Determining the right level of loss using the historical loss method

The historical loss method uses an annualized average net charge-off rate incurred during a prescribed time period as a proxy for estimating future losses. The loss rate is derived either from losses incurred from the institution’s own portfolio or those incurred by a peer bank or a pool of peer banks. In the latter case, the data is typically derived from an analysis of recent call report data.

Historical loss is most commonly used by smaller institutions or by other institutions with statistically small portfolio segments. When considering the historical loss method, three items should be considered: the availability of data, the loss horizon and portfolio segmentation. 

The challenge with historical losses isn’t accessing the data, but rather transferring it for reporting purposes. Many banks and credit unions have either aggregated portfolio performance data in periodic spreadsheet exports, or they will reference historical data available in call reports. Yet, documenting that information into a transparent calculation can be difficult. 

Loss horizons can be challenging to develop because the horizon for each segment must be able to generate an annualized historical loss rate that captures losses to be incurred in the next year. How many years must be included, then, in the horizon? One, two or even five? 

Although it should be based on an institution’s own loss experience, a typical period spans either two or three years. During periods of economic growth or recession, financial institutions may consider a shorter horizon, as loss rates during stable times may be lower than normally expected. During stable times, a longer time horizon would be considered to incorporate higher loss rates that could occur if the portfolio quality changed during a recession. 

While institutions may take economic fluctuations into account, guidance states that a consistent loss horizon should be applied to the Allowance for Loan and Lease Losses (ALLL), making these periodic adjustments to the loss horizon difficult to defend. Instead, institutions should select a loss horizon that best fits their loss experience and consider making changes to qualitative factor adjustments as warranted.

Another critical consideration is whether to apply a common historical loss across all segments of the portfolio or to use unique loss rates for each segment. There should be a “…historical loss rate for each group of loans with similar risk characteristics in its portfolio based on its own loss experience for loans in that group…” according to recent guidance. Unless each segment has behaved similarly and has similar risk, each segment will require a unique loss rate. 

Furthermore, each segment may use a unique time horizon if its loss experience differs from the rest of the portfolio. Properly segmented pools using reflective loss horizons will be more accurate than a common loss horizon or loss rate.

Smaller institutions or smaller portfolio segments lend themselves to a historical loss calculation, given the availability of data and ease of use. Historical loss is generally sufficient, provided it incorporates a loss horizon that captures losses incurred from the appropriate economic environment and risk profile. 

In calculations encompassing mergers and acquisitions, institutions should follow FAS 91 guidance, which can incorporate existing historical loss methodologies or a different rate if reasons can be documented and justified.

To learn more about how to measure loss in the allowance, view our webinar on How to Determine the Right Measure of Loss.