I will argue that the structure of Fannie and Freddie, and the interaction of such with other elements of our financial system particularly the Basel Capital Accords, resulted in a dangerous leveraged system of mortgage finance that would have resulted in systemic failure with even modest levels of credit risk.
Before we contemplate the future, let us review key elements of the past.
As it relates to mandated capital, that is a financial company’s cushion of equity that protects both creditors and determines insolvency, Fannie and Freddie had the very unusual situation of having their minimum capital determined by statute rather than regulation, as is the case with federal depositories.
Under the 1992 Act, the GSEs were required to maintain 250 basis points of capital for on-balance sheet risks and 45 basis points for their off-balance sheet guarantee business.
Amendments made to the 1992 Act in 2008 under HERA later gave the regulator the flexibility to establish temporary capital increases by order and permanent increases by regulation. As the Housing and Economic Recovery Act of 2008 was passed only months before the conservatorship, new minimum capital standards have not been promulgated.
I will also note as a matter of legislative history, the ability for the regulator to increase minimum capital was an obstacle to achieving GSE regulatory reform in 2004 and 2005.
Lacking a consensus to increase GSE capital reform efforts failed until 2008.
Also worth noting is that the 1992 Act established a system of risk-based capital standards for the GSEs. Despite the relatively low levels of required minimum capital, the ultimate risk-based standards were even lower and never a binding constraint.
To put this into perspective, let’s start with some basic numbers. Year end 2007, Fannie Mae’s core capital was just over $45 billion and Freddie Mac’s was just under $38 billion. So between then about $83 billion in core capital.
As a percent of assets and outstanding guarantees, Fannie Mae’s core capital ratio was 1.51 percent, and Freddie Mac’s was 1.74. Another way of saying this is that each company was leveraged approximately 60 to 1.
Given that the foreclosure rates among prime loans were in excess of 3 percent from 2009 to 2012, the failure’s of Fannie and Freddie would have resulted even if they hadn’t made any subprime loans, which of course they did.
Given that I see some small chance that Fannie and Freddie come out of this crisis intact, I believe that Federal Housing Finance Agency would eventually issue minimum capital standards considerably in excess of those currently in regulation.
So one point is that both Fannie and Freddie were guaranteed to fail due to the extremely high levels of leverage of which Congress believed they should operate.
But that’s only part of the picture. It would be bad enough if Fannie and Freddie simply displayed such leverage on their own, but they also helped to drive increased leverage in the banking system.
Recall that under the Basel bank capital rules, the minimum required capital is a function of the risk weight applied to various categories of assets.
If a bank were to hold a whole mortgage on its balance sheet, the risk weight would be 50%. So if the minimum is weighted capital of 8%, actual capital equals that 8% times the 50%, yielding an actual capital of 4%.
If however the bank were to hold a GSE MBS the corresponding risk weight is 20%, yielding an actual capital of 1.6%, which is obviously a lot less than 4%.
Given this difference it wasn’t usual for a bank to sell say a thousand mortgages to Fannie Mae and purchase from Fannie the MBS holding those thousand mortgages. The bank would re-acquire the interest rate risk, but level the credit risk with Fannie Mae.
As an aside this is an important point to emphasis, in general the interest rate risk, which is really what’s special in the 30 year fixed rate mortgage, flowed to the holder of the MBS and did not remain with the GSEs.
Let’s step back for a moment and take a system-wide approach. Again we take the example of a bank selling mortgages to Fannie and buying back the MBS holding those mortgages. The bank now holds 1.6 percent capital against that risk and Fannie holds 45 bp.
Combined the financial system is holding 2.5 percent capital against the risk. This amounts to an almost 40 percent reduction in system-wide capital behind mortgage due to the interaction of the GSEs and the Basel Accords.
To put this in perspective, my rough calculations are that if the GSEs had not existed and banks had held the same amount of whole loans on their books, we would have had over $200 billion more in capital in our financial system.
This effect also shows up in the declining trend of bank risk-weighted assets to total assets, witnessed from 2000 to 2006, which might have made the system look safer but masked an increasing amount of leverage.
Not surprisingly these perverse incentives also drove large bank holdings of GSE securities. When the financial crisis hit, depositories were holding an amount of GSE MBS and debt equal to 150 percent of their Tier 1 capital levels.
The inter-connectedness of Fannie and Freddie with the rest of our financial system was not an accident but one of deliberate design.
Before the implementation of Basel 1, the market share of securitized mortgages began to level of around 30%. After Basel I implementation we witnessed a reversal in that trend which eventually doubled.
One of the lessons here is that however Fannie and Freddie are reformed or whatever replaces them, we must give consideration to how the changes interact with the rest of our financial regulatory system.
As one justification for the existence of the GSEs is the expansion of homeownership, it is worth nothing that the national homeownership rate leveled off around 64 percent by 1969, at which time the market share of the GSEs was in low single digits. So to be clear, the expansion of Fannie and Freddie has delivered essentially nothing in the way of expanded homeownership.
Nor have they narrowed the homeownership gap between white and African-American households. At the height of the bubble in 2007, homeownership rate for whites was 76.5 percent, while that for African-Americans was 54 percent, leaving a gap of 22.5 percent.
In 1910, before the creation of Federal Housing Administration, Fannie Mae, or Freddie Mac, that gap was 23.5 percent. In more than one hundred years, the difference in white and African-American homeownership rates has decline a whole 1 percent. I must note that the gap had narrowed to 18.8 percent by 1980—before the massive growth of our agency-driven secondary mortgage market.
As we are thinking about the future of Fannie and Freddie, were they to survive, there is at least one important accounting change that could dramatically impact their leverage.
Despite the fact that their guarantee business kept them on the hook for credit losses, the GSEs have historically been allowed to consider these guarantees as off-balance sheet obligations, allowing a much longer capital charge.
While in conservatorship, the Federal Accounting Standards Board made significant changes to the treatment of off-balance sheet interests. While this change was not done with the GSEs in mind, it did force the GSEs to consolidate their guarantee business.
As their capital requirements have been suspended, the impact has only been on appearances. Were the GSEs to leave conservatorship or have their capital requirements re-instated, then I believe there is a very good chance that they guarantee business will see its capital requirements increase from 45 basis points to 250 basis points.
If we go back to our system-wide capital comparison, the new accounting treatment could result in a combined bank-GSE capital of 410 basis points, which is slightly above the 400 basis points required for the holding of whole loans.
So while I believe some degree of GSE market dominance resulted from their greater leverage, it is difficult to quantify how much.
What is certain is that the previous model would lose some competitive advantage by higher capital charges.
As legal curiosity, I want to quickly raise another issue related to GSE capital. Section 1303 of the 1992 Act places an interesting limitation on what constitutes core capital.
Let me quickly read this brief limitation: “the core capital of an enterprise shall not include any amounts that the enterprise could be required to pay, at the option of investors, to retire capital instruments.”
Essentially capital must be perpetual to count as “core” capital. Why does this matter?
Well Section 1117(C) of the Housing and Economic Recovery Act of 2008, which provides the authority for Treasury’s preferred share injections lists several restrictions, including “temporary authority”, “limits on maturity”, “need for preferences or priorities”. “repayment”. So its not clear to me how one can really reconcile requirements of the Housing and Economic Recovery Act of 2008 that any support be temporary and repaid with the 1992 Act’s requirement that core capital be perpetual.
This leads me to conclude that Fannie and Freddie both lack any core capital and under the requirements of the Housing and Economic Recovery Act of 2008 must be placed into receivership.
Interestingly enough, the Office of Management and Budget seems to agree with me on the first part, having in its budget releases since conservatorship, clearly stated it considers Treasury’s preferred shares to be temporary in nature. So somebody here is trying to pull a fast one.
That said, I have no expectation that Federal Housing Finance Agency will actually follow the law here, so the threat of a receivership is essentially zero, even if in my view that is the route we should take.
I’ve talked exclusively about capital on the part of financial institutions. The first trip-wire, if you will, is the borrower. Generally borrowers will not default if they have considerable equity in the home.
After the 2001 increase in GSE housing goals, we witnessed a steady increase in the loan-to-value of first time buyers. Both GSEs also first issued zero and low downpayment products around this time.
This is also borne out looking at the financial disclosures of the GSEs. Between 2004 and 2007, the percent of their books comprising loans with LTVs in excess of 90 percent doubled, from around 8% to 16%.
One can certainly debate the role of the GSEs in leading the market in low-downpayment lending, as Federal Housing Administration was doing low down long before the GSEs, but they clearly helped to contribute to increased leverage on the part of borrowers.
So while I believe there were several factors behind the crisis, the increased leverage behind our mortgage market, both on the part of borrowers and financial institutions were important contributors.
The GSEs played a role on both accounts.
Ultimately we would have a more stable mortgage finance system were both borrowers and financial institutions to have more equity behind those mortgages. Given recent actions by the regulators on the QRM rule as it relates to downpayments, I believe the politics push too strongly against downpayments to expect our future mortgage system to be one where borrowers have greater equity.
If we eventually set Fannie and Freddie free, there are several reasons to believe they will, at least at first, be required to reduce their leverage. Of course this could erode over time and politicians and the public forget that housing prices do occasionally decline.
If there is no future for Fannie and Freddie, then the details really matter. A deposit based bank balance sheet model, while still highly leveraged would imply less leverage. A particular concern I have is that the adoption of a government run insurance fund would likely increase leverage, even relative to Fannie and Freddie. Thank you and look forward to comments and questions.