Stakeholders in the financial services ecosystem are familiar with the term “incurred loss”. It is a widely accepted and prevalent concept. However, with the advent of Current Expected Credit Loss Model (CECL), “expected loss” model is set to replace the “incurred loss” model.
For the uninitiated, CECL is the new accounting standard issued by Financial Accounting Standards Board (FASB) for the recognition and measurement of credit losses for loans, lease, guarantees, trade receivables, and debt securities. Although, early adoption is slated to start from 2019, entities registered with US Securities and Exchange Commission must implement it by 2020, while non-registered institutions have a two-year window.
One of the major changes ushered in by CECL has been discussed earlier. In order to understand the implication of this shift, let’s define both incurred loss and expected loss. The former is on account of an event that resulted in impairment to the loan. Usually, for purpose of incurred loss model impairment is measured in pools of loans and is based on historic annualized charge-off rates.
Expected loss ‘futuristic’ and is taken into account when the event, affecting the loan, has not occurred but there’s a probability. Unlike incurred loss model which is based on annual loss rate, expected loss model is on life of loan or life of portfolio loss rates. The following is a comparative chart highlighting the differences between the expected loss model and the incurred loss model –
|Incurred loss model||Expected loss model|
|Event causing impairment has occurred||Event may not have occurred, but there’s a likelihood|
|Based on annual loss rates||Based on life of portfolio loss rates|
|Reflects current losses||Reflects risk in portfolio for both current and future expected losses|
|Typically, based on pool of loans and doesn’t take into account product mix, loan terms (variable/fixed rate), etc.||Expected to vary based on products, borrower credit quality, loans terms (variable/fixed interest rate)|
|Considers contract life including estimated renewals||Considers only contractual life|
|Limited or no reference to historic data||Expected to refer to historic data include paid-off and charged-off loans|
|No impact on loan origination process||Impacts loan origination process as expected loss based on life loan is to be estimated. Factors influencing the loss such as collateral, LTV etc. may be closely scrutinized.|
With transition to CECL being imminent, the following requirements are key to actioning it –
The transition roadmap for CECL starts with rearticulating loan origination process and credit risk assessment. Subsequently, the data warehouse has to be modified to accommodate past, present, and future estimation data for facilitating loss calculations.
The development and testing of various models is in turn, dependent on multiple qualitative and quantitative factors. It’s incumbent on organizations to analyze and ascertain data requirements for various models, like historical loss, peer analysis, among others. This entails bringing about changes to regulatory reports, disclosures (past due status, non-accrual status), and compliance reports.
Limitations & Advantages of CECL
Post implementation, CECL is expected to pose certain challenges for organizations. For example, recording expected losses at the time of origination could affect processes and controls. The incurred loss model worked without historical data, whereas it’s essential for the expected loss model. Therefore, non-availability of the data during initial year(s) of implementation is an impediment.
Currently used models such as probability of default (PD), loss given default (LGD) have be modified to assess ‘life of loan’. In order to ascertain remaining age of loans, portfolios might be divided into multiple portions, leading to as many estimates.
Since, CECL takes into consideration factors like product mix, rate structure, and macro-economic indicators, various models have to be developed to estimate risks and calculate allowance. Organizations would not only have to maintain an audit trail of loss calculations, but also manage overlapping requirements of CECL, IFRS9, BASEL, CCAR, etc.
Despite the aforementioned challenges, CECL will streamline credit loss expectations by factoring in future expected loss to determine capital requirements effectively. The adoption of CECL will also drive efficiency in estimation and allowance booking. Finally, by emphasizing on historical data and data retention, there will be greater reliability on estimates.