Among the less tractable problems of the financial system identified by the FSI is how to address the risk posed by financial institutions that are “too big to fail”. It would seem likely that the final report will include some recommendation on this matter.

One possibility is for the legal separation of businesses thought risky from those thought safe: Volcker, Vickers and even the old Australian trading versus savings bank distinction are mentioned (see page 3-19 of the Interim Report). Another is a more advanced resolution regime which would keep a failing bank going while imposing the costs of failure on creditors of the bank without contribution from government or the public at large. This might be done by turning the creditors into shareholders (a “bail-in”). It is time to debate whether these ideas are worthwhile.

So far Australian regulation has approached the problem of “too big to fail” by measures to reduce the risk of the failure occurring. Banks have been made subject to more oversight and higher capital requirements. Like heavier and heavier armour on a battleship, this improves safety but may compromise efficiency. Nor can it ever eliminate the risk of failure. Even the best built battleship may have berthed at Pearl Harbour.

Another approach is to place business lines thought worthy of special protection in a separate entity, sequestered from the risks in other business lines conducted by the bank. This is another “armour-plating” technique. The degree to which it reduces the risk of failure for the sequestered business depends on how that business is defined: is it mere payment and the most basic deposit taking? Does the business include any element of credit origination? The approach is likely to come at some cost, as the report acknowledges. It cannot ensure immunity from risk and in some situations it may increase it, for instance where it results in a loss of the reduction in credit risk that might have been obtained through netting of obligations, had the business all remained in the one legal entity.

So what then can be done to address the situation where, despite all the preventative measures, a significant bank is on the brink of failure? The public good will generally favour keeping the institution going in some form: suspending payment and the withdrawal of credit disrupts the economy and brings ruin to innocent individuals and businesses. The fear of either of those things happening can paralyse business activity.

Some might say a public good justifies the use of public funds to support the bank. The failure of a major Australian bank would most likely have some effect on every Australian household: the household may not bank with the distressed bank, but it is likely to be reliant to some degree on a payment, delivery or service from someone who does. Yet that is not the philosophy of the Basel Committee. A good case can be made that the perception that government intervention would resolve any difficulty may encourage a financial institution to take on risks that have imperilled it in the first place.

So if not from the public at large, where is the necessary support to come from? The bank at this stage may not be able to access the capital markets. Should fellow banks provide support? That works for a small failure, but with a large bank that carries the risk of a weak bank dragging down others and perhaps the system as a whole. Ordinary shareholders and other capital instrument holders, including subordinated debt holders, may lose some or all of their investment, but in the situation we are talking about that may already have happened. That leaves only unsubordinated creditors.

Unsubordinated creditors come in many classes: depositors, bond creditors, holders of bills of exchange or promissory notes, derivatives counterparties, trade creditors, employees, the Tax Office, the Reserve Bank. Any law mandating a bail-in has to consider which classes of creditors are affected and to what extent.

At this point three general propositions may be noted. First, as noted above, it is critical to keep the flow of payments functioning. Payments are made from current deposit accounts. To some degree, these need to be insulated. Second, it hardly seems appropriate for a resolution regime to bail-in creditors on whom the law has conferred a preferred position in the event of a liquidation, at least until after the lower ranking creditors have been bailed-in. A third general principle might be that it is important to protect the position of essential suppliers, so that the bank can continue to operate.

Here some awkward facts in the existing legal regime intrude: a great many creditors enjoy a preferred position over general creditors if an Australian bank fails: holders of protected accounts in Australia have a priority over general creditors for payment out of the bank’s Australian assets. Claims of the Reserve Bank are preferred for payment out of the bank’s Australian assets. Holders of covered bonds have their specific security (and to compromise that would be to defeat the purpose of the covered bond). Holders of derivatives under a close-out netting contract can net their positions and the collateral by exercising rights of close out (and to interfere with that would raise profound systemic issues). These arrangements have been added piecemeal to over time. Each has its justifications but the scheme as a whole lacks clarity of concept and certainty of application.

But if all the existing preferred creditors are excluded from the bail-in regime, and trade creditors are also immune in the interests of the bank continuing to trade, that throws the whole burden of the bail-in on a smaller base of creditors, principally foreign depositors and holders of bonds and bills. These are the investors on whom the nation depends to fund the gap between deposit savings and investment. It may be said that they currently accept the risk of coming behind protected account holders and other preferred creditors in the event of a bank’s failure. But it is a different question to ask what price they will charge if, after only holders of capital instruments, they are expected to bear the whole burden of recapitalising a distressed Australian bank? And what further price will they charge if the principles on which the bail-in would operate are left unclear or uncertain?

Developing a more sophisticated bank resolution regime may be a preferable alternative to ever higher capital charges or structural separation of businesses. Any regime must be fair, clear and coherent and that is likely to require some serious policy choices, including re-examining how classes of creditors of a bank are defined and treated in a liquidation as well as in the resolution regime.