‘Too big to fail’ is still a threat

Opinion: the problem is not only the subsidy for bank risk-taking, it is also the likelihood of disasters

No solvent government will allow its banking industry to collapse. Leveraged institutions whose liabilities are more liquid than their assets are vulnerable to panics. In a panic, it will be hard to distinguish illiquidity from insolvency. These three points shape my views: the state stands behind banking even though it might not stand behind individual institutions.

One of the obstacles to making the bearing of losses by creditors credible is “too big to fail” – the challenge posed by banks that are individually systemic.

A question about post-crisis regulation is whether this risk is gone. The answer is no. Mark Carney, the governor of the Bank of England and chairman of the Financial Stability Board, agrees that "firms and markets are beginning to adjust to authorities' determination to end too-big-to-fail. However, the problem is not yet solved."

No it is not, as a chapter on banks in the latest Global Financial Stability Report from the International Monetary Fund shows. "Subsidies rose across the board during the crisis but have since declined in most countries," it concludes. "Estimated subsidies remain more elevated in the euro area than in the US. . . All in all, however, the expected probability that systemically important banks [SIBS]will be bailed out remains high in all regions." Moreover, in another crisis, the necessary subsidies might jump once again.


One reason is that the banking sector has tended to become even more concentrated. Furthermore, the total assets of a number of big banks have continued to soar: institutions with assets of $2 trillion are common. Such banks remain highly interconnected, though the extent of this might have diminished recently.

Another reason is that the subsidies are still large. The IMF notes there are three different ways of assessing the subsidy. The first is from the difference between the interest rates on bonds issued by SIBs and non-SIBs. The second is a “contingent claims analysis”. The third comes from the analysis by rating agencies of the gap between the standalone rating and one allowing for state support.

Comparisons between interest rates on bonds paid by SIBs and non-SIBs are quite tricky. Nevertheless, the contrast between US institutions with comparable leverage ratios (ratios of total assets to equity) reveals that the SIBs have a funding advantage. This confirms that the subsidy does indeed endure.

The second approach is based on comparing observed spreads on credit-default swaps (a form of insurance on bonds) with fair-value spreads derived from prices of equities. The CDS spreads – unlike data from equities, whose owners are unlikely to be protected – take account of the probability of distress and the likelihood and size of government support. This method shows huge support during the crisis, which then declines in the US and rises in the UK and euro zone.

The third approach is taken directly from estimates by rating agencies. This too shows that the subsidies are large, though slightly declining after the crisis.

Reforms introduced after the crisis aimed at reducing the likelihood of state support seem to have increased the perceived riskiness of SIBs, just as the imposition of lower leverage ratios reduced it. Both outcomes were desirable. Yet the implied subsidies remain large – as high as $312 billion in the euro zone, on one approach. In terms of the funding cost advantage to SIBs the subsidies are at least 15 basis points in the US; 25-60 basis points in Japan; 20-60 basis points in the UK; and 60-90 basis points in the euro zone.

It is hard to conceive of a good argument for subsidising these banks against smaller competitors. The only plausible argument is that banking systems dominated by a few large institutions might be more stable than competitive ones. This is not ridiculous: a well-organised cartel might be more stable than a large number of small competing banks. Yet if so this would clearly come at the expense of customers and the economy. It would be far better – and feasible – to take another route.

So what is to be done? The IMF suggests three options: restrict the size and activities of banks; reduce the probability of distress; and lower the probability and size of any bailout if a bank becomes distressed. Of these, the second is best.

While ringfencing of retail activities makes good sense, governments should not decide what private businesses do. On the third, commitments to limit the probability and scale of a bailout in the event of a systemic crisis are usually not credible. Given this, the best policy is the second: reducing the likelihood of a crisis. That can be best achieved by raising capital requirements and ensuring maximum transparency of balance sheets. This is even more vital if the aim is to safeguard the economy and the solvency of governments.

Unfortunately, despite efforts in that direction, bank leverage remains too high. The US has now proposed a (non-risk-weighted) equity ratio of 5 per cent for large bank holding companies. But it is too easy for a bank’s assets to lose 5 per cent of their value. Funding by equity should be at least 10 per cent of the balance sheet and ideally more. At the very least, as the IMF suggests, equity should be raised until all measures of the subsidy are zero.

Yet it is always the system, stupid. It would be quite wrong to suppose the chief problem is individual banks that are too important to fail. – (Copyright The Financial Times Limited 2014)