Author: Jon Kozlowski, firstname.lastname@example.org
At most community financial institutions, deal pricing is typically set by loan officers, and approved by loan committees, in a process that essentially lets the institution’s competitors price their deals; experienced lenders generally have a good sense for what rate/fee proposals will work on different types of deal structures in their markets, and look to meet or beat what competitors are likely to bid on the same deal. One of the biggest revelations institutions implementing a pricing model for the first time have is how frequently competitors misprice deals. This could mean either “giving it away” and not being adequately compensated, or realizing there were quality credit opportunities the institution could have bid more competitively on and still enjoyed strong profitability.
Loan and deposit pricing models vary in their methodology, but all attempt to quantify the various costs and risks of extending credit or raising funding in the current rate environment and generate projected rates of return. A conceptually sound and analytically robust pricing model can bring pricing errors to light. Most of these errors fall into one of several common categories:
- One size fits all pricing
Many of the costs of extending credit are proportional to a loan’s size. However, operating expenses are inelastic. While in many cases originating and servicing a $100 thousand loan is less than a $1 million loan, it is not 1/10 the cost. This leads to progressively higher required rates/fees to maintain a consistent rate of return on the portfolio.
The mark-ups achieved on smaller versus larger loans varies widely by institution, but in most cases don’t approach what is required for consistency in rates of return. This is unavoidable to some extent: “relationship pricing” is often used to justify sub-optimal pricing at the loan level to ensure the retention of valuable customer relationships. But while this may often be justifiable, the larger point here is that many institutions are largely ignoring this critical variable entirely.
- “Personalizing” Assumptions
When implementing a pricing model and quantifying the variables for the first time, the temptations are to think of these in terms of the institution’s experience: its funding costs, its credit experience, and its operating costs. This is a mistake, as pricing assumptions should always reflect marketplace norms for competitors in the institution’s footprint. To not do so would lead to systemic over- or underpricing based on whether the institution is worse, or better, than its competitors.
To illustrate, let’s assume an institution suffers from unusually high operating expenses. The consequences of using these in their pricing model could only be that because of these higher costs, the institution will need to charge higher rates/fees than its competitors to recoup them. The likely result is getting out-bid on most deals. Customers cannot be expected to be sympathetic to the institution’s cost problems, and be willing to pay more. Conversely, if the institution out-performs its peers by whatever metric, the result of embedding this experience as a pricing assumption will lead to overly aggressive bidding, ending in passing along the benefits of this out-performance to its customers rather than to its shareholders.
- Under-Valuing Customer Options
Most financial products contain embedded customer options. Interest rate floors and caps, as well as the ability to prepay a loan or call a time deposit, are common examples. The value of these options are often either ignored or quantified in rudimentary ways, e.g., a 3%-2%-1% penalty structure for prepayment within one, two or three years. While any attempts at quantification are (usually) better than none, most institutions under-value the risk these options create for financial loss.
The problem with most customer options is that they are notoriously difficult to quantify. Most institutions don’t have the resources to do such analyses in-house, nor do many pricing models perform this level of analysis. But since such valuations are being performed daily in the financial markets, an acceptable solution may be to receive indications on option valuation from the institution’s securities dealer or other service providers.
- Rosy Scenarios
Calculating a projected return necessarily requires making numerous assumptions. Some of these require individual judgment calls: for example, how long will a commercial mortgage with a 20-year term really stay on the books, or how much will a line of credit be utilized? Human nature being what it is, there is usually a bias towards the optimistic when making these assumptions, leading to unrealistically high projected returns, and ultimately underpricing. There are two potential solutions to this problem. First, with some factors management can provide standardized guidance for how to handle subjective variables. The second is an audit process to provide assurance of the reasonableness of assumptions used in preparing proposals.
- Stale Assumptions
All of the assumptions going into a rate of return calculation are subject to changing conditions. Assumptions such as capital requirements, loss experience and operating expenses should be evaluated and adjusted at least annually. The trend for all of these assumptions has been upward (more capital, higher expenses) so the more stale these assumptions become the more of an upward bias they place on results, leading to underpricing. As rate of return goals are another important input in a pricing model, these should be reevaluated along with the other inputs.
Evaluating pricing decisions in a financial model will always be an inherently subjective, and thus error-prone, process. The ultimate insurance against costly errors being made is the knowledge and objectivity of the practitioners building, calibrating and using it, as well as the reasonable and effective controls and guidelines put in place by the executives who will ultimately be held responsible for the results.