Much press in the past decade has been devoted to the conceptreferred to as risk- based loan pricing, or RBL.

Most financial institutions claim they are engaged in some formof RBL. Turns out that most of these institutions could well befooling themselves at their peril.

Risk-based loan pricing when done properly has measurablepositive returns on investment (as much as 35 basis pointsimprovement in ROA). And there lies the crux of the debate – notmany RBL models actually use measurable, stochastic methods toprice consumer loans.

Let's look at some elements necessary for an RBL program to:

• Improve the ability to profitably offer loans to borrowersacross the full spectrum of FICO scores;
• Decrease subsidies high FICO score borrowers typically providelow FICO score borrowers, thereby improving the ability to providecompetitive rates to all borrowers regardless of FICO;
• Assure receipt of interest sufficient to cover the increasedrisk in making lower FICO grade loans;
• Provide for increased interest income to significantly enhanceROA,

Almost all financial institutions have attempted some sort ofpricing system that adequately prices loans according to creditgrade (risk). Most of these models lack accurate measurement of thecosts distinct to each credit grade. Since these models are absentaccurate cost identification, at best they could be referred to astiered pricing.

Empirical, stochastic methods are necessary to assure a returnadequate to truly offset the risk and costs associated withlower-grade loans and to assure the higher-grade borrowers are notsubsidizing riskier loans. Pricing loans using tested empiricalmethods assures all costs associated with lending are identifiedand allocated according to the unique risks and costs each gradeposes.

To assure accurate pricing of loans according to risk,stochastic methods need to be utilized that assure:

• All costs associated with loan programs are quantified andapplied to rates;
• Costs are accurately measured and assigned to each credit gradeindependently to assure risk is quantified;
• The replacement cost of money is measured and assigned toappropriate loan terms;
• Profit margins (in excess of costs) are measured accordingly asrates are set and evaluated;
• Potential losses and cross grade subsidies are identified andcorrected.

By identifying, quantifying, and applying costs unique to eachcredit grade, loan performance and earnings will be maximized. Thefour key costs associated with lending are:

• Cost of funds (consistent across grades);
• Loan processing and management (mostly consistent acrossgrades) ;
• Collection expenses (variable across grades);
• Charge-offs (variable across grades).

Since the crux of accurately pricing loans according to risk isdependent on distributing costs meticulously, the following processis necessary for each key cost:

Cost offunds:

COF (deposits and borrowed funds) is calculated as a blendedrate. COF is the starting point for RBL.

Processing andmanagement:

Activity-based costing (ABC) or some equivalent method should beused to identify costs, both direct and indirect associated withthe lending process. With costs identified by loan type, astatistical process can be used to convert costs into an effectiveinterest rate.

Collections:

Again using ABC, a statistical model using past experience anddata is developed to identify collections costs incurred accordingto each credit grade at the time of loan origination. These costsare also converted to an interest rate.

Charge-offs:

Using a three year average, a statistical model is used todivide charge-off costs among credit grades (credit grade atorigination). These costs are then converted to interest rates thatcan be applied to each unique credit grade.

Once it has been established what the interest rate needs to beto break even on each credit grade and type of loan, the financialinstitution next needs to establish what margin they want to add tothe break-even point.

This cost + margin is the starting point for setting rates oneach grade and type of loan. Financial institutions can then use awell-designed, robust spreadsheet using this data to perform “whatif” pricing strategies.

Finally, it would be well to bear in mind the direction given byregulatory agencies regarding risk-based lending processes. Theyuse a concept referred to as “reasonableness” in determining thevalidity of a financial institution's RBL and its adherence toregulations.

Besides frequent testing to determine the accuracy of originalassumptions that were used to establish an institution's RBL, itmust be shown that credit pricing and decisions are supportedby:

• Actual experience with loans having similarcharacteristics;
• An empirically derived, demonstrably sound statisticalanalysis;
• Industry-wide data available from outside credit reportingservices;
• A well-documented estimate of the servicing, counseling andcollection costs.

Financial institutions may learn too late that setting loaninterest rates according to risk-based loan pricing schemes usingany methods other than those that are stochastic and statisticallyvalidated, could well face legions of risks and problems.

Dennis Child is a retired credit unionCEO in Logan, Utah, who has been associated with Thompson Consultingand Training in Boise, Idaho, for 25years.