By: Tom Hannagan
This post continues my discussion of the reasons for going through the time and trouble to analyze risk-based pricing for loans. I mentioned before that the second general major justification for going through the effort to risk-adjust loan pricing as a normal part of the lending function is financial.
I thought it might help put this into perspective by offering rough numbers that relate to risk-adjusted profit performance, bottom line earnings and expand on the premise that risk has a cost. Lending, in the leveraged/banking sense, involves credit risk, market (interest rate) risk and operational risk. The fourth area, the risk of unexpected loss, is covered by capital. Unmitigated risk will eventually impact earnings and common equity. The question is when and by how much? It’s important to understand that the cost of risk mitigation efforts depend on the various risk characteristics of the bank’s loans and loan portfolio.
The differential cost of market risk
As an example, a floating rate loan that reprices every month involves little market risk, requiring little if any expense to offset. Compare it to a five-year fixed rate, interest-only loan that involves greater exposure to market risk. That risk costs something to offset. The difference in annualized marginal funding cost ranges widely depending on the steepness of the yield curve on the date the loan is closed. The difference between Federal Home Loan Banks 30-day rates and five-year bullet funding today, for instance, is close to 200 basis points. If risk-based loan pricing models don’t reflect this difference by using a matched marginal funding cost, the bank is voluntarily assuming some or all of the market (or interest rate) risk. Multiply an implied 200 bps risk-based funding cost difference by $100M in average loan balances and the implied annualized additional risk-free funding expense is $2,000,000. Multiply that by the average life of the portfolio to get the full risk-adjusted cost difference that the bank is assuming. And that’s just for the market risk.
The implied cost of credit risk
A loan with a pass risk rating of ‘2’ involves a lower likelihood of defaulting than a loan with a pass risk rating of ‘4.’ The lower risk (grade 2) loan, therefore, involves less of an Allowance for Loan Lease and Losses reserve requirement and an implied lower provisioning expense than the higher risk (grade 4) loan.
- Depending on the credit regimen and net loss experience of a given bank, the difference in the implied annualized expected loss due to credit risk could be 40 bps or more.
- Multiply the implied 40 bps credit risk cost difference by $100M in average loan balances and the implied annualized additional risk-adjusted credit expense is $400,000.
- Multiply that by the average tenor of the portfolio to get the full risk-adjusted cost difference to the bank.
The implied difference in administrative (or operations) expenses
These expenses include all mitigated (insured) operational risk. An owner occupied commercial mortgage is normally much less expensive to monitor than a line of credit backing a construction project. Those cost differences often range into several thousand dollars per annum.
- If, in our example of the $100M portfolio, our average credit is $400K, then we have around 250 loans.
- These loans multiplied by $3,000 in fully-absorbed annual non-interest expense differences would amount to $750K. A competent risk-adjusted loan pricing effort would take this cost difference into account.
- Again, multiply that yearly amount by the average life of the portfolio to get the full cost difference that the bank is incurring.
In reality, the three sample portfolios above would not overlap perfectly. The total actual assets from the above examples would lie between $100M and $300M. However, the total pretax cost difference of these three sample risk-based costs adds up to $3.15M per annum. The after-tax negative impact on risk-adjusted earnings is therefore about $2M yearly. So, the impact on ROA would be between 2.00% (if the three portfolios overlapped perfectly, for $100M in total assets) down to .67% (if there was no overlap, for $300M in total assets). This is a huge difference in earnings, on a risk-adjusted and fully cost-absorbed basis.
Finally, the amount of risk-based capital needed to back loans with differing risk characteristics, for purposes of unexpected loss, can be substantially different. This can be looked at as a difference in the implied cost of capital or in the performance ratio of ROE.
- In a simple application, the implied required equity might range from say 6% on the lower-risk loans up to 8% for moderate risk (average pass grade risk rating).
- If the portfolio in question is earning 1% ROA, the difference in risk-based equity would result in an ROE of either 12.5% for the higher risk loans versus 16.7% for the lower risk loans.
- The differences in fully risk-based ROE, or RAROC, could easily be more dramatic than this.
As stated before, if these differences are not “priced” into the loans somehow, the bank is not getting paid for the risk it is incurring or it is charging the lower risk borrowers a rate that pays for the added risk expenses of the higher risk borrowers. The business risk to the bank then becomes losing the better clients over time rather than attracting the riskier deals.
An economic look at performance
We are not talking in terms of “normal” accounting practices or “typical” quarterly reporting periods. We do use general ledger numbers to start the analysis process by relying on actual balances, rates and maturities. But, GAAP doesn’t address risk. So the risk adjustments are a more “economic” look at performance. Eventually, the risk reduction approach and the GL-based results will even out. The question is not “if” risk will eventually surface, but when and how it will manifest itself in GL results. We’ve seen a lot of this in the news the past eighteen months – and there’s likely more to come as the economy is in a downturn phase.
Going through the effort is worth it
Once risk is created by making a loan or placing a bet, someone owns it. The reason to go through the effort to price loans (and relationships) on a fully risk-adjusted basis is to understand the impact of risk at the only point in time when you can do something about getting paid for it – at the time the loan is agreed upon. After that, the bank is pretty much along for the ride. Risk-adjusted pricing is smart banking. It not only puts some teeth in the bank’s already existing risk management policies, it is justifiable to the client and it makes sense to most lending officers.