This article was first published on Garp.org and can be found here.
The conservatorships of government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac — one of the last major unresolved matters of the financial crisis – have recently received heightened attention. Policymakers are now awaiting an Administration announcement on GSE reform, which many expect to feature privatization of each company.
However, while releasing the GSEs from conservatorship and recapitalizing them is a reasonable and pragmatic solution to perpetuating a policy vacuum on this issue, it misses a golden opportunity to once and for all address a number of problems that amplified risk in housing finance during the mortgage boom – and that remain today.
Overlapping market competition between the GSEs and the Federal Housing Administration (FHA) often leads to adverse selection against the weaker FHA, posing risk to the agency and, ultimately, the taxpayer. Moreover, the GSE duopoly, the cyclical nature of mortgage banking and increased market power of the largest originators are inherently procyclical, yielding excessive risk-taking over time. A more encompassing set of housing finance reforms could put the system on a much surer footing in the future.
Risks of the GSE Duopoly
Prior to conservatorship, the economic framework under which the GSEs operated was unique to their franchise. Calling it a duopoly oversimplifies the actual conditions under which the companies competed. At its core, Fannie and Freddie (full disclosure: I worked at both) are virtually indistinct in terms of their business models. Both firms compete for the same business and sellers (loan originators), subject to regulatory oversight.
A major contributor to greater risk-taking in housing finance lies in the competition between both GSEs. The only real levers each company has to compete on are price, product and/or service. During the boom period leading up to the crisis, the origination market became increasingly concentrated in the hands of a small number of very large originators: e.g., Countrywide Financial and Washington Mutual – two more of my former employers.
Originator concentration remains an issue today. In fact, with the increased presence of large, lightly-regulated nonbank originators, things may have actually gotten worse.
The stories of significant guarantee fee price concessions being made by the GSEs during the boom, in return for greater market share agreements, are almost legendary. In addition to underpricing credit risk on a swath of risky loans bought by the GSEs during the boom, guarantee fees never directly priced in the impact of defective loan manufacturing process on credit risk.
Representations and warranties, moreover, put lenders on the hook to repurchase defective loans from GSES only after a lending contract has been breached. As a result, even today, reps and warrants do not effectively incent lenders to clean up their processes in advance of a problem.
A step in the right direction would be to require lenders to submit to a process-quality assessment that can be priced as an additional component into guarantee fees. This would allow risk-based pricing of seller origination processes, improving in advance processes that have trouble keeping up with elevated product risks during booms.
The GSEs compounded the guarantee fees dilemma by accepting products with greater risk-layering and nontraditional features – particularly during the last few years of the boom. At that time, loosening credit standards on mortgage products was an effective, but ultimately risky, way to build market share.
The secondary mortgage market was also impacted by the advent of automated underwriting systems (AUS) and valuation models, which ushered in a new way to compete by vastly reducing underwriting and collateral valuation turn-around times in the loan origination process. While AUS and AVM technology provides tremendous risk management capabilities, lenders largely view them as labor-saving technologies.
In retrospect, in the lead-up to the crisis, all the elements of excessive risk were evident: competition between two companies in a highly commoditized market; concentration of market power by a small number of very large lenders; a frenzied economic mortgage boom, and weak regulatory oversight. In the aftermath of the crisis, none of the GSE reform plans put forth addressed these abnormalities.
In a highly cyclical market such as mortgage banking, to preserve market share and firm relevancy, competitive forces naturally lead to some combination of price, product and/or service concessions. This race to the bottom during the boom also took place among private mortgage insurance companies. In an era of the common security platform for GSE mortgage securities, some of these incentives to take greater risk have been mitigated, but not completely – and this will hold particularly true if both GSEs are released from conservatorship.
After the crisis, policymakers recognized that having dual mortgage securities providers didn’t really make sense and actually posed significant costs and risks to the system and the GSEs. So, as a natural extension to that logic, why does the secondary mortgage market today need two sets of pricing, product and service paths? Instead of two AUS processes, why not one? Why not a single seller-servicer guide, rather than two?
Creating a common underwriting, servicing, product and pricing platform would remove drivers of excessive risk-taking during the next boom.
Market Overlap Risks and Potential Fixes
Compounding the duopolistic competition of the GSEs for housing finance risk is the fact that the GSEs and FHA compete for some of the same loans. An essential part of a mortgage lender’s secondary marketing desk is best-execution pricing. Basically, what this amounts to is choosing the best outlet for a mortgage.
At times, this may come down to selling the loan to a GSE or to the government as an FHA loan. In many instances, it can be shown that the FHA is the best execution option for higher-risk loans – e.g., loans with lower FICO scores and higher loan-to-value (LTV) ratios. The GSEs, on the other hand, are typically the best option for lower-risk loans. As seen during the crisis, when the FHA’s insurance fund capital ratio fell below the Congressionally-mandated limit of 2 percent (which is itself far too low a threshold for the inherent risk of FHA loans), this dichotomy creates an inherent adverse selection problem for FHA and taxpayers.
Over the years, the role of the FHA has blurred. One of the nagging issues for FHA is there are no income limits for their loans. In theory, a higher income borrower in a high-cost city could go the FHA route for a mortgage, which is not exactly consistent with the original purpose of the FHA. Given the explicit guarantee by the federal government over FHA loans, and the critical importance of ensuring fair access to credit for certain homeowners, the FHA’s mission should be to be the backstop of affordable housing finance to low-income borrowers.
One way to eliminate the market overlap between FHA and GSE loans, and to clarify and strengthen FHA’s posture in the market (ending the adverse selection problem), would be to impose median income-based limits that demarcate FHA from GSE-eligible business. At the same time, we should eliminate all of the GSEs’ affordable housing goals.
In recent years, nearly a quarter of all single-family mortgages securitized by the GSEs were classified as low-income from an affordable housing goals perspective. Congressional mandates on affordable housing imposed on Freddie Mac and Fannie Mae were ostensibly a concession for the GSE charter. These requirements further blur already murky roles in providing mortgage financing between government and the private sector, and ultimately introduce unnecessary risks to housing finance over the long-term.
The FHA, meanwhile, should be allowed to set premiums that fully allow it to attain the level of risk management capabilities consistent with an agency overseeing an insurance fund over $1 trillion. Moreover, they should be permitted to engage in credit risk transfer (CRT) arrangements, similar to the GSEs. These initiatives would significantly protect taxpayers in the long-run.
Recapitalizing and releasing the GSEs from conservatorship is the right solution to the GSE policy inertia that has gripped the secondary market since the crisis. After all, the GSE model worked well for three quarters of a century.
However, more sustainable reforms would significantly reduce risk to the housing finance system and to taxpayers by taking this plan a bit further. Creating a common framework for delivering guarantee fees, products and service from the GSEs to the market should be a key part of an expanded plan.
Imposing median income-based limits demarcating FHA from GSE-eligible loans, eliminating affordable housing goals for the GSEs and building the lender mortgage process risk into GSE pricing are among the other housing finance reforms that could bear fruit.