Last year, and again this year, Rep. Richard Baker, R-La., introduced legislation to reform the regulation of Fannie Mae and Freddie Mac. In response to these initiatives and to growing public concern about the risk their potential insolvency would pose to taxpayers, both government-sponsored enterprises have offered to issue some subordinated debt, supposedly to foster market discipline and increase their effective capital.
In theory, higher insolvency risk would be punished by higher interest costs on the unprotected subordinated debt, thus creating an incentive to limit risk, and the added capital buffer would absorb losses that otherwise would be borne by taxpayers.
But the new debt offerings do not cultivate market discipline nor create a cushion for losses. To be effective, such debt must be credibly unprotected and of sufficient size. The subordinated debt issued by Fannie and Freddie fails on both counts.
Fannie and Freddie offered to build up and maintain a sum of equity and subordinated debt equal to at least 4% of on-balance-sheet assets. When completed within the promised three years, this would create a subordinated debt layer of roughly 2% of on-balance-sheet assets, which is even thinner when measured against the sum of on- and off-balance-sheet asset risks. Even this small amount could be withdrawn if it became costly to maintain, since there is no regulation requiring it.
Along with our colleagues on the Shadow Financial Regulatory Committee, we have advocated a 2% minimum subordinated debt requirement for large banks, and we view this as adequate to bring market discipline into the bank regulatory process. But for Fannie and Freddie the proportion would have to be much larger. In a recent committee statement (No. 171), we recommended a regulatory requirement on the order of 10% of the sum of on-balance-sheet and off-balance-sheet mortgage assets and obligations.
Why should the minimum amount be larger for Fannie and Freddie? For banks, the effect of subordinated debt is magnified by strong preexisting regulations. A high sub-debt servicing cost would produce a direct market-determined deterrent to bank risk taking. It would also signal bank regulators who could respond by raising deposit insurance fees and requiring additions to bank capital (using prompt-corrective-action powers). These actions would magnify the direct effect of market discipline.
But Fannie and Freddie do not pay insurance fees for their implicit protection and are not subject to prompt corrective action. Any discipline imposed on Fannie and Freddie through subordinated debt results only from higher debt service costs, implying that a much larger minimum requirement is needed.
But even that larger amount would be useless unless it were credibly subject to loss. But there is no reason to believe that subordinated debt would be any less protected than Fannie and Freddies senior debt. The yield differentials between subordinated and senior GSE debt are tiny and mainly reflect liquidity differences in secondary markets rather than a difference in default risk.
The Federal Financial Institutions Examination Council gave this subordinated debt a 20% risk weighting for purposes of bank capital regulation (the same rating given senior GSE debt), making clear that it does not regard the debt as bearing significant default risk.
This makes sense. Fannie and Freddies subordinated debt holders can avoid losses even if Fannie and Freddie become insolvent. Under the current rules, Fannie and Freddie can only temporarily restrict subordinated-debt interest payments, and only in limited circumstances.
Interest payments are suspended only if the GSEs capital is severely impaired or if its capital is somewhat impaired and it has formally asked to draw on its line from the U.S. Treasury. Thus, credit from the Treasury (especially if provided without a formal request) could be used to repay interest on subordinated debt. Furthermore, even if interest payments were temporarily suspended, all principal and accrued interest would be payable at the debts maturity date. That is scant protection against the implicit bailing out of debt holders.
The imitation subordinated debt that Fannie and Freddie have issued could be made real by requiring that all subordinated debt payments be suspended when the institution has drawn on its Treasury credit line or has impaired capital to any degree and by further requiring that any permanent government fund transfers to these entities be matched by dollar-for-dollar writedowns of the subordinated debts face value.
To forestall preemptive flight, subordinated debt should also be of long maturity. Only outstanding debt that has more than a year of remaining maturity should count toward the subordinated debt requirement. And to ensure a continuing flow of market information about Fannie and Freddies riskiness, regular offerings of such debt (at least quarterly) should be required. To ensure arms-length purchasing of subordinated debt, Fannie and Freddie should be prohibited from making markets in these instruments or writing derivative contracts with subordinated debt holders that mirror the risks of these debts.
Full privatization of Fannie and Freddie is the best solution to the problems posed by their mixing of private profitability and public protection. Regulation is a second-best alternative. Credible prudential regulation (preferably by the Treasury rather than the Fed, as proposed in the Baker bill), including mandatory subordinated debt that credibly bears risk, would go a long way toward controlling the risks these companies pose to taxpayers.
Mr. Calomiris and Mr. Wallison are co-directors of the Project on Financial Deregulation at the American Enterprise Institute, a think tank in Washington.