See our 2019 Risk Rating Default Assumptions.
Recommendations for Annual Loss, Credit Capital and Guarantor Factors
Background
As a new service in 2018, PrecisionLender (PL) provides industry-based recommendations for the three key inputs used to calculate economic capital. These factors are:
- Annual Loan Loss Percentage (sometimes referred to as Probability of Default)
- Economic Capital Percentage
- Guarantor Factor Percentage
Ideally, each client will provide unique portfolio data annually which PL uses to perform a credit migration study. The results of this study are then used to make updates to that specific client’s key factors.
Some of our clients have found it difficult to provide this data on an annual basis and have asked if we could share “Industry” data as an alternative. To meet this need we have complied credit migrations performed over the past several years to create “industrywide” groupings. The grouping specific to a client’s situation can be used as an alternative to client specific credit migration results. PL will continue to review these recommendation sets and the resulting groupings annually.
Scope
This article does not address any other assumptions beyond the three key inputs listed above. PL also provides guidance for Minimum Capital and Minimum Annual Loss amounts. Unmitigated Capital factors, to cover market and operational risk can be determined by publicly available Call Report information. For most of our community bank clients an Unmitigated Capital factor of 1% of asset balance is our usual recommendation. In addition, PL provides Usage Given Default (UGD) recommendations across the various risk rating templates.
Please let us know of any questions that may come up in incorporating these recommendations, and we will be happy to help. Please note, if your instance of PL is not currently configured to use a “multi-factor” approach (i.e. separately entering risk rating, collateral and guarantees) for capturing credit risk the standard assumption sets included herein should NOT be used. Please contract PL for further assistance.
Overview
PL presents seven categorizations of risk rating groups. Our research indicates that the key drivers of the migration-based credit risk assumptions are, by commercial & agricultural loan count:
- The most common risk rating used
- And the first “watch” category
The most common risk rating is determined by looking at the distribution by risk grade across the active commercial loan portfolio. We generally define a watch loan as a performing loan where the lending institution has credit or operational concerns about the borrower. If this watch loan were a new borrower with no existing relationships that bank would typically deny the loan based on these concerns.
As such, we have organized industry credit migration results into templates that may assist you in updating your credit risk assumptions.
PL notes the recommendations are based on a sample of over 35 financial institutions. These institutions have a median asset size of approximately $1.5 billion and are located across the United States.
Methodology
For each of the seven standard risk rating sets, PL aligned the credit migration results for Credit Capital, Annual Loss and Guarantor Factors across the financial institutions included in the sample data. Next, we evaluated “how” the representative institutions employ both the most commonly used and watch categories. Each of the component credit migrations employed Basel III methodology to determine the Annual Loss, Credit Capital and Guarantee Factor results. Hence, these recommendations are derived from the Basel III best practice methods.
PL aligned common risk rating meanings across the data set.
2018 Recommendations
PL recommends that the credit risk assumptions follow the term structure from 12 months to 120 + months. For the purposes of input in to the solution, we display the 12, 60, and 120+ month categories and interpolate for points in between using the interpolation method you have previously selected. You may always add duration points to the assumption tables.
A sample recommendation set and the UGD recommendations are presented below.
Usage Given Default Recommendations
