Recently, the FDIC updated the fee structure for its Deposit Insurance Fund. Though there is no doubt that strong efforts are needed to replenish the fund — its value dropped below the previous 15-year low-water mark last year and has continued its plummet into negative territory — the current fee schedule places an undue burden on small banks, which are effectively subsidizing larger banks.
The DIF is supported by fees on insured banks, which are assessed by a risk-based schedule — that is, "based on an institution's probability of causing a loss to the Deposit Insurance Fund due to the composition and concentration of the institution's assets and liabilities, the amount of loss given failure and revenue needs of the DIF" according to the Federal Deposit Insurance Act.
We believe the discussions around the assessment of the fee have focused too narrowly on the probability of an institution causing a loss to the fund, with little attention being paid to capturing the severity of such a loss at thresholds that would cause the fund the most pain i.e., the fees assessed should reflect a premium for the larger banks that, under duress, present the most solvency risk to the fund.
Such an approach would explicitly adjust for "name concentration," a term commonly used by credit portfolio managers to describe the risk that a large borrower will default and thereby lead to a noticeably large loss for the bank because of the sheer size of the exposure.
Since there is a higher likelihood that this large borrower would contribute to a possible tail event or "worst case" loss, it should attract a higher amount of capital as a percent of exposure.
This name concentration risk needs to be mitigated by hold limits (i.e., maximum lending amounts to individual borrowers) or compensated through risk-based pricing (i.e., an add-on to the credit spread).
The same should be the case for the DIF, which has name concentration of its own because of the insurance policy it issues around the deposit base of the largest financial institutions in the country.
Risk mitigation via hold limits is likely a nonstarter since it would mean significantly less insurance coverage for large banks, possibly leading to liquidity and funding challenges. However, the second tactic of risk-based pricing is certainly a real possibility.
The fee-assessment process for the DIF is very similar to the risk-based pricing process for banks. The goal is the same: properly compensate the institution for the risk that each entity brings to that institution, considering two elements:
• The amount that the institution expects to lose on the entity based on the entity's financial strength (the entity's balance sheet and income statement).
• A capital charge required to compensate the institution for a possible catastrophic, worst-case loss because of name concentrations (sheer size of the exposure).
The first component of the risk cost, the expected loss, is easy to compute. It is simply the probability of the entity defaulting, multiplied by the loss given default: EL% = PD% x LGD%.
For example, if the likelihood of a bank defaulting is 1% and the loss in the event of default is 25%, the EL% would be 25 basis points.
This first component of expected loss is where the focus has been.
Now, consider a case where two banks, X and Y, both have the same exact financials, so, their likelihood of failing is the same.
Using the example above, their EL% will be 25 basis point. However, X is one of the largest banks insured by the fund and contributes 5% of the total deposits insured by the fund, while Y is about a hundred times smaller.
Indeed, the four largest banks account for about 40% of the total deposits at FDIC-insured institutions.
A common point of confusion is that X is already paying more into the fund than Y on a total dollar basis (because of a larger deposit base).
However, we argue that X needs to actually pay more on a per-dollar-of-deposits basis to account for the name concentration risk it brings to the fund.
This second component can be quantified by measuring the contribution of the bank to the tail risk or "worst case" loss to the fund.
It is best captured in a credit portfolio model that summarizes this contribution in a metric called economic capital, which is a function of the bank's size and the solvency standard that the fund wants to achieve.
This number is then multiplied with the hurdle rate, which is the return rate taxpayers require for investing their capital with the institution, to get the capital charge: CC% = EC% x Hurdle%.
Suppose that the EC% reflective of the contribution of X is 2% and of Y is 0.5%, and the hurdle rate is 5%. This would translate into a CC% of 10 basis points for X and 2.5 basis points for Y.
As a result, the total fee assessment in this hypothetical example of two banks with the same financial strength would be 35 basis points for X and 27.5 basis points for Y.
These results illustrate the concept of name concentration and its potential to significantly alter the "fair economic price" that banks should pay to the fund.
The two prerequisites in implementing such an approach are strong credit portfolio analytics and data.
We believe that the Office of Financial Research can help address the issue of data.
Having dynamic, bank-level data will improve the FDIC's risk assessment, giving it a better indication of which banks are more likely to fail, where concentrations are building and when systemic risk is increasing.
The result is a better assessment of the risks to the DIF that will help it to appropriately charge for deposit insurance.