Bill Berliner: The Risk Retention Debate
Originally published in the February 2011 issue of Asset Securitization Report
Considerable uncertainty surrounds the recently postponed issuance of the mortgage securitization regulations required by the Dodd-Frank Act (DFA). In particular, two critical decisions involve the required risk retention contained in the DFA. The forthcoming directives will clearly have a major impact on the competitive and economic structure of the mortgage industry in the future.
The two major issues are 1) what types of loans will be exempt from the requirements (so-called qualifying residential mortgages, or QRMs) and 2) how the risk retention requirement will actually be applied. The former decision will focus on the breadth of the QRM definition. Wells Fargo recently created a stir with a proposal that QRMs should have a maximum LTV of 70%, a much narrower definition than advocated by other industry participants, including the MBA.
Under the Wells proposal, numerous loans with LTVs lower than the historical benchmark of 80% would require risk retention even if they can be securitized through the GSEs. Since securitizers would have to hold capital against these loans, a 70% standard would make them more expensive to originate, which would in turn be passed on to consumers in the form of higher rates. (I’d estimate that the increased capital cost probably translates into an incremental 50 basis points in rate, all else equal.)
In addition to its direct impact on mortgage rates, a 70% LTV threshold would serve to reduce competition in the industry. Small and mid-size mortgage bankers would either need to be better capitalized or forced to act as correspondent lenders that sell their production in loan form to better-capitalized entities. This would make their pricing increasingly dependent on the large lenders, making them less competitive and ultimately limiting consumers’ choices.
While the DFA states that a minimum of 5% of non-qualifying loans in a securitization must be retained, the form of the risk retention piece has not yet been determined. The possible definitions of the risk retention position could be either “vertical” (i.e., holding a pro-rata portion of the loans being securitized) or “horizontal,” which is essentially a first-loss tranche of the securitization. The horizontal definition would make the assets being retained much more leveraged and risky. For a $100 million securitization, for example, 5% losses on the collateral would wipe out the entire $5 million risk retention piece if the horizontal definition is used, but result in only a $250,000 writedown using the vertical definition.
The horizontal definition has been prominently supported by Redwood Trust, which is a large REIT. The economic model for many REITs has been to securitize their production while holding the subordinated classes, which has the same effect as holding a horizontal risk retention tranche. This strategy makes economic sense because REITs pay no corporate income tax as long as 90% of their earnings are distributed to shareholders.
If the horizontal model is adopted, I expect that the market for non-qualifying loans will be jointly dominated by REITs (which would securitize their production and leverage their tax advantage) and large banks that would simply hold these loans in portfolio. This would ultimately result in a two-tiered loan market with major differences in rates between qualifying and nonqualifying loans.
In my opinion, the potential for a series of bureaucratic decisions to profoundly impact the economics of mortgage lending highlights how far the flawed “skin in the game” argument has been removed from its original purpose of insuring quality underwriting. I favor as broad an interpretation of the QRM standard as possible, as well as a vertical risk retention definition, simply because these will have the smallest potential impact on the competitive structure of the mortgage lending industry.