PayNet Advanced Score of Scores ( PASS )

Transcription

1 PayNet Advanced Score of Scores ( PASS ) for Banking Taking the Risk Out of Small Business Lending November PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 1

2 Presenters Adam White Credit Process Director PayNet, Inc. Tom Ware Senior Vice President Analytics & Product Development PayNet, Inc. 2

3 PayNet Advanced Score of Scores ( PASS ) Agenda Introduction Overview of PASS model Common Challenges and Solutions with PASS Benefits of Combining Scores Performance & Impact statistics of PASS PayNet Overview Benefits of PASS & Summary 3

4 PayNet Advanced Score of Scores ( PASS ): A Score of Scores that allows lenders to leverage the scores/data they already use today PayNet MasterScore v2 Commercial Score A model which easily incorporates other scores, data and industry information Empirically Derived Statistically Sound score for lenders that may lack resources, time or sufficient historic data to build a statistical model Consumer Score PASS Demographic Info 4

5 Common Challenges: Underwriting smaller transactions is inefficient, expensive and therefore probably unprofitable High level of manual intervention requires significant resources and leads to increased operational risk Limited scores/data utilized in the scoring & decisioning platform Reluctance to change process, data sources or system for new models Inability to grow portfolio within your risk appetite 5

6 PASS Solutions: Enhance efficiencies and decision strategies with a single score Enhance Efficiencies Expand app-only threshold without sacrificing quality Increase predictiveness with a blended score for better risk discrimination Improve P&L with reduced losses and increased originations through additional lift Easy Deployment PASS Expand Threshold Easy deployment of PASS custom model provides for limited disruption Improve P&L Increase Predictive Lift 6

7 Combining Scores Taking the Risk Out of Small Business Lending PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 7

8 The Power of Combining Scores: Score A Very Low Low Below Average Above Average High Very High 4.6% 19.6% 9.5% 5.2% 2.3% 1.0% 0.3% Low 9.4% Score B Below Average 4.2% Above Average 1.7% High 0.5% 8

9 The Power of Combining Scores: Joint Odds Table: Score A Very Low Low Below Average Above Average High Very High 4.6% 19.6% 9.5% 5.2% 2.3% 1.0% 0.3% Low 9.4% 22.7% 13.0% 8.1% 4.0% 2.4% 0.5% Score B Below Average 4.2% 15.3% 6.6% 4.9% 2.6% 1.4% 0.4% Above Average 1.7% 12.5% 5.6% 2.1% 1.4% 0.5% 0.4% High 0.5% 8.1% 3.2% 1.1% 0.6% 0.2% 0.1% 9

10 The Power of Combining Scores: Joint Odds Table: Score A Very Low Low Below Average Above Average High Very High 4.6% 19.6% 9.5% 5.2% 2.3% 1.0% 0.3% Low 9.4% 22.7% 13.0% 8.1% 4.0% 2.4% 0.5% Score B Below Average 4.2% 15.3% 6.6% 4.9% 2.6% 1.4% 0.4% Above Average 1.7% 12.5% 5.6% 2.1% 1.4% 0.5% 0.4% High 0.5% 8.1% 3.2% 1.1% 0.6% 0.2% 0.1% 10

11 The Power of Combining Scores: Joint Odds Table: Score A Very Low Low Below Average Above Average High Very High 4.6% 19.6% 9.5% 5.2% 2.3% 1.0% 0.3% Low 9.4% 22.7% 13.0% 8.1% 4.0% 2.4% 0.5% Score B Below Average 4.2% 15.3% 6.6% 4.9% 2.6% 1.4% 0.4% Above Average 1.7% 12.5% 5.6% 2.1% 1.4% 0.5% 0.4% High 0.5% 8.1% 3.2% 1.1% 0.6% 0.2% 0.1% 11

12 Trade-off: Hit Rate vs. Predictiveness Predictiveness Hit-Rate The green data source above is most Predictive, but has a lower Hit Rate If only using one data source, forced into trade-off But if both, get the higher Predictiveness AND higher Hit Rate 12

13 PASS Model Performance & Impact Taking the Risk Out of Small Business Lending PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 13

14 PayNet Advanced Score of Scores (PASS) Lorenz Curve For transactions in the model development sample 100% 90% 80% Score of Scores Single Business Score Random Performance over Single Business Score Cumulative Percent of Bads 70% 60% 50% 40% 30% Single Business Score ROC K-S PASS % 10% Score of Scores ROC = 0.820, K-S = Single Business Score ROC = 0.735, K-S = Increase in Lift 36% 39% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Cumulative Percent of Population 14

15 Financial Impact over Single Business Score At a 70% Approval Rate: Could Increase Originations by 16% without increasing defaults P&L Benefit: $4,800,000 OR Could Reduce Losses by 26% without reducing approval rate Originations - $1B Loss Rate - 1% Current Losses - $10M Originations - $1B Increase in Orig - 160M Life Cycle Margin - 3% P&L Benefit: $2,600,000 15

16 Historical Bad Rates by PASS Score For transactions in the model development sample 50% 47.8% 45% 40% 35% 30% 28.6% Bad Rate 25% 20% 21.9% 16.5% 15% 10% 5% 9.0% 8.1% 2.2% 1.7% 1.4% 0.3% 0.0% 0% Bad Rate 47.8% 28.6% 21.9% 16.5% 9.0% 8.1% 5.8% 4.8% 2.2% 1.7% 1.4% 0.3% 0.0% 5.8% 4.8% 16

17 Historical Bad Rates by PASS Score For transactions in the model development sample 50% 47.8% 45% 40% 35% 30% 28.6% Bad Rate 25% 20% 21.9% 16.5% 15% 10% 5% 9.0% 8.1% 2.2% 1.7% 1.4% 0.3% 0.0% 0% Bad Rate 47.8% 28.6% 21.9% 16.5% 9.0% 8.1% 5.8% 4.8% 2.2% 1.7% 1.4% 0.3% 0.0% 5.8% 4.8% 17

18 Historical Bad Rates by PASS Score For transactions in the model development sample 50% 47.8% 45% 40% 35% 30% 28.6% Bad Rate 25% 20% 21.9% 16.5% 15% 10% 5% 9.0% 8.1% 2.2% 1.7% 1.4% 0.3% 0.0% 0% Bad Rate 47.8% 28.6% 21.9% 16.5% 9.0% 8.1% 5.8% 4.8% 2.2% 1.7% 1.4% 0.3% 0.0% 5.8% 4.8% 18

19 PayNet Advanced Score of Score (PASS) for Banking: Formula 19

20 PayNet Overview Taking the Risk Out of Small Business Lending PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 20

21 Overview of the PayNet Database One of the largest databases of private financial obligations: Over 24 million term debt contracts Over $1.5 trillion in financial obligations A complete spectrum of industry sectors: Services Manufacturing Retail Trade Transportation Agriculture Construction Wholesale Trade Finance Public Administration Mining 21

22 Term vs. Trade Term Debt Payment History provides a better indicator of Credit Quality and Capacity vs. Net 15/30 data Transaction Size Payment Frequency Term Long-Term Debt Larger Transaction Size Indicates Credit Capacity of borrower Regular Periodic Payments Consistent Cash Flow needed to Repay Payment size indicates ability to service debt Multi-Year Commitment Indicates Credit Quality through the economic cycle Short-Term Payables (Net 30) Smaller Transaction Size Lack of information about overall credit capacity Variable Payments Can be reduced quickly through cost reductions Month-to-Month Commitment Indicates operating activity (working capital) not leverage or cash flow Can a bank be comfortable making a term loan based on a cellular phone service payment history? Like Predicts Like 22

23 PayNet s Credit Scores Overview PayNet MasterScore v2 is PayNet s Third Generation of Credit Scores First generation completed in 2002 Hundreds of financial institutions have used PayNet scores Over the years, PayNet credit scores have been used to help decision millions of transactions worth over 200 billion dollars 23

24 MSv2: 29 Scorecards Healthcare Professional Transportation Construction Other Industries General Medical & Surgical Hospital Physician, Optometrist, Dentist & Veterinarian Chiropractor & Physical Therapist Social Services Banking, Insurance & Investment Accountant & Engineer Consultant, Lawyer & Other Professional Real Estate Agent & Loan Broker Transportation w/o Delinq. for General Lenders Transportation with Delinq. for General Lenders Transportation 5+ Yr. in Bus. for Captive/Specialty Lenders Transportation 0-5 Yr. in Bus. for Captive/Specialty Lenders Large Construction (top 3%) for General Lenders Small/Medium Construction for General Lenders Construction for Captive/Specialty Lenders Mega Borrower* Major Borrower (H/C > $100m) Medium Borrower (H/C $30m-$100m) Micro Borrower (H/C < $30m) Non-Profit Retail Manufacturing Agriculture *Includes the largest borrowers in i ndustries where large borrowers demonstrate very low default rates School, University & Non-Profit Food, Drug, Convenience & Miscellaneous Retail Manufacturing Large/Corporate Farm (H/C > $150m) Government Building Materials, Furniture, Electronics & Clothing Retail Printing & Copy Center Small/Hobby Farm (H/C < $150m) Motor Vehicle Dealer Forestry & Logging

25 MSv2: 135 Key Factors Impacting the Score Positive Impact Recent/overall prompt payments Consistent delinquency & credit quality patterns Negative Impact Recent/frequent delinquency Erratic/inconsistent delinquency or credit quality patterns 20 variables: Delinquency now, delinquency in past years, the portion of accounts that have been delinquent, overall avg. DPD, payment recency 8 variables: Variability of delinquency, seasonal payment patterns, trends in repayments practices & overall credit quality No serious delinquency history History of serious delinquency 11 variables: Extent, frequency, & severity of serious delinquency No defaults or derogatory public records Defaults and/or derogatory public records 9 variables: Bankruptcy, repossession, litigation, write-offs, magnitude of liens & judgments Long-time borrower Newer borrower 5 variables: Time in business, how long the business has been borrowing Large business size with extensive borrowings Lower risk borrowing practices Borrowings paid down Lower risk industry, state, and/ or local economic conditions Small business size with limited borrowings Higher risk borrowing practices Borrowings not paid down Higher risk industry, state, and/ or local economic conditions 16 variables: Borrower high credit & extent of financing & repayment, annual sales revenue, # of employees, incorporation status 23 variables: Credit utilization, # of lenders used, type of lenders used, length of lender relationships, pre-payment, portion of originations w/ new lenders, recent inquiries & borrowing activities, transaction structures, guarantees 6 variables: Absolute & relative extent of paydown on existing & prior transactions 37 variables: Borrower s industry, current performance in the borrower s state, current & long-term state & local economic & employment conditions 25

26 PASS Model Summary Taking the Risk Out of Small Business Lending PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 26

27 Benefits of PayNet Advanced Score of Scores Single, Comprehensive Blended Score Leverages, builds upon, and extends the credit scores your institution is currently using Easy implementation with limited disruption PayNet MasterScore v2 Commercial Score Provides extra lift for reduced losses and increased originations Consumer Score PASS Demographic Info Allows for increased thresholds and improved efficiencies due to its strong predictive power 27

28 PASS: What You Get! Custom Blueprint of the PASS model for you to implement the scorecard manually or in an automated environment Complimentary custom Retro/Swap Analysis performed by PayNet Analytics Consulting by the PayNet credit scoring team to assist in the implementation 28

29 THANK YOU Adam White Credit Process Director (847)

30 Appendix Taking the Risk Out of Small Business Lending PayNet Inc. PayNet and PayNet AbsolutePD are registered trademarks of PayNet, Inc. PayNet AbsolutePD Outlook is a service mark of PayNet, Inc. 30

31 PASS Calibration: 20 Points Doubles Odds, 650 = 10:1 Odds of G:B Score Good:Bad Odds X : 1 Default Rate =B/(B+G) Score Good:Bad Odds X : 1 Default Rate =B/(B+G) Score Good:Bad Odds X : 1 Default Rate =B/(B+G) Score Good:Bad Odds X : 1 Default Rate =B/(B+G) Score Good:Bad Odds X : 1 Default Rate =B/(B+G) % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % Bad/Default Definition: (within 36 mo. of orig.) - 2x60 dpd - Restructuring - Non-accrual 31