Nyberg: Financing of deposit insurance – a central banker's perspective
OBS: Talet publiceras endast på engelska
Introduction
Some of you may be wondering why on earth a central banker would have an opinion on the financing of deposit insurance. But to the extent that central banks are concerned with financial stability, there is actually an important reason for the central bank to take an interest in the overall design of deposit insurance.
Deposit insurance and financial stability
In the EU as in most other developed economies in the world, deposit insurance is seen primarily as a form of consumer protection. Bank failures are few and far between and bank crises involving the entire system are something that most of us have only read about or heard about at a conference. Under these generally stable circumstances, the role of deposit insurance is limited to reimbursing the relatively few customers who are unfortunate enough to have deposits in the (always) very small banks that once in a while have to be closed down.
When large banks show signs of being in serious trouble, on the other hand, authorities are often very keen to avoid an outright insolvency and a subsequent failure. Instead, the authorities’ usual procedure is to support the ailing bank – through injections of liquidity or capital – while at the same time enforcing some kind of restructuring, including replacing the management, merging the weak bank with a stronger competitor, raising capital through the sale of non-core operations, etc. Hence, deposit insurance payouts for large bank failures or systemic banking crises are practically unheard of.
But it was not always like this. At the time when deposit insurance was introduced in the United States in 1933, the main purpose was “to restore public confidence in the nation’s banking system”. It was created in response to the thousands of bank failures that occurred in the 1920s and early 1930s, culminating in a national bank holiday in 1933. In essence, the federal deposit insurance was motivated by financial stability concerns – to stop the bank runs and enable the banking system to start functioning normally again. More recently, many emerging economies in Latin America and Asia have had similar experiences – to avoid a complete breakdown of or to restore the confidence in the banking system, authorities have had to convince the public that their deposits will not be lost.
In other words, even though in today’s Europe deposit insurance in 99 per cent of the cases is about consumer protection, ultimately, it is also an important building block in the safety net underpinning the stability of the financial system.
The nature of deposit insurance
Given that society seems to want deposit insurance, we need to think about how it should be organised and who should pay for it. Because someone will – there is no free lunch. I will now try to give you some tentative answers to these questions. I should also say that these ideas are not the unique intellectual property of the Riksbank but are very much in line with the recommendations given this spring by a Swedish government committee on reforming the deposit insurance system.
Today, deposit insurance in the EU is provided either by cooperatives where the member banks insure each other or, more frequently, by a government authority. In many cases, there is a fund in place and a scheme for how surviving banks would pay for failing banks. In this sense, deposit insurance systems in the EU often have both public and private characteristics. When small banks fail, the funds in place or the cooperative arrangements can be expected to be adequate. In the case of a really large bank going bankrupt or even a system-wide banking crisis, however, there is no doubt that the ultimate guarantor is the state, that is, the tax payers – either by supporting the deposit insurance system or, more likely, by saving the banks from failure through injections of capital or liquidity.
Still, we should always ask ourselves if the market could not solve this on its own without public interference. After all, insurance companies sell insurance against various other forms of “low probability-high impact” events like terrorist attacks and earthquakes. Why couldn’t insurance companies also sell insurance protecting depositors against bank failures?
Well, beyond political considerations such as depositors also being voters, in the end, it boils down to the fact that the state cannot avoid responsibility for the payment system. Since the market understands that the state has a very strong interest in the functioning of the payment system, it is almost impossible for the state to convince the market that it wouldn’t – in one way or another – protect depositors. Thus, if deposit insurance – whatever the institutional arrangements – ultimately rests on a state guarantee, where the risk is borne by the tax payers, it seems sensible that tax payers should be made aware of this. This speaks in favour of an explicit state guarantee that clearly shows the distribution of responsibilities, risks and costs. Which then brings us to the next question: what is a reasonable price for the tax payers to demand from the banks for providing them with deposit insurance?
The financing of deposit insurance
As a starting point, it is important to recognise that deposit insurance has a value for banks, since it lowers their funding cost. And we have already concluded that it implies a (potential) cost for the state. Even if we have rejected the idea of private deposit insurance, it is useful to think about how the market, in that case, would have priced this insurance. The market would have charged a premium, which size would depend on the amount covered, the length of the insurance period and the probability of the bank failing. Personally, I think that the principles of private insurance, with some exceptions, should also be applied in the design of a public deposit insurance system. This would have mainly two implications.
First and most importantly, the premiums that banks pay should be risk-based, meaning that a high-risk bank should pay a higher fee than a safe bank. This provides the banks with sound incentives for risk management and reduces some of the moral hazard inherent in all guarantees. But risk-based premiums are also economically sound from the tax-payers’ perspective, since a higher risk bank is more likely to go bankrupt and therefore more likely to “make use” of the insurance.
While I am a strong proponent of risk-based premiums, it is important to admit that, in practice, it can be quite difficult to calculate the risk of a bank. I will not here go into the details and pros and cons of various models and approaches. But I would like to make some general points regarding risk calculation.
To begin with, I strongly argue for a simple approach which is easy for the banks to understand and predict. The value added from developing something very complex but scientifically more precise is likely to be very marginal. Second, I would advocate authorities to make use of the risk and capital assessment that is anyway conducted by the supervisor. With the implementation of Basel II considerable resources will be invested in developing this supervisory assessment and it seems a pity to develop a parallel system. Also, with two different systems authorities run the risk of sending contradicting signals.
Finally, even if premiums should be risk-based, this does not imply that they should be market-based. The state insurer differs from a private insurer in at least two fundamental ways. As we are all painfully aware, the state can raise taxes. Hence, the state cannot default in the same way as a private insurer and therefore does not need to compensate itself by charging a risk premium. The only thing that really matters to the state insurer is the expected loss for each bank. That there is high uncertainty around such expectations is not a problem for the state, as long as the expectation is correct in the long-run. In addition, the state does not need to make a profit but should, as already noted, only be concerned with covering its expected long-run cost. Thus, the risk-based premiums will be lower when charged by the state than when charged by a private insurer.
Second, banks should pay their premiums in advance (or ex ante). The most important reason for this is that otherwise it is not possible to have risk-based premiums. But there are also some other good reasons for this. Paying in advance means that the premium payments are smoothed over time whether there has been a failure or not. In contrast, only paying once a failure or crisis has occurred means that banks have to increase payments when the banking system might already be in a fragile state. Alternatively, it might prove difficult in practice for the authorities to recoup all the costs of an insurance payout in the aftermath of a crisis.
Finally, in some sense ex ante payment of premiums is also fairer since it implies that all banks – even the ones which will later fail – pay for the insurance. If premiums are paid ex post the remaining banks will have to pay for the failures of their competitors.
What to do with the premium income?
The state may use the premium income in two ways. Either it can use it to build up a deposit insurance fund ready to be used in the event of a payout case, or it can let the premium income flow into the state’s coffers just like any other government fee or tax. Given the long-term and potentially large liabilities of deposit insurance, however, it is probably advisable that the premiums do not flow directly into the state budget to be freely spent by the current government. Therefore, the Swedish deposit insurance committee proposed that the premium income would flow into a separate account where it can only be used to fund the government deficit. This will mean that in good times, premium income from the banks will contribute positively to the state’s finances by decreasing the borrowing need, while in bad times when banks fail, the state will have to increase borrowing in order to fund eventual insurance payouts.
In all the ex ante paid systems that I am aware of there is also a fund. A fund may be practical since the deposit insurance authority then can manage minor cases of payouts without having to borrow money from the government. But it is difficult to see why this would be enough to motivate the setting up of a fund. In the event of major failures the government would be involved in any case. And having a fund in place also entails costs, in terms of fund management. If we do not believe that the government-appointed fund manager can systematically beat the market, which we don’t, there is really no strong case for having a fund.
Cross-border aspects in the EU
In the EU at present, there are more deposit insurance models than there are member states, since some countries have more than one system. These systems differ in all possible aspects: scope and amount of the cover, whether premiums are paid ex ante or ex post and whether there is a fund in place and, if so, how large it is.
Up to now, EU banks have conducted cross-border retail banking almost exclusively in subsidiaries as opposed to branches. In the few and relatively minor cases of retail banking being run in a branch, the differences in level and scope have been handled by the possibility of topping-up. However, as cross-border retail banking has continued – albeit slowly – to grow in importance, the interest in deposit insurance has recently been rising. The sources for this rising interest are, however, quite different.
On the one hand, in the perspective of the banking sector and the Commission, deposit insurance may be both a potential barrier to further integration and a threat to the level playing field.
Some large EU banks, of which one has its headquarters just 200 metres from here, have expressed discontent with the current regime on mainly two accounts.
Firstly, the existence of funds and the link between the size of the fund and the level of the fees give rise to competitive effects when entering and exiting different systems. This may discourage transformation to branches and – since branches may be a more cost-efficient way to run foreign retail operations – create an additional barrier to cross-border consolidation. Let me give you an example. When a large bank exits one system, that system might become “over-funded” leading the authorities to lower fees for the bank’s competitors which remain in the system. At the same time, the large bank joins (or adds more deposits to) another system, which means that this system might become under-funded, leading the authorities to raise the fees. The net effect is that the bank ends up with a higher total cost for deposit insurance, while some of its competitors get a lower cost.
Secondly and most obviously, different levels of fees threaten the level playing field between banks competing for the same deposits. Topping-up only works in one way: it allows the foreign bank that enters to get the same level of insurance as the host country banks. However, when the entering bank has a better or cheaper insurance than the host country banks, there is no way for the host country banks to compensate for this. Hence, backed with generous or cheap deposit insurance from their home system, some banks have a distinct advantage to expand into other EU countries.
I believe that this criticism should be taken seriously and I welcome the review of the directive that the Commission is currently undertaking. At the same time, I would not overemphasize the competitive and level playing field aspects of deposit insurance. They are important in principle, but in practice and relative to, for example, national differences in capital regulation they are rather marginal.
The perspective on deposit insurance of some national regulators and central banks, on the other hand, has been quite different and focussed on the burden sharing between home and host countries. What are the wider implications of tax payers in one country insuring substantial amounts of deposits in another country? It has been noted that it may be politically very difficult to make payouts from one country to another in the aftermath of a large bank failure. It could also be questioned if it is reasonable that, after a large bank failure, the authorities in the host country just refer its depositors to the home country authorities, without taking any responsibility. And, of course, there is always the question of whether the failure of some large cross-border EU banks will be too big to manage for a single EU member state.
Conclusion
Both kinds of cross-border aspects – level playing field as well as burden sharing – are indeed tricky. Still, I would like to note that if we all agreed to charge risk-based ex ante fees and scrap the funds, we would take a major step towards solving many, but not all, of these problems. The competitive effects from exiting and entering systems would be significantly reduced. Banks would pay premiums depending on their own risk and the amount of insurance received. There would still be a need for topping-up, as there would still be differences in the level and scope between systems, but at least the banks covered by a more generous insurance would have to pay a higher premium. Regarding the burden-sharing concerns, one could still doubt the capacity of an individual small country to deal with a large cross-border bank failure, but at least the country’s tax payers would be aware of the liability and compensated for the expected cost.
That's what one might call killing two birds with one stone – better national systems that would at the same time work better on the EU level.
Thank you.
