Andrew Davidson: Four Steps Forward: Streamline, Share Risk, Wrap, and Mutualize

 April 19, 2016

September will mark eight years since Treasury Secretary Paulson announced the conservatorship of Fannie Mae and Freddie Mac and a brief “‘time out,’ where we have stabilized the GSEs (government-sponsored enterprises) while we decide their future role and structure.” During those eight years, there have been many proposals for GSE reform, but no resolution.

Much of this delay, I believe, is because of misinformation about the role the GSEs played in the financial crisis and flawed ideas about the economics of mortgage-backed securities (MBS). In particular, critics of the GSEs have suggested that the GSEs and their affordable housing goals were the primary cause of the financial crisis and that the mortgage market could function without a government guarantee.

The first claim, that the GSEs caused the financial crisis, has been widely debunked but continues to weigh heavily on GSE reform. Our analysis clearly shows that the problems in the mortgage market arose primarily from lax underwriting and high loan-to-value lending in the non-GSE mortgage market that was fueled by ratings arbitrage in the collateralized debt obligation market. For example, increases in home prices and subsequent declines were largely correlated with states that had a significant amount of option arms and high loan-to-value subprime loans. States without those products experienced much smaller home price bubbles and declines. While the GSEs contributed to these trends and had other severe flaws, they were the tail, not the dog. In fact, a substantial portion of the GSE losses can be traced to reduced-documentation high-loan-balance loans that did not qualify for the housing goals.

Because of this false narrative, congressional and administration reform proposals have focused on “winding down” the GSEs. This is difficult to do because the GSEs are essential to our housing finance system. Because the GSEs cannot be eliminated quickly, these proposals tend to have long timelines for implementing alternatives and phasing out the GSEs. Furthermore, proposals based on gradually eliminating the GSEs cannot assure that the proposed alternative arrangements can be established and achieve the necessary functional scale.

Because of the second claim, that a government guarantee is unnecessary, Congress considered proposals to fully eliminate the government guarantee and sought alternative methods to maintain the liquidity and functioning of the mortgage market and the TBA (to-be-announced) securities market. However, only a government guarantee fully separates the credit risk of mortgages from the interest-rate risk and prepayment risk of MBS. This allows an extremely liquid secondary market and an active forward market to function. Proposals to replicate this liquidity and functionality without a government guarantee are wishful thinking.

While the GSEs were not the fundamental cause of the housing finance crisis in 2007 and the more severe financial crisis in 2008, they contributed to the meltdown, and there were serious flaws in the structure of the GSEs that allowed private gain at public expense. The GSEs placed shareholder gain over risk management and were severely undercapitalized. GSE reform is needed to address the structural inadequacies of the pre-2007 system.

To eliminate the risk of the GSEs exercising monopolistic power and being too big to fail, several proposals (including the Johnson-Crapo reform bill) suggested creating a system of competitive guarantors. Unfortunately, it is unlikely that such a system would work. While many guarantors are needed to create competition, there are strong arguments for having fewer guarantors to promote standardization. If there are many guarantors, each guarantor would have slightly different guidelines for sellers and servicers which would lead to a race to the bottom as the players compete for market share or would require a government entity to establish rules. Such government-set rules would likely be static and could be gamed by unscrupulous originators. (The history of seller-financed down payments is a good lesson in the efficacy of the government setting the rules for loan programs.) Even if there is not a race to the bottom, it is very likely that a dominant player will emerge. The dominant guarantor would likely charge monopoly profits that would increase costs to borrowers and threaten the profitability of originators and servicers.

Instead of shutting down the GSEs, eliminating the guarantee, or creating new entities, a realistic approach to GSE reform would be to strip the GSEs down to the functions that are essential to promote standardization, liquidity, and access to credit, and adopt the best governance structures for those functions. This can be achieved in four steps: streamline, share risk, wrap, and mutualize. While some of these actions could be taken by the Federal Housing Finance Agency (FHFA) and the administration, the second, third, and fourth steps would require congressional action to fully implement.

I. Streamline

The GSEs should only perform functions related to their guarantee and securitization businesses. Their retained portfolio business should be substantially reduced or eliminated. This is already in progress. (Without having GSEs hold MBS, there could be greater volatility in pricing mortgages. This impact is currently mitigated by the large holdings of MBS by the Federal Reserve.)


  • Reduce the retained MBS portfolios
  • Allow GSE to maintain portfolios for cash window and repurchases of seriously delinquent loans
  • Establish a plan to sell off nonperforming assets
    • Securitize and sell reperforming assets in senior and subordinated transactions
    • Recognize that such sales could produce losses
  • Establish a common securitization platform or other shared functions, if it can be cost effective

II. Share Risk

The GSEs have demonstrated over the last two years that they can “reinsure” a significant portion of their credit risk into the market through several risk-sharing approaches, reducing taxpayer risk and the concentration of credit risk in the economy. For example, our analysis shows that Freddie Mac has shed approximately two-thirds of the risk on its target loans in its 2014 book of business. To continue this process, it is important to allow for the expansion of the investor base for credit risk transfer (CRT) products.


  • Establish reinsurance and credit risk–sharing programs for up to 75 percent of risk on new GSE loans
  • Develop risk retention and up-front risk sharing for originators on a pooled or specific basis
  • Establish capital rules that encourage the use of risk sharing but adequately address counterparty risk
  • Change Internal Revenue Service and Securities and Exchange Commission rules to establish risk-sharing transactions as real estate mortgage investment conduits and good assets for real estate investment trusts
  • Address bank capital rules that reduce liquidity in secondary markets by requiring excessive amounts of capital to hold CRT bonds
  • Address Commodity Futures Trading Commission rules that limit the use of credit-linked note structures

III. Wrap

The GSE MBS should receive an explicit government wrap like Government National Mortgage Association securities and should be closely regulated. In exchange for the government guarantee, the GSEs would pay a fee to the government that would defray the potential cost of the guarantee and fund affordable housing or other national housing goals.


  • Establish a Government National Mortgage Association program to wrap GSE MBS or new government wrap
  • Charge an explicit fee and housing affordability fee
  • Hold the GSEs liable for losses up to a cap by cohort or vintage
  • Government will establish eligible loan criteria, capital, and risk-sharing requirements to access wrap; Treasury will oversee this
  • Establish a fund to cushion losses
  • Maintain GSE access to agency market to maintain cash window and repurchases of delinquent loans from MBS pools

IV. Mutualize

Mutual ownership of the GSEs by mortgage originators offers the opportunity to align incentives and protect taxpayers from future losses while limiting the incentives to siphon the benefits of government guarantees to private shareholders. (If we need an acronym, we could call these NMMs, or national mortgage mutuals. The two entities could be named the Federal National Mortgage Mutual, or FNMM, and the Federal Home Loan Mortgage Mutual, or FHLMM.)

The mutuals would also implement the national duty-to-serve responsibilities of the secondary market. The regulator of the mutual can track the availability of credit to borrowers throughout the country and can require the mutual to have its members serve those communities. The mutual can decide what incentives would be necessary to achieve that goal, rather than having a government entity create incentives for private companies.

If the mutual is unable to raise capital in a stressed environment, Treasury could take on a portion of the risk-sharing role at a high (but not stifling) required return. This would ensure that the mutual would seek private capital first and that the government would earn an adequate return if and when it supports the market.


  • Create a mutual structure, owned by originators, with activity-based funding
  • Retain FHFA as regulator of the mutuals and the Home Loan Banks
  • Require profits to be paid to members as dividends
  • Require mutual to implement national duty-to-serve obligations, subject to regulatory oversight
  • Allow Treasury  to provide a backstop to risk sharing for new vintages (the target return on investment could be the 10-year Treasury rate plus 20 percent on a blended basis across the CRT stack)

There are three other possible governance structures for the streamlined entities: competitive private stock corporations, regulated utilities, and government corporations. Stock corporations are supported by those who favor “recap and release,” but this was the pre-crisis structure that led to private gain at public expense. The reason for using mutuals rather than stock companies is that the mortgage-guarantee business is likely a natural monopoly. This arises from the need for standardization and because of investor preferences for a large liquid issuer. A single security and government standards may create competition, but there is no assurance that they will succeed. The mutual structure accommodates a market structure of very few guarantors. A variant on this is to regulate the released entities as utilities. While this would be better than hoping for a competitive equilibrium, a privately owned utility would constantly seek to expand its footprint, reduce its capital, and increase its allowed return on equity. Government corporations would maintain the status quo, whereby taxpayers bear the risk of loss and mortgage availability would be subject to political winds swinging from excessively tight to excessively lax.

The transformation of Fannie and Freddie into issuer-owned mutual insurance companies also provides a clear and relatively simple transition plan. Table 1 compares the governance options.

While congressional action is needed to complete the transformation of the housing finance system, there is much that can be done under conservatorship. A primary goal of FHFA, Treasury, and the administration should be to establish and publish risk and capital guidelines for the GSEs. These guidelines should address

  •  the amount of risk in the loan portfolio;
  • the value of future guarantee-fee income;
  • reductions in risk because of risk sharing;
  • counterparty risk, asset haircuts (e.g., mortgage-insurance, deferred taxes); and
  • target returns, including cross-subsidization guidelines.

These capital guidelines would allow more consistent oversight of the GSEs and would provide legislators with a clearer picture of the risk remaining in the entities. They could also help evaluate the probability of exhausting the Preferred Stock Purchase Agreement credit line.

FHFA should also promote the creation and expansion of loan-originator risk-sharing programs. These programs could reflect originator-specific risk or pooled risk from several originators. Such programs could become the basis for originator activity–based capital requirements and impose the necessary skin-in-the-game discipline on the origination process. To some extent, the GSEs already have some of this via their claim on servicing rights and through representations and warranties. Originator risk sharing would not be a net cost to the originators, as it would provide originators another source of income or offset guarantee fees.

The congressional actions required to implement the plan do not need to be combined into a single GSE reform bill (although that would be nice) but could be implemented sequentially with the ability to adjust the plan as events unfold.

First, Congress would need to facilitate the expansion of the risk-sharing investor base. There are several needs along this line, including changing Internal Revenue Service and Securities and Exchange Commission rules to facilitate CRT investment by mortgage real estate investment trusts. Commodity Futures Trading Commission rules also limit the use of credit-linked notes, which could further expand investor participation and protect taxpayer interests. Congress also needs to encourage and enable regulators to address bank capital rules that impose excessive capital requirements on CRT investment. Bonds created by the GSEs under the regulatory oversight of the FHFA should not be treated like the subprime and collateralized debt obligation investments that contributed to the crisis.

Second, Congress would need to make the implied guarantee of the GSEs explicit and establish a wrap fee and housing affordability fee structure associated with access to that guarantee. Treasury or another regulator such as FHFA or Government National Mortgage Association would establish the loan eligibility and first-loss capital requirements to access the guarantee. Such a guarantee should lower mortgage rates without increasing the risk to taxpayers because the government has demonstrated that it would stand behind the GSE MBS. Such legislation would also allow the restructuring and eventual elimination of the Preferred Stock Purchase Agreements.

Third, Congress would need to authorize the establishment (or reestablishment, in the case of Freddie Mac) of a mutual structure for the GSEs. This legislation would need to include guidelines and authority for FHFA to oversee these institutions and enforce duty-to-serve requirements.

GSE reform may continue to languish until more of the public and Congress examine Fannie Mae and Freddie Mac objectively and practically rather than through philosophical and political prisms. An objective analysis would reveal that an effective housing finance system can be created from the existing flawed system with just a few steps that rely on economic structures with known features and manageable risks.

Andrew Davidson is a financial innovator and leader in developing financial research and analytics. He has worked extensively on mortgage-backed securities product development, valuation, and hedging. He is the founder and president of Andrew Davidson & Co. Inc., a firm specializing in the application of analytical tools to investment management.

He was instrumental in creating the Freddie Mac and Fannie Mae risk-sharing transactions STACR and CAS. These transactions allow Freddie and Fannie to attract private capital to bear credit risk, even as they remain in government conservatorship. Davidson is also active in other dimensions of GSE reform and has testified before the Senate Banking Committee several times. He also helped establish the Structured Finance Industry Group and served on the executive committee at its inception.

Davidson is coauthor of several books including, most recently, Mortgage Valuation Models: Embedded Options, Risk and Uncertainty. He received a BA in mathematics and physics from Harvard University and an MBA in finance from the University of Chicago.