Following the financial crisis, there has been a lot of discussion about what went wrong in securitisation markets. Much of this work has focused on the conflicts of interests and moral hazard problems that occur at the various links in the securitisation chain (see Keys et al. 2012, Gorton and Metrick 2012). In response, academics, policymakers and practitioners alike have debated potential reforms. Some key elements of the Dodd-Frank reforms involve requirements of the level and type of credit risk retention by the deal sponsors. Prior to these mandates, it is important to understand how these markets work on their own and how market participants use financial contracting solutions to overcome the informational problems in the securitisation market. However, there is very little empirical work in this area. In Begley and Purnanandam (2013), we establish some stylised facts about the tranche structure of deals in the private-label residential mortgage-backed securities (RMBS) markets, and then examine the role that the equity tranche played in mitigating information problems in this market during the period leading up to the financial crisis.
Tranche structure: Facts and determinants
For our study, we use a carefully assembled representative sample of about 200 private-label RMBS deals (i.e., deals not affiliated with Fannie Mae or Freddie Mac), which are backed by about 500,000 loans. This sample constitutes about 12% of this market during 2002 and 2005. Each deal contains a pool of mortgages, whose cash flows are tranched into a ‘waterfall’ senior-subordinated manner and flow first to the senior-most AAA-rated tranches, then to the mezzanine tranches, and then finally to the junior-most equity tranche.
The average deal in our sample is made up of 90.4% AAA-rated (senior) tranches, 8.4% mezzanine tranches, and 1.2% equity tranche. Moving from 2002 to 2005, the level of AAA-rated tranche decreased from 92.6% to 88.2%, and the level of equity tranche increased from 0.7% to 1.6% of the deal. In comparison, prior work has shown the level of unrated equity tranche to be 3-4% in commercial MBS (Stanton and Wallace 2011), and 11% in collateralised loan obligations (Benmelech and Dlugosz 2009).
Perhaps not surprisingly, we find in regression analysis that many of the observable credit risk factors, such as FICO score, loan-to-value (LTV) ratio, and geographic diversity of the pool, explain well the relative sizes of the AAA-rated and mezzanine tranche in a deal. However, these factors do not explain well the variation in the size of the equity tranche. Instead, we find that a strong determinant of the size of the equity tranche is the level of information asymmetry between the deal sponsors and investors, as proxied by the proportion of loans in the underlying pool with no documentation. One standard deviation increase in the portion of no-documentation loans is associated with a 60% increase in the size of the equity tranche in the median deal. In sum, while observable credit risk factors explain the variation in the size of rated tranches, the size of the unrated, equity tranche, is driven by asymmetric information concerns.
The information content of the equity tranche size
What was the role of the equity tranche in this market?
- Some commentators argue that the equity tranche was simply ‘toxic waste’ that could not be initially sold by the sponsors.
- On the other hand, several theoretical models of security design suggest that the larger size of the equity tranche reveals the sponsors’ positive information about the quality of the underlying assets (e.g. Leland and Pyle 1977, DeMarzo and Duffie 1999, DeMarzo 2005).
Before placing regulations on this piece of the securitisation as proposed by the Dodd-Frank legislation, it is important to understand if and how the design of this part of RMBS deals mitigated contracting frictions.
To evaluate the information content of the equity, we relate the size of the equity tranche at the time of the deal creation to the future performance of the underlying mortgage loans.
- If a larger equity tranche is associated with poor quality deals, then we should see these mortgage pools have higher ex-post foreclosure rates.
- Alternatively, if a larger equity tranche conveys favourable private information about the mortgage pool, then we should see better ex-post performance of these loans.
- Moreover, if a role of the equity tranche is to provide information about the unobserved quality of the underlying loans, then the relationship should be stronger for relatively opaque deals, where the information problems are greater.
To make meaningful inference about the performance of these deals, we create a performance benchmark that allows us to control for observable characteristics of a deal. In particular, we use a matching strategy to create a ‘matched pool’ for each deal in our sample. For each loan in a given pool, we find a similar loan, which is virtually identical in terms of FICO score, LTV ratio, state, interest rate type, origination date, etc., but is not in that particular pool. Thus, the actual and matched pools are similar on the key observable dimensions including potential geographical correlation structure. However, while the actual pool contains the private information of the deal sponsor, the matched pool, by construction, does not. The difference in the foreclosure rates of the actual and matched pools yields a measure of ‘abnormal default.’
Consistent with the notion that a higher level of equity tranche provides positive information about the underlying mortgages, we find that deals with a higher-than-median equity tranche have substantially lower abnormal default rates. The results are concentrated within more opaque deals, as proxied by above-median levels of no-documentation loans in the pool. Opaque deals with a high equity tranche have abnormal foreclosure rates that are about 20-25% lower than low equity tranche deals. In an average pool of about 3,000 loans, this translates at about 100 fewer foreclosures amounting to about $25 million in loan principal.
To further study the channel of information asymmetry in these transactions, we focus on the subsample of deals where the sponsor is also the originator of the loans in the deal. It is these cases where the information advantage of the sponsor over the buyer is likely to be greatest, and thus the informational role of the equity tranche is likely to be most prominent. Indeed, we find that the results are the strongest for this subset of deals.
Finally, we investigate whether this information about unobserved quality was incorporated into the prices of these securities. If the market interpreted these signals of high unobserved quality, then we should expect to see higher offering prices for the more senior tranches in deals with larger equity tranches. When examining the relationship between at-issuance yields and the size of the equity tranche we find that, conditional on credit rating, a larger equity tranche is associated with lower yield spreads (higher prices) for the more senior tranches. The effects are particularly strong for opaque deals, and for more informationally sensitive non-AAA tranches.
In summary, there has been much debate about the role of the equity tranche and the role that its size and retention should play in regulating RMBS markets moving forward. Our study provides some stylised facts about how these markets were functioning during the period leading up to the crisis, and shows that the size of the equity tranche provided information about the unobserved credit quality of the underlying mortgages in private-label RMBS deals. As predicted by some of the basic models of contracting under asymmetric information, we show that a larger equity tranche is related to superior ex-post default performance.
Begley, T., & Purnanandam, A. (2013), “Design of Financial Securities: Empirical Evidence from Private-label RMBS Deals,” Working Paper.
Benmelech, E., & Dlugosz, J. (2009), “The alchemy of CDO credit ratings,” Journal of Monetary Economics, 617-634.
DeMarzo, P. (2005), “The pooling and tranching of securities: A model of informed intermediation,” Review of Financial Studies, 1-35.
DeMarzo, P., & Duffie, D. (1999), “A liquidity-based model of security design,” Econometrica, 65-99.
Gorton, G., & Metrick, A. (2012), “Securtization,” NBER working paper.
Keys, B., Piskorski, T., Seru, A., & Vig, V. (2012), Mortgage Financing in the Housing Boom and Bust. Housing in the Financial Crisis, University of Chicago Press.
Leland, H., & Pyle, D. (1977), “Informational asymmetries, financial structure, and financial intermediation,” Journal of Finance, 371-387.
Stanton, R., & Wallace, N. (2011), “CMBS subordination, ratings inflation, and regulatory-capital arbitrage,” Working Paper.