Fixing What Works

Serious proposals for housing finance reform must have clearly defined objectives against which they can be evaluated, and must be derived from a clear-eyed analysis of the causes of the previous crisis so that by addressing and fixing those causes we minimize the chance of a similar crisis happening again.

My proposed reform objective is the following: to create a capital markets-based secondary market mechanism capable of financing at least $1 trillion of 30-year fixed-rate mortgages annually throughout the business cycle, at the lowest cost to homebuyers consistent with an agreed-upon standard of taxpayer protection. The emphasis on homebuyer cost is deliberate. Low-, moderate-, and medium-income homebuyers suffered the most during the 2008 crisis, and received no significant relief from the government. These same families have seen little growth in their incomes during the recovery, so it should be a policy priority to provide them with the greatest possible access to mortgage credit at the lowest possible cost.

Fixing the correct problem is the second essential element of mortgage reform. Immediately after the crisis Fannie Mae and Freddie Mac were singled out as its primary cause. Based on indisputable data, however, we now know that this was not true. To the contrary, Fannie and Freddie were the most disciplined sources of mortgage finance in the years leading up to the crisis. During the summer of 2008 the serious delinquency rate on the single-family loans owned or guaranteed by the companies was about one-third the serious delinquency rate of other prime lenders, and less than one-tenth that of subprime lenders. The subsequent performance of Fannie’s and Freddie’s loans was equally superior: the loss rates on their single-family loans from 2008 to 2015 averaged less than 50 basis points per year—about one-third the average loss rate on comparable mortgages held by banks and less than one-fifth the loss rates on loans financed with private-label securities.

We also now understand why Fannie and Freddie had to take $187 billion in senior preferred stock from the Treasury Department. It was not because of operating losses. Through 2011, the companies’ business revenues—net interest income, guaranty fees, and other income—exceeded their combined credit losses and administrative costs. Their draws of senior preferred stock were made necessary by $151 billion in noncash expenses (plus $36 billion in dividend payments) booked by their conservator, the Federal Housing Finance Agency (FHFA), based on highly pessimistic estimates of future losses. The large majority of those losses did not materialize, and as a consequence Fannie and Freddie had enough income to pay Treasury $158 billion—more than all of the $151 billion in non-cash expenses taken earlier—in just 18 months, beginning in the fourth quarter of 2012. The companies never needed the $187 billion “bailout” they received from the government.

When invented narrative is replaced with verifiable fact, Fannie and Freddie cease to be a “failed business model” that must be wound down and replaced; they instead become valuable resources that must be built upon and improved. My proposal—outlined below—does that. It makes fundamental changes to Fannie and Freddie in three key areas: relationship with the government, capital, and regulation. It also preserves the companies’ ability to support affordable housing, and can be implemented administratively.

Relationship with the government. Experts generally agree that the role of the government in the charters of Fannie and Freddie is too ambiguous, and that the balance of benefits tilts too far in the direction of the companies. Moving to a “utility model,” with limited returns and a more focused business purpose, addresses both issues. In the model I propose, Fannie and Freddie would remain shareholder-owned but would agree to accept (1) a cap on the average return they could target in their guaranty fee pricing (I suggest 10 percent after-tax), (2) restrictions on the size and use of their portfolios (limited to 10 percent of outstanding credit guarantees, and to purposes ancillary to the guaranty business), and (3) standards for minimum and risk-based capital determined by administration policymakers with percentages set and imposed by FHFA. In an “exchange for consideration,” the government would commit to provide temporary support to the companies should their capital ever prove insufficient (which by design would be highly unlikely).

Homeowners and the government each would benefit from this arrangement. The government backstop would produce the lowest possible yield on the companies’ mortgage-backed securities, benefiting homeowners, while the government would limit its risk—and control moral hazard—thorough rigorous capital standards, close regulation and supervision of Fannie and Freddie, and caps on their returns.

There are many advantages to the government’s supporting utility-like companies rather than the companies’ securities. Individual pools of securitized mortgages have limited diversification, and can experience much higher loss rates than the companies that issue them. Even in normal times guarantees on securities will require the government to make unrecoverable payments to investors. Worse, if the government guarantees only securities but not the issuing companies, in a crisis there may be no surviving entities to issue new securities and keep the system from collapsing. Having the government stand behind companies keeps the system intact, and allows the government to recover any outlays after the crisis has passed.

Capital. With strict limits on their portfolios, Fannie and Freddie will be taking one type of risk (credit) on one high-quality asset (residential mortgages) in one country and one currency. Proposals for Fannie and Freddie to adopt the Basel III bank capital standards therefore contradict the principle that capital must be related to risk. Large multinational banks can take many types of risks on many types of assets (including very risky ones) in countries and currencies around the world. Giving Fannie and Freddie bank-like capital requirements without bank-like asset powers would doom them to failure.

Fortunately, there is a proven way to set capital standards for a company that deals in a single, homogenous asset type: require that company to hold enough capital to withstand a defined, worst-case stress scenario. In my proposal, administration policymakers would pick that scenario. I recommend that they require Fannie and Freddie to hold sufficient capital to survive a 25 percent nationwide decline in home prices over five years. Even though such a price decline did happen between 2006 and 2011, both major factors that precipitated it—very risky mortgage types like no-documentation loans or interest-only ARMs with teaser rates now prohibited by the Consumer Financial Protection Bureau (CFPB) and the dominance of a financing method, private-label securitization, that placed few limits on the risks of the mortgages it accepted—will be absent in the future, making the chance of a repeat of the previous episode vanishingly small.

Fannie’s prior experience suggests how it would have to capitalize against a future 25 percent home price decline. With the loans it had in 2008, and using its guaranty fees (but none of its portfolio or other income) to help absorb credit losses, Fannie would have needed less than 2 percent capital to survive the previous crisis. And if we remove from the data the loan types no longer permitted by CFPB regulations—which accounted for roughly half the company’s post-crisis credit losses—it could have survived with only about 50 basis points of capital.

If FHFA confirms these results, it should set Fannie and Freddie’s minimum capital ratio at 2 percent, then specify a supplemental risk-based standard that imposes capital requirements by product type and risk category (defined at a minimum by paired combinations of loan-to-value ratios and credit scores). FHFA would grade the companies’ business as it comes in, and require them to hold the greater of the risk-based or minimum capital amounts. All of the companies’ capital would have to be retained earnings or common or preferred stock.

Fannie and Freddie could make their minimum standard binding by holding down the risk of the mix of business they acquire. With 2 percent capital and a 10 percent target return, Fannie and Freddie’s average charged single-family guaranty fee would be about 40 basis points, which—after the 4.2 basis point affordable housing fee and the 10 basis point payroll tax fee (through October 2021)—would be a little over 50 basis points to the borrower. At this level the companies could use cross-subsidization effectively to attract a broad range of business, including affordable housing loans. Should the risk mix of the companies’ business rise, their risk-based standard would cause their capital and average guaranty fees to rise as well.

Regulation. After a stress standard for the companies has been chosen, FHFA will need to analyze Fannie and Freddie’s credit performance during the prior housing market collapse to determine the percentage of minimum capital—for the types and characteristics of loans the companies are permitted to acquire today—that would allow them to comfortably withstand that stress. FHFA would use that same data to determine the stress capital percentages by product type and risk category used to calculate required risk-based capital.

Once the minimum standard and the risk-based requirements are in place, Fannie and Freddie would be permitted to price their business as they saw fit—including using cross-subsidization—as long as their guaranty fees in the aggregate were consistent with no more than a 10 percent return on capital. FHFA would monitor the companies’ pricing, and if it found their average fees to be too high it could take whatever remedial action it deemed appropriate. FHFA would track each company’s business and calculate its required risk-based capital on a quarterly basis, with adjustments as warranted for any risk-sharing transactions they do.

 Affordable housing. Fannie and Freddie’s role in supporting affordable housing is limited by the fact that they only can purchase or guarantee the loans lenders originate. Despite this, FHFA should set affordable housing goals for the companies. FHFA also should have the power to impose penalties for failing to meet those goals, but only if the percentage of affordable business Fannie or Freddie does falls short of the percentage originated by lenders that year.

FHFA should not increase the amount it requires Fannie and Freddie to contribute to affording housing funds beyond the 4.2 basis points mandated by legislation. Fees for affordable housing imposed only on the companies are an excise tax on the secondary market. Should Congress wish to increase support for affordable housing through additional fees, it should levy them on all mortgages. This would raise more money—or raise the same amount at a lower fee rate—and not favor primary market over secondary market financing.

 Implementation. The above changes could be effectuated through administrative action, as were the 2008 Preferred Stock Purchase Agreement and its amendments. With the written consent of the boards of directors of Fannie and Freddie, FHFA as conservator would make binding commitments on behalf of the companies, and Treasury and FHFA would make binding commitments on behalf of the government.

Before these reforms could take effect, the government would need to settle all of the lawsuits against it for its treatment of Fannie and Freddie before and during the conservatorships. It likely will take rulings adverse to the government’s current position to trigger that settlement. Assuming such rulings are forthcoming, Treasury should cancel the warrants it holds for 79.9 percent of the companies’ common stock, allow them to use proceeds from the reversal of the net worth sweep to repay their senior preferred stock, and retroactively replace the 10 percent dividend on that stock with a more reasonable 1 percent markup over the cost of the funds Treasury borrowed to give the companies the $187 billion they did not need.

Treasury is prohibited by the “Jumpstart GSE” legislation from liquidating Fannie and Freddie’s senior preferred stock before January 2018. Until then, FHFA should stop paying dividends on it, and notionally credit the companies with the amount of capital they will have when the stock is repaid, to assist them in planning for their recapitalization.