Skip to main content


In the aftermath of the 2007-08 Global Financial Crisis, regulators encouraged many of the world’s largest banks to hold a new type of regulatory instrument with the goal of improving their safety and soundness. The regulatory instrument was known as a “CoCo,” short for contingent convertible bond. CoCos are neither debt nor equity. They are something in between, designed to give the bank a shot in the arm during times of stress. Many of the largest international banks have issued CoCos worth hundreds of billions of dollars. After more than ten years—a decade that includes the collapse of Credit Suisse in Switzerland—this regulatory experiment has failed.

We leverage insights from economic theory to show that CoCos were doomed from the beginning for two reasons. First, from a finance perspective, providing more equity only stabilizes a wobbling bank in normal times before the market and depositors ask questions about the bank’s health. Once they start asking questions and the bank faces a liquidity crisis (i.e., a bank run), having more equity on the bank’s balance sheet becomes meaningless. Only more liquidity can save the bank from complete collapse. Second, from a game theory perspective, controlling the public availability and flow of information is crucial in times of stress. If the market and depositors can ascertain which bank is weak or how much financial trouble that bank is in, a liquidity crisis will ensue, and that bank is as good as gone. The stigma effect is lethal. Ironically, the trigger mechanism built into CoCos can send a public signal that a bank is on its deathbed. It allows the market and depositors to differentiate between the weak and the strong, nearly guaranteeing a bank’s failure.

Is the regulatory experiment salvageable? We offer a set of reform proposals consistent with our theoretical insights. We argue, foremost, that the trigger mechanism should be used early, well before a liquidity crisis begins. We also recommend that the mechanism should protect a bank in poor financial health by sending as little information to the market as possible. That may require a greater reliance on regulators’ discretion and a simultaneous trigger across several banks to prevent the market from identifying which bank(s) may be in trouble. To be sure, we are cleareyed that our proposals come with costs, which we describe at length. If regulators conclude that the costs are too high and our proposals are too difficult to implement in practice, they should end the experiment altogether. The status quo is a regulatory fiction.

Related Working Papers

Scroll to Top