The high-risk business model of universal banks helped bring about the financial crisis. Universal banks rely on cheap funding from deposits and shadow banking liabilities to finance their speculative activities in the capital markets. By combining deposit-taking and short-term borrowing activities with underwriting, market making, and trading in securities and derivatives, universal banks create a strong probability that problems that occur in one sector of the financial industry will spread to others.

In order to prevent such contagion, federal regulators have powerful incentives to bail out universal banks and protect their depositors and shadow-banking creditors. That is precisely what occurred during the financial crisis. To prevent future bailouts, Title II of the Dodd-Frank Act requires the Federal Deposit Insurance Corp. to liquidate a failed systemically important financial institution and to impose losses on shareholders and creditors.

This liquidation-only mandate threatens the TBTF subsidy and presents a direct challenge to the universal banking model. To meet that challenge, Wall Street has developed and promoted the "single point of entry" plan as its preferred strategy for resolving failed megabanks. This plan would guarantee future bailouts while imposing the costs on ordinary investors and taxpayers. Unfortunately, while U.S. and foreign regulators have not yet formally adopted this plan, it appears very likely that they will do so.

Under the single point of entry plan, only the parent holding company of a failed SIFI would be placed in receivership. Its operating subsidiaries would be kept open for business. Short-term creditors of the holding company and all creditors of the operating subsidiaries would be fully protected. Protected creditors would include uninsured depositors and shadow-banking creditors with close connections to Wall Street, such as holders of commercial paper and securities repurchase agreements.

The single point of entry plan relies on a two-part funding model to protect Wall Street creditors. First, each SIFI's holding company would be required to issue long-term "bail-in" bonds. The FDIC would convert bail-in bonds into equity when a SIFI fails, thereby imposing losses on bail-in bondholders as well as shareholders.

To avoid the risk of contagion, SIFIs would not sell bail-in bonds to other large financial institutions. Instead, megabanks would sell bail-in bonds to non-systemic investors, including retail mutual funds and pension funds. Bail-in bonds, which pay somewhat higher yields than conventional senior debt, are tempting investments in an ultra-low-rate environment. However, retail mutual funds and pension funds do not have the ability to evaluate the complex and opaque risks inherent in those bonds. And ordinary investors would be the primary losers when bail-in bonds are converted into equity.

If bail-in bonds do not provide enough funding to recapitalize a failed megabank and its operating subsidiaries, the FDIC would borrow the rest from the Treasury Department through the Orderly Liquidation Fund. Because the fund currently has a zero balance, OLF loans would be backstopped by taxpayers. Under Wall Street's plan, OLF loans would ensure full protection for short-term creditors of the failed bank's holding company and all creditors of the operating subsidiaries.

At the end of the resolution process, a new SIFI would emerge with very few structural changes. Thus, contrary to Title II's liquidation-only mandate, the single point of entry plan would reorganize failed SIFIs and guarantee bailouts for Wall Street creditors. Society — in the form of taxpayers as well as ordinary investors who bought bail-in bonds — would be left holding the bag once again.

Given the strong likelihood that regulators will accept the single point of entry strategy, policymakers must adopt three major reforms to reduce the TBTF subsidy inherent in it.

First, regulators should prohibit megabanks from selling bail-in bonds to ordinary individuals, retail mutual funds or pension funds unless those bonds are expressly designated as subordinated debt that is junior to all general creditor claims. This prohibition would prevent SIFIs from misleading ordinary investors by selling them bail-in bonds that are described as "senior" to subordinated debt (and therefore pay lower interest rates) but actually contain complex and high-risk "triggers" for conversion into equity. Bail-in bonds that purport to be "senior" to subordinated debt should be sold only to accredited investors.

Second, Congress should require SIFIs to pay risk-based premiums over a period of several years to pre-fund the OLF at a level of $300 billion or more. This requirement is justified by the fact that federal officials provided almost $300 billion of capital through the Troubled Asset Relief Program to the 19 largest banks as well as insurer AIG during the financial crisis. The pre-funded OLF should be used to cover the costs of resolving failed SIFIs after the FDIC has written off investments by shareholders, subordinated debt holders, and accredited holders of bail-in bonds.

Forcing megabanks to pre-fund their own bailout money would protect taxpayers from bearing the costs of resolving failed SIFIs. In addition, a well-designed, risk-based schedule for OLF premiums would encourage SIFIs to follow more prudent operating strategies and adopt less complex business structures.

As part of their OLF premiums, big banks should pay fees on their uninsured deposits and shadow banking liabilities, since both would be fully protected under the single point of entry strategy. The required fees, which should be similar to risk-based deposit insurance premiums, would encourage SIFIs to establish more stable, longer-term funding structures.

Third, federal agencies should adopt long-delayed incentive compensation rules under Section 956 of Dodd-Frank. Those rules should require big banks to pay at least half of their total compensation to senior executives and other key employees, including risk managers and traders, in the form of contingent convertible bonds (CoCos). These employees should be compelled to hold their CoCos, without any hedging, for several years after their employment ends.

CoCos would expose insiders to immediate losses — without any need for clawbacks — if their banks fail during their employment or the post-employment holding period. They would therefore encourage key insiders to adopt sustainable, long-term business strategies that are more closely aligned with the interests of long-term creditors, the FDIC, and taxpayers.

These reforms would not eliminate the TBTF subsidy. However, they would compel megabanks to internalize at least some of the systemic risks they create. They would also support other initiatives by regulators — including capital, liquidity and living will requirements — that are designed to persuade SIFIs to reduce their size and complexity.

Arthur E. Wilmarth Jr. is a professor of law at George Washington University. This essay is based on a his forthcoming article in the Georgia Law Review.