Developing Historical Losses

Dorsey Baskin / Graham Dyer

Historical experience is the basis for ALLL estimates;whether for incurred or expected losses, it is the starting point. Historical loss is usually measured using the charge-off rate over time, that is, the confirmation of losses that happened earlier in time. The breakdown of the loan portfolio for purposes of measuring historical losses must relate to and support the other components of the ALLL estimation methodology.

How long should the historical charge-off accumulation period be?

Three recommendations:

  • It should be at least as long as the loss discovery period (LDP) for the loan category being evaluated. Different risk grades will have different LDPs. The measurement of experience is specific to the factors that drive credit risk and how they affect credit risk over time, and should be tied to the loan type and support the discovery, measurement and application of adjustment factors.
  • It should not be much longer than the average loss discovery period. The bank should examine the correlations that gave rise to those losses and how they relate to the portfolio.
  • Variable charge-off accumulation periods should be avoided. Change adjustment factors, not the historical charge-off accumulation period.

Other suggestions:

  • Concentrations: If charge-off experience is being tracked for the commercial loan portfolio by risk grade and a growing portion of the portfolio is a specific type of loan, breaking out that loan type from the others is advantageous.
  • Risk categories: Use categories of risk to develop historical charge-off experience, but give consideration to how many risk grades are needed. Reevaluate the balance between the cost of more risk categories versus the added value.
  • Seasoning: The use of seasoning categories is advisable, especially for growing or shrinking loan populations.

A key to understanding the historical charge-off rate is that it needs to be calculated in a manner that it relates to historical economic conditions and underwriting policies. That is, the historical experience needs to be calculated so that correlations between the conditions that influenced the rate of loss and the resulting charge offs can be identified. The objective of finding these correlations is so the bank can look into the recent past (and for CECL, the future)and estimate how conditions and factors will translate into adjustments to the historical average charge-off rate. Development over time of quantified correlations between economic and other conditions and the resulting historical charge-off experience is the best empirical, objective and documentable means available to compute a reasonable estimate of the ALLL, under the current incurred loss or the coming expected loss model.