Statement No. 171 (revised)

For Information Contact:

Charles W. Calomiris
(212) 854-8748
Peter J. Wallison
(202) 862-5864

Last year, Congressman Richard Baker, Chair of the House Subcommittee on Capital Markets, introduced legislation to reform and improve the regulation of Fannie Mae and Freddie Mac. That legislation was analyzed in detail by the Shadow Committee in Statement No. 164 (September 25, 2000). The initial Baker bill was a modest effort in relation to the problem it sought to address, but nevertheless drew only slight support in Congress. This year, Chairman Baker has introduced a new bill (H.R. 1409), but it too has significant shortcomings.

The two key provisions of Mr. Baker's new bill are its designation of the Federal Reserve Board as the proposed regulator for Fannie and Freddie, and the authority it grants the Board of Governors of the Federal Reserve System to approve all new activities by these organizations. In the Shadow Committee's view, the Federal Reserve Board is not an appropriate regulator for Fannie and Freddie. As an agency that values its independence, the Fed may not take on the difficult political fights that will be necessary to restrain Fannie and Freddie's growth. It would be better, instead, to turn the regulation of Fannie and Freddie over to the Treasury Department, which not only has expertise in the regulation of financial institutions (through its involvement in the Office of the Comptroller of the Currency and the Office of the Thrift Supervision), but also bears the principal risk of financial difficulties at Fannie and Freddie.

In addition, the authority that Chairman Baker would grant to the Fed or any other regulator provides inadequate guidance. The charters of Fannie and Freddie are very general, and have allowed them to expand their activities beyond the secondary mortgage markets. This represents a derogation of the original intentions of Congress. It does little good for Chairman Baker's bill to provide authority for a regulatory agency to approve new activities by Fannie and Freddie without defining what the scope of their activities should be. The Baker bill should amend these GSE charters to make clear that they must devote their attentions solely to the secondary mortgage markets.

Despite its lack of initial support in Congress, the Baker bill did stimulate a reaction from Fannie and Freddie. In response to Congressman Baker's initiatives and to growing public concerns about the risks the GSEs pose for taxpayers, Fannie Mae and Freddie Mac offered to issue a small amount of subordinated debt in the marketplace, suggesting that this would foster more market discipline and increase their effective capital (that is, decreasing the proportion of their funds raised by senior, implicitly protected debt issues). The initial issuances of this debt have now taken place, and whether they create market discipline or significantly increase Fannie and Freddie's capital can now be evaluated.

Contrary to the claims of the housing GSEs, the new subordinated debt offerings do not cultivate market discipline or constitute an effective increase in their capital. In order to be effective in creating market discipline, a subordinated debt must be credibly unprotected and of sufficient size. The subordinated debt issued by the GSEs under the current arrangement fails on both counts.

First, there is no reason to believe that this subordinated debt will be any less protected than senior debt. The small yield differentials between subordinated and senior GSE debt observed thus far probably reflect primarily liquidity differences in secondary market trading of the offerings rather than a difference in their perceived risk of default. Indeed, the Federal Financial Institutions Examination Council (which represents the various bank regulatory authorities) gave these instruments a 20% risk weighting for purposes of bank capital regulation (the same rating given to senior GSE debt), making clear that it does not regard them as bearing any significant risk of default.

Second, the amount of subordinated debt that Fannie and Freddie offered to issue is inadequate. This year, the GSEs have issued $3 billion in five- and ten-year subordinated debt, which amounts to less than 0.5% of on-balance sheet assets. They also offered to build up and maintain a sum of equity and subordinated debt equal to at least 4% of on-balance sheet assets. This amounts to a new subordinated debt layer of financing that when achieved will be paper thin relative to on-balance sheet assets (roughly 2%), and even thinner when measured against the sum of on- and off-balance sheet asset risks. Furthermore, there is no guarantee that even this small amount would remain outstanding if the GSEs decided to withdraw it; the new debt issues are voluntary, and the GSEs could decide to repurchase them or fail to roll them forward at maturity.

The Shadow Committee continues to advocate full privatization of the housing GSEs, as described in Statement No. 164, as the best solution to the problems posed by the mixture of private profitability and public protection enjoyed by the housing GSEs. Moreover, we continue to advocate numerous other regulatory reforms for the GSEs, and limits on the expansion of their powers, as also described in Statement No. 164. Nevertheless, as an interim measure, we believe that imposing a minimum subordinated debt requirement on the GSEs, if it were properly designed, could substantially discipline these entities and reduce the contingent liabilities borne by taxpayers because of the potential for a costly bailout of the housing GSEs. But, to make a subordinated debt requirement effective demands careful consideration of the way that subordinated debt might, and might not, impose effective discipline on issuing institutions.

In Statement No. 160 (March 2, 2000), the Shadow Committee described in detail how an effective subordinated debt requirement could be constructed as part of a capital requirement for large commercial banks. Although some aspects of that proposal are relevant in designing an effective subordinated debt requirement for the GSEs, important differences in the regulation of banks and GSEs lead to differences in the details of effective rules governing subordinated debt issues. Banks, unlike the GSEs, pay insurance fees on their insured liabilities, and in principle those fees should vary with the risk of loss to the insurance fund. Also, banks are subject to prompt corrective action, as provided in FDICIA. Thus, as we envisioned in Statement No. 160, subordinated debt issues by banks not only provide direct market discipline (through the debt service cost on uninsured debt), but indirect discipline, since the information provided in the subordinated debt market can be used by regulators either in the pricing of deposit insurance or to trigger regulatory interventions under prompt corrective action. Those regulatory actions magnify the effects of the direct discipline imposed by the interest cost on subordinated debt.

In contrast, the GSEs do not pay insurance fees for their protection, and they are not subject to clearly specified guidelines for prompt corrective action. Thus, under current rules, any discipline imposed on the GSEs from subordinated debt offerings is almost entirely direct--that is, it results from the incentive for prudent management of risk that derives from higher debt service costs on subordinated debt.

That difference in the role of subordinated debt in the GSE and banking contexts implies that the amount of subordinated debt required of GSEs should be substantially larger than that envisioned in our bank subordinated debt proposal. The Committee believes that, to ensure that direct market discipline provides an effective deterrent to imprudent risk taking by the GSEs, the minimum ratio of subordinated debt should be set at 10% of total assets (defined as the sum of assets held and guaranteed).

Is there a cost imposed on the GSEs by this requirement? To the extent that the market will judge their subordinated debts as risky, the GSEs will be forced to pay higher interest and higher underwriting fees to place subordinated debt. But the only cost imposed on them would result from asking them to bear most of the costs of the risks they create rather than pass those costs on to taxpayers (who currently provide free implicit protection).

To be effective, however, required subordinated debt must be credibly subject to loss. If loans from the Treasury or other injections of government funds are expected to insulate subordinated debtholders from loss, then subordinated debt cannot provide any ex ante market discipline. Thus, any effective subordinated debt requirement must ensure that government funds are not used to bail out subordinated debt holders. This could be achieved by a simple rule that requires that all subordinated debt payments must be suspended when an institution has drawn on its credit line from the government or has impaired capital, and that requires that any permanent fund transfers to these entities be matched by dollar-for-dollar write downs of the face value of subordinated debt. Under the current rules, the GSEs must restrict subordinated debt payments in very limited circumstances. Debt interest payments are suspended only if the GSE's capital is impaired and it has formally asked to draw on its line from the Treasury. If capital falls to less than 125% of the critical capital level (1.75% of on balance sheet assets and off balance sheet obligations) interest payments must be suspended whether or not the GSE has used its Treasury line. Even then, principal and accrued interest will be paid at the maturity date of the debt. This is scant protection against the implicit bailing out of subordinated debt holders by the government.[1]

Effective subordinated debt must also be of sufficiently long maturity; only outstanding debts of greater than one year of remaining maturity should count toward the 10% subordinated debt requirement. To further enhance the information signaled to the prudential regulator by the market, it would also be useful to require regular offerings of subordinated debt (say, at least one offering per quarter). And limits must be placed on Fannie and Freddie's ability to make markets in these instruments or to write derivative contracts with debtholders that mirror the risks of these debts (as discussed in Shadow Statement No. 160).

Together credible limits on the scope of the GSEs' activities, the vesting of enhanced regulatory authority over them in the U.S. Treasury, and the creation of a credible minimum subordinated debt requirement would substantially strengthen market and regulatory discipline and limit the risks posed to taxpayers.

Notes

[1] This paragraph was revised from the initially released statement.