The Administration has made clear that it intends to address housing finance reform as soon as financial regulatory reform is concluded. To many, this means the Administration will release its long-awaited plan for the future of Fannie Mae and Freddie Mac, the two housing government-sponsored enterprises (GSEs) whose combined losses could result in a $1 trillion bailout. Reform would be incomplete, however, were it not to include reform of the “other” housing GSE, the Federal Home Loan Bank (FHLB) system.
The FHLBs are a system of independently operated banking cooperatives headquartered in 12 regions that correspond roughly to those of the Federal Reserve System. Rather than being governmental in nature, like the Fed banks, the FHLBs are “instrumentalities” of the federal government that are owned by the private depository institutions in their districts. When banks acquire common equity in their local FHLB, they become “members” which permits them to obtain access to low-cost funding in the form of “advances.” Advances are loans made by the FHLBs to member banks that provide funds used to extend mortgages to homebuyers. These advances are fully collateralized by qualifying bank assets, which protect the FHLB should the member bank default on its advance.
The FHLBs have never incurred a credit loss on an advance. This is quite a remarkable track record that many have used as evidence of the inherent conservatism of FHLB operations. In truth, it reflects the FHLBs’ overcollateralization practices. The FHLBs do not price loans for risk. Instead, they secure their advances through priority liens on member bank assets. While these liens are generally applied to the mortgages the advances are used to finance, the FHLBs can also place a blanket lien on “all or specific categories of a member’s assets.” Also, the FHLBs can over-collateralize advances so that a $7 million advance to a bank can be secured with $10 million of Treasury securities. The ability to substantially over-collateralize positions provides the FHLBs with a “super creditor” status that all but eliminates any possibility of loss. When a member bank fails – sometimes directly as the result of losses on mortgages funded through FHLB advances – the FHLBs enjoy a claim on any of the additional eligible collateral in the failed bank and the right to demand prepayment from the FDIC of outstanding advances.
For example, when IndyMac failed in 2008, it had $10 billion of outstanding advances from FHLBs and another $19 billion of the bank’s liabilities. When IndyMac’s assets proved insufficient to meet its outstanding obligations (i.e. the assets were worth far less than the combined $19 billion of deposits and advances), all of the losses were apportioned to the Deposit Insurance Fund (DIF) because of the FHLB’s super creditor status. Before the FDIC could even take control of an institution’s assets, it had to first pay-off the FHLB’s outstanding advances. The FDIC estimates that the FHLB’s super lien caused the DIF to incur $9 billion of losses on the resolution rather than the $4 billion it would have incurred had IndyMac relied instead on market-based funding sources. The “super lien” of the FHLB creates obvious moral hazard because the FHLB has no incentive to monitor the risk-taking of its member borrower. The FHLB didn’t care what kind of mortgage loans IndyMac was originating because their performance had no impact on whether the FHLB would get paid back. This moral hazard is particularly visible when looking at differences in FHLB system assets through time and on a cross-sectional basis. From the end of 2005 to the end of 2008, FHLB assets increased by 35% to $1.35 trillion, with advances growing by 50% during this time to $928.6 billion. From December 2008 to March 31, 2010, FHLB assets fell by 28% with advances declining by 38% to $572 billion. The FHLB provided low-cost funding to member banks to fuel the mortgage bonfire on the way up and has reduced its funding of new mortgages just as swiftly on the way down. This isn’t the fault of the FHLBs; it just shows that the system allows for unlimited amounts of low-cost funding when member banks want to make mortgages, while doing nothing to ensure consumers have reliable access to mortgage finance when member banks want to cut their mortgage exposure.
Secondly, the boom in FHLB assets and advances was not spread evenly across 12 regions. As you might expect, assets and advances climbed and later fell most dramatically at FHLBs where the housing bubble – and problems with non-traditional mortgages – was most acute. The San Francisco FHLB – serving California, Arizona, and Nevada – saw its total assets and advances increase by about 70% from 2004-2007 to $323 billion and $251 billion, respectively. As of March 31, 2010, the San Francisco FHLB’s balance sheet had shrunk by 46% to $173.8 billion with advances falling to $112 billion, a decline of more than 55% in 27 months. While no FHLB’s asset growth and contraction was as dramatic as San Francisco’s because no other FHLB’s footprint was exclusively bubble states, the Atlanta FHLB’s Florida exposure is also visible from the data. In the four years ending in December 2008, assets grew by 56% to $208.56 billion while advances grew by 73% to $165.9 billion. As of March 31, 2010, assets were only $146.3 billion (off 30% in 15 months) while advances had fallen by 36% to $105.47 billion (larger than the decline in the rest of the system net of San Francisco). Again, the data show that rather than ensure adequate mortgage finance is available across America, the FHLBs channeled capital to fund nontraditional mortgages in sand states during the bubble years and then reduced lending in these states at a faster than average rate as the bubble collapsed.
Beyond willingness to extend advances to banks without concern about loan quality, moral hazard is also introduced by the “joint and several liability” of FHLB obligations. To fund advances and other assets, the FHLBs issue “Federal Agency” debt – the same implicitly guaranteed, AAA-rated debt issued by Fannie Mae and Freddie Mac. However, these obligations are not issued by each FHLB individually. Instead, the centralized Office of Finance issues “consolidated obligations” (COs) on behalf of all 12 FHLBs. The rating agencies embrace COs because they reduce the probability of default by requiring well-managed FHLBs effectively to subsidize weaker siblings. More pointedly, rather than 12 institutions with $50 billion to $250 billion of assets, the COs create a single $1 trillion institution that is “too big to fail.” In total, about $146 billion of COs are issued each month (down from $273 billion in 2008) and the total outstanding stock of COs are $875 billion – about $100 billion more than Fannie Mae’s outstanding debt (page 55; Fannie also issues $2.4 trillion in guaranteed MBS).
The most obvious reform, therefore, is to close the Office of Finance and require that each FHLB fund itself independently. This would create 12 institutions where failure would be an option, which would likely end implicit credit market subsidies. The second reform would be to increase capital ratios at individual FHLBs as advances grow. This would force member banks to contribute more marginal capital to its FHLB to fund balance sheet growth. The additional capital would reduce the subsidy provided by the low-cost advances by requiring additional cash outlays from banks making greater use of advances. This should slow lending in “hot” housing markets and reduce the probability of a bubble. Finally, the FDIC should take additional steps to penalize the use of FHLB advances above a threshold amount. The cost of the FHLB system manifests itself through higher loss rates on the FDIC bank resolutions. Therefore, the banks whose resolutions are made more costly due to heavy reliance on FHLB advances – like IndyMac – should be required to pay higher insurance premiums. This would reduce the funding advantage of FHLB advances.
There is a temptation to focus housing finance reform exclusively on Fannie Mae and Freddie Mac, given their size and eye-popping losses. Lawmakers should not lose sight of the problems presented by the “other” GSE. During the bubble, banks like IndyMac could finance their investments in bad mortgages by relying exclusively on government sources of funding – deposits insured by the FDIC and low-cost advances provided by the FHLBs. When institutions can fund their investments exclusively through non-market channels, the system becomes prone towards crisis. Who knows how many bad mortgages would never have been funded if the provider of the capital had a financial incentive to see how its money was being used? If lawmakers are serious about preventing another crisis, they must reform the FHLB system.
Christopher Papagianis is the Managing Director of e21 and was Special Assistant for Domestic Policy to President George W. Bush.