The new CECL standards fundamentally change the allowed loan and lease loss (ALLL) calculation for GAAP-reporting institutions.
In a previous blogpost we discussed the fundamentals of CECL in more detail and how these may affect financial institutions. In this post, we discuss the benefits of strategically leveraging the new requirements rather than just trying to meet the minimum standard. Specifically, we focus on our preferred methodology for estimating lifetime credit losses of financial assets, the probability of default/loss given default (PD/LGD) method.
Many (if not most) institutions traditionally rely on aggregated estimated measures of credit loss. Such methods lack precision due to no distinction being made among the credit risks of different products, borrowers and firms, and credit environments. These techniques may also rely heavily on the currently classified assets and may not be measuring other risks that may exist in the portfolios that have not yet emerged. While these methods may produce estimates that are adequate, loan loss reserves may not adjust appropriately as the product mix on the balance sheet and the credit environment change.
Imprecise loan loss estimation is inefficient in a number of ways. First, overestimation of credit loss will affect income and capital due to higher than needed reserves. Second, obviously underestimation of credit loss may create another layer of problems and leads to insufficiently hedged credit exposure for the institution.
Ideally, institutions’ allowed loan and lease loss (ALLL) reserves will match their actual credit loss. In addition, imprecise estimates of risk can lead to inefficient loan pricing and underwriting. Therefore, a more robust implementation with a more precise ALLL estimation represents an opportunity to reduce the aforementioned inefficiencies and improve institutions’ bottom lines.
What is PD/LGD and What Does it Require?
As the name implies, PD/LGD is an estimation of the probability of default (PD) for each individual loan and the consequent loss following default (LGD). When done properly, the results of PD/LGD estimation explain the credit risk inherent in each loan based on the borrower-, loan- and time-specific characteristics of the loan. When the probability of default is estimated over the life of the loan, PD/LGD complies with the new CECL requirements.
Appropriately estimating the PD and LGD requires regression modeling techniques using a bank’s historical credit loss data coupled with measures of borrower, credit-product and credit-environment attributes. Thus, PD/LGD requires more granular data than many smaller institutions currently retain and more complex modeling than some institutions traditionally engage in. However, the necessary data, technological and governance changes resulting from CECL implementation present both an opportunity and incentive to enhance the banks analytical capabilities.
In order to calculate ALLL using PD/LGD, institutions must first define default. The next step is estimating the probability of default (PD). Modeling default as a function of borrower characteristics, product characteristics, market and industry conditions, and the economic environment and outlook yields a means of predicting the likelihood of default in the future. The results from this estimation explain, for example, how credit score or collateral type affects the probability of default.
Next, loss given default (LGD) is estimated using similar techniques. LGD is the percentage of the loan’s value that the bank loses in the event of default and after recoveries.
The fourth step is determining the bank’s exposure at default (EAD). For term loans, EAD is simply the outstanding balance of the loan. However, for other credit products, like lines of credit, EAD calculation may be more complicated.
The product of PD, LGD and EAD yields the bank’s expected credit losses. Using PD/LGD to estimate the lifetime credit loss loan satisfies the new CECL requirements for lifetime credit loss estimation. More importantly, though, PD/LGD more precisely estimates the ALLL than methods commonly used by community banks today, thus providing very valuable business intelligence to the senior and executive management.
How Will PD/LGD Benefit My Institution?
Isn’t there a path of less resistance for community banks implementing CECL? In short, yes. Representatives from the Fed, the FDIC, the CSBS, the FASB and the SEC conducted a webinar (access here) in February of 2018 discussing methodologies available to community banks for CECL implementation. Institutions can modify the current loan loss estimation methodology to estimate lifetime loan losses rather than the shorter, often annual, estimations used currently.
The above is obviously one path and may be appropriate for very small and/or non-publicly traded institutions that have no near-term expansion plans. However, it is very likely “the bar will be raised” over time with respect to what is compliant and what is not. In addition, it should be stressed that CECL compliance will be evaluated by both agency examiners and auditors who may very well have different interpretations. So a minimalist approach may be more, “kicking the can down the road,” which is seldom wise as it is usually more costly in the long run and may be risky. In contrast, a proactive and planned approach will be more beneficial and carries far less risk.
The PD/LGD estimation of loan losses is a more precise estimation technique that will yield benefits for risk and capital management, pricing and profitability. We detail some of these, below:
The most obvious benefit of a more precise estimate of loan losses is improved risk management. PD/LGD is, after all, a methodology used to satisfy CECL standards, which are intended to ensure adequate reserves for, and that investors have adequate information about, credit risk. Insight into the factors driving credit losses can improve the underwriting and collections functions of the bank.
PD/LGD leads to greater efficiency in the management of capital. Using loan-level, lifetime estimates of credit loss to calculate the ALLL yields more appropriate levels of loan loss reserves than the results of alternative methods. In some cases, institutions’ ALLLs will increase because they have been underestimating their credit risk. In other cases, institutions’ ALLLs will decrease because they have been overestimating their credit risk. Regardless of the outcome, improvements to estimation precision results in reserve levels that more accurately match institutions’ credit risk, lower costs and/or reducing risk.
Loan pricing is probably the most significant and direct enhancement that can be gleaned from PD/LGD methodology. The PD/LGD quantifies how borrower, loan and environmental characteristics affect credit losses. This information can then be incorporated into loan pricing. For example, if customers with credit scores below 600 are associated with a five percent increase in the probability of default relative to other credits, the additional risk and capital costs can be accounted for through higher rates. Likewise, if specific forms of collateral result in greater LGDs on average, then prices should increase to reflect the risk and carrying costs associated with these loans.
Ultimately, profits increase when risk and capital are managed more efficiently and products are priced more accurately. Therefore, banks should view CECL implementation as an opportunity for improvement rather than just a requirement for compliance.
How to cite this blog post (APA Style):
Premier Insights. (2018, July 18). Leveraging CECL to Enhance Efficiency and Profitability [Blog post]. Retrieved from https://www.premierinsights.com/blog/Leveraging-CECL-To-Enhance-Efficiency-and-Profitability.