Deposit insurance will not help to stabilise the banking industry. (Photo: AFP)
In a classic 2005 paper from the highly authoritative Journal of Monetary Economics, Asli Demirgüç-Kunt and Enrica Detragiache investigated the question "Does Deposit Insurance Increase Banking System Stability?" Their answer, based on an empirical study of a large panel of countries from 1980 to 1997, was that it does not; in fact, as they put it: "explicit deposit insurance tends to be detrimental to bank stability". This was a well-established notion amongst money and banking academics. When I used to lecture in money and banking, our standard textbooks made much of the failures of deposit insurance in the Savings and Loan crisis in the US in the 1980s. I had actually developed the somewhat naive idea that everyone whose opinion counted in the US and UK understood this. That was, after all, why in the UK we never had any form of deposit insurance until 1979, even thereafter only had what was required of us by EU directives, and even up to 2007 only the first £2,000 of deposits was insured. (There was 90 per cent insurance of the next £33,000, but since even in the epic US banking collapses of the 1930s, depositors typically recovered more than 80 per cent, a 90 per cent level of insurance only in practice provided very marginal additional protection of a few per cent. I didn't like it; but it wasn't appalling.)
Given how destructive deposit insurance was – I thought – understood by informed opinion to be, you will appreciate that I was aghast and taken aback that in September 2007 the government introduced blanket 100 per cent deposit insurance in the UK. Subsequently, though the blanket protection has, mercifully, been removed (with the failure of the Southsea Mortgage and Investment Company this June, where around 5 per cent of the depositors were not insured or otherwise bailed out), the deposit insurance threshold has been raised to £85,000. I hope that can be taken down to around £10,000, one day, but for now it wreaks its damage – not as bad as unlimited insurance, but damaging nonetheless.
I was reminded of this today when reading the latest Bank of England Quarterly Bulletin. For in the Demirgüç-Kunt and Detragiache 2005 paper, they suggest that (in addition to the loss of incentive for depositors themselves to shop around for low-risk banks) a key reason why deposit insurance damages financial stability is that the higher and the more explicit are deposit insurance thresholds, the more tempting it is for the government to bail out other creditors such as bondholders. In the Quarterly Bulletin, there is an article on "Bank resolution and safeguarding the creditors left behind". This notes that, under the current rules, when a bank goes bust its depositors will typically be transferred to another bank, along with some matching assets. But under current insolvency law, depositors rank equally with bondholders. Transferring away the assets to match the deposits might leave bondholders worse off than they would be under standard insolvency proceedings. So to make sure that no creditor left behind is worse off, the state-backed deposit insurance scheme (the Financial Services Compensation Scheme) pays the bondholders the money it would have paid to insure depositors, had they lost out. So as a direct consequence of deposit insurance, the state ends up responsible for paying other creditors.
Because the state is responsible for paying other creditors if it lets the bank go into solvency, the temptation to bail out the bank in advance, with the off-chance that it trades its way out of the trouble and the government actually makes money rather than losing it, will often be overwhelming.
The Vickers report contained two useful measures to help with this. The first is the principle called "depositor preference". That means that, in insolvency, depositors rank ahead of bondholders. So the government would not make any bondholder worse off by selling off the deposits with matching assets – and the obligation mentioned above would vanish. Secondly, and intimately connected with this, is the concept of a bail-in, whereby bank bonds are semi-automatically converted into equity in a failed bank, reducing the temptation for the government to bail them out.
These measures will help. But they will not solve the ultimate problem. Deposit insurance – a measure intended by its fans to reduce financial instability by reducing the likelihood of bank runs – actually achieves the opposite. Bank runs are not a consequence of depositors having a rational fear of losing significant amounts of their capital – depositor capital losses in failed banks are only ever very small. Bank runs are the consequence of depositors fearing they will lose access to their money (they lose liquidity, not capital). Deposit insurance works, to the extent that it does, because when there is deposit insurance governments are much more likely to bail banks out, keeping them running, so depositor liquidity access is guaranteed. But the implicit promise to bail banks out destroys financial stability, by inducing banks to take big risks, leveraging themselves up and making high-risk loans. By reducing the likelihood 0f bank runs, deposit insurance also means that bank staff receive much higher pay and much greater bonuses than they would otherwise – high financial sector pay is a direct consequence of deposit insurance, as detailed in the famous Journal of Finance paper by Diamond and Rajan, "A Theory of Bank Capital". It is not a mystery why bank bailouts lead to high bank pay and high bank bonuses – theory told us, all along, that that was a consequence.
There should be totally insured deposits – just not in fractional reserve banks. Until the mid-1980s we used to have savings banks, 100 per cent backed by government bonds. The government could insure these, without limit, without damaging financial stability. If the government wants to ring-fence some parts of banks in response to the Vickers Review, it should ring-fence an old-style savings bank inside every fractional reserve bank. Then we could have deposit insurance without inducing high pay, high bonuses, high leverage, high risk-taking, and financial instability – and all this without either destroying universal banking (an industry in which the UK excels) or excessive, invasive regulation.
