Skip to main content

How Much Capital Is Enough?

This post is the second in our Housing America Series, aimed at elevating the need for housing policy reforms—from housing finance to affordability—as we move through a pivotal federal election and transition.

If there is one principle uniting the at-times competing Democratic and Republican visions for housing finance reform, it is that the private sector should assume the vast majority of credit risk and be compensated for it. This compensation necessarily requires “pricing in” the cost to hold capital, which is needed to buffer losses in a severe economic downturn.

As the Federal Housing Finance Agency (FHFA) has demonstrated, of the three components of the GSEs’ guaranty fee, the capital cost portion is by far the largest.1 The other two components price in expected credit losses in normal economic times and general and administrative expenses. According to the FHFA, while the combined cost to the GSEs of the latter two components is 11 basis points, the cost of holding capital against unexpected credit losses ranges from 28 to 115 basis points, depending upon assumptions about the required level of capital and the after-tax return on capital.2

Because the cost of capital has real impacts on the price of mortgage credit and access to homeownership, one of the most important policy questions that lawmakers must answer is: “How much capital is enough?”

Planting a Stake in the Ground

A reasonable way to tackle the capital issue is determining what it would have taken for the GSEs to have survived the recent financial crisis. One analysis from the Urban Institute demonstrates that using the GSEs’ 2007 experience, a 4-to-5 percent capital buffer would have been sufficient to enable them to avoid conservatorship. In a 2013 paper, the Center for Responsible Lending (CRL) also found that Fannie Mae and Freddie Mac would have needed approximately 4 percent capital to withstand the 2008 financial crisis, a level more than eight times higher than their statutory minimum requirements. While Senators Bob Corker (R-TN) and Mark Warner (D-VA), and later Senators Tim Johnson (D-SD) and Mike Crapo (R-ID), started from the same data point in their housing finance reform legislation, they doubled down on taxpayer protection by setting minimum capital requirements at ten percent, which they recognized was “more than twice the amount Fannie and Freddie lost during the crisis.”

Independent analysts were quick to point out the extreme caution that a ten percent capital requirement represented, as well as the potential impact such a high requirement could have on the cost of a mortgage. A Goldman Sachs report noted, “More than 1 in 5 borrowers across the country would need to default at a severity of 50 percent in order for a first loss layer of ten percent to be penetrated.”3 They said this scenario was way beyond the GSEs’ actual loss experiences in the recent crisis, even factoring in the performance of non-traditional mortgages such as negative amortization and Alt-A loans.

The minimum ten percent top-line capital requirement quickly became non-negotiable, and attention shifted to the issue of how broadly capital should be defined. Drafters appropriately identified the need for some flexibility, recognizing that if capital were too narrowly defined, mortgage credit would be priced out of the reach of ordinary families. Drafters settled on a broad definition which included in the numerator not just common equity but also “instruments and contracts that will absorb losses before the Mortgage Insurance Fund [such as] reinsurance, letters of credit, and future guarantee fees to be earned by the guarantor after accounting for the stress scenario.”

The minimum ten percent top-line capital requirement quickly became non-negotiable, and attention shifted to the issue of how broadly capital should be defined.

Adopting a reasonable interpretation of how regulators might allow the ten percent capital requirement to be satisfied—through a combination of four percent common equity and six percent subordinated debt and future guaranty fees—Mark Zandi and Christian deRitis of Moody’s Analytics estimated the Johnson-Crapo legislation would raise average mortgage rates by about 40 basis points relative to the current system. This was consistent with the administration’s internal estimates.

Given the administration’s reluctance to engage directly on the pricing issue, a leaked memo from Freddie Mac contributed to the impression that housing finance reform would be bad for the middle class and affordable homeownership. According to the memo, if Johnson-Crapo’s ten percent capital requirement had to be met entirely by common equity (requiring a 16.25 percent after-tax return), the average cost of a mortgage could rise by as much as 200 basis points relative to the status quo. In short, while opening the door to other forms of capital besides high-cost common equity, Johnson-Crapo’s treatment of the composition of capital satisfied few, leading to criticism by Senate Banking Committee Republicans and Democrats alike.

Looking Ahead

In the 27 months since the Johnson-Crapo bill was marked up in committee, Congress has failed to unite around a common vision for housing finance reform. Yet, during this period, many observers have shared their own reform plans that contain a wide spectrum of ideas on capital requirements.

Alex Pollack, a senior fellow at the R Street Institute favors a form of recapitalization and re-privatization of Fannie Mae and Freddie Mac that would designate the GSEs as Systemically Important Financial Institutions (SIFIs) and impose on them the same 5 percent of total assets minimum capital requirements as applies to other SIFIs.

Tim Howard, a former CFO of Fannie Mae, favors maintaining the GSEs as shareholder-owned institutions but treating them more like utilities by capping their average annual return at ten percent after-tax. He points out that loans no longer allowed by the Consumer Financial Protection Bureau accounted for roughly half of Fannie’s post-crisis credit losses and, if these loans had been removed from Fannie’s 2008 book of business, the company could have survived the crisis with only about 50 basis points of capital.

In the 27 months since the Johnson-Crapo bill was marked up in committee, Congress has failed to unite around a common vision for housing finance reform.

Another recent paper proposes to merge Fannie Mae and Freddie Mac and move their core infrastructure into a government corporation. To protect taxpayers against losses, this plan, whose authors include experts affiliated with the Clinton campaign, would require a total of 8.5 percent of loss-absorbing capital to be wiped out before the federal backstop would kick in. This capital buffer would consist of three layers: The GSEs would be required to syndicate 3.5 percent of first-loss credit risk into the capital markets. Should losses exceed that threshold, there would be 2.5 percent of fixed dividend securities (e.g., preferred stock) to absorb these losses. A mortgage insurance fund with 2.5 percent of capital built up over time would serve as the final buffer before taxpayers would be tapped.

Finally, while it is still being fleshed out, the Milken Institute’s multi-issuer Ginnie Mae-centered proposal would convert the GSEs into guarantors mutually owned by their respective seller-servicers. Co-authors Michael Bright and Ed DeMarco would require that issuers of mortgage-backed securities obtain 500 basis points of first-loss credit protection in order to qualify for a full faith and credit Ginnie Mae wrap. In addition, the two guarantors and other credit enhancers would have to hold 2-3 percent of capital to meet their loss-absorbing obligations. Should a guarantor fail, losses would be paid for by a mortgage insurance fund before the taxpayers would have to step in.


At the end of the day, in determining what the capital requirements in a reformed system should be, lawmakers must strive to find the sweet spot between protecting taxpayers against another government bailout and providing broad and affordable access to mortgage credit. CRL’s Mike Calhoun and Sarah Wolff also remind us that most pricing analyses of reform proposals speak to changes in average costs relative to the current system, but that “average price estimates obscure significant variation” among borrowers. Greater amounts of capital and higher expected rates of return on that capital will cause borrowers with lower credit scores and higher loan-to-value mortgages to pay sometimes significantly higher rates.

In short, because capital is not a free good, a belt and suspenders approach to ensuring against any future government losses by layering capital requirement upon capital requirement or limiting first-loss risk protection to high-cost common equity can pose serious access and affordability challenges. Such an approach may be tantamount to creating a fully privatized secondary market system and the higher costs and reduced access that come with it. That’s why it is critical to conduct more deeply empirical work around capital requirements before lawmakers begin the next round of legislative debates on housing finance reform.

1 The calculation excludes the temporary 10 basis point fee imposed in 2011 to pay for the extension of the payroll tax reduction.

2 The cost of holding 200 bps of capital at a 9 percent after-tax return is 28 bps, while the cost of holding 500 bps of capital demanding a 15 percent after tax return is 115 bps.

3 Goldman Sachs, The Mortgage Analyst, Credit Strategy Research, Assessing proposed capital requirements for mortgage guarantees, March 20, 2014.

Read Next

Support Research Like This

With your support, BPC can continue to fund important research like this by combining the best ideas from both parties to promote health, security, and opportunity for all Americans.

Give Now