Remarks as Prepared for Delivery
Edward J. DeMarco
Acting Director
Federal Housing Finance Agency
Federal Reserve Bank of Chicago
49th Annual Conference on Bank Structure and Competition
Chicago, Illinois
May 9, 2013

Thank you for inviting me to speak here today. The Bank Structure Conference is renowned for tackling important issues related to our Nation’s financial system and regulatory infrastructure. This year’s topic, the regulation of systemically important financial institutions is clearly one of great interest and importance.

For years, concerns were raised about the potential for Fannie Mae and Freddie Mac to pose a systemic risk to our financial system. They were large companies, interconnected throughout our financial system, subject to limited market discipline, integral to the operation of the mortgage market, and subject to a bifurcated regulatory regime that had inadequate powers.

In July 2008, legislation was enacted to improve regulatory oversight and address some of these existing concerns. However, by then we were well into the housing crisis and unfortunately a short time later those concerns proved correct. In September 2008, both companies were placed into conservatorships and the Treasury Department provided financial support to avoid the systemic impact that their failure would have created.

Today, I am going to focus on the fundamental choices that policymakers face in considering housing finance reform. I hope we have learned from our past experiences with Fannie Mae and Freddie Mac, and that we can develop a housing finance system that has appropriate oversight, relies on private capital, and significantly reduces the systemic risk that was inherent in the old model.

Before turning to my subject, let me say a word about Fannie Mae and Freddie Mac today. From their just-released first quarter results, it is clear they are each beginning to show regular, strong profitability. This, of course, is a good thing. I want to acknowledge the hard work of the management teams and staff at both companies in this turnaround. At the same time, I want to acknowledge the importance of substantial and sustained government support. I also note that the companies are making substantial dividend payments to Treasury, but the structure of the Treasury support means the $187 billion drawn from Treasury is not reduced by these payments.

Role of the Government in Housing Finance

At the most fundamental level, the key question in housing finance reform is what, and how large, should be the role of the federal government?

Let me start with a brief review of some basic economics that many of you are familiar with, but at times gets lost in the policy debate on housing. A typical way of considering the government’s role in the marketplace is in the context of a potential market failure. Broadly speaking, in housing finance there are at least two potential market failures that are often considered; each may lead to an under-provision of mortgage credit.

A potential market failure could arise in housing finance if market participants have undue or unnecessary concerns about the ongoing stability and liquidity of mortgage credit in a purely private market across various economic environments. If this view prevails in the housing market, less credit will be provided than would be the case in the absence of this type of uncertainty. The government response to this type of potential market failure could take a number of approaches, ranging from establishing standards and greater transparency for the market; to providing liquidity or credit support under certain market conditions; to providing a government guarantee that largely eliminates uncertainty.

Another potential market failure is what is often thought of as the positive externality associated with homeownership. In this view, the benefits of homeownership extend beyond the individual household to the broader aspects of society, hence if left solely to the market the number of homeowners will be less than optimal. A common government approach to increase market demand is to provide some type of subsidy or other assistance to encourage or facilitate such consumption. Direct subsidies to assist with down payments, lower the cost of mortgage credit, or ease the eligibility terms for a mortgage are methods of delivering subsidies through the housing finance market. Other government policies beyond the housing finance market also are used. Prominent among these is the mortgage interest tax deduction.

The country’s housing finance system evolved over the years to address both of these issues. I could spend much time going over that history, but for simplicity let me make two points.

First, FHA, VA, and other direct government credit guarantee programs have historically targeted certain borrowers that may lack access to credit or where the public policy goal was to subsidize financing to encourage homeownership. Ginnie Mae aids this process by providing a securitization vehicle to broaden access to capital markets for certain government credit programs.

Second, the Government Sponsored Enterprises (GSEs or Enterprises) Fannie Mae and Freddie Mac were established to promote liquidity and provide stability in housing finance on a national basis. They accomplished this task by linking mortgage originators to capital markets. They were established as relatively unique entities that were chartered by Congress, owned by private shareholders, provided a specific set of benefits not available to other companies, and given a public mission.

The general notion of GSE status resulted in an important benefit, the ability to fund their operations at rates lower than other private sector companies. Some amount of this benefit was passed on to borrowers in terms of lower rates, but over the years questions were raised in terms of how much of that benefit remained with management and shareholders. Over time, other requirements were added to their mission, through mandates like the affordable housing goals, which were intended to recapture some of those benefits for the public.

While there does not seem to be broad agreement in Washington on many issues, there does seem to be relatively broad agreement that the Fannie Mae and Freddie Mac model of the past posed systemic risk and ultimately failed. That model relied on investors providing funding for housing at preferential rates based on a perception of government support. That perception ultimately turned out to be correct and has resulted in Fannie Mae and Freddie Mac drawing $187.5 billion in funds from Treasury as of December 31, 2012.

Today’s Single-Family Housing Finance Market

With the collapse of the private securitization market in 2007, and its limited recovery to date, we are left with a single-family market dominated by government support. At the end of 2012, there was $10 trillion in single-family mortgage debt outstanding. About 13 percent was guaranteed through direct government programs, 52 percent was owned or guaranteed by Fannie Mae and Freddie Mac, and the remainder not guaranteed by the federal government. The non-government share largely reflects pre-crisis private-label securities.

On a flow basis, the numbers are much different. Measured by securities issuance, the proportion supported by the government is over 90 percent.

I think it is clear that the government’s role in housing finance will not recede to zero. FHA and other traditional government credit programs are typically what are used to address credit market failures or to achieve public policy goals, and I would expect that to continue. Whether the targeting, structure, and other aspects of these programs should be reconsidered will be an important part of the housing finance reform debate. Deciding what part of that market should be served by the traditional government credit programs is essential to framing the debate for the government’s role in the rest of the market.

For the remainder of the market, I perceive a consensus among policymakers that we need to get private capital back into the market. In this context, most people are referring to private capital absorbing the credit risk that is supported by the federal government through the financial support currently provided to Fannie Mae and Freddie Mac.

There are two broad ways that private capital could be employed to take on credit risk. One is through issuers of mortgage-backed securities (MBS) or other financial institutions guaranteeing the credit on those securities by maintaining appropriate levels to capital. I will refer to this as an “issuer-based approach.” Another is for securities to be issued or structured in such a way that market participants themselves advance the capital to absorb credit risk, or what I will refer to as a “securities-based approach.” Both approaches could be used in the near-term to attract private capital. However, in considering structural reform of our housing finance system, each approach has benefits and questions that need to be fully considered.

Issuer-Based Approach

An issuer-based approach is generally associated with a financial institution guaranteeing principal and interest repayment to investors. In this model, the issuer’s guarantee is backed by its shareholders’ capital. While not necessarily part of an issuer-based approach, typically this approach assumes a further credit enhancement in the form of a government guarantee on the securities issued.

In general, proposals to establish an issuer-based future structure with a government guarantee have some common elements. These proposals would rely on federally chartered or approved entities that would have the ability to issue explicitly guaranteed securities. The entities would be regulated for capital adequacy and overall safety and soundness, and there would be a payment of an insurance fee to the federal government for the use of the explicit guarantee. These types of proposals make clear that the government guarantee would only apply to securities issued, and not to the equity of actual issuers. And they often assume private mortgage insurance would be required at the loan level for low down payment mortgages.

While that sounds like a novel concept, it is similar to the Fannie Mae and Freddie Mac model. Pre-conservatorship, the Enterprises were said to benefit from an implicit guarantee. There was a safety and soundness regulatory structure, but it did not have a full range of powers. In the end, the securities, both debt and MBS, issued by the Enterprises were provided financial support by Treasury, while the shareholders have not benefited from that support.

Would replacing Fannie Mae and Freddie Mac’s implicit guarantee with an explicit one, improving the regulatory structure, and charging for the government guarantee resolve all the shortcomings and inherent conflicts in the old Enterprise model, or would it produce its own problems?

Clearly if the securities offered in a reformed housing finance market have a government guarantee, those securities will be priced favorably and have a high degree of liquidity to reflect that guarantee. This would ensure a reliable source of funds to the housing market.

However, this approach would not provide market-based pricing of credit risk of the underlying mortgages. In these structures, much like the banking system and deposit insurance, private sector capital through equity investment would stand in a first loss position, with a government guarantee that was funded through an insurance premium being available to cover other losses.

This type of structure requires a significant amount of regulatory safety and soundness oversight to protect against the moral hazard associated with providing a government guarantee. It also relies heavily on federal regulators rather than private investors to measure risk and set the capital needed to absorb losses across a wide array of possible economic environments. While such an outcome has certain merit and some attractive features, a number of issues need further exploration.

Pricing the government’s risk exposure. Some have recommended setting an actuarially fair price for the government guarantee. The concept of actuarially fair comes from the insurance market where risks such as life expectancy can be reasonably calculated. Does that concept apply to risk in the mortgage market where the likelihood of catastrophe cannot be calculated with statistical confidence? If that leads to an underpricing of risk by the government, would future losses be passed on to market participants in an ex post structure much like deposit insurance, possibly at a time when mortgage institutions could least afford it?

Concentrated, undiversified exposure to residential mortgage credit risk. MBS issuers in such a structure would have concentrated and undiversified exposure to residential mortgage credit risk. Is that an appropriate way to structure a key component of our financial system, and would it pose systemic risks to our financial system?

Principal-agent problem between stockholders and taxpayers. As with the GSE model, competing interests will always be an issue when public benefits are provided to a limited number of private entities. Can regulatory oversight effectively mitigate this problem?

Government allocation of credit within the housing sector. If the government provides explicit credit support for the vast majority of mortgages in this country, it would likely want a say with regard to the allocation or pricing of mortgage credit for particular groups or geographic areas. If the government wants to subsidize homeownership, is this type of structure the appropriate tool?

Over-allocation of credit to housing. Explicit credit support for all but a small portion of mortgages, on top of the existing tax deductibility of mortgage interest, would further direct our nation’s investment dollars toward housing. It would also drive up the price of housing, other things being equal. Is this an appropriate outcome for the overall economy?

Viability of the business model if risk pricing and the regulatory structure is set at appropriate levels. One of the reasons that the GSE model seemed to work from a private shareholder perspective was in addition to the funding advantage, there also were significantly lower capital requirements relative to banks. If safety and soundness oversight is structured appropriately and the regulatory capital requirements are calibrated appropriately and evenly across regulated financial institutions, will there be enough incentive for private shareholders to establish such a company?

An issuer-based approach with a government guarantee is one option for the future of housing finance. While it is somewhat similar to the past, we have experience setting up government insurance structures and establishing regulatory regimes to address some of these challenges. We also have experienced the failures of those regimes in the past. We have learned from those experiences and improvements can be made. But I think before we go in this direction we should fully consider another approach.

Securities-Based Approach

The other approach to consider is what I will refer to as a securities-based approach. Like an issuer-based approach, this concept should be familiar. In a securities-based approach, as opposed to credit risk being absorbed by the equity of the securities issuer, credit risk would be absorbed through capital markets.

To a large degree, this is what the old private-label MBS market did. We know that even though the old private-label MBS market functioned in a way to distribute credit risk, in the end it was not durable and led to a misallocation of credit risk in its own right. But was the problem with the concept or the lack of infrastructure, standards, and rules to govern the market? In my own view it was not the concept, but the lack of an appropriate infrastructure.

Given what we learned from the housing crisis, what will it take to get investors back to absorbing and pricing credit risk? Considering the issue from the point of view of an investor, there would seem to be a number of key areas in the securitization process where it is important to have certainty or the ability to enforce rights if we expect private capital to price credit risk. In short these would include:

• Standard data definitions – if a more liquid market is going to develop, investors are going to want standard data terminology to be able to price across different risk characteristics.

• Loan quality assurances – investors are going to want to have greater certainty that the loans being delivered into a security actually meet the defined characteristics, and if they do not, what are the remedies and who ensures those remedies are enforced.

• Servicing standards and protocols – investors are going to want to know the rules of servicing, how default servicing will be handled, and what their rights are to address poor performance.

• Data disclosure – investors are going to need ongoing access to a considerable amount of data on individual mortgages to accurately price credit risk.

To a large degree, what I am describing is a standard-setting approach that would replace some of the standard-setting that the Enterprises undertake today. Replacing the Enterprises would be a regulatory regime or a market utility that sets those standards. This model need not rely on a government guarantee to attract funding to the mortgage market, but rather would look to standardization and rules for enforcing contracts.

What the securities-based approach would be establishing is a market infrastructure for the pricing and capitalizing of mortgage credit risk. If such an approach were established, it would allow for the broader disposition of credit risk across capital market investors, as opposed to the equity investors in the issuer-based approach.

A securities-based approach would allow for institutions that specialize in mortgage credit risk to continue playing a role to the extent that investors see this option as a valuable risk transfer tool. For example, a reformed mortgage insurance industry, with enhanced capital and clear contract terms, could attract an even broader more stable capital base to this approach.

Like the issuer-based approach, the securities-based approach also has issues that need to be considered.

More volatility in the price of mortgage credit than with a government guarantee. Without a government guarantee there would be more volatility for certain classes of mortgages. For example if economic conditions deteriorated in certain areas of the country, or if borrowers had lower down payments or other risk characteristics, as market conditions deteriorated, credit in those sectors would tighten. But this type of volatility is a form of market discipline as the market is signaling higher potential risk. In the end, how much volatility are we willing to accept?

Amount of capital available to absorb the credit risk currently undertaken by the Enterprises. In today’s mortgage market, investment structures have developed around the existence of government support. Would those investment structures change if government support was not broadly available and a transparent market structure existed? In other words, if current investors in Fannie Mae and Freddie Mac MBS stop investing in MBS, what asset class would they invest in and would other investors be attracted to take their place?

A non-government guaranteed secondary market might not be available for all borrowers. To establish a liquid non-government guaranteed market there would seem to be a need to have greater homogeneity in borrower characteristics in a given MBS pool. I think such a market would broadly cover the bulk of the business that the Enterprises undertake today, but such a market might not be available to all borrowers currently served by the Enterprises. With greater transparency in requirements, it would give borrowers a clear sense of the qualification requirements. Traditional government guarantee programs would still exist to meet various policy goals.

For borrower characteristics that do not fit neatly into the secondary market, we need to find a way to get insured depository institutions back into the business of funding mortgages. Understanding individual borrowers and special circumstances is at the heart of the financial intermediation function of banks, thrifts, and credit unions. This issue deserves further exploration. I would add that the Federal Home Loan Banks provide banks, thrifts, and credit unions access to funding across the maturity spectrum to assist in financing mortgages on balance sheet. Overall, would this type of market structure meet the housing policy needs of our Nation?

Finally, there still may be ongoing concerns either with how a securities-based approach might operate in a time of stress or whether enough capital could be attracted to the market even in more normal times. In a time of stress with a standardized, transparent market and clarity of rights, it would be possible for an existing guarantor, like Ginnie Mae, to play a temporary role. And such a market structure should make the provision of liquidity by the Federal Reserve or other liquidity providers easier to evaluate. Such an approach would be akin to the more limited role of government in the market proposed as one of the options in the Administration’s white paper on housing finance reform a few years back.

If the amount of capital in normal times remains a concern, a government guarantee could be attached at the appropriate point, with the private sector taking a first loss exposure through the security structure. While this would broaden the base of capital and provide a more stable funding source, the use of a government guarantee would raise some of the same questions noted under the issuer-based approach: appropriate pricing and attachment point for the government guarantee; use of the guarantee to allocate credit in the housing market; and over allocation of resources to housing.

FHFA’s Strategic Plan

One of FHFA’s goals in our Strategic Plan for the Operation of the Enterprise Conservatorships is to build a new infrastructure for the secondary mortgage market. A key motivation behind this goal is that the Enterprises’ existing proprietary infrastructures are not effective at adapting to market changes, issuing securities that attract private capital, aggregating data, or lowering barriers to market entry. The ultimate goal is to develop a new securitization model that will have benefits beyond the current Enterprise business model and be able to accommodate any choice that policymakers make for housing finance reform, whether an issuer-based, securities-based, or other approach.

Some of the individual components associated with building a new secondary market infrastructure include the following:

Developing a common securitization platform. The platform would eventually replace the back office functions of the Enterprises’ current securitization infrastructure. We have outlined an approach where the focus of the platform could be on functions that are routinely repeated across the secondary mortgage market, such as issuing securities, providing disclosures, paying investors, and disseminating data. These are all functions where standardization could have clear benefits to market participants.

Considering elements of a model contractual framework. Similar to the securitization platform, the focus of this effort is to identify areas where greater standardization in the contractual framework would be valuable to the mortgage market of the future. A great deal of work has already been done in this area by market participants. As the Enterprises move forward with risk-sharing transactions such as those I will describe shortly, the development of transactional documents will provide a real time test of a new standardized contractual framework for transactions where the private sector is absorbing credit risk.

Data Standards. A solid foundation of data standards is important regardless of the future direction of housing finance reform. The Enterprises—working through an industry procedure established by the Mortgage Industry Standards Maintenance Organization (MISMO)—have moved this process forward through the Uniform Loan Delivery Dataset, Uniform Appraisal Dataset, and work that is beginning on the Uniform Mortgage Servicing Dataset.

Servicing and Other Standards. FHFA led efforts to develop servicing standards through the Servicing Alignment Initiative, and continues to look to other areas where standards could be useful, such as our recently announced efforts on force-placed insurance.

Finally, we have taken steps to bring private capital back to the mortgage market. Since conservatorship, Enterprise guarantee fees have steadily increased and now are around 50 basis points, about double what they were before conservatorship. A key motivation behind increasing Enterprise guarantee fees is to bring their credit risk pricing closer to what would be required by private sector providers.

In terms of risk sharing, we have set a target of $30 billion of unpaid principal balance in credit risk-sharing transactions in 2013 for both Fannie Mae and Freddie Mac. We have specified that each Enterprise must conduct multiple types of risk-sharing transactions to meet this target. For example, we expect to see transactions involving: expanded mortgage insurance; credit-linked securities; senior/subordinated securities; and perhaps others structures. The goal for 2013 is to move forward with these transactions and to evaluate the pricing and the potential for further execution in scale. What we learn in 2013 will set the stage for the targets for 2014, and I fully expect to move from a dollar target to a percentage of business target at some point in the future. These transactions may also be seen as pilot tests of the securities-based approach that I just described.

Conclusion

Lawmakers in Washington agree on the need to draw private capital back into the mortgage market but no consensus on how to do so has formed.

I have offered here two broad approaches designed to draw private capital into the market to replace Fannie Mae and Freddie Mac. In each approach, private investors would be compensated for pricing and bearing mortgage credit risk ahead of a government guarantee, if such a guarantee exists at all.

One approach, the issuer-based approach, utilizes a set of specially chartered financial institutions to pool capital from the company’s shareholders. This capital would provide the basis for a guarantee of principal and interest to MBS holders. The corporate guarantee could be further backstopped by a taxpayer guarantee, for which investors would pay a government-set fee.

The second approach, the securities-based approach, relies on standardization and transparency to attract capital. Under this approach the structure of the security itself is employed to pool capital to absorb losses. The simplest structure is a senior-subordinated security structure where the investors in the subordinate piece absorb credit losses ahead of the senior security holders. A sufficiently large subordinate security should insulate the senior investors from credit losses in all, or almost all, credit outcomes.

In either approach, the point is to pool capital sufficient to bear mortgage credit risk while efficiently pricing such risk.

The approaches differ from each other in the institutional and regulatory frameworks required and in the potential systemic risk that may result. In either approach, an ultimate government backstop could be deployed, although the securities-based approach has more flexibility in how that might be done.

I hope my framing of this important policy question spurs more in-depth analysis of these options. I welcome your feedback and, more importantly, encourage your participation in producing analytics that will help inform policymakers of the choices facing them.

Thank you again for inviting me here today.