ALLL: Back to Basics Part I • What is the ALLL? • Incurred vs Expected • ASC 310-10 • ASC 450-20: Segmentation
Part II • ASC 450-20: Quantitative
• ASC 450-20: Qualitative
ASC 450-20: Quantitative Normally, an institution should determine the historical loss rate for each group of loans with similar risk characteristics in its portfolio based on its own loss experience for loans in that group. While historical loss experience provides a reasonable starting point for the institution’s analysis, historical losses, or even recent trends in losses, do not by themselves form a sufficient basis to determine the appropriate level for the ALLL. 2006 Interagency Policy Statement on the Allowance for Loan and Lease Losses
ASC 450-20: Quantitative Basic Calculation for Historical Loss Rates (Charge-offs – Recoveries) Balance
ASC 450-20: Quantitative Homogeneous Historical Loss Calculation Many institutions choose to average the aggregate net charge-offs (chargeoffs-recoveries) on an annual or quarterly basis over a defined period of time (the look-back period) and then divide this total by the average balance over the same period. If this is done using quarter-by-quarter balances, the final result should be multiplied by 4 to produce an “annualized” loss rate.
ASC 450-20: Quantitative Homogeneous Historical Loss Calculation • Net Charge-offs: Charge-off and recoveries that occurred in the quarter or year listed for loans in this segment • Balance: Average (or ending) total balance for the quarter or year listed in this segment
ASC 450-20: Quantitative Quiz Question: Why are ASC 450-20 allowances calculated using loss rates determined based on annual loss rates?
Loss Emergence Period Loss Emergence Period
Incurred Loss Event
Loan Origination
Bank Confirmation Of Loss
Bank Recognition Of Deterioration
Loss Emergence Period • Should reflect the best estimate of the average distance between incurred loss events and confirmation of losses • May (and probably should) vary between different loan pools or segments
Homogenous Historical Loss Rates Advantages
Disadvantages
• Easy to implement
• Slow to react to changes
• Data is easy to acquire
• Doesn’t differentiate loss rates based on risk metrics
ASC 450-20: Quantitative Cohort or Migration Analysis Tracking groups of loans based on a particular criteria (“cohorts”) over a series of periods and then analyzing the eventual net losses associated with that group of loans. This establishes a loss rate which can then be applied to loans matching those same criteria going forward.
Cohort or Migration Analysis
• Basic calculation is the same: Net Charge-Offs divided by Balances • However, the losses are aggregated by matching them up to the corresponding periods in the cohort rather than simply the quarter in which they occurred
Cohort or Migration Analysis • Basic calculation is the same: Net Charge-Offs divided by Balances • However, the losses are aggregated by matching them up to the corresponding periods in the cohort rather than simply the quarter in which they occurred • The losses will be accumulated back to the corresponding periods based on your Loss Emergence Period
Cohort or Migration Analysis
Step 1: Choose Cohort Criteria
Step 2: Determine Beginning Cohort Period
Cohort or Migration Analysis Step 3: Determine Initial Cohort Populations Step 4: Identify Net Charge-Offs over Loss Emergence Period to loans in initial cohort population
Cohort or Migration Analysis Step 5: Divide Net Charge-Offs by initial cohort population balance Step 6: Apply cohort loss rates to loans matching cohorts in current portfolio
Cohort or Migration Analysis
In an advanced cohort analysis, the results of several different beginning cohort periods would be averaged in order to achieve a better result.
ASC 450-20: Qualitative While historical loss experience provides a reasonable starting point for the institution’s analysis, historical losses, or even recent trends in losses, do not by themselves form a sufficient basis to determine the appropriate level for the ALLL. Management should also consider those qualitative or environmental factors that are likely to cause estimated credit losses associated with the institution’s existing portfolio to differ from historical loss experience… 2006 Interagency Policy Statement on the Allowance for Loan and Lease Losses
ASC 450-20: Qualitative In an ideal world, quantitative models would produce perfectly accurate allowance results without any adjustments necessary. We don’t live in an ideal world, so any quantitative allowance model will require correction based on current conditions and recent trends which may not be reflected in your quantitative model.
ASC 450-20: Qualitative
What factors should I be looking at?
ASC 450-20: Qualitative • Changes in Lending Policy
• Changes in Credit Concentration
• Changes in Nature/Volume of Portfolio
• Changes in Economic Conditions
• Experience; Ability; Depth of Mgmt
• Collateral Value Changes
• Past Due; Non-Accrual; Graded Loans
• Competitive; Legal; Regulatory
• Quality of Loan Review System
Environment
ASC 450-20: Qualitative Key Considerations • Every factor may not be appropriate for every situation/segment/institution • Don’t double up and adjust for a factor that may already be taken into account in your quantitative model (e.g. risk rating changes) • Qualitative factors are the “margin for error” on your qualitative model
ASC 450-20: Qualitative Rating Scale
Basis Point Adjustment
• Reduces guesswork
• More flexible
• Easier to defend
• Zero initial setup
• More difficult to manipulate
• Regulatory low-hanging fruit
• Initial setup can be complex
• More management discretion
Qualitative Rating Scale Setup • Determine “margin of error” for quantitative model • Identify desired qualitative adjustment factor • Determine “weight” of each factor
• Identify key performance indicators for each factor • Track factor KPIs over time to adjust factor ratings
Qualitative Factor KPIs • Some will be easy (e.g. economic factors), some may be very “fuzzy” • Continue to reevaluate over time and add/remove as necessary • Short to mid-term trends may be more important than point in time measurements