Written By: Aaron Lenhart
The role of a banker is becoming increasingly complex–gone are the days when a calculator and a vault were the tools of the trade. Today, customer funds can be transferred with smart phones, and many complicated back-end processes are made easier with automation. As result, the complexities for many institutions involve credit-risk management.
Bolstering risk management practices is a top concern for many bankers, and strengthening their allowance for loan and lease losses (ALLL) methodology plays a big part in that effort.
One way financial institutions can potentially improve their reserve methodology is to employ a more robust and accurate analysis for the general reserve (FAS 5) loss-rate estimate. Probability of default/loss given default (PD/LGD) is widely recognized as one of these more sophisticated methodologies.
PD/LGD has become an increasingly popular method used for calculating the ALLL because of its specific mentions in Basel II, Basel III and the coming CECL model proposed by the FASB. But this method is primarily popular among large financial institutions that have entire teams dedicated to compliance.
Generally, smaller institutions – including community banks and credit unions – typically use rolling charge offs to estimate the incurred losses in their portfolio. Since the rolling charge off (the historic loss rate) method is a more basic calculation, the data required is more easily accessible, and validation is more straightforward because it reflects the institution’s historical experience.
But easier isn’t necessarily better. PD/LGD calculations can offer a more sound and defensible ALLL. Implementing this methodology for the ALLL is typically viewed as a step toward more robust risk management, and it can give management deeper insight into what factors drive losses at the institution. Those insights can help the institution improve underwriting standards or collection processes, for example, to address either PDs or LGDs.
Calculating Expected Losses
Formulaically, PD/LGD is simple.
PD x LGD x EAD = Expected Losses
- PD (probability of default): the average percentage of borrowers who default over a certain period of time
- LGD (loss given default): the percentage of exposure to a bank if the borrower defaults
- EAD (exposure at default): an estimate of the outstanding amount, or exposure to the bank, if a borrower defaults
PD x LGD calculates the expected loss rate. The calculation of PD x LGD x EAD generates the total dollar amount of expected losses.
While the formula may be simple, the institution must first solve for the values to be used for PD, LGD and EAD.
Probability of Default (PD)
To determine probability of default (PD), an institution must settle on the definition of a “default” and determine over what length of time it should be measured. One example definition of default is 90 days past due, but other indications may be used as well, including possible definitions outlined in Basel II. Typically a one-year time period is used to asses PD. Once default is defined, PD can be measured and applied using various risk criteria. Risk rating is one common way to apply PDs, but other methods can use risk level, days past due or FICO.
Loss Given Default (LGD)
Like PD, there are a number of ways loss given default (LGD) can be measured. One way is to determine the percentage of loss by facility or collateral type. LGD estimates could also be driven, or influenced, by product type, industry or geography. Also like PD, LGD can be difficult to determine internally. Although inputs can be calculated using simple historical averages, once inputs are calculated they often require regression modeling, which can be challenging without appropriate resources. Consequently, financial institutions may find it easier to source data externally. If PDs or LGDs are acquired externally, an institution should document the relevance of the inputs when applied to its own portfolio.
Exposure at Default (EAD)
Generally, exposure at default (EAD) can be thought of as the borrower’s balance. However, further adjustments may be warranted. EAD could be the value of the financial asset today, or, depending on the product type, it could be lower or higher. The calculation may be further complicated by the economic environment.
The main benefit to financial institutions using PD/LGD is the simple calculation despite the complex inputs: the ASC 450-20 (FAS 5) general reserve can be easily determined within simple models that create directionally consistent expected loss numbers. That consistency is appealing to bankers, auditors and regulators.
Ultimately, the level of granularity with a PD/LGD analysis can allow a banker to estimate the ALLL on a loan-by-loan basis using the data that is most relevant and indicative of future losses for the institution. As the industry transitions from the current incurred loss model to an expected loss model (CECL), the advantages of the PD/LGD methodology could be even more pronounced.
As institutions of all sizes work to better manage their risk, PD/LGD is likely to become a more widely used approach to the ALLL calculation. Employing this methodology can provide key insights into credit losses and help strengthen an institution’s overall risk management program, giving bankers more power for granular risk analysis.
Originally published in October 2015 by Financial Managers Society. Used with permission.
Original link: http://fmsperspectives.com/2015/10/21/how-to-improve-your-alll-reserve-methodology/