CECL Overview and Analytics Catalyst Corporate Webinar

Transcription

1 CECL Overview and Analytics Catalyst Corporate Webinar 8/31/2017 Sponsored by: 1 PRESENTERS Ian Dunn, EVP Products Rachel Messick, Data Scientist 1

2 Introduction Allowance for Expected Credit Loss The main objective of the Accounting Standards Update is to provide financial statement users with more decision useful information about an institution s expected credit losses by requiring consideration of a broader range of reasonable and supportable information. 3 More Discussion and Disclosures Two biggest changes: 1. Life of Loan Loss Estimates 2. Forecasts Key Steps Allowance for Expected Credit Loss 1. Properly Segment the Portfolio 2. Individually Review 3. Collectively Review (Base Loss Methods) 4. Collectively Review (Q&E and Forecast Adjustments) 5. Stress Scenarios 4 2

3 Data Requirements 5 Data Requirements Origination Date Maturity Date Term Rate Balance Original Balance Credit Limit (if applicable) Senior Balance (if applicable) Original Credit Score Credit Score DTI (not required, but useful) Days Delinquent Original Value (if applicable) Borrower/Collateral Address (if applicable) Borrower/Collateral City (if applicable) Borrower/Collateral State (if applicable) Borrower/Collateral Zip Code (if applicable) Auto Make (if applicable) Auto Model (if applicable) Auto Year (if applicable) Direct/Indirect Indicator (if applicable) Gross Charge off Balance Gross Charge off Date Recovery Balance Recovery Balance Date Closed Date 6 3

4 Data Requirements By Loss Method Static Pool Vintage DCF/PD (Res. RE) DCF/PD (Auto) DCF/PD (Card) DCF/PD (Student) Loan ID C,H C,H C C C C Loan Type C,H C,H C C C C Orig. Date C,H C,H C C C C Maturity Date C,H C,H C C C C Term C,H C,H C C C C Rate C C C C Balance C,H C,H C C C C Orig. Balance C,H C C C Senior Balance C C Credit Limit C Orig. Cred. Score C C C C Credit Score C C C C DTI C Days Delq. C C C C Orig. Value C C Borr./Coll. Address C * C * C * C * Borr./Coll. City C * C * C * C * Borr./Coll. State C * C C C Borr./Coll. Zip Code C * C * C * C * Auto Make C * Auto Model C * Auto Year Direct/Indirect Net Charge Off Balance H H Charge off Date H H Close Date H H C C Collectively Review Loss Methods 8 4

5 Base Loss Methods Allowance for Expected Credit Loss 1. Static Pool Loss Rate Method 2. Vintage Loss Rate Method There are many CECL compliant loss methods. Visible Equity will focus on these methods, with potentially more to come in the future 3. Discounted Cash Flow (w/pd) 4. Probability of Default Method 5. Roll Rate Method 9 Case Study Loss Methods 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% Institution A ($100M) Institution B ($200M) Institution C ($300M) Institution D ($400M) Institution E ($500M) Institution F ($1B) Institution G ($2B) Vintage Static Pool DCF PD 1 Yr. C/O 5

6 Review Question #1 11 Static Pool: Steps 1. Determine remaining contractual life, L, of portfolio 2. Isolate pool of loans that were active at least L years ago 3. Sum up total losses that occurred in the pool during L year period 4. Divide total loss by balance of pool at starting point 6

7 Static Pool: Example New Auto Direct (5 year term) Period Beginning Balance Closed Balance Charge-off Balance Ending Balance ,000, ,000 5,000 7,500, ,500,000 1,475,000 25,000 6,000, ,000,000 4,485,000 15,000 1,500, ,500,000 1,427,000 3,000 70, ,000 70, Total 48,000 Loss Rate: $48,000 / $8,000,000 = 0.6% 2017 Active Balance: $10,000,000 Expected Lifetime Loss = $10,000,000 x 0.60% = $60,000* Static Pool: Pros and Cons Pros 1. Simple Calculations 2. Similar to current industry standard Cons 1. Data availability 2. Only using small part of history 7

8 Vintage: Big Idea Years 0.18% 0.35% 0.1% 0.05% 0.01% 0.0% 6 Year Loan Lifetime Loss Rate: 0.69% Vintage: yearly rate creation Loss Amount in Year Following Origination ($) New Auto Direct Denom. Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Balances (Denominator) Yearly Loss Amounts (Numerator) 8

9 Vintage: yearly rate creation Loss Amount in Year Following Origination ($) Denom. Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 New Auto Direct ,000,000 5, ,500, , ,200,000 10,000 5, ,500, , ,000, , , ,000,000 15,000 5,000 2, ,200,000 3, , , , ,000, ,000 3,000 2,000 1, ,000 5, , ,000 2, , Vintage: yearly rate creation Loss Amount in Calendar Year Following Origination ($) Denom. Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 New Auto Direct % 0.35% 0.1% 0.05% 0.01% 0% ,000, % ,500, % ,200, % 0.23% ,500, % ,000, % % ,000, % 0.5% 0.2% ,200, % , % % ,000, % 0.3% 0.2% 0.1% , % % , % % 0 0 Apply to current 2017 vintage balance: $2,600,000 * 0.69% = 17,940 9

10 Vintage: Big Idea Years 0.18% 0.35% 0.1% 0.05% 0.01% 0.0% 6 Year Loan Lifetime Loss Rate: 0.69% Vintage: Big Idea Years X X X X X 0.35% 0.1% 0.05% 0.01% 0.0% 6 Year Loan X Lifetime Loss Rate: 0.51% 10

11 Vintage: Big Idea Years 0.4% 0.15% 0.06% 0.015% 0.0% 6 Year Loan Lifetime Loss Rate: 0.63% Vintage: yearly rate creation Loss Amount in Year Following Origination ($) Denom. Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 New Auto Direct , , , , , ,000 2, , , ,000 3,000 2,000 1, , ,

12 Vintage: yearly rate creation Loss Amount in Year Following Origination ($) Denom. Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 New Auto Direct % 0.15% 0.06% 0.015% 0% % % % % % % 0.29% % % % 0.38% 0.25% 0.13% % % 0 0 Apply to current 2016 vintage balance: $1,600,000 * 0.63% = 10,080 Pros and Cons Vintage Loss Rate Pros 1. Uses all of history 2. Stable estimates Cons 1. Involved calculations 2. Different from current approaches 12

13 DCF-PD: Big Idea DTI Unempl oyment Size House Prices Rate What factors affect loan risk? LTV Age Credit Score Risk DCF-PD: Big Idea Goal: translate risk factors into monthly probabilities of default/prepay 13

14 Probability of Default: Cox Competing Hazard Model Inputs Equation Monthly P(Default) Monthly P(Prepay)? Use historical data The Specifics Calculations (LOAN ID: ): Month Orig. Balance ($) Balance ($) Credit Score Cur. LTV (%) Unemployment (%) % Change HPI (By Q) Estimated PD (%) Estimated PP (%) % 0.01% 1.0% % 0.011% 1.2% % 0.028% 1.5% % 0.039% 1.9% % % 3.0% % % 3.1% % % 4.2% % % 6.7% LTV = Loan-to-Value HPI = Home Price Index Q=Quarter PD = Probability of Default 14

15 The Specifics Calculations (LOAN ID: ): Mont h Estimated PD (%) Estimated PP (%) Payment Balance Coll. Value Cost To Sell (%) Net Proceeds Expected Cash ($) Disc. Exp. Cash ($) % 1.0% 1, , , % 62, , , % 1.2% 1, , , % 62, , , % 1.5% 1, , , % 62, , , % 1.9% 1, , , % 62, , , % 3.0% 1, , , % 67, % 3.1% 1, , , % 68, % 4.2% 1, , , % 68, % 6.7% 1, , , % 68, Total 75, PD = Probability of Default LGD = Loss given default $76,000 $75,000 = $1,000 Probability of Default: Pros and Cons Pros 1. Only need current data 2. Economic Q&E and forecast adjustments seamlessly integrated Cons 1. Unfamiliar 15

16 Review Question #2 31 Forecasts 32 16

17 Motivation Reasonable & Supportable Forecasts When developing an estimate of expected credit losses on financial asset(s), an entity shall consider available information relevant to assessing the collectability of cash flows. This information may include internal information, external information, or a combination of both relating to past events, current conditions, and reasonable and supportable forecasts. FASB, ASU Motivation Defaults vs. Unemployment (MSA) 17

18 Motivation Defaults vs. House Price Changes (MSA) Generating Forecasts Industry-produced: Federal Reserve National Unemployment 18

19 Generating Forecasts Industry-produced: Federal Reserve National Unemployment 4.5% Generating Forecasts Industry-produced: Federal Reserve Nat. Unemployment? 19

20 Generating Forecasts Industry-produced: relationship between MSA and Nat. Generating Forecasts Industry-produced 20

21 Generating Forecasts Statistical Generating Forecasts Statistical Forecast using statistical methods (ARIMA, Gaussian Process, Etc.) 21

22 Generating Forecasts Statistical Applying Forecasts How do you do it? It depends Implicit PD/DCF Explicit Static Pool Vintage 22

23 Review Question #3 45 THANK YOU FOR MORE INFORMATION PLEASE CONTACT: Nathan Quade Kathy Gensler