The Fall and Rise of Debt in Bank Capital Structures

In the aftermath of the financial crisis much criticism was directed at the low level of equity in banks’ capital structures and, by implication, the high level of debt funding (leading to a high debt to equity or leverage ratio). The rules of the Basel Committee on Banking Supervision (BCBS), the international standard setter, required that only 2% of banks’ assets be financed by shareholders’ equity (essentially the excess of assets over liabilities assessed on book values and attributable to the ordinary shareholders - referred to in the Basel jargon as ‘Common Equity Tier 1’). This level of excess had to be kept within the bank and not distributed to shareholders. Although national regulatory authorities were free to add to that minimum, it was clear that only a relatively modest decline in the valuation of banks’ assets might mean that this minimum was no longer met. Outside a systemic crisis, this threat to the bank’s operating licence could be addressed by a capital raising or by retaining a higher level of profit within the bank. In a systemic global crisis, where asset values declined suddenly and precipitously, this measured response was not adequate across the board, and many banks had to be bailed out with taxpayers’ money.

Not surprisingly, the immediate post-crisis focus was on the Basel minimum requirements. CET1 was raised from 2% to 4.5% and various ‘buffers’ were added on top. All internationally active banks were required to hold a capital conservation buffer of 2.5% CET1 (though this could be reduced in the down-turn of the economic cycle) and systemically important banks were require to hold a further 2.5% CET1 in the case of the largest ones. National regulators have made greater use of their powers to move the minimum number upwards.

More surprising was the role of subordinated debt in the Basel rules. Pre-crisis, the minimum CET1 was 2% but the overall Basel minimum capital figure was said to be 8%. Much of the gap between the two numbers banks were free to cover by issuing subordinated debt. (Also available were some types of preference shares as well as more exotic equity instruments.) The presence of debt in capital requirements is, of course, really odd, at least to a run-of-the-mill corporate lawyer. Taking on debt increases assets, of course, but to an equivalent amount it increases liabilities. So, the net asset position of the bank is not improved.

Pre-crisis bank regulators seem to have convinced themselves, nevertheless, that subordinated debt could be permitted to count towards capital, on the grounds that the primary purpose of capital was to protect bank depositors and so reduce the probability of a bank run. Debt lower in the hierarchy than deposits could perform this role, as well as equity. And debt was more attractive to banks than equity because of the tax shield for debt interest which many jurisdictions provide (ie interest comes out of pre-tax revenues whilst dividends come from taxed income). What this argument ignored was the fact that debt carries losses only once the bank is placed into insolvency. Then indeed subordinated debt will be written off before deposits, but until insolvency the debt-holders’ claim against the bank is legally intact. In the crisis, all governments concluded that the damage caused to the real economy through the loss of lending capacity if a bank went into insolvency was likely to be more than the damage done to the public finances through a bail-out. So, most banks were bailed out rather than put into insolvency, with the result that the loss-absorbing capacity of subordinated debt was not triggered. Taxpayers bailed out subordinated debt-holders as well as depositors.

The response of the Basel authorities was not to take subordinated debt out of the permitted categories of capital but to tweak the definition of eligible subordinated debt. To count now the subordinated debt must be capable of being written off or converted into equity by the public authorities when breach of the minimum capital requirements is likely but before insolvency is reached, for example, when the bank is bailed out. Writing down or converting the subordinated debt at this earlier point means it will bear losses before the taxpayer injects funds into the bank.

One might have expected debt to disappear entirely from the Basel minimum capital requirements, but the attraction of the debt shield to banks, with the associated argument that removing it would increase the costs of bank borrowing, seem to have preserved it in this modified form as part of the Basel minimum. However, there is a distinct sense of reluctance in the post-crisis treatment of subordinated debt in the Basel rules. Banks are not required to have subordinated debt but are permitted to do so in partial discharge of the minimum capital requirements, as had been the case previously. The buffers, which are a novel feature of the post-crisis reforms, by contrast must be met with CET1 only. For large, systemic banks the CET1 minimum is nearer 10% post-crisis, as compared with the pre-crisis requirement of 2%.

Nevertheless, a minimum equity figure at this level has been criticised as too low by a range of authors: public commissions established by government (such as the Vickers Commission in the UK), policy advisers acting within regulatory bodies (such as the Bank of England) and independent academics. Most have proposed minimum equity requirements about twice the new Basel minimum cum buffers, whilst Professors Admati and Hellwig have argued vigorously for minimum equity requirements of over 50% of assets (at least when assets are valued on a risk-weighted basis). The banks responded with a reiteration of the arguments which had been deployed within the Basel process in relation to the banning of subordinated debt. A large increase in the equity requirements would in effect require the banks to replace a substantial proportion of their long-term senior debt (bank bonds) with equity, and, since equity is more expensive for banks than debt (largely because of the tax shield), the costs of borrowing would go up. The two sides fought this argument to a standstill. Although neither side delivered a knock-out blow intellectually, it was noticeable that no regulator increased the pure equity requirements above the Basel minimum plus buffers by the multiple suggested by the critics.

So matters might have remained. The debt/equity dichotomy (good equity, bad debt, or vice versa, depending on your point of view) might have continued to frame the discussion, but for developments in another part of the field of banking regulation. In a second reform phase attention moved from increasing capital requirements (designed to reduce the probability of bank failure) to designing bank resolution procedures capable of handling failing banks without putting them into a standard insolvency procedure. Bailing-out is one such resolution procedure but its dependence on taxpayer funding makes it deeply unpopular. In the hunt for an alternative, the notion of restructuring the capital of a failing bank moved centre stage – the so-called bail-in. In this mechanism a bank which is in financial distress has its liabilities reduced by the write-off of parts of its debt and, possibly, its equity position restored by the conversion of that debt into equity. This mechanism has the potential to restore a failing bank to viability – but by imposing the costs of resolution on the debt-holders rather than the taxpayer.

With bail-in came the rehabilitation of debt. Bail-in has a chance of working only if the bank carries a substantial amount of debt; writing off trivial amounts of debt is unlikely to produce viability. Ironically, the more highly leveraged a bank is, the more likely bail-in is to work. Second, bail-in favours the holding of substantial amounts of subordinated debt. The prospect of bailing in depositors is virtually a guarantee that they will run at the first hint of trouble, thus increasing the chances of a bail-out. The debt which is in the front line for bail-in must be subordinate to the deposit debt of the bank. Subordinated debt is attractive in the bail-in process precisely because it will still exist as a liability at the time of resolution, unlike shareholders’ equity which will have declined substantially by the point.

Instead of a debt/equity dichotomy, equity and subordinated debt subject to write off and/or conversion are put together as loss absorbing instruments. This approach is most evident in the resolution proposals of the Financial Stability Board. The FSB focusses on what it calls Total Loss Absorbing Capacity (TLAC). In 2014 it consulted on TLAC level of between 16 and 20% of risk-weighted assets, thus responding to the Basel critics. Unlike those critics, however, the FSB envisages TLAC being made up of a combination of shareholders’ equity and bail-in debt. Further, the composition of the TLAC is not to be left to individual banks. Rather, the FSB envisages that at least one third of TLAC will be made up of bail-in debt.

With the shift of focus from capital requirements aimed at reducing the probability of bank failure to capital requirements formulated to facilitate the resolution of failing banks, debt has moved from regulatory suspicion to regulatory promotion – a singular piece of rehabilitation in a period of some five years. I have charted this developed and analysed the conditions under which bail-in can be expected to work in a forthcoming article: The Fall and Rise of Debt: Bank Capital Regulation after the Crisis to be published this year in a special edition of the European Business Organization Law Review (an earlier version of which is available here).