset-backed securities (ABS). Despite the growth and size of the market, no research has been done, to our knowledge, on underwriting fees in this market. The topic of underwriting fees has been previously examined in debt markets. For example, Livingston and Miller (2000) find an inverse relationship between underwriter prestige (market share) and underwriting fees charged in corporate debt markets. Burch, Nanda and Warther (2004) find that loyalty (repeat business) leads to lower fees for common stock offers, but found the opposite holds true for debt offers. In this paper, we investigate whether the relations between underwriter prestige and underwriting fees and between loyalty and underwriting fees that have been found to exist in other debt markets exist in the ABS market.

The topic of underwriting fees for ABS is of interest for two main reasons. First, ABS is a relatively new financial market leading to relatively little prior research being done on it. Second, at least part of the financial crisis in 2009 has arguably been attributed to the growth in the market of these types of securities. To better understand ABS, may help better understand potentially one of the influencing factors of the financial crisis.

ABS are inherently different from other debt instruments since they are collateralized by specific receivables. The growth of the ABS market can be attributed to two main factors: demand by investors in search of spread (above "safer" fixed income securities such as government and corporate debt) and supply by lenders wishing to offload receivables. Using a proprietary database from Bloomberg LP this paper provides an overview of the ABS market that is more detailed than the existing literature. Further, using a methodology similar to that used in Livingston and Miller (2000), this study explores the relationship between ABS underwriter prestige and underwriting fees. This study presents two key findings. First, the relation between underwriter prestige and underwriter fees is found to be positive and statistically significant, indicating that more prestigious underwriters charge higher fees. Second, the analysis identifies a positive relation between underwriter fees and loyalty, indicating that the more an issuer uses the same underwriter, the higher the fees that are charged.

The rest of this paper is organized as follows. Section II provides an overview of the market for ABS. Section III provide a literature review and hypothesis formulation. Section IV describes the methodology, Section V describes the results, and Section VI concludes.

II. Overview of the Market for ABS

In this section, we first provide an overview of the market for ABS, and then take a more detailed look at ABS collateral types, underwriters, ratings and weighted average life. The source of all data is the Bloomberg Data License product. All non-private placement U.S. asset-backed securities issued from January 1st. 1999 through December 31, 2006 are included in our analysis. The period selected represent a period in which the data are richest and most complete, and spans a number of economic cycles.

Since its inception in 1985 with First Boston's sale of lease-backed notes by Sperry Lease Finance Corporation, the U.S. ABS market has grown considerably (van Eck 1995). In addition to the sizeable U.S. market, there are markets in Europe and Japan. Growth in ABS extends to other markets such as Korea, Taiwan, and Greece (Lester, Asaria and van der Linden 2002), (Park, Han, and Kim, 2002) (Pergamalis 2003). Table 1 shows the U.S. asset-backed securities market is large and growing as compared to the U.S. corporate bond market.

The growth of the market can be attributed to two main factors: demand by investors in search of spread above "safer" fixed income securities such as government and corporate debt, and supply by lenders wishing to off-load receivables. As the market grows, so does the ever increasing different types of ABS that are created by Wall Street. Investors in ABS are typically institutional investors of fixed income securities seeking portfolio diversification and/or higher yields. Individual investors also indirectly invest them as many bond funds will hold ABS.

ABS are financial instruments whose cashflows are "backed by" installment loans or other receivables. An issuer of an ABS forms a trust that consists of loans, generally characterized by some common factor; for example, automobile loans. From this trust, the issuer will create a series of classes (tranches) of securities that make up the deal. These classes receive their cashflows from the trust. The timing of the cashflows to an individual class depend on the priority of the class within the overall structure and the payment behavior of the underlying loans. The deal structure and rules associated with priority of cashflows impacts the payments received by the ABS investor.

The payment behavior of the underlying loan holders is important as well. Loan payments are generally differentiated by normal scheduled payment vs. prepayments. Further, prepayments are often differentiated by either partial prepayment or full prepayments. Scheduled payments are the expected monthly payment the borrower agrees to pay on a monthly basis that includes both principal and interest. The borrower generally has the right to either partial prepay, pay additional principal above and beyond what is required each month or can full prepay, pay off the loan in its entirety. In either case, the additional cashflow paid by the borrower which in turn is paid into the trust will often be "passed-through" to the investor of the ABS. This means the ABS expected cashflows and actual cashflows can vary greatly depending on the prepayment behavior of the borrowers.

Within the market of ABS, there are numerous deal types--overall characteristic of the loans or collateral that backs a particular ABS issue. Some of the more common deal types of ABS are home equity loans, home equity line of credits, automobile loans, and credit card receivables. Each of these deal types as well as the numerous others have subtle differences, but the common element is the cashflow paid by the loan borrower or credit card holder is ultimately used to pay the ABS investor. Please see Table 2 for details of outstanding and new issuance of ABS, by various deal types.

In addition to defining an ABS by the deal type, specific classes that comprise the overall deal structure are often defined by class or tranche descriptors. These descriptors are designed to provide the investor in a specific class a general understanding of the payment schedule/structure of the particular class they are investing in and how their bond relates to the overall deal structure. Each deal has associated rules for how to distribute the cashflow received from the underlying collateral. As the ABS market developed so did the complexity of the associated payment rules and in turn various terms/class descriptions used by the market. In some instances, class payment rules require the use of multiple descriptors to accurately provide description of the cashflow payment structure. The prospectus will often include some of the more market accepted descriptors as part of the description of the ABS.

One of the distinguishing features of ABS in contrast to other fixed income securities is the concept of prepayments. Unlike traditional corporate bonds that generally pay interest during the term of the bond and then pay the principal at maturity or if a bond is called, ABS pays principal along with interest throughout the term of the bond. In addition, the principal component of the cashflow is a function of the prepayment behavior of underlying collateral which greatly impacts the overall term of the ABS. Although ABS are assigned a maturity date at issue, ABS rarely will remain outstanding until their legal maturity date. In both MBS and ABS, market participants generally use another value to measure "term" of an MBS/ABS--weighted average life (WAL) which is defined as time weighted average time of receipt of principal. When an ABS is issued, it is quite common for the lead manager of the deal to provide an original WAL value which is calculated using an assumed average prepayment rate.

Similar to corporate bonds, ABS are often quoted in terms of basis points spread to a corresponding benchmark security. For example, an ABS with an original WAL of 5 years will often be priced as a spread to the U.S. 5 year treasury. Because the market for ABS is typically over the counter and cashflows are highly dependent on prepayment behavior of the underlying borrowers, there is much variability in terms of the price or value of a given ABS. Overall, there is less price transparency in ABS market as compared to other fixed income markets such as government bonds, corporate bonds, and municipal bonds.

Much like other fixed income securities, ABS deals are brought to market by an underwriter who is responsible for structuring the deal and bringing the deal to market. For the underwriting services, the underwriter will generally receive a fee that is based on percentage dollar amount of the individual class amounts. It should be noted that the underwriting firm will likely also receive compensation not only for their underwriting services but also in the form of commissions as they sell the bond into the primary market.

The ABS market experienced significant growth during late 1990's and early/mid 2000's. The growth in the ABS market coincided with the growth in the number of loans provided to subprime borrowers. Sub-prime borrowers is a term associated with borrowers with lower credit scores and generally considered more likely to potentially default on a loan. The increase in loans led to an increasing number of securitized securities such as ABS. However, beginning in late 2006 the market for ABS changed as these borrowers began to default on their loans. As borrowers began to default or became delinquent in payments, the values of ABS securities diminished significantly. This decrease in turn led to an overall downturn in the ABS issuance. Issuance in 2007 was $759 Billion, nearly $200 Billion less than the $943 Billion issued in 2006. We next turn to a more detailed overview of ABS collateral types, underwriters, ratings and weighted average life.

1. Collateral Type

ABS are typically structured by collateral type. Appendix A provides a table of the deal type classifications used by Bloomberg to classify ABS deals. These classifications are generally considered "industry standard" and commonly used not only by Bloomberg, but by the market in general when classifying particular ABS deals. There is significant variety in the types of loans/receivables that are packaged together when structuring an ABS ranging from automobile loans to receivables from utilities such as electricity companies.

Although there are numerous collateral types and ABS, the dollar amount issued varies significantly across different collateral types. Table 3 presents dollar issuance by deal type for all nonprivate placed U.S. ABS issued between 1999 through 2006. For example, ABS backed by home equity loans during the period dominates issuance at $2.3 Trillion followed by automobile loan ABS with $716 Billion, credit card ABS with $486 Billion and student loan ABS with $285 Billion. The remainder of the issuance is largely fragmented across nearly 30 other loan/receivable types.

2. Underwriter

Table 4 provides ABS underwriter rankings of the top 20 underwriters ranked by dollar amount underwritten during the period 1999-2006. A few observations can be noted. First, some underwriters such as Credit Suisse and Lehman were either top underwriter or in the top 5 underwriters for each during the period 1999-2006. Second, there are some underwriters that show significant growth in market share during the period--two notable examples are Barclays and Countrywide that had little or no underwriting activity prior to 2001, but show substantial and consistent growth year to year for years thereafter.

Table 5 presents the average underwriting fees charged for the same top 20 underwriters during the period 1999-2006. For example, consider Lehman Brothers. During 1999 through 2002, Lehman Brothers charged higher than average or average fees. However, during the periods of the most rapid growth in terms of issuance, 2003 through 2006, the fees they charged were lower than average.

3. Ratings

Credit ratings are often assigned to ABS. Table 6 shows dollar issuance of ABS by year and credit rating. Over half of the total dollar issuance during the period 1999-2006 has credit rating of A- or higher with the majority of it being rated AAA. The remaining issuance is relatively evenly distributed across the other, lower ratings.

Table 7 shows average underwriting fee by year and credit rating. ABS rated AAA typically have lower than average underwriting fee during each individual year, while lower rated securities such as BBB have higher than average underwriting fees. This suggests that underwriters charge higher fees for lower rated securities, suggesting relative difficulty in marketing lower quality securities.

4. Original Weighted Average Life

Table 8 shows ABS by weighted average life and amount issued during the sample period used in the study. During the sample period, nearly a third of all issuance had original WAL between 2.5 and 3.5 years and over 70% of the issuance had original WAL between .5 and 5.5 years.

Table 9 shows ABS by weighed average life and average underwriting fee. This data does not indicate a relationship between original weighted average life and average underwriting fees. This contrasts with the findings of Livingston and Miller (2000) who found a positive relationship between term/maturity and underwriting fees.

III. Literature review and hypothesis formulation

1. ABS literature

DeMarzo and Duffle (1999) note that issuers of ABS need to evaluate two costs when determining optimal security design: The opportunity cost related to holding assets with lower returns than those that could be sought if the issuer securitized these assets and used the capital raised to invest in higher-return assets, and the potential negative impact of including these lower returning assets in the securitization and the potential impact of lower demand for the security. DeMarzo and Duffie develop a framework for evaluating optimal security design.

Han and Lai (1995) note that securitization has been successful in markets such as mortgages and asset-backed loans; it has not been as successful for insurance products. They offer three reasons this has been the case: 1) it is more costly to securitize unstable cashflows from insurance products into fixed income securities, 2) regulations do not make it conducive to do so since regulators do not permit to take the securities assets/liabilities off their balance sheet, and 3) insurers have other ways to diversify their portfolio thereby reducing the need/attractiveness of securitization.

Plantin (2004) develops a model to gain insight into why firms issue asset securitization deals into separate classes or tranches. He points out that many ABS structures such as CDOs are split into senior and junior classes. He suggests that investment banks that sell these securities generally target different types of investors for each piece: the senior pieces are generally sold to less sophisticated, retail institutions, while the junior pieces generally go to more sophisticated investors who have the knowledge and resources to analyze these securities.

One the biggest challenges with ABS is the ability to properly value/assess the risk of the securities. A number of papers focus on this topic such as Heidari and Wu (2004). Based on the results of a survey of market participant, they compare these ideal attributes to those of six models of major MBS dealers. They find that five of the six fall short of meeting the desired attributes. In addition, they find high correlation among these five suggesting potential herding among MBS analysts.

Adelson (2003) points out some of the risks associated with ABS/CDO markets. In addition to risks such as prepayment risk, liquidity risk, he raises another source of risk - model risk. Model risk is the "risk that a model does not describe reality well enough to produce reliable results." Antonov and Raevsky (2003) develop a model for the modeling of credit risk that can be used for ABS portfolios.

Childs, Ott and Riddiough (1996) develop a model for the pricing of commercial mortgage-backed securities. CMBS are securities backed by non-residential mortgages, for example, mortgages for businesses, apartment complexes, etc. One of their findings involves the relationship between pool size and tranche value. They find that 5 to 10 distinct mortgages are required to realize most of the effects of asset diversification.

Additional research is related to prepayment and credit risk. Hetfield and Sabarwal (2004) find that prepayments on subprime loans increase with loan age. However, they do not find prepayments affected as much by current market interest rates. Default rates are much more sensitive to aggregate shocks than are prepayment rates such as increases in unemployment. They also find significant differences in the default and prepayment rates faced by different subprime lenders. Lenders charging the highest interest rates experience the highest default rates, but also experience somewhat lower prepayment rates. They believe that there are substantial differences among subprime borrowers, and that different lenders target different segments of the subprime market.

Lacour-Little and Chun (1999) explore the relationship between third party loan originators and prepayment behavior. They point out those third parties, such as mortgage brokers, have economic incentives to encourage borrowers to refinance and, accordingly, their actions may affect asset values. They find that loans originated by third parties are significantly more likely to prepay. Moreover, third party loans are about three times as sensitive to refinancing incentives, compared to retail loans.

Lucas, Goodman, and Fabozzi (2004) examine how rating agencies calculate default rates on structured finance securities. They point out a number of the limitations of the methodologies used by S&P and Moody's. They conclude by offering a new calculation which uses as a base the calculations of S&P and Moody's but is modified to address at least some of the weaknesses of the two rating agencies methodologies.

Ammer and Clinton (2004) examined the impact of credit rating changes on the pricing of asset-backed securities. Using a sample of 1300 rating changes by Moody's or S&P, they find that rating downgrades tend to be accompanied by negative returns and widening spreads with the average effects being stronger than those for corporate bond rating changes. This suggests that ABS investors appear to rely more on rating changes as a source of negative changes than bond investors. In terms of effects of rating upgrades, the authors found very little in terms of market reaction to these events.

Additional studies have focused on financial innovation, the process or decisions involved in developing new financial products such as ABS. Silber (1983) provides an overview of financial innovation. He cites three main factors that most commonly lead to the development of newer financial products such as mortgage-backed securities. The first is that firms innovate to "lessen the financial constraints imposed on firms." The second is technology. Finally, the third main source he cites is legislative, which he points out, was the key factor in the development of the mortgage-backed securities market.

Boot and Thakor (1993) explore the issue of security design and why firms issue multiple claims on an asset, many classes when securitizing loans or mortgages rather than a single security. They develop a complex model that suggests that firms split securities into two types: information insensitive and information sensitive. They argue that informed traders will focus on the information sensitive securities and will move the security closer to its fundamental value, thereby increasing the issuer's total expected revenue.

2. Underwriting fees for IPOs

Underwriting fees are the fees paid by issuers of financial securities to financial firms that take on the responsibility of the marketing of debt or equity to the financial market. A number of studies have focused on the underwriting fees for IPOs. Chen and Ritter (2000) investigated why underwriting fees/spread for U.S. IPOs tend to cluster towards 7% which is much higher than fees in non-U.S, markets. They argue that the relative high average spread and clustering are due to "strategic pricing." They argue that investment bankers maintain the higher fees and compete for IPO business on other grounds. By avoiding competing on the grounds of fees, investment banker's year-end bonuses are not in jeopardy.

Carter and Manaster (1990) examined IPOs and underwriting fees. They identify a relationship between underwriter prestige (as measured by relative position an underwriter is listed in the announcement of a pending public offering) and returns of IPOs. They argue that underwriter prestige is a signal to investors as to the relative riskiness of an IPO. They find that prestigious underwriters are generally associated with IPOs with lower returns / lower risk offerings.

Hansen (2001) explores potential reasons why IPOs fees in the U.S. tend to be 7%. His evidence suggests that collusion between underwriters is not the source, but rather underwriters compete for business by other factors such as reputation, placement abilities, and the degree to which IPOs are underpriced. In a similar study, Barondes, Butler, and Sanger (2000) focused on IPOs whose underwriter fees differed (higher or lower) than the typical seven percent. They examined the relationship between these deviations and offering prices of the IPOs. They found that the lower (higher) the fees paid, the lower (higher) the offering price of the IPO. They offer marketing efforts of these IPOs by the underwriter is a function of the amount paid to them in terms of fees.

Additional studies have focused on underwriting fees in non-U.S, markets. Torstila (2001) explored what determines IPO spreads in Europe. He found IPO spreads by European issues in Europe are significantly lower than by European issuers in the U.S. In addition, IPOs listed jointly in U.S. and Europe generally have higher spreads than issues listed only in Europe.

Bajaj, Mazumdar, Chen, and Sarin (2003) examine IPO underwriting spreads during the period 1980 to 1998. They found that median size of IPOs have tripled during this time. Further they find that the more recent IPOs involved riskier firms. They also find that clustering of IPO fees existed in earlier periods often at higher rates then 7%.

James (1992) explored whether underwriter fees in IPOs are associated with relationship-specific assets or setup costs. More specifically, he examined how underwriting fees are affected by expectations that the firm will issue additional shares in the future. He finds that underwriter fees are significantly lower in IPOs when firms issue additional shares and use the same underwriter than when firms do not continue with the same underwriter or do not issue again.

Carter, Dark, and Singh (1998) explore the relationship between underwriter reputation and performance of IPO stocks. Using 3 different measures of underwriter prestige, they find that reputation is significantly related to initial return of the IPO; there is less short-run underpricing with more prestigious firms vs. less prestigious. In addition, they find that on average the long-run market-adjusted returns (3-year holding period) are less negative for IPOs underwritten by more prestigious underwriters. Lastly, of the three measures used to measure underwriter prestige, they find the Carter-Manaster (1990) measure serves best to measure underwriter prestige.

Hebb and MacKinnon (2000) examine the IPO valuation comparing those IPOs underwritten by non-commercial banks versus commercial banks. They find greater uncertainty in the true value of IPOs underwritten by commercial banks versus non-commercial banks. They offer as a potential reason for this uncertainty as the market's perception of a potential conflict of interest by the commercial bank. (2)

3. Underwriting fees for debt

Livingston and Miller (2000) examined investment bank reputation and the underwriting of debt. Using both the Carter and Manaster metric and percentage of total dollar issued of debt to measure underwriter prestige, they find that higher prestige firms charge lower underwriting fees. In addition, offering yields tend to be lower and offering prices higher for more prestigious underwriters. Livingston and Jewell (1998) explore the issue of underwriter spread for industrial bonds that have split ratings. They find that if the bond has an investment grade rating from both S&P and Moody's, even if they differ, the underwriting fees are effectively the same. However, if the bond is rated below investment grade by one or both of the agencies, underwriting fees are effected and are generally between the spreads for the higher rating and lower rating.

Burch, Nanda and Warther (2004) explore the issue of underwriting relationships and fees charged. They find that loyalty (repeat business) lead to lower fees for common stock offers, but the opposite for debt offers. In addition, they find that firms that change to higher quality (higher reputation using the Carter-Manaster metric) face lower fees for both equity and debt offerings.

Butler (2007) examined municipal bond underwriting and whether the choice of using a local underwriter--underwriter with ongoing business in the same state as the municipal bond--appears to influence underwriting fees. He found that local underwriters charged lower fees and bonds were offered with lower yields as compared to non-local underwriters. Further, he found these benefits existed greater for bonds with lower credit quality and bonds not rated by the rating agencies.

Santos and Tsatsaronis (2002) researched the effects of the introduction of the euro on the underwriting of corporate bonds. Their study identified two results. The first is that with the introduction of the euro, the average underwriter fee decreased--pre vs. post euro. This presumably was due to the increased competition amongst firms in the region. Their study also examined when choosing an underwriter who were they more likely to choose: an underwriter from their home country with whom they are more likely to have an ongoing relationship or larger investment houses that have a more global placing capacity for their bonds. They found firms migrating towards larger international investment banking houses.

Yasuda (2003) explores the issue of underwriting of corporate bonds and what effect entry of commercial banks has had on the market for underwriting services. The research focuses on examining two scenarios of coexistence of commercial and investment banks in the market for underwriting services.

* Both underwriters and investment banks fetch the same price for the security (there is not differentiation in certification ability)

* Commercial banks fetch a higher price for the security, and investment houses discount their fees to the level where issuing firms are indifferent to underwriter (there is a differentiation in certification ability).

Altinkihc and Hansen (2000) examine underwriting fees for bond offerings and SEO (seasoned equity offerings) and how fees vary by issue size. They suggest that these should exhibit a U-shaped curve (fees as a function of issue size). The logic behind this assertion is that initially there are certain fixed costs with underwriting and therefore as size increase, the fees will decrease. However, as size increases to a certain point, the fees begin to rise due to placement cost--the costs associated with the difficulty the underwriter may experience in trying to place larger issues. They find evidence of fees curve is U-shaped both for equities and bond offerings.

Two papers are most notable as they pertain to the current study. The first is that of Butch, Nanda and Warther (2004). They found that loyalty (repeat business) leads to lower fees for offers for common stock which is consistent with James (1992). However, they found the opposite for debt offers. This leads to the question for our study will underwriting fees for repeat business, subsequent ABS issuance by the same issuer using the same underwriter, be akin to IPO and lead to lower underwriting fees or be similar to that of debt markets and lead to higher underwriting fees.

The second paper most notable for the current study is that of Livingston and Miller (2000) which examines bank reputation/prestige and underwriting fees for debt issues. They find that higher prestige firms charge lower underwriting fees. Our study will use methodology closely related to that of Livingston and Miller as we ask the question, do more prestigious underwriter's charger lower fees for ABS?

4. Hypothesis formulation

Although asset-backed securities and corporate debt share some commonalities in that they are part of the overall fixed income/debt market, a number of differences exist. First, by definition, asset-backed securities and corporate debt differ in that ABS are collateralized by a package of loans, where corporate debt are securities issued by corporations. Second, the literature on ABS suggests that valuing ABS presents some unique challenges that do not exist for corporate bonds. Lastly, the ABS market has experienced periods of rapid growth, such as the period 2002-2005 where it nearly doubled. During the same period, the corporate debt market remained relatively constant. In this paper we investigate the following: will the inverse relationship between underwriter prestige and underwriting fees identified by Livingston and Miller for corporate debt hold true for the asset-backed securities market? Alternatively, will the rapid growth in the market instead lead to less prestigious/ smaller underwriters charging lower fees as a means to gain market share, leaving top ranking underwriters charging relatively higher fees? In consideration we accordingly propose the following hypothesis and alternative hypothesis:

Hypothesis 1: Consistent with that of U.S. corporate bonds and identified by Livingston and Miller (2000), there exists an inverse relationship between underwriter prestige and underwriter fees paid by issuers of asset-backed securities.

Alternative Hypothesis 1: The ABS market has shown rapid growth in market size between 1999 through 2006 which may attract more underwriters to the market willing to accept lower underwriting fee and thereby accepting higher internal costs to capture business. This suggests that there exists a positive relationship between underwriter prestige and underwriting fees paid by issuers of asset-backed securities as newer entrants into the underwriting of these securities attempt to gain market share.

Prior research in both IPOs and corporate bonds focused on loyalty of firms using the same underwriter. Burch, Nanda and Warther (2004) found that loyalty (repeat business) leads to lower fees for common stock offers that which is consistent with James (1992). However, they found the opposite for debt offers. Livingston and Miller (2000), however, using their variable count found an inverse relationship between loyalty and underwriting fees for corporate bonds counter to the findings of Burch, Nanda, and Warther (2004). The underwriting process for ABS is quite similar to that of corporate bonds.

In consideration we accordingly propose the following hypothesis and alternative hypothesis:

Hypothesis 2: Similar to corporate bonds, there exists an inverse relationship between loyalty (repeat business) and underwriting fees. This could be explained by that a underwriters due diligence costs" may be lower due to repeat business/same issuer.

Alternative Hypothesis 2: Similar to IPOs, there exists a positive relationship between loyalty (repeat business) and underwriting fees.

IV. Methodology

Our methodology is similar, though not identical, to Livingston and Miller (2000), who use OLS to explore relation between underwriter prestige and underwriter spread/fees in their study of nonconvertible debt. Our definition of Prestige variable is consistent with Livingston and Miller. Our methodology differs in the following ways. First, weighted average life is used instead of maturity. Second, we do not include callability as a variable, as callability is much less of a factor in pricing of ABS. Third, we include dummy variables for each of the four largest collateral types in terms of dollar issuance (auto, credit cards, home equity and student loans). Fourth, we include the original deal size. Fifth, we include dummy variables for the class descriptor for the ABS. Finally, a sixth variable for short-term loyalty is included. The following four equations are estimated in our study:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

US : [[alpha].sub.0] + [[beta].sub.1] x AAA + [[beta].sub.2] x Mezzanine + [[beta].sub.3] x Subordinated + [[beta].sub.4] X DealAmount + [[beta].sub.5] x ParAmount + [[beta].sub.6] x Prestige + [[beta].sub.7] x WAL + [[beta].sub.8] x Count + X'YEAR_DUMMIES + [epsilon] (4)

US is the underwriter spread (in basis points) which is provided typically in the prospectus or other documentation provided by underwriter when ABS is issued. AAA is a dummy variable equal to one if the ABS is rated AAA by S&P. Underwriting spreads vary and typically are higher for classes that are more difficult to sell i.e. mezzanine bonds. Since the credit crisis that began in 2007, credit ratings have been questioned as to whether they are reliable proxies for risk. However, during the sample period of this study, credit ratings were generally accepted as a proxy of risk and therefore we include them in the study. We expect that higher credit rated securities would have lower underwriting fees.

Mezzanine and Subordinated are dummy variables representing class descriptors mezzanine and subordinated. These were included to see if different class descriptors, which are assigned to help identify variability in expected cashflows, impacts underwriting fees. Indeed, class descriptor may represent an alternative proxy of risk. One may expect that class descriptors associated with higher variability in cashflows, namely mezzanine and subordinated, are associated with higher underwriting fees.

Deal Amount is the log of original deal amount. ABS are commonly issued not individually, but as part of an overall deal structure. We include this variable to see if any relationship exists between overall deal size and underwriting fees.

Auto, CreditCard, HomeEquity, and StudentLoan are dummy variables representing collateral types auto, credit cards, home equity and student loans. The type of collateral may serve as an alternative proxy for risk, and also controls for cross-collateral type effects. For example, deals backed by collateral types that are smaller in terms overall issuance such as ABS backed by boat loans may be more difficult to bring to market than deals backed by collateral types with larger amount of issuance such as auto, credit card, home equity, and student loans.

ParAmount is the log of size/par amount of the class. Livingston and Miller (2000) found larger bond issues tend to have higher underwriting fees. We include this variable to see if this exists within ABS market as well. Prestige is calculated by taking the market share of each underwriter during the entire sample period (1999-2006). WAL is the original weighted average life. It is more appropriate to use weighted average life as opposed to maturity date due to the inherent strong likelihood the security will not remain outstanding until maturity. It is quite common in the market to use weighted average life for this purpose. Similar to the finding in Livingston and Miller (2000) that longer term bonds tend to have higher underwriting fees, one expect to see a similar relation between WAL and underwriting fees for ABS.

Loyalty is dummy variable for short term loyalty which is set to 1 if the underwriter used was the same as the one used for prior issuance. This metric is used by Burch, Nanda, and Warther (2004). Count is the log of the number of securities of a particular issuer that were underwritten by the same underwriter during the sample period. A similar metric was used in Livingston and Miller (2000), and represents another metric for loyalty. Burch, Nanda and Warther (2004) found no relation using this metric. Livingston and Miller (2000), however, did find a relation in terms of lower underwriting fees.

YEAR DUMMIES is a vector of dummy variables for the year of the given observation. We include these dummies to investigate whether year of issuance impacts underwriting fees.

Equations 1 and 2 include Loyalty while Equations 3 and 4 include Count. Equations 1 and 3 include dummy variables for four ABS deal types, Auto, CreditCard, HomeEquity, and StudentLoan while Equations 2 and 4 do not. Equations 1 through 4 are initially estimated across the entire pooled cross sectional time series and then estimated separately, by year. The annual regressions exclude the year dummies. Further, by the annual regressions Prestige was calculated year by year. For example, the prestige for 1999 was calculated by using only 1999 issuance. This was then used as a variable when evaluating influence on fees in the following year.

V. Results

The results for the pooled cross sectional time regressions estimation of Equations 1 through 4 is reported in Table 10. The results for the annual regression estimations for Equations 1 through 4 are reported in Tables 11 through 14, respectively.

Hypothesis 1 and Alternative Hypothesis 1 focused on the relationship between underwriting fees and underwriter prestige. Hypothesis 1 stated that consistent with U.S. corporate bonds and identified by Livingston and Miller (2000), there exists an inverse relationship between underwriting fees and underwriter prestige. Due to the rapid growth in market size during the sample period 1999-2006, we offered that this may attract more underwriters willing to accept lower underwriting fees to capture business. We offered Alternative Hypothesis 1, there may exist a positive relationship between underwriting fees and underwriter prestige. As newer entrants into the underwriting of these securities attempt to gain market share, they accept lower underwriting fees than more prestigious underwriters.

To test Hypothesis 1 and Alternative Hypothesis 1, we used one of the same metrics used by Livingston and Miller (2000) to determine prestige--the proportion of market share for an underwriter during the sample period. In our estimation of all four equations, we found the coefficient to be positive. In 3 of the 4 equations, we found prestige to be statistically significant at .05 level while in one equation we found it to be statistically significant at .1 level. Of the 28 annual regressions, 17 of them are positive and statistically significant, while 2 are negative and statistically significant. Collectively, these results are supportive of Alternative Hypothesis 1.

Hypothesis 2 and Alternative Hypothesis 2 focused on underwriter loyalty to see if there was any benefit in terms of lower underwriter fees for an issuer to use the same underwriter for subsequent issues. Livingston and Miller (2000) used a measure they called count--defined as the number of bond underwritten by the same underwriter for a given issuer of ABS during the sample period. Butch, Nanda, and

Warther (2004) used an alternative measure of loyalty which was a dummy variable assigned 1 if the underwriter used in current bond also served as underwriter in previous issuance. We use both metrics in our estimations. In Equations 1 and 2 we use the Burch, Nanda, and Warther measure of Loyalty. In both equations we find the coefficient to be positive and statistically significant at .05 level. In Equations 3 and 4 we use the Livingston and Miller measure of Count and also find the coefficients to be positive and statistically significant at .05 level.

We also evaluate loyalty year by year. Of the 28 annual regressions, we found all measures of loyalty to be positive and statistically significant at the .01 level in all but 6 of the 28 samples. Collectively, these findings suggest there is no benefit, in fact potentially a detriment, to using the same underwriter--the more loyal, the higher the underwriting fees. This provides evidence in support of Alternative Hypothesis 2.

Significant coefficients associated with AAA are negative and statistically significant. The negative coefficients suggest AAA rated securities have lower underwriting fees vs. non-AAA securities. This finding is consistent with Livingston and Miller. The coefficients associated with Subordinated and Mezzanine are both statistically significant at .05 level across all estimations. Both have positive coefficients which suggest securities with subordinated or mezzanine class descriptor have higher underwriting fees vs. non-subordinated and non-mezzanine securities. This finding is expected since subordinated and mezzanine securities are lower tier within the overall deal structure. Since class descriptors are an ABS/CMO concept, this finding is unique to this particular financial market.

The coefficients associated with DealAmount are positive and statistically significant at .05 level in all estimations of the pooled cross sectional time series, suggesting that the larger deal size, the larger underwriting fees on the individual securities. Since the concept of an overall deal is unique to ABS/CMO, this finding is unique to this particular financial market.

Dummy variables for the four largest deal types (auto, credit cards, home equity and student loans) were tested. With the exception of CreditCard, the coefficients associated with these variables are statistically significant with negative coefficients in Equation 3, but not statistically significant Equation 1. Collectively, this suggests a question able relationship between deal type and underwriting fees.

The coefficient associated with ParAmount is negative statistically significant at .05 level in all four pooled cross sectional time series regressions, a result that holds for most of the annual estimations as well This suggests that the larger the size of the security, the lower the underwriting fees charged. This is inconsistent with Livingston and Miller (2000) who did not find a statistically significant relationship between par amount (or what they defined as proceeds) and underwriting fees for U.S. corporate bonds.

The coefficient associated with WAL is statistically significant with positive coefficients in all four pooled cross sectional time series regressions, a result that holds for most of the annual estimations as well. This result is consistent with Livingston and Miller (2000) that found similar results for U.S. corporate bonds between maturity and underwriting fees.

VI. Conclusions

The contributions of this study are two-fold. The first is a contribution to the overall literature on asset-backed securities. Despite being developed in the early 1980's and issuance nearly the same in terms of dollar amount of U.S. corporate bonds, relatively little research exists for ABS. Much of the existing research pertaining to ABS is related to the issues of financial innovation, optimal security design, and some of the inherent risks associated with ABS and valuing ABS. The current study provides a detailed descriptive analysis of the ABS market.

The second contribution of the study is to the existing literature on the factors that influence underwriting fees. Underwriting fees have been researched in a number of financial markets. Most literature pertains to IPOs (Chen and Ritter 2000; Hansen 2001; James 1992) among many others. There also exists a fair amount of research on corporate bonds and underwriting fees (Livingston and Miller (2000), Burch, Nanda, Warther (2004), Santos and Tsatsaronis (2002), among others. The current study extends the literature on underwriting fees, exploring them for a different market--asset-backed securities.

One of the implications of the current study is related to the positive relationship between underwriter prestige and underwriter fees. This suggests that issuers of ABS are charged higher overall underwriting fees for using a more prestigious underwriter. This result is inconsistent with the evidence by corporate bonds (Livingston and Miller, 2000) which finds an inverse relationship between underwriter prestige and underwriting fees. One potential explanation for the positive relationship found in ABS is that with relatively rapid growth in terms of issuance during the sample period the ABS market attracted newer entrants into the underwriting of ABS. These newer entrants charged lower overall fees as a means to get into the ABS underwriting business.

The second implication of the current study pertains to underwriter loyalty and underwriting fees. This suggests that issuers are penalized by having higher underwriting fees when they are loyal, i.e., use the same underwriter, in subsequent issues. This result is consistent with the findings of Burch, Nanda, and Warther (2004) that provide evidence of a similar relationship between underwriting fees and loyalty for corporate bonds using their metric of loyalty. However, using a different metric to measure loyalty--count--Livingston and Miller (2000) provide evidence that an inverse relationship exists between loyalty and underwriting fees for corporate bonds. Unlike corporate bonds in which the issue of loyalty offers inconsistent results, our results using both metrics of loyalty yielded the same results for ABS--issuers are penalized by using the same underwriter.

This study's primary limitation was the years selected for the study. Although eight years worth of data was used, it's certainly possible the results found during this period may not be applicable to future periods. Also due to recent events beginning in 2007 during the sub-prime crisis, issuance in ABS has decreased dramatically. This fact may further influence fees charged for ABS in the future. Another limitation of both this study as well as earlier studies is sample selection bias. Underwriting fees in both the ABS and corporate bond markets are not always disclosed.

The current study opens the door for additional research in a number of areas. First, the study provides evidence that there are still additional research opportunities in the area of underwriting fees. Prior research on underwriting fees focused primarily on corporate bonds and equities. This paper extended the literature to include U.S. asset-backed securities. As the financial markets continue to evolve and newer security types emerge, it lends itself naturally for the opportunity to research underwriting fees in these markets. For example, the market for non-U.S. ABS continues to grow. This may provide additional research opportunities in the area of underwriting fees.

In addition to underwriting fees, this paper extended the current literature on ABS. As noted, the current literature on ABS is relatively limited as compared to other financial markets. There are many additional opportunities to research whether relationships found to exist in other financial markets exist in ABS market.

Appendix A: Deal Types This following are deal type used by Bloomberg to classify ABS deals. Source: Bloomberg.

* ABS: backed by various type loans.

* AUTOS: backed by automobile loans

* BOATS: backed by boat loans

* BU.S.INESS: backed by business equipment loans

* CARDS: backed by credit card receivables

* CBO: backed by various bonds

* CDO: backed by various debt obligations

* CONSUMERS: backed by various consumer/ personal loans

* CREDIT LINK: backed by credit-linked receivables

* EQUIPMENT: backed by equipment leases/loans

* FILM: backed by film/motion picture receivables

* HEALTHCARE: backed by healthcare receivables

* HOME EQTY: backed by home equity loans

* HOME IMP: backed by home improvement loans

* MANUFCT HM: backed by manufactured home loans.

* PLANES: backed by airplane loans

* RE-SEC: backed by other ABS deals

* RV: backed by loans for recreational vehicles

* STUDENTS: backed by student loans

* TAX LIENS: backed by tax liens

* TRADE: backed by trade receivables

* UTILITY: backed by utility receivables

References

Adelson, M. H. 2003. "CDO and ABS Underperformance: A Correlation Story." Journal of Fixed Income 13(3): 53.

Altinkihe, O. and R. Hansen. 2000. "Are there Economies of Scale in Underwriting Fees? Evidence of Rising External Financing Costs." Review of Financial Studies 13(1): 191-218.

Ammer, J. and N. Clinton. 2004. "Good News is No News? The Impact of Credit Rating Changes on the Pricing of Asset-backed Securities." International Finance Discussion Papers (809).

Antonov, A. and D. Raevasky. 2003. "Effective Solutions of the Quasi-Multi Period Model for Large Credit and ABS Portfolios." Working Paper.

Bajaj, M., S. Mazumdar, A. Chen, and A. Sarin. 2003. "Competition in IPO Underwriting: Time Series Evidence." EFA 2003 Annual Conference Paper No. 707.

Barondes, R., A. Butler, and G. Sanger. 2000. "IPO Spreads: You Get What You Pay For." Working Paper.

Boot, A. and A. Thakor. 1993. "Security Design." The Journal of Finance 48(4): 1349-1378.

Burch, T. R., V. Nanda, and V. Warther. 2004. "Does it Pay to be Loyal? An Empirical Analysis of Underwriting Relationships and Fees." Journal of Financial Economics.

Butler, A. 2007. "Distance Still Matters: Evidence from Municipal Bond Underwriting." Review of Financial Studies.

Carter, R. and S. Manaster. 1990. "Initial Public Offerings and Underwriter Reputation." Journal of Finance 45(4): 1045-1067.

Carter, R., F. Dark, and A. Singh. 1998. "Underwriter Reputation, Initial Returns, and the Long-run Performance of IPO Stocks." Journal of Finance 53(1): 285-311.

Chen, H.-C. and J. Ritter. 2000. "The Seven Percent Solution." Journal of Finance LV(3): 1105-1131.

Childs, P., S. Ott, and T. Riddiough. 1996. "The Pricing of Multiclass Commercial Mortgage-backed Securities." Journal of Financial and Quantitative Analysis 31 (4): 581-603.

Cliff, M. and D. Denis. 2003. "Do IPO Firms Purchase Analyst Coverage with Underpricing?" Working Paper.

DeMarzo, P. and D. Duffie. 1999. "A Liquidity-Based Model of Security Design." Econometrica 67(1): 65-99.

Han, L.-M. and G. C. Lai. 1995. "An Analysis of Securitization in the Insurance Industry." The Journal of Risk and Insurance 62(2): 286-296.

Hansen, R. 2001. "Do Investment Banks Compete in IPOS? The Advent of the '7% Plus Contract'." Journal of Financial Economics 59:313-346.

Hebb, G. and G. MacKinnon. 2000. "Valuation Uncertainty and IPO's: Investment Bank versus Commercial Bank." Working Paper.

Heidari, M. and L. Wu. 2004. "What Constitutes a Good Model? An Analysis of Models for Mortgage-Backed Securities." Working Paper.

Heitfield, E. and T. Sabarwal. 2004. "What Drives Default and Prepayment on Subprime Auto Loans?" Journal of Real Estate Finance and Economics 29(4).

James, C. 1992. "Relationship-Specific Assets and the Pricing of Underwriter Services." Journal of Finance 47(5): 1865-1885.

Lacour-Little, M. and G. H. Chun. 1999. "Third Party Originators and Mortgage Prepayment Risk: An Agency Problem?" Journal of Real Estate Research 17.

Lester, T., M. Asaria, and U. van der Linden. 2002. "Securitization: Korea and Taiwan follow Japanese lead." International Financial Law Review 21(10): 22.

Livingston, M. and J. J. Jewell. "Split Ratings, Bond Yields, and Underwriter Spreads for Industrial Bonds." Journal of Financial Research 21(2): 185-204.

Livingston, M. and R. Miller. 2000. "Investment Bank Reputation and the Underwriting of Nonconvertible Debt." Financial Management (Summer).

Loughran, T. and J. Ritter. 2002. "Why Don't Issuers Get Upset About Leaving Money on the Table in IPOs?" Review of Financial Studies 15(2): 413-443.

Lucas, D. J., L. S. Goodman, and F. Fabozzi. 2004. "Default Rates on Structured Finance Securities." Journal of Fixed Income 14(2): 44.

Park, S. M., M. Han, and Y. Kim. 2002. "Structured finance." International Financial Law Review: 85.

Pergamalis, G. 2003. "Greece unveils new securitization law." International Financial Law Review: 1.

Ritter, J. 1991. "The Long-Run Performance of Initial Public Offerings." Journal of Finance 46(1): 3-27.

Santos, J. and K. Tsatsaronis. 2002. "The Cost of Barriers to Entry: Evidence from the Market for Corporate Euro Bond Underwriting." Working Paper.

Silber, W. L. 1983. "The Process of Financial Innovation." American Economic Review 73(2).

Tinic, S. 1988. "Anatomy of Initial Public Offerings of Common Stock." Journal of Finance 43(4): 789-822.

Torstilla, S. 2001. "What Determines IPO Gross Spreads in Europe?" European Financial Management 7(4): 523-541.

van Eck, T. H. 1995. "Asset-Backed Securities." The Handbook of Fixed Securities: 583.

Yasuda, A. 2003. "Do Bank-Firm Relationships Affect Bank Competition in the Corporate Bond Underwriting Market?" Working Paper.

Notes

(1.) We thank Richard Ottoo, Elena Goldman, Susan Hume, and Michael Ehrlich for many useful comments and suggestions. We thank Maggie Huang for research assistance. We thank Bloomberg LP for providing the data. All errors remain our responsibility.

(2.) Another line of literature related to IPOs focused not on the underwriting fees but the issue of underpricing, when the issue price is significantly lower than the closing price on the first day or days of trading. See Ritter (1991), Tinic (1988), Loughran and Ritter (2002) and Cliff and Denis (2003).

by David Puskar, Bloomberg L.P. E-mail: dpuskar@bloomberg.net

Aron A. Gottesman, Lubin School of Business, Pace University. E-mail: agottesman@pace.edu
TABLE 1.
Issuance in $ Billions for U.S. corporate bonds and
non-private placed ABS

 U.S. Corporate U.S. Asset-Backed
Year Bonds Securities

2006 1,138 839
2005 866 826
2004 873 673
2003 848 505
2002 669 442
2001 789 365
2000 524 259
1999 529 227

TABLE 2.
Outstanding and new issuance in $ Billions for global ABS

 2001 2001 2002 2002

Year Outstanding New Outstanding New

Card 270.1 71 295.1 69.9
Auto 168.9 97.1 188.2 107
Home Equity 184.9 90.7 228.4 121
Manufacturing 52.8 7.2 49.2 4.6
Student 52.4 11.8 67.1 25.7
Other 330.4 106 384.9 119
Total 1059 384 1213 447

 2003 2003 2004 2004

Year Outstanding New Outstanding New

Card 307.1 68.1 300.1 53.6
Auto 192.5 95.3 1,115 175.3
Home Equity 277.5 164 342.8 219
Manufacturing 40.4 0.6 35.1 1.2
Student 94.5 37.4 123.7 45.2
Other 466.8 169 691.8 341
Total 1379 535 1,669 742

 2005 2005 2006 2006

Year Outstanding New Outstanding New

Card 297.4 67.9 300.2 65
Auto 193.4 115 194.8 92.1
Home Equity 414 258 501.4 199
Manufacturing 30.5 0.7 26.5 4.4
Student 162.9 64.3 199.2 24.6
Other 995.4 527 1,456 267
Total 2,094 1,032 2,678 1,236

TABLE 3.
Non-private placement U.S. ABS by Deal Type in $ Millions

 1999 2000 2001 2002 2003

ABS 2,007 2,980 8,498 6,131 9,807
AUTOS 61,752 82,345 95,085 105,699 96,013
BOATS 1,285
BU.S.INESS 988 543 1,365 939 2,958
CARDS 40,234 54,533 76,665 71,534 65,395
CBO 493 70
CDO 264 46
CMO
CONSUMERS 27
CREDIT LINK 50 275
EQUIPMENT 11,047 13,757 8,095 6,040 12,271
FILM 179
HLTHCARE 124 850 700
HOME EQTY 70,247 67,740 132,152 206,161 275,097
HOME IMP 871 290
MANUFCT HM 15,673 10,227 10,987 5,830 604
MBB
MUNICIPAL
N.A.
PLANES 3,455 7,029 7,100 526 3,670
PYMT RIGHTS 127
RE-SEC 154 417 920 13,163 184
RV 1,950 280 2,932
STUDENTS 9,321 17,131 12,188 23,570 38,175
SWAP TRU.S.T 415
TAX LIENS 261 157 278 317 30
TRADE 346 344 110 638 622
UTILITY 7,852 1,000 8,416 1,167 500
Grand Total 227,568 259,735 365,886 442,880 505,603

 2004 2005 2006 Grand Total

ABS 6,194 10,558 15,843 62,019
AUTOS 80,446 108,503 86,530 716,373
BOATS 1,285
BU.S.INESS 2,676 2,652 1,036 13,158
CARDS 53,181 59,084 66,341 486,966
CBO 563
CDO 5 1,071 1,386
CMO 12,608 12,608
CONSUMERS 27
CREDIT LINK 133 2,682 509 3,648
EQUIPMENT 9,374 12,888 14,446 87,919
FILM 179
HLTHCARE 101 1,775
HOME EQTY 454,507 536,886 568,945 2,311,735
HOME IMP 0 1,160
MANUFCT HM 685 589 201 44,797
MBB 100 100
MUNICIPAL 800 800 1,600
N.A. 1,403 1,403
PLANES 1,118 2,473 1,120 26,492
PYMT RIGHTS 200 327
RE-SEC 14,729 12,908 42,475
RV 403 5,565
STUDENTS 48,672 68,224 68,133 285,415
SWAP TRU.S.T 415
TAX LIENS 50 66 1,159
TRADE 350 2,410
UTILITY 790 5,361 1,922 27,008
Grand Total 673,313 826,081 838,900 4,139,966

TABLE 4.
ABS issuance by top 20 underwriters in $ Millions

 1999 2000 2001 2002 2003

Lehman 31,302 28,604 35,408 39,570 46,006
Credit Suisse 35,986 24,628 49,186 42,790 52,765
Bank of America 11,084 15,152 48,200 64,807 49,639
Countrywide 440 6,718 11,682 30,406 28,589
JP Morgan 7,639 9,025 45,137 41,213 44,418
Citigroup 33,025
Deutsche Bank 1,553 32,124 46,094
Merrill Lynch 19,217 14,875 8,296 15,427 18,972
RBS Greenwich 1,449 6,602 28,896
Morgan Stanley 1,357 22,440 31,059
Bear Steams 12,310 13,767 24,736 15,740 19,586
Salomon 27,703 37,310 34,521 41,063 19,669
Goldman Sachs 13,548 10,239 5,245 6,759 6,916
Barclays 2,578 9,047 9,057
Banc One 2,539 1,063 15,377 22,749 29,518
UBS 1,930 6,185
Deutsche Bk AB 703 23,784 33,595 3,918
Wachovia 145 6,059 2,894
CS
Greenwich 7,254 6,323 17,019 19,808 874
Grand Total 227,568 259,735 365,886 442,880 505,603

 2004 2005 2006 Grand Total

Lehman 65,979 80,829 91,646 419,345
Credit Suisse 65,061 77,118 817 348,350
Bank of America 48,808 52,578 54,225 344,493
Countrywide 82,102 79,022 65,293 304,252
JP Morgan 35,038 43,079 50,547 276,096
Citigroup 74,146 70,529 70,482 248,182
Deutsche Bank 43,335 46,574 45,770 215,450
Merrill Lynch 46,964 43,336 48,206 215,293
RBS Greenwich 51,463 71,151 54,962 214,523
Morgan Stanley 46,519 58,077 48,477 207,928
Bear Steams 31,893 39,253 39,688 196,972
Salomon 160,266
Goldman Sachs 19,140 35,235 56,820 153,902
Barclays 13,419 29,086 42,223 105,409
Banc One 6,581 77,828
UBS 12,587 22,939 19,084 62,725
Deutsche Bk AB 62,000
Wachovia 14,691 20,497 11,956 56,242
CS 55,600 55,600
Greenwich 51,278
Grand Total 673,313 826,081 838,900 4,139,966

TABLE 5.
Average percentage ABS underwriting fees for Top 20 Underwriters

 1999 2000 2001 2002 2003 2004 2005

Lehman 0.37 0.38 0.39 0.39 0.30 0.29 0.26
Credit Suisse 0.30 0.37 0.30 0.35 0.24 0.25 0.24
Bank of America 0.36 0.30 0.29 0.25 0.27 0.25 0.24
Countrywide 0.25 0.40 0.54 0.44 0.77 0.63 0.70
JP Morgan 0.25 0.25 0.27 0.27 0.27 0.26 0.24
Citigroup 0.29 0.27 0.29
Deutsche Bank 0.33 0.28 0.25 0.25 0.23
Merrill Lynch 0.30 0.33 0.31 0.32 0.39 0.35 0.32
RBS Greenwich 0.24 0.36 0.28 0.27
Morgan Stanley 0.29 0.26 0.27 0.24 0.21
Bear Steams 0.30 0.25 0.32 0.27 0.31 0.24
Salomon 0.31 0.30 0.36 0.28 0.30
Goldman Sachs 0.26 0.29 0.25 0.49 0.22 0.36 0.27
Barclays 0.38 0.21 0.25 0.23 0.24
Banc One 1.15 0.31 0.24 0.24 0.28 0.30
UBS 0.30 0.26 0.27
Deutsche Bk AB 0.25 0.27 0.30 0.34
Wachovia 0.25 0.36 0.30 0.34
CS
Greenwich 0.28 0.44 0.45 0.34 0.25

Grand Total 0.32 0.32 0.33 0.30 0.34 0.35 0.35

 2006 Total Amount Avg

Lehman 0.26 419,345 0.34
Credit Suisse 348,350 0.29
Bank of America 0.22 344,493 0.26
Countrywide 0.68 304,252 0.66
JP Morgan 0.22 276,096 0.25
Citigroup 0.29 248,182 0.28
Deutsche Bank 0.27 215,450 0.26
Merrill Lynch 0.38 215,293 0.35
RBS Greenwich 0.24 214,523 0.27
Morgan Stanley 0.22 207,928 0.23
Bear Steams 0.25 196,972 0.29
Salomon 160,266 0.31
Goldman Sachs 0.25 153,902 0.27
Barclays 0.22 105,409 0.23
Banc One 77,828 0.29
UBS 0.25 62,725 0.26
Deutsche Bk AB 62,000 0.29
Wachovia 0.26 56,242 0.30
CS 0.24 55,600 0.24
Greenwich 51,278 0.39

Grand Total 0.34 4,139,966 0.34

TABLE 6.
Dollar issuance of ABS by year and S&P credit rating

 1999 2000 2001 2002 2003

AAA 25,336 35,506 86,631 162,646 250,621
AA+ 60 67 522 1,555 3,309
AA 316 389 1,126 3,024 9,816
AA- 26 19 103 211 408
A+ 457 375 1,762 1,622 1,885
A 348 890 3,666 5,909 10,257
A- 56 209 1,511
BBB+ 800 12 318 1,959
BBB 171 1,380 2,448 5,563 7,870
BBB- 49 77 149 627 1,211
BB+ 500 3,625 824 22
BB 6 689 325 723 164
BB- 227 77 322 84
B+ 1,126 570 0 125 0
B 438 75 86 491 143
B- 1,548 257 711 43
CCC+ 14 189 78 63
CCC 2,067 1,796 187 396 39
CCC- 194 470 1,095 85
CC 448 15
D 1,775 1,287 653 101
N.A. 32,380 31,730 42,972 57,060 29,123
NR 160,233 183,390 219,233 200,948 187,157
Grand Total 227,568 259,735 365,886 442,880 505,603

 2004 2005 2006 Grand Total

AAA 476,898 620,109 658,438 2,316,186
AA+ 11,280 20,144 22,063 58,999
AA 15,932 21,481 23,562 75,646
AA- 3,201 6,731 7,775 18,474
A+ 5,033 9,077 8,470 28,681
A 13,622 14,510 11,175 60,376
A- 4,858 5,914 4,255 16,802
BBB+ 4,235 5,867 6,187 19,378
BBB 8,246 9,194 9,143 44,015
BBB- 3,105 4,621 1,922 11,761
BB+ 201 571 1,091 6,834
BB 219 729 4,259 7,115
BB- 4 5 59 777
B+ 0 2 262 2,085
B 24 0 4,754 6,011
B- 114 2,673
CCC+ 344
CCC 12 2,827 7,323
CCC- 1,843
CC 463
D 15 427 4,258
N.A. 26,549 53,124 42,806 315,745
NR 96,264 52,844 28,463 1,128,531
Grand Total 673,313 826,081 838,900 4,139,966

TABLE 7.
Average ABS percentage underwriting fee by year and S&P credit rating

 1999 2000 2001 2002 2003 2004 2005 2006

AAA 0.36 0.31 0.33 0.28 0.27 0.23 0.22 0.21
AA+ 0.63 0.35 0.51 0.36 0.42 0.44 0.32 0.29
AA 0.50 0.40 0.44 0.40 0.42 0.46 0.36 0.36
AA- 0.45 0.58 0.49 0.49 0.43 0.50 0.46 0.44
A+ 0.25 0.38 0.35 0.37 0.60 0.54 0.49 0.46
A 0.60 0.47 0.49 0.43 0.47 0.49 0.49 0.49
A- 0.50 0.47 0.81 0.46 0.61 0.61
BBB+ 0.40 0.63 0.72 0.53 0.66 0.58
BBB 0.61 0.63 0.65 0.54 0.70 0.57 0.58 0.50
BBB- 1.11 0.75 0.59 1.05 0.64 0.67 0.54
BB+ 0.60 0.29 0.25 0.74 0.80
1313 0.46 0.60 0.38 1.25 0.25 0.60
BB- 0.10 0.63 0.34 0.65 1.58
13+ 0.35 0.45 0.50
B 0.33 0.48 0.46 0.61
B- 0.39 0.45 0.37 0.50
CCC+ 0.44 0.18 0.64
CCC 0.44 0.39 0.64 0.56 0.26
CCC- 0.67 0.34 0.35 0.63
CC 0.70
D 0.56 0.73 0.65 0.62 1.00
N.A. 0.36 0.39 0.37 0.35 0.46 0.29 0.25 0.23
NR 0.28 0.28 0.28 0.23 0.21 0.19 0.15 0.13
Grand Total 0.32 0.32 0.33 0.30 0.34 0.35 0.35 0.34

 Grand Total

AAA 0.24
AA+ 0.35
AA 0.39
AA- 0.46
A+ 0.49
A 0.48
A- 0.58
BBB+ 0.61
BBB 0.58
BBB- 0.67
BB+ 0.68
1313 0.58
BB- 0.46
13+ 0.37
B 0.60
B- 0.39
CCC+ 0.47
CCC 0.32
CCC- 0.42
CC 0.70
D 0.67
N.A. 0.36
NR 0.25
Grand Total 0.34

TABLE 8.
ABS by original weighted average life in $ Millions

 1999 2000 2001 2002 2003

<.5 9,571 14,014 13,755 17,964 16,514
.5<1.5 30,281 32,735 38,727 48,441 56,934
1.5<2.5 29,776 37,723 37,436 56,827 62,948
2.5<3.5 46,903 65,263 125,935 150,672 173,683
3.5<4.5 13,527 10,971 13,943 16,199 16,002
4.5<5.5 17,419 33,622 40,501 35,348 64,363
5.5<6.5 6,368 3,453 4,578 7,754 16,704
6.5<7.5 8,578 11,729 14,925 10,593 11,960
7.5<8.5 1,437 1,355 1,211 3,461 5,235
.8.5<9.5 2,477 1,779 3,632 1,788 1,237
9.5<10.5 2,665 2,911 3,767 3,697 5,346
10.5<11.5 1,558 2,043 2,494 2,031 3,521
11.5<12.5 1,201 1,417 910 1,122 1,205
12.5<13.5 293 824 951 340 452
13.5<14.5 281 492 409 26 698
14.5<15.5 341 5 13 170
15.5<16.5 4 4 0
16.5<17.5 17 15 966 403
17.5<18.5 8 7 95
18.5<19.5 270 11
19.5<20.5 380 60
20.5<21.5 20
25.5<26.5 52
26.5<27.5 32
28.5<29.5
29.5<30.5
N.A. 54,863 39,388 61,949 85,421 68,270
Grand Total 227,568 259,735 365,886 442,880 505,603

 2004 2005 2006 Grand Total

<.5 15,344 22,481 24,197 133,841
.5<1.5 89,052 158,837 193,907 648,914
1.5<2.5 91,820 131,865 156,263 604,658
2.5<3.5 224,989 214,266 143,196 1,144,906
3.5<4.5 17,552 24,292 47,769 160,255
4.5<5.5 55,255 93,836 96,417 436,761
5.5<6.5 21,426 15,867 15,639 91,788
6.5<7.5 25,358 30,649 27,340 141,133
7.5<8.5 6,062 16,008 15,462 50,232
.8.5<9.5 3,348 5,232 10,001 29,494
9.5<10.5 7,392 11,733 10,439 47,950
10.5<11.5 3,752 1,804 3,077 20,280
11.5<12.5 1,324 3,427 2,627 13,234
12.5<13.5 113 1,939 1,513 6,427
13.5<14.5 508 2,121 3,621 8,156
14.5<15.5 1,513 3,556 7,257 12,854
15.5<16.5 207 1,999 703 2,917
16.5<17.5 578 749 170 2,898
17.5<18.5 188 989 631 1,919
18.5<19.5 550 832
19.5<20.5 100 540
20.5<21.5 20
25.5<26.5 52
26.5<27.5 32
28.5<29.5 676 676
29.5<30.5 0 0
N.A. 107,532 83,654 78,119 579,196
Grand Total 673,313 826,081 838,900 4,139,966

TABLE 9.
Average ABS percentage underwriting fee by original weighted average
life

 1999 2000 2001 2002 2003 2004 2005 2006

<.5 0.14 0.13 0.13 0.13 0.12 0.12 0.12 0.12
.5<1.5 0.19 0.18 0.18 0.18 0.18 0.18 0.18 0.18
1.5<2.5 0.26 0.27 0.25 0.27 0.28 0.26 0.21 0.25
2.5<3.5 0.29 0.27 0.26 0.27 0.26 0.24 0.22 0.20
3.5<4.5 0.34 0.39 0.52 0.41 0.81 0.61 0.68 0.32
4.5<5.5 0.38 0.38 0.41 0.40 0.47 0.53 0.44 0.47
5.5<6.5 0.44 0.61 0.59 0.45 0.48 0.27 0.47 0.41
6.5<7.5 0.48 0.42 0.45 0.37 0.37 0.53 0.46 0.43
7.5<8.5 0.59 0.34 0.41 0.29 0.32 0.25 0.25 0.24
8.5<9.5 0.45 0.56 0.38 0.39 0.41 0.26 0.24 0.25
9.5<10.5 0.51 0.56 0.54 0.51 0.44 0.37 0.32 0.33
10.5<11.5 0.43 0.48 0.46 0.44 0.38 0.36 0.36 0.34
11.5<12.5 0.49 0.44 0.56 0.59 0.38 0.33 0.31 0.30
12.5<13.5 0.40 0.47 0.59 0.48 0.29 0.41 0.31 0.46
13.5<14.5 0.45 0.50 0.41 0.34 0.33 0.32 0.31
14.5<15.5 0.50 0.50 0.73 0.48 0.37 0.25
15.5<16.5 0.27 0.24
16.5<17.5 0.42 0.50 0.35 0.24 0.35 0.37 0.33
17.5<18.5 0.20 0.33 0.31 0.33
18.5<19.5 0.25 0.31
19.5<20.5 0.40 0.65
20.5<21.5
25.5<26.5 0.25
26.5<27.5
28.5<29.5
29.5<30.5
N.A. 0.34 0.31 0.29 0.26 0.29 0.28 0.26 0.26
Grand Total 0.32 0.32 0.33 0.30 0.34 0.35 0.35 0.34

 Grand Total

<.5 0.13
.5<1.5 0.18
1.5<2.5 0.25
2.5<3.5 0.25
3.5<4.5 0.51
4.5<5.5 0.46
5.5<6.5 0.43
6.5<7.5 0.46
7.5<8.5 0.30
8.5<9.5 0.33
9.5<10.5 0.44
10.5<11.5 0.40
11.5<12.5 0.39
12.5<13.5 0.41
13.5<14.5 0.35
14.5<15.5 0.38
15.5<16.5 0.26
16.5<17.5 0.33
17.5<18.5 0.30
18.5<19.5 0.29
19.5<20.5 0.53
20.5<21.5
25.5<26.5 0.25
26.5<27.5
28.5<29.5
29.5<30.5
N.A. 0.29
Grand Total 0.34

TABLE 10.
Underwriting fees--full period analysis

Independent Variable US US

Intercept 0.70055 ** 0.767986 **
 -8.388164 (10.35501)
AAA -0.023476 ** -0.018663 **
 (-2.594656) (-2.132015)
Mezzanine 0.075074 ** 0.076529 **
 (6.170144) (6.658604)
Subordinated 0.197144 ** 0.187523 **
 (15.77486) (15.81513)
Deal Amount 0.022627 ** 0.024246 **
 (4.857101) (5.630582)
Auto 0.017201
 (1.340346)
Credit Card -0.007355
 (-0.45365)
Home Equity 0.028464
 (2.414706)
Student Loan 0.015939
 (0.945704)
Par Amount -0.059351 ** -0.063594 **
 (-15.47095) (-18.49084)
Prestige 1.455502 ** 1.452287 **
 (14.03731) (14.02468)
WAL 0.008641 ** 0.007966 **
 (6.500939) (6.891374)
Loyalty 0.148386 ** 0.148779 **
 (25.23054) (25.83672)
Count

1999 0.0103 0.006956
 (0.630703) (0.433347)
2000 0.037779 ** 0.033018 **
 (2.825256) (2.502031)
2001 0.031504 ** 0.027193 **
 (2.429743) (2.119839)
2002 0.010687 0.007649
 (0.905397) (0.652873)
2003 0.044663 ** 0.041355 **
 (4.030266) (3.771524)
2004 0.00915 0.009419
 (1.001979) (1.032641)
2005 0.008753 0.008815
 (1.06862) (1.077612)
# of Observations 9204 9209
Adjusted R-squared 0.305714 0.30517

Independent Variable US US

Intercept 0.660729 ** 0.486393 **
 (9.024349) (7.41368)
AAA -0.014146 -0.023758 **
 (-1.796523) (-3.100573)
Mezzanine 0.098982 ** 0.083624 **
 (9.237869) (8.245917)
Subordinated 0.215614 ** 0.222635 **
 (19.82468) (21.48809)
Deal Amount 0.014717 ** 0.017696 **
 (3.590823) (4.670519)
Auto -0.018403
 (-1.773387)
Credit Card -0.01667
 (-1.203752)
Home Equity -0.062119 **
 (-6.324422)
Student Loan -0.033503 **
 (-2.355085)
Par Amount -0.043233 ** -0.039234 **
 (-12.72165) (-12.68719)
Prestige 0.179105 * 0.238656 **
 (1.904005) (2.542794)
WAL 0.011605 ** 0.011838 **
 (10.03838) (11.6708)
Loyalty

Count 0.000186 ** 0.000178 **
 (46.54663) (46.38599)
1999 0.083306 ** 0.094771 **
 (6.807137) (7.910568)
2000 0.061172 ** 0.074281 **
 (5.257324) (6.462479)
2001 0.055068 ** 0.068142 **
 (4.834924) (6.056174)
2002 0.035127 ** 0.043721 **
 (3.361887) (4.209006)
2003 0.053801 ** 0.061949 **
 (5.408508) (6.276172)
2004 0.010349 0.010769
 (1.259876) (1.309623)
2005 0.013856 0.014312 *
 (1.874233) (1.934175)
# of Observations 9989 9994
Adjusted R-squared 0.38761 0.38463

* Significant at the .10 level

** Significant at the .05 level

The above table shows the regression of underwriter fee (the
dependent variable in basis points) on a number of independent
variables that affect the marketability of ABS. The sample includes
non/private placed US ABS from 1999/2006. AAA is a dummy variable
assigned 1 if bond is rated AAA, zero otherwise. The variables
Mezzanine Bond and Subordinated Bond are dummy variables assigned 1
if bond includes a class descriptor of mezzanine or subordinated
respectively, zero otherwise. Deal Amount is the log of deal amount/
/dollar amount of entire deal of which the specific class/bond is
included in. Dummy variables Auto, Credit Card, Home equity and
Student Loan are variables assigned 1 if deal is of given type,
zero otherwise. Dummy variables are assigned for each year included
in the study. Par amount is the log of par amount for the bond.
Prestige is the proportion of market share that the underwriter had
over the study period. WAL is weighted average life of the bond at
issuance and serves as a measure of maturity/term of bond. Loyalty
is a dummy variable assigned 1 if underwriter used in current bond
also served as underwriter in previous issuance. Count is defined
as the number of bond underwritten by the same underwriter for a
given issuer of ABS during the sample period.

TABLE 11.
Underwriting Fees and Loyalty including Deal Type--Year by Year
Analysis

Year 1999 2000 2001

Independent
Variable US US US

Intercept 0.381 *** 0.595 ** 0.5787 ***
 (3.0300) (2.3977) (3.8764)
AAA -0.0454 *** 0.0243 0.0059
 (-2.7636) (0.8895) (0.4234)
Mezzanine 0.1733 *** 0.1726 *** 0.0897 ***
 (10.6514) (5.4976) (4.6866)
Subordinated 0.2072 *** 0.1873 *** 0.1425 ***
 (11.7161) (5.8756) (7.2838)
Deal Amount -0.0099 0.0193 0.0063
 (-1.2876) (1.3243) (0.7475)
Auto -0.0205 0.0234 -0.0351
 (-1.4486) (0.8813) (-1.5035)
Cards -0.1112 *** -0.0253 -0.0706 ***
 (-5.5595) (-0.7686) (-2.8384)
Home Equity 0.0279 ** 0.0514 ** -0.015
 (2.0665) (1.9854) (-0.735)
Students -0.0604 ** 0.0036 -0.1009 ***
 (-1.9939) (0.0727) (-3.1404)
Par Amount -0.0009 -0.0449 *** -0.0279 ***
 (-0.1394) (-3.9149) (-3.7232)
Prestige 0.1256 0.4606 * 0.1771 *
 (1.4057) (1.9346) (1.68)
WAL 0.0265 *** 0.0116 *** 0.0198 ***
 (12.3228) (3.818) (8.1437)
Loyalty 0.0141 0.0124 0.0373 ***
 (1.5017) (0.6867) (3.3301)
# Observations 628 670 706
Adjusted 0.6221 0.3899 0.4827
R-squared

Year 2002 2003 2004

Independent
Variable US US US

Intercept 0.9645 *** 0.5389 *** 0.5662 ***
 (3.5776) (2.7018) (2.7822)
AAA 0.0108 -0.0718 *** -0.056 *
 (0.4968) (-3.4231) (-1.6713)
Mezzanine 0.0461 0.0363 -0.011
 (1.178) (1.2441) (-0.2658)
Subordinated 0.1324 *** 0.2849 *** 0.2255 ***
 (3.8285) (9.5392) (5.2834)
Deal Amount 0.0455 *** 0.0185 * 0.0565 ***
 (3.3139) (1.6628) (4.4672)
Auto 0.015 0.0478 0.0368
 (0.383) (1.2064) (0.8361)
Cards 0.0797 0.0264 0.0603
 (1.6933) (0.543) (1.1175)
Home Equity 0.1008 *** 0.091 ** 0.0303
 (2.479) (2.3927) (0.741)
Students -0.0158 0.0997 ** 0.0609
 (-0.3121) (2.0163) (1.1717)
Par Amount -0.0948 *** -0.0433 *** -0.091 ***
 (-8.2542) (-5.4224) (-9.1559)
Prestige 0.5874 ** -0.3257 1.9272 ***
 (2.3515) (-1.2408) (10.4074)
WAL 0.0072 0.0117 *** -0.0006
 (1.6199) (3.184) (-0.1801)
Loyalty 0.0879 *** 0.2316 *** 0.2036 ***
 (4.4802) (17.1008) (13.5823)
# Observations 864 1551 2255
Adjusted 0.3733 0.3856 0.3307
R-squared

Year 2005

Independent
Variable US

Intercept 1.0553 ***
 (4.1878)
AAA 0.0122
 (0.312)
Mezzanine 0.0957 **
 (2.111)
Subordinated 0.1701 ***
 (3.6019)
Deal Amount 0.0411 ***
 (3.0294)
Auto 0.0016
 (0.0341)
Cards 0.0041
 (0.0683)
Home Equity -0.0938 **
 (-2.1067)
Students 0.051
 (0.9717)
Par Amount -0.1024 ***
 (-9.3927)
Prestige 3.0328 ***
 (10.0582)
WAL -0.0009
 (-0.2471)
Loyalty 0.1949 ***
 (12.481)
# Observations 1642
Adjusted 0.3451
R-squared

 * Significant at the .10 level

 ** Significant at the .05 level

 *** Significant at the .01 level

The above table shows the regression of underwriter fee (the
dependent variable in basis points) on a number of independent
variables that affect the marketability of ABS. AAA is a dummy
variable assigned 1 if bond is rated AAA, zero otherwise. The
variables Mezzanine Bond and Subordinated Bond are dummy variables
assigned 1 if bond includes a class descriptor of mezzanine or
subordinated respectively, zero otherwise. Deal Amount is the log of
deal amount//dollar amount of entire deal of which the specific
class/bond is included in. Dummy variables Auto, Credit Card, Home
equity and Student Loan are variables assigned 1 if deal is of given
type, zero otherwise. Dummy variables are assigned for each year
included in the study. Par amount is the log of par amount for the
bond. Prestige is the proportion of market share that the
underwriter had during the Year listed. For example Year 1999 table
represents 1999 Prestige values applied to 2001 issuance. WAL is
weighted average life of the bond at issuance and serves as a
measure of maturity/term of bond. Loyalty is a dummy variable
assigned 1 if underwriter used in current bond also served as
underwriter in previous issuance.

TABLE 12.
Underwriting Fees and Loyalty excluding Deal Type--Year by Year
Analysis

Year 1999 2000 2001

Independent
Variable US US US

Intercept 0.776 *** 0.8731 *** 0.813 ***
 (7.4818) (4.3988) (6.6326)
AAA -0.0392 ** 0.0283 0.0024
 (-2.3351) (1.0615) (0.181)
Mezzanine 0.1608 *** 0.1621 *** 0.0913 ***
 (9.7347) (5.4044) (5.0299)
Subordinated 0.1561 *** 0.1582 *** 0.1201 ***
 (9.488) (5.7418) (6.9235)
Deal Amount -0.0131 * 0.0186 0.004
 (-1.8087) (1.4274) (0.5658)
Par Amount -0.019 *** -0.0567 *** -0.0391 ***
 (-3.249) (-5.7488) (-6.3947)
Prestige 0.1721 * 0.2348 0.1895 *
 (1.9495) (1.10914) (1.866)
WAL 0.0245 *** 0.0103 *** 0.0163 ***
 (11.8112) (3.5949) (7.8303)
Loyalty 0.0117 0.0036 0.0423 ***
 (1.2089) (0.2059) (3.9883)
# Observations 628 670 706
Adjusted 0.5943 0.3874 0.474
 R-squared

Year 2002 2003 2004

Independent
Variable US US US

Intercept 1.3387 *** 0.6268 *** 0.5274 ***
 (5.4417) (3.478) (2.9357)
AAA 0.0102 -0.0598 *** -0.0614 **
 (0.4802) (-2.9174) (-1.9941)
Mezzanine 0.1056 *** 0.0458 * -0.0145
 (2.9432) (1.6626) (-0.3809)
Subordinated 0.1532 *** 0.2699 *** 0.2291 ***
 (4.7274) (9.3143) (5.4949)
Deal Amount 0.0248 ** 0.0222 ** 0.0559 ***
 (2.0579) (2.17) (4.7705)
Par Amount -0.0901 *** -0.0483 *** -0.0865 ***
 (-8.8059) (-6.6109) (-9.8189)
Prestige 0.4731 * -0.4718 * 1.9144 ***
 (1.9208) (-1.8363) (10.4759)
WAL 0.0075 ** 0.0132 *** 0.0015
 (2.0513) (4.145) (0.5513)
Loyalty 0.1111 *** 0.2318 *** 0.2037 ***
 (5.9494) (17.512) (13.8607)
# Observations 864 1551 2260
Adjusted 0.3659 0.3829 0.3311
 R-squared

Year 2005

Independent
Variable US

Intercept 0.5295 **
 (2.3573)
AAA -0.0291
 (-0.767)
Mezzanine 0.0422
 (0.9649)
Subordinated 0.1715 ***
 (3.6478)
Deal Amount 0.0517 ***
 (4.0184)
Par Amount -0.0864 ***
 (-8.6769)
Prestige 2.7686 ***
 (9.3056)
WAL 0.004
 (1.3611)
Loyalty 0.184 ***
 (11.9804)
# Observations 1642
Adjusted 0.3374
 R-squared

* Significant at the .10 level

** Significant at the .05 level

*** Significant at the .01 level

The above table shows the regression of underwriter fee (the
dependent variable in basis points) on a number of independent
variables that affect the marketability of ABS. AAA is a dummy
variable assigned 1 if bond is rated AAA, zero otherwise. The
variables Mezzanine Bond and Subordinated Bond are dummy variables
assigned 1 if bond includes a class descriptor of mezzanine or
subordinated respectively, zero otherwise. Deal Amount is the log
of deal amount--dollar amount of entire deal of which the specific
class/bond is included in. Par amount is the log of par amount for
the bond. Prestige is the proportion of market share that the
underwriter had during the Year listed. For example Year 1999
table represents 1999 Prestige values applied to 2000 issuance.
WAL is weighted average life of the bond at issuance and serves
as a measure of maturity/term of bond. Loyalty is a dummy variable
assigned 1 if underwriter used in current bond also served as
underwriter in previous issuance.

TABLE 13.
Underwriting Fees and Count including Deal Type--Year by Year Analysis

Year 1999 2000 2001

Independent
Variable US US US

Intercept 0.5092 *** 0.5126 ** 0.5502 ***
 (4.3688) (2.4653) (3.9818)
AAA -0.0424 *** 0.066 *** 0.0169
 (-2.7811) (2.7901) (1.3666)
Mezzanine 0.1751 *** 0.1681 *** 0.1083 ***
 (11.3598) (5.7954) (5.7537)
Subordinated 0.2011 *** 0.1872 *** 0.1747 ***
 (12.7849) (6.3971) (9.2661)
Deal Amount -0.014 * 0.017 -0.0007
 (-1.9533) (1.3031) (-0.0845)
Auto -0.0227 * 0.0006 -0.0617 ***
 (-1.8865) (0.0266) (-3.0835)
Credit Card -0.1074 *** -0.0489 -0.1072 ***
 (-5.8878) (-1.6454) (-4.9552)
Home Equity 0.0194 0.0066 -0.0612 ***
 (1.5752) (0.2829) (-3.4005)
Student Loan -0.0582 ** -0.043 -0.1529 ***
 (-1.9992) (-0.9291) (-5.7023)
Par Amount -0.0029 -0.0377 *** -0.0176 **
 (-0.4993) (-3.582) (-2.458)
Prestige 0.0967 0.5788 *** 0.2155 **
 (1.1956) (2.8451) (2.1584)
WAL 0.0265 *** 0.0136 *** 0.0219 ***
 (13.0862) (4.9388) (10.0958)
Count 0 0.0001 *** 0 ***
 (0.3328) (4.4085) (5.0457)
# of Observations 729 749 776
Adjusted 0.6256 0.4033 0.5117
 R-squared

Year 2002 2003 2004

Independent
Variable US US US

Intercept 0.88 *** 0.4112 ** 0.3121
 (3.7126) (2.258) (1.6191)
AAA 0.0078 -0.0622 *** -0.0503
 (0.4089) (-3.2573) (-1.6149)
Mezzanine 0.0235 0.05 * 0.0142
 (0.6837) (1.8708) (0.3681)
Subordinated 0.142 *** 0.3033 *** 0.2253 ***
 (4.6458) (11.076) (5.6633)
Deal Amount 0.0397 *** 0.0161 0.0656 ***
 (3.2617) (1.588) (5.5188)
Auto 0.0061 0.0368 -0.0447
 (0.1784) (1.0554) (-1.1087)
Credit Card 0.0697 * 0.0473 0.0635
 (1.7197) (1.0931) (1.2681)
Home Equity -0.0249 -0.0124 -0.0888 **
 (-0.6947) (-0.3688) (-2.3508)
Student Loan -0.0503 0.0179 0.0139
 (-1.1416) (0.4179) (0.2908)
Par Amount -0.0805 *** -0.03 *** -0.077 ***
 (-7.9165) (-4.0769) (-8.2489)
Prestige -0.0972 -0.3943 * 0.4623 **
 (-0.4402) (-1.6556) (2.3481)
WAL 0.0124 *** 0.013 *** -0.0003
 (3.1668) (3.9529) (-0.0895)
Count 0.0002 *** 0.0002 *** 0.0002 ***
 (14.7364) (24.3328) (21.0229)
# of Observations 905 1596 2324
Adjusted 0.4761 0.4675 0.3895
 R-squared

Year 2005

Independent
Variable US

Intercept 0.8957 ***
 (3.8435)
AAA 0.0086
 (0.2378)
Mezzanine 0.079 *
 (1.881)
Subordinated 0.1735 ***
 (3.9694)
Deal Amount 0.0543 ***
 (4.2914)
Auto 0.0102
 (0.2379)
Credit Card 0.1186 **
 (2.2324)
Home Equity -0.0983 **
 (-2.491)
Student Loan 0.055
 (1.1644)
Par Amount -0.0979 ***
 (-9.7265)
Prestige 0.021
 (0.0645)
WAL -0.0032
 (-0.9874)
Count 0.0002 ***
 (20.2979)
# of Observations 1675
Adjusted 0.4197
 R-squared

* Significant at the .10 level

** Significant at the .05 level

*** Significant at the .01 level

The above table shows the regression of underwriter fee (the
dependent variable in basis points) on a number of independent
variables that affect the marketability of ABS. AAA is a dummy
variable assigned 1 if bond is rated AAA, zero otherwise. The
variables Mezzanine Bond and Subordinated Bond are dummy variables
assigned 1 if bond includes a class descriptor of mezzanine or
subordinated respectively, zero otherwise. Deal Amount is the log
of deal amount//dollar amount of entire deal of which the specific
class/bond is included in. Dummy variables Auto, Credit Card, Home
equity and Student Loan are variables assigned 1 if deal is of given
type, zero otherwise. Dummy variables are assigned for each year
included in the study. Par amount is the log of par amount for the
bond. Prestige is the proportion of market share that the
underwriter had during the Year listed. For example Year 1999 table
represents 1999 Prestige values applied to 2000 issuance. WAL is
weighted average life of the bond at issuance and serves as a
measure of maturity/term of bond. Count is defined as the number of
bond underwritten by the same underwriter for a given issuer of ABS
during the sample period.

TABLE 14.
Underwriting Fees and Count excluding Deal Type--Year by Year Analysis

Year 1999 2000 2001

Independent
Variable US US US

Intercept 0.8362 *** 0.6106 *** 0.725 ***
 (8.6112) (3.5232) (6.2007)
AAA -0.0362 ** 0.0692 *** 0.0086
 (-2.347) (3.0259) (0.7068)
Mezzanine 0.1646 ** 0.1536 *** 0.0976 ***
 (10.5934) (5.574) (5.5345)
Subordinated 0.158 ** 0.1597 *** 0.1452 ***
 (10.7407) (6.2392) (8.5943)
Deal Amount -0.0163 ** 0.0218 * -0.0011
 (-2.4282) (1.8815) (-0.1628)
Par Amount -0.0186 *** -0.048 *** -0.0294 ***
 (-3.462) (-5.2568) (-4.7982)
Prestige 0.1603 ** 0.552 *** 0.2381 **
 (1.9744) (2.9104) (2.3556)
WAL 0.0244 *** 0.0128 *** 0.0195 ***
 (12.52) (4.9128) (10.3799)
Count 0 0.0001 *** 0 ***
 (1.2206) (4.633) (4.8798)
# Observations 729 749 776
Adjusted 0.6035 0.4028 0.4875
 R-squared

Year 2002 2003 2004

Independent
Variable US US US

Intercept 0.8542 *** 0.3017 * 0.1127
 (3.9522) (1.8301) (0.6324)
AAA -0.0066 -0.0741 *** -0.0861 ***
 (-0.3527) (-3.9783) (-2.973)
Mezzanine 0.0324 0.0325 -0.0135
 (1.0195) (1.2978) (-0.3775)
Subordinated 0.1725 *** 0.3096 *** 0.2455 ***
 (6.0282) (11.6756) (6.2927)
Deal Amount 0.0292 *** 0.018 * 0.0565 ***
 (2.777) (1.9307) (5.1285)
Par Amount -0.0677 *** -0.0256 *** -0.0589 ***
 (-7.3879) (-3.77) (-7.0436)
Prestige -0.0014 -0.2934 0.4624 **
 (-0.0064) (-1.2527) (2.3487)
WAL 0.0115 *** 0.0129 *** 0.0054 **
 (3.5604) (4.5219) (2.0966)
Count 0.0002 *** 0.0002 *** 0.0002 ***
 (15.4522) (24.7537) (20.6272)
# Observations 905 1596 2329
Adjusted 0.4713 0.4662 0.3843
 R-squared

Year 2005

Independent
Variable US

Intercept 0.3469 *
 (1.6707)
AAA -0.0435
 (-1.228)
Mezzanine 0.031
 (0.7597)
Subordinated 0.1909 ***
 (4.3706)
Deal Amount 0.0568 ***
 (4.7382)
Par Amount -0.0716 ***
 (-7.7467)
Prestige -0.1615
 (-0.4926)
WAL 0.0027
 (0.9771)
Count 0.0002 ***
 (19.3452)
# Observations 1675
Adjusted 0.4059
 R-squared

* Significant at the .10 level

** Significant at the .05 level

*** Significant at the .01 level

The above table shows the regression of underwriter fee (the
dependent variable in basis points) on a number of independent
variables that affect the marketability of ABS. AAA is a dummy
variable assigned 1 if bond is rated AAA, zero otherwise. The
variables Mezzanine Bond and Subordinated Bond are dummy variables
assigned 1 if bond includes a class descriptor of mezzanine or
subordinated respectively, zero otherwise. Deal Amount is the log
of deal amount//dollar amount of entire deal of which the specific
class/bond is included in. Dummy variables are assigned for each
year included in the study. Par amount is the log of par amount for
the bond. Prestige is the proportion of market share that the
underwriter had during the Year listed. For example Year 1999 table
represents 1999 Prestige values applied to 2000 issuance. WAL is
weighted average life of the bond at issuance and serves as a
measure of maturity/term of bond. Count is defined as the number of
bond underwritten by the same underwriter for a given issuer of ABS
during the sample period.
COPYRIGHT 2012 Omicron Delta Epsilon
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Puskar, David; Gottesman, Aron A.
Publication:American Economist
Article Type:Report
Date:Sep 22, 2012
Words:15036
Previous Article:The equation of exchange: a derivation.
Next Article:The effects of alcohol use on high school absenteeism.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters