In the aftermath of the recent financial crisis many were convinced that banks had too little capital. The low levels of capital contributed to the contraction of bank lending and excessive risk taking. The response to the crisis was to raise the minimum capital requirement and simultaneously increase the quality of bank capital. Accepting bankers’ claims that raising equity can be ‘expensive’, Basel III also allows banks to partially meet the increased equity requirements with certain types of contingent convertible bonds (CoCos).
CoCos are regulatory instruments which enhance the issuing bank’s loss-absorption capacity. A CoCo is a coupon-paying bond which, depending on its loss-absorption mechanism, must either convert into shares of equity or be written off at the trigger effect. The trigger event is associated with the depletion of the bank’s capital below a certain level. The trigger event is typically defined in terms of the bank’s Common Equity Tier 1 (CET1) ratio. High-trigger CoCos, with a trigger level above 5.125% in terms of the CET1 ratio, are meant to automatically reduce the leverage of the issuing bank to avoid the disruption of its operations. Such CoCos may qualify for Additional Tier 1 (AT1) capital which is a part of the bank’s Tier 1 capital, thus receiving an equity-like treatment from the regulators. Low-trigger CoCos, on the other hand, serve as securities which can be bailed-in, thus counting as Tier 2 capital. Moreover, in most jurisdictions CoCos are treated as debt for tax purposes, thus providing the benefit of tax shields.
Given the equity-like treatment by regulators and debt-like treatment by fiscal authorities, it is somewhat surprising that AT1 CoCos are not as popular as one would expect. For example, in the European Economic Area, which is characterised by a homogenous regulatory framework, less than 30% of all publicly traded banks have issued AT1 CoCos. This number is substantially smaller for privately held banks in the same area. The existing theoretical literature on CoCos provides insights into the trade-offs for banks to issue them.
On the one hand, absent agency problems, issuing CoCos always improves a bank’s loss-absorption capacity and, hence, lowers its default (Albul et al. 2013, Pennacchi et al. 2011 and McDonald 2013). This is because, similar to equity, CoCos add an additional buffer to the balance sheet of the banks. Therefore, abstracting from agency problems and taxes, banks’ incentives to issue CoCos or equity are the same.
On the other hand, a number of papers argue that CoCos could induce agency problems such as debt overhang and risk shifting (Berg and Kaserer 2015, Chan and van Wijnbergen 2017, Goncharenko 2018, Hilscher and Raviv 2014, Koziol and Lawrenz 2012 and Martynova and Perotti 2018). Intuitively, when the contractual terms of a CoCo are such that the bank’s losses are at least partially shared between CoCo and equity holders, then the CoCo induces an anticipated positive wealth transfer to equity holders at the triggering event. Under such conditions, equity holders benefit from conversion and therefore have little incentive to reduce leverage which would help to avoid the trigger. In particular, they have little incentive to issue additional equity, a situation we refer to as ‘debt overhang’.
We argue that the debt overhang problem of CoCos is particularly problematic given current capital regulations. Under the current regulatory framework, it is expected that banks may have to issue equity well before the trigger is reached. Indeed, most equity-like CoCos have trigger levels set at 5.125 % in terms of the CET1 ratios, while at the same time Basel III requires banks to hold the CET1 ratio in the range of 7 to 8.5 %.1 Thus, given that banks usually hold an additional voluntary buffer on top of the minimum requirements, it is very likely that they would experience deleveraging pressure when their CET1 ratios are still above 7% to 8.5%, or well before any CoCo trigger would be reached. Therefore, anticipating the future need for recapitalisation, banks can find it suboptimal to issue CoCos today because of future debt overhang.
In a recent study (Goncharenko et al. 2018), we use a simple contingent claim model to examine the debt overhang of CoCos. We show that for a given CET1 ratio, a bank with liabilities consisting of straight debt and CoCos will suffer from more debt overhang than a bank with liabilities consisting of straight debt only. In this scenario, the existing equity holders of banks with outstanding CoCos will have lower incentives to voluntarily issue equity, and if these banks are forced by regulators to re-capitalise, the existing equity holders suffer greater losses. Moreover, we show that this relative debt overhang increases in bank asset volatility. Intuitively, higher asset volatility is associated with a higher likelihood of conversion and a higher CoCo coupon, both of which increase the value of the wealth transfer induced by CoCos at the triggering event. Thus, our theoretical analysis suggests that bank asset volatility relates inversely to a bank’s decision to issue CoCos. Banks with higher asset volatility are less likely to issue CoCos as they anticipate higher costs associated with the debt overhang that these securities will induce once issued.
Next, we proceed with the empirical analysis of the determinants of CoCo issues to test the predictions of our model. In particular, we test whether bank asset volatility relates inversely to the decision to issue CoCos. Our main dataset consists of the CoCo issues by publicly held banks from the European Economic Area which took place during 2010-2017. We restrict ourselves to the analysis of AT1 CoCos since only this type of CoCo enjoys equity-like treatment. Consistent with our model’s prediction, we find strong empirical evidence that banks with lower asset volatility are indeed more likely to issue CoCos compared to their counterparts with more volatile assets. In particular, our estimation predicts that with a 1% increase in banks asset volatility, the odds of CoCo issuance fall by about 1.22%.
We further supplement our analysis of AT1 CoCo issues with an analysis of common equity issues. First, we do not find statistical evidence that bank asset volatility determines the likelihood of common equity issue. Second, the joint analysis of AT1 CoCos and equity issues suggests that the likelihood of a CoCo issue decreases in bank asset volatility relative to the likelihood of equity issuance. These findings are consistent with our theoretical model.
Our analysis provides empirical evidence of the agency problems of CoCos which are pointed out in the theoretical literature. From a positive perspective, our analysis shows that banks take into account the future debt overhang when forming their capital structures. From a normative perspective, our analysis suggests that the design of CoCos should be revised if they are to enjoy an equity-like treatment. We suggest that, at least, the minimum trigger level for AT1 CoCos should be raised.2 Moreover, in the current form, the use of such securities should not be promoted by using tax subsidies or state bailouts (for example, as part of Portugal’s bailout by the EU and IMF in 2012, its banks were recapitalised through CoCos). Although the banks which issue such CoCos are the ones that suffer the least from the debt overhang induced by these securities, the consequence of this debt overhang can be quite dramatic in times of financial distress. Thus, during the next financial crisis, banks with outstanding CoCos may have even fewer incentives to recapitalise, which could make the next credit crunch even worse.
Albul, B, D M Jaffee and A Tchistyi (2013), “Contingent convertible bonds and capital structure decisions”, working paper.
Berg, T and C Kaserer (2015), “Does contingent capital induce excessive risk-taking?”, Journal of Financial Intermediation 24: 356-385.
Chan, S and S van Wijnbergen (2017), “CoCo design, risk shifting and financial fragility”, ECMI working paper 2.
Goncharenko, R (2018), “Fighting fire with gasoline: CoCos in lieu of equity”, working paper.
Goncharenko, R, S Ongena and A Rauf (2018), “The agency of CoCos: Why contingent convertibles aren’t for everyone”, CEPR Discussion Paper 13344.
Hilscher, J and A Raviv (2014), “Bank stability and market discipline: The effect of contingent capital on risk taking and default probability”, Journal of Corporate Finance 29: 542-560.
Koziol, C and J Lawrenz (2012), “Contingent convertibles. Solving or seeding the next banking crisis?”, Journal of Banking & Finance 36: 90-104.
Martynova, N and E Perotti (2018), “Convertible bonds and bank risk-taking”, Journal of Financial Intermediation 35: 61-80.
McDonald, R L (2013), “Contingent capital with a dual price trigger”, Journal of Financial Stability 9: 230-241.
Pennacchi, G (2011), “A structural model of contingent bank capital”, Federal Reserve Bank of Cleveland working paper 10-04.
 This level includes the minimum CET1 of 4.5%, Capital Conservation Buffer of 2.5%, 1 to 2.5 % for being a Global Systemically Important Bank, and additional national buffers for Global Systemically Important Banks.
 The Prudential Regulation Authority of the Bank of England now requires the minimum trigger level to equal 7% in terms of CET1 as opposed to the 5.125% minimum level found in Basel III. This, however, is motivated by the greater assurance that these capital instruments can be fully converted to CET1 while the bank is still a growing concern, not by any aspiration to mitigate future debt overhang.