One of the most interesting sessions in Chicago during the last ASSA Annual Meeting was about ending too big to fail. Neel Kashkari presented the Minneapolis Fed plan and Markus K. Brunnermeier and Randall Kroszner gave their view on the plan.
The plan involves a massive increase in common equity requirements for systemic banks, to 23.5% in the first instance, potentially going up to 38%. Kashkari’s rationale for the plan is clear: the resolution framework, which should make bondholders absorb losses before considering a bailout, will not work in a crisis. Therefore, the total loss absorbing capacity of the bank should be formed of common equity.
For the resolution framework to work it needs to be credible. If nobody believes it will work—i.e., if nobody believes that the junior bonds will absorb losses if the bank enters into resolution—prices will not incorporate this possibility. You need a critical mass of market participants to anticipate the losses. If that is not the case, once a supervisor tries to trigger losses on bondholders we can have an extreme re-pricing of other bonds, potentially leading to a negative contagion effect.
So is the resolution framework credible? Unfortunately, we might only find out in a crisis. Even if some bondholders absorb losses in some cases—for instance, Banco Espirito Santo—these are typically not full-blown crises but instead individual failures. And even in such cases, the political appetite does not seem to be imposing losses on bondholders, such as with Monte Paschi de Sienna.
If the resolution framework is not credible, Kashkari argues, we should make sure that all loss-absorbing capacity is made up of common equity rather than junior and senior bonds. Hence, for systemic banks, the equity requirement should be around 23.5% of risk-weighted assets, similar to the phased-in TLAC requirements plus Basel III buffers (18% + 2.5% CCoB + GSIB buffer + CCyB).
The requirements would continue to increase up to 38% as long as the bank continues to be systemic. Interestingly, the net benefits of the first step (23.5%) are higher than those of the second step (38%). If the cost-benefit analysis is credible, they should then stop at 23.5%, but Kashkari justified going up further by claiming that systemic crisis and bailouts have broader costs than just output losses: for instance, political polarisation and social unrest. For this reason, he said, they wanted to decrease the probability of such crisis below 10% in the next 100 years.