What happens if your bank fails? Decoding India’s deposit insurance scheme and ways to address its flaws

The deposit insurance scheme as it exists currently suffers from several practical and conceptual problems. From a depositor’s point of view the biggest problem is not so much the absolute amount of coverage but the lag with which the insurance company, the Deposit Insurance and Credit Guarantee Corporation (DICGC), pays up. In FY19 for instance, it took on average four years for DICGC to settle claims. But note that this period of four years is from the date of deregistration of a bank and not the date on which a bank defaulted (or the RBI imposed restrictions on deposit withdrawals). Imagine an insurance company settling health insurance claims several years after a fire caused damage or a medical procedure. At a 7 percent interest rate, a five-year delay means that the effective deposit insurance coverage is just over Rs 70,000 rather than the Rs 100,000 that the law says is the coverage. Not to mention the emotional trauma of waiting. If there is one single thing that needs to be changed about the deposit insurance scheme is the timing of insurance payment. DICGC should pay depositors when there is a default, not when the RBI orders the bank into liquidation. This is in keeping with the spirit of what a deposit insurance scheme truly implies.

But there are conceptual issues with the deposit insurance scheme as well. For one, the deposit insurance premium is the same for all banks. It does not differentiate between banks based on the underlying risk. An HDFC Bank or the State Bank of India pays the same premium for deposit insurance as a cooperative bank. The stronger banks thus overpay and cross-subsidise the weaker banks. There is also another type of cross-subsidy – one where the large depositors subsidise the smaller depositors. This is because while the insurance amount is capped at Rs100,000, the insurance premium paid to DICGC is on the entire amount of deposit.

But by far the biggest conceptual issue with the current scheme is around the pricing of the risk of a bank failure. The current insurance premium is 0.1 percent of deposits. This is not based on some statistical model nor based on the historical experience of bank defaults. The premium does not change based on the state of the business cycle (it is neither pro nor countercyclical). And as discussed above, the premium is the same irrespective of the underlying risk. This does not mean that the premium being charged is low. All it means is that we do not know whether the premium being charged, in aggregate, is too high or too low or about fair. For example, the current insurance pool with DICGC is around $14 billion as against a total insured base of about $500 trillion – so the insurance pool is around three percent of insured deposits. The insured deposit base of the most vulnerable section of the banking system, the cooperative banks, is more than 3.5x the insurance pool with DICGC. Is this adequate to cover the scenario of a banking crisis (on the liability side)?

A related issue is the perverse incentives that underpricing of risk creates. If I am a small bank enticing depositors with 100-200bps higher interest rates, I am potentially adding risk to the system. But if I pay the same cost as the least risky entity in the system, my behaviour is being encouraged. And this problem will only worsen if insurance coverage is made more liberal. Akin to what happens when you make the car safer for the driver which perversely makes him drive less conservatively. It is in this context that an overly liberal deposit insurance scheme is counterproductive. The insurance coverage, in terms of monetary limit, although unchanged since the 1990s, is reasonably liberal. As per data from DICGC, the current deposit insurance scheme fully covers the deposits of over 90 percent of the depositors in terms of number and almost 30 percent of outstanding deposits in terms of value. This seems liberal enough in absolute terms, though the coverage has fallen in absolute terms.

So, all we need to do is link deposit insurance premiums to the underlying risk and have the deposit insurance being paid out immediately on default rather than on liquidation. Does that fix most of the problems with our deposit insurance scheme? Not quite. For, one it is unclear whether the DICGC has the necessary skills to do a proper risk-based assessment of banks. This is simply a reflection of the fact that DICGC has not been called upon to do a proper risk assessment of banks to date. And even if DICGC builds the capability, a key constraint will be the fact that it is a subsidiary of the RBI. Its risk assessment is likely to correlate strongly with what the RBI does. There is probably not going to be a lot of independence.

An elegant solution to deal with this problem and to address the issue of the (perceived) modest insurance coverage is to allow depositors to buy deposit insurance directly from the market. In effect creating a market for deposit insurance either as an insurance product or a financial market product akin to a CDS. Under this policy, depositors will pay a premium for insuring their deposits and they would get a pay-out whenever there is a default on the part of the financial institution - effectively whenever there is an RBI action restricting payment of deposits. Allowing depositors to buy insurance directly addresses several of the issues with current deposit insurance scheme:

First, it results in markets monitoring the risk of a financial institution – in addition to the RBI. And two people monitoring the riskiness is better than just one. And markets, while imperfect, generally spot problems earlier than regulators or rating agencies or policymakers.

Second, this places an explicit number on the riskiness of a deposit reducing the potential interest rate arbitrage. So, for instance, if a cooperative bank is offering 100bps higher interest rate than say an HDFC Bank but if the deposit insurance cost with that bank is also, hypothetically say, 100bps higher than with an HDFC Bank, then it is clear that the extra interest rate being offered is entirely due to higher riskiness inherent in those deposits. A depositor is then better placed to choose which institution to park his savings with.

Third, by creating a situation where a depositor can insure 100 percent of his deposits, it reduces the pressure on the regulator (RBI) to bail out depositors by merging a failing financial institution with another. This increases the possibility that financial institutions might be simply allowed to fail with all stakeholders taking losses. The bail-in clause that derailed the FRDI bill might then no longer be a political hot potato.

How will such a scheme work in practice? While it is ideal that state-provided deposit insurance is replaced with a depositor purchased, market-priced, deposit insurance, that is unlikely to happen. What though can happen is that the state or the regulator provides a floor level of deposit insurance applicable to all depositors (say at current level). Above that threshold, depositors would be able to buy insurance from the market. The government and the regulator would need to work to develop this market. This will probably require that initially the additional cover is provided by DICGC itself (but with the depositor paying the price for the additional coverage). But simply having an explicit cost associated with insuring deposits will be a good start to making everyone appreciate the potential underlying risk.