Ever since the great financial crisis, financial reformers have pushed banks to prevent future taxpayer bailouts by adopting “bail-in” mechanisms.

One increasingly popular strategy, encouraged by the Basel III framework for banking regulation, is for banks to issue “contingent convertible” bonds that convert to ordinary stock if the bank’s capital reserves get too low.

It’s a way of forcing the bondholders to shore up the bank’s capital by converting its debt into equity.

The problem is that these “Co-Cos” are all pegged to regulatory definitions of capital, and banks are masters at gaming that system. By regulatory standards, Citigroup and other giant institutions were all officially “well-capitalized” at the time they had to be bailed out. When Lehman Brothers declared bankruptcy, igniting a global meltdown, it had twice as much capital as it was required to have. (And as an investment bank, Lehman was actually using a more conservative valuation scheme than the one used by commercial banks.)

“Regulatory requirements get worse the longer they last,” says Jeremy Bulow, professor of economics at Stanford Graduate School of Business. “Market-based capital requirements aren’t perfect, but they improve over time. They have a self-correcting element.”

To stop the gaming and provide real protection, Bulow and Paul Klemperer of Oxford University have proposed a market-based bail-in, triggered by a drop in the bank’s share price.

A low share price means investors have lost confidence in the bank’s financial strength, and the market judgment will sometimes be more accurate than the regulatory capital valuation. Not requiring firms to conserve or add cash when the market price falls puts the taxpayer at risk if the market value is right and the regulatory value is wrong. During the crisis, regulators did not respond to market signals until things had deteriorated so badly that bailouts were the only option to save the banks.

Bulow and Klemperer propose that banks replace their traditional bonds with “equity recourse notes,” or ERNs. They first sketched out the idea in an article with Jacob Goldfield, a former senior partner at Goldman Sachs and former chief investment officer for George Soros’ Fund Management and Quantum funds.

An ERN would be like a contingent-convertible bond. But if the bank’s share price fell to a predefined trigger point, perhaps one quarter of its price at the time the bond was issued, then interest and principal payments would be made in shares, valued at the trigger price. If the share price later recovers, future payments would again be made in cash.

Bulow argues that ERNs have several big advantages over Co-Cos. The most important is that they convert automatically: The trigger isn’t based on either the judgment of financial regulators or on a bank’s regulatory capital. The conversion would be based on the bank’s stock price, which is not subject to the same kind of manipulation as regulatory capital.

A second advantage is that the ERNs convert more gradually than Co-Cos. Conversions would be relatively small, because they would occur only for the current payments due on ERNs, and only for those ERNs that had been issued when the share price was at least four times above the current level.

The advantage of automatic and gradual conversion, say Bulow and Klemperer, is that regulators’ path of least resistance will be to passively permit it.

By contrast, an important concern about Co-Cos is that they would all convert at once, which could become a panic-inducing event of its own. If regulators are the ones who have to decide the time has come, the all-or-nothing nature of that decision could make them reluctant to pull the trigger. And if the conversion is triggered by a bank’s regulatory capital, banks are likely to use every trick in the book to keep their official capital above the trigger point.

Do banks really manipulate regulatory capital, and is it that different from market measures?

Yes. Andrew Haldane, the Bank of England’s executive director for financial stability, offered a striking illustration last year. Haldane noted that the ratio of “risk-weighted” assets to total assets at major international banks shrank by almost half between 1993 and 2011. As Haldane put it, that meant there were only three possible explanations: Bank assets had become half as risky, bank managers had become twice as good at risk management, or the banks were gaming the system. Haldane put his money on gaming.

A classic problem for raising capital, called “debt overhang,” is that when a firm has suffered significant losses and its debt has become risky, then it needs to raise equity to repair its balance sheet. But, at such times, issuing new equity will automatically transfer wealth from shareholders to the creditors or the debt insurers (who are the taxpayers in the case of banks), because the infusion of new capital makes the bank’s debt safer. As a result, managers often avoid issuing new shares, preferring instead to either cut back on new investment or manipulate regulatory capital. In fact, one reason the banks needed rescuing was that they did not issue enough new shares when they still had the opportunity.

Bulow argues that equity recourse notes would relieve some of the angst about debt overhang. When a bank’s share price has fallen significantly, the bank could issue new ERNs with a proportionally lower conversion price, even as old ERNs started to convert. The newer notes would be senior to the older ones, and less expensive to issue than new junior debt or new equity. That would make it easier for banks to repair their balance sheets, and — at the same time — encourage the banks to do more lending in bad times.

If bondholders are forced to take on more risk, won’t they demand a higher return from the banks? Yes, says Bulow. But they would be embracing the risks head on, rather than handing them over to taxpayers.

View the Research Papers

Equity Recourse Notes: Creating Counter-cyclical Bank Capital (May 2014)

Market-Based Bank Capital Regulation (September 2013)