How to Improve Forecasting Under the New Credit Loss Standard

This article first appeared at FEI Daily.

The new credit loss standard, commonly referred to as CECL (Current Expected Credit Losses) or ASC 326, marks a significant change to how companies account for credit losses. ASC 326 consolidates guidance from various areas into a single model that requires an estimate of expected credit losses over the contractual life of a financial asset to be recorded on Day 1. While companies with extensive activities involving receivables—such as financial institutions—will be the most significantly impacted by the new standard, CECL will affect all companies that have financial assets measured at amortized cost on their balance sheet.

In order to comply with the new standard, companies are required to include reasonable and supportable forecasts in their measurement process. While forecasting in general is not a new concept, incorporating credit loss forecasts is new with ASC 326.

Here are some steps that companies can take to improve forecasting under the new credit loss standard.

Pooling of assets with similar characteristics

ASC 326 will require entities to group financial assets together based on similar risk characteristics for the purposes of measuring expected credit losses. Examples of risk characteristics are the type of collateral, the borrower credit rating, loan-to-value ratio, contractual life of the asset, aging of the asset, type of borrower, industry of borrower or geography of the asset. To increase the accuracy of the estimate of credit losses, companies should refine these pools by using as many layers of risk characteristics as possible. This will allow companies to streamline their forecasting to align with the specific asset pools and produce more relevant forecasts.

This will lead to a more thorough understanding of risks within the portfolio, which risks are already considered in the historical loss data, and how to most precisely measure those risks that are not currently included, therefore requiring an adjustment.

Gather and validate historical data early

In order to measure the credit loss, companies must understand historical losses by asset pool. Management teams should evaluate the completeness and accuracy of the data that will be used in evaluating the historical losses by asset pool. It may also be helpful to try multiple look back periods to determine the most appropriate approach for the asset pool.

The average credit lifecycle typically lasts around 7 years. Although the length of credit cycles may be similar, the extent of credit losses within each credit cycle can vary significantly by asset type. Some organizations may find it difficult to aggregate accurate and meaningful historical loss data at a sufficiently granular level for their current pools of financial instruments for the past seven years; those that can will only have historical loss experience for one credit cycle, which has been unique compared to other historical credit cycles. Companies should put in the time early on to determine what historical loss data is readily available and useful. Additionally, companies may have acquired other companies, implemented new systems, or sold new products and services to a new set of customers. Any of these situations can impact the availability of accurate historical loss information. Once a company has determined the extent of historical credit loss history that is available, judgments can be made on how best to utilize that loss information when determining expected credit losses under CECL. Companies that only have reliable historical loss data for a shorter period, for instance the past three years, will have to make more adjustments since most companies have experienced extremely low levels of credit losses during this current portion of the credit cycle. If companies don’t have adequate historical loss data, they may need to supplement with data from third-party sources.

Factor in forward-looking data

ASC 326 requires the use of forecasts in the measurement process that are both reasonable and supportable. This means taking into account current economic conditions and forecasts while also determining what forward looking time horizon is deemed to be supportable in order to make appropriate adjustments to the historical loss data. As part of this process, companies can consider entity-, borrower-, or asset-specific factors. Companies should also evaluate their current forecasting methodology to determine if the data used provides a reasonable basis for adjusting estimated credit losses. As these current conditions and forecasts are typically unreliable beyond a couple of years, companies will need to determine a methodology for reverting to historical loss rates over the remaining contractual life of a financial asset beyond when the forecasts are deemed reliable. Reverting to historical loss rates is required, but a specific methodology is not required. Companies can use an immediate reversion, a straight-line method over the entire contractual life of the financial assets or another reasonable method. The method chosen should be consistent and provide the most accurate measurement of credit losses for a pool of financial assets.

The extent of the forecasting process will vary by the contractual life of the asset as well as the methodology selected. Some methodologies do not naturally build quantitatively. When this occurs, management typically utilizes significant judgment and applies qualitative factors. These qualitative factors are often heavily scrutinized by auditors, given the judgement that goes into estimating the quantitative impact.  Additionally, some companies have indicated that they will purchase economic forecast data from reputable third-party firms. While this may provide enough data, interpreting the data and applying it to the individual pools of assets, or determining the quantitative amount of the qualitative factor, could be difficult.

CECL was established as a way of providing visibility and transparency for investors and stakeholders if and when another financial crisis occurs. Since it is impossible, however, to predict future losses with absolute certainty, companies should remain prudent in their forecasting to avoid significant variance and inaccuracies.