While the authors of the paper, Ingo Fender of the BIS and Janet Mitchell of the National Bank of Belgium, support the concept of risk retention, their mathematical model of the incentives created suggest that different formats would incentivise more "screening" of borrowers in the underwriting process in different circumstances.
"In particular, while increasing effort relative to the case of non-retention, having the originator or arranger retain the equity tranche of a securitisation may lead to lower screening effort than other retention schemes," they write in The future of securitisation: how to align incentives?
Consequently, they argue, regulators should focus on disclosure of the scale and nature of risk retention, allowing investors to exert market pressure on originators to use the optimum form for a given deal.
The fairly simple model assumes a highly granular pool (so that idiosyncratic risk is diversified away) comprising "good" and "bad" borrowers. Borrowers are assumed to behave differently depending on a systematic risk factor that has two states. When "an unfavourable state of nature" applies, all bad borrowers default while some good borrowers will also default with a given probability. Conversely in an unfavourable state, no good borrowers default while bad borrowers default with a given probability.
Given these assumptions, and the assumptions that the originator’s decision to securitise and the level of screening it employs is driven by profit maximisation (ignoring reputational effects), the incentives for screening depend on both expected economic performance and the size of the retained portion relative to the portfolio.
"If a downturn is likely ... and the equity tranche is thin enough to be depleted if the downturn materialises ... the originator will expect zero payout," wrote Fender and Mitchell. "Knowing this prior to loan origination, when screening effort is chosen, reduces its incentives to exert effort. In contrast, both mezzanine tranche and vertical slice retention will tend to generate positive originator payouts."
Conversely, when the probability of a downturn is low or the equity tranche is relatively thick, equity tranches provide more incentives to exert effort in screening borrowers.
n Rating agencies hit the CMBS sector with further downgrades this week.
Fitch cut the three most junior tranches of FCC Proudreed Properties 2005, backed by French properties, by three notches each. The downgrades were driven by an estimated market value decline of 23%, resulting in a loan to value ratio of 73%.
Moody’s, meanwhile, downgraded the ‘M’ and ‘B’ tranches of Business Mortgage Finance 5, the small ticket CMBS for Commercial First, on the basis of rising arrears and loan losses. The ‘M1’ and ‘M1’ tranches were cut from A2 to A3, while the ‘B1’ and ‘B2’ notes were cut from Ba1 to B1. Some 38.3% of the pool is in arrears of at least one month, while three month arrears stand at 29.5%.
There were no surprises when Moody’s put the senior tranche of Windermere XII and White Tower 2006-3, already rated Baa1, on review for downgrade. Fitch also placed White Tower on review.
Finally, Moody’s put the Aaa rated ‘A2’ tranche of Perseus (ELoC 22) on review for downgrade because of the refinancing risk created by the deal’s short maturity and property value declines.
n Moody’s has downgraded all outstanding tranches of Caja Madrid’s high loan to value residential mortgage securitisations Madrid RMBS I, II and III, with the exception of the ‘A2’ notes of the second deal.
The downgrades result from an increase in the rating agency’s loss expectations after it conducted a loan level analysis of delinquent and defaulted mortgages. In the case of Madrid RMBS III’s ‘A2’ tranche, the downgrade was also prompted by the possibility of high arrears triggering a switch to pro-rata amortisation within the ‘A’ class.