Washington | London | Amman | Johannesburg | Mexico City | Ramallah | Islamabad Risk Rating and Credit Scoring for SMEs March 27, 2012
Washington | London | Amman | Johannesburg | Mexico City | Ramallah | Islamabad
Risk Rating and Credit Scoring for SMEs
March 27, 2012
Introduction
DAI is a global development consulting agency, with 40 years of experience and offices in Washington D.C., London, Amman, Johannesburg, Mexico City, and Ramallah
With experience in 150 countries, DAI implements international development projects across the sectors of Corporate Services, Economic Growth, Environment and Energy, Governance, Health and Security
Introduction
DAI has extensive experience in management, economic and financial consultancy services to business and government clients throughout the world, focussing particularly on the area of MSME Lending and Leasing
These projects have included aspects of credit scoring and credit risk analysis, mutually reinforcing business strategies, structures, products, policies, and procedures and how to strengthen core functions, such as lending, risk management, marketing, internal control, and IT/MIS
DAI’s Experience in SME Credit Scoring
31 Financial Institutions
Key Benefits of Credit Scoring
Credit Scoring provides a consistent, quantitative estimate of borrower risk
Relative risk allows for differentiation in:
• the loan approval process• loan conditions and pricing• collection activities
Scoring leads to process automation (efficiency) and improved risk measurement (quantification) and management (consistency)
Working with Limited Data
Data availability affects how we weigh factors
Factor selection always involves expert judgment
Assuming we develop the best scorecard possible given our data and resources constraints, all scorecards are
again “equal” in that they must be monitored, periodically validated, and adjusted or re-developed
Scorecard Development
The greatest challenge is not statistical or technical ("accuracy") but rather human and organizational ("practicality")
How Scoring Works
Scoring models assume the future will be like the past– Based on historic data when available– Based on organizational experience in all cases
Scoring models should include the same set of key financial and non-financial risk factors that banks analyze subjectively
Assign points for the different risk characteristics – the point total for any given client is its “score”
How Scores Estimate Probability of Default
Group scores into some number risk classes
Evaluate borrower performance over time
Historic performance by risk class becomes the probability of default estimate for future periods
Example of Scoring Policy Table
Risk Class Decision Policy Interest Margin Historic PD
1 Branch Approval 1.25 0,00%
2 Branch Approval 1.75 0.08%
3 Branch Approval 2.25 1.9%
4 Branch Approval 3.00 3.6%
5 Central Review 5.00 6.8%
6 Exception Only 7.00 9.12%
7 Reject N/A 18.4%
Example SME Loan Process Without Scoring
Client Completes Application
Visit Client’s Place of Business
Check Credit Report / Minimum Criteria
Create Loan Memorandum
Present Loan to Credit Committee
30% 70%
Example SME Loan Process With Scoring
Client Completes Application
Visit Client’s Place of Business
Check Credit Report / Minimum Criteria
Further Review / Documentation at Next Approval Level
10% 60%Scoring
30%
Short-form documentation
Example of Time and Money Savings
Reduce time documenting 70% of loans
Reduce time spent by Credit Committee reviewing 70% of loans
Systematically eliminate the riskiest 10% of applicants
Scorecard Deployment
A web-based online “single-entry, multiple use” application processing system is the most appropriate long-term model deployment solution.
SME Lending Scorecards Summary
1. No “right” set of factors
2. Factors should generally be those that the Bank considers the most important when subjectively deciding to issue or not issue a loan
3. If the factors make sense individually, then as a group of factors, the scorecard will be able rank risk
SME Lending Scorecard Example Factors
CAPACITYTotal SalesLoan Size as % of SalesTotal Debt/EquityCurrent RatioInventory TurnoverInterest CoverageDebt Coverage Ratio
CHARACTERYears with BankBusiness Credit HistoryOwner's Personal Credit HistoryYears Experience in BusinessType of Legal Entity
CAPITALClient Contribution to FinancingAverage Bank Account Turnover / Turnover from Income StatementAverage Bank Account Balance
COLLATERALLoan To Value RatioType of CollateralPresence of Additional Guarantees
CONDITIONSSector RiskKey Buyer/Supplier Dependencies
How Statistical Scoring Works
For each potential factor, count the number of good and the number of bad contracts for across each possible “category”
Look for meaningful patterns of increasing/decreasing risk
Assign point weights equal to the concentration of bad loans per category
Models Rank Applicants by Risk
Single Factor Cross Tabulation: Risk = “Bad Rate”
Example Variable “Requested Loan to Turnover”
Higher Ratio is Higher Risk
4,000 Loans, 200 are “Bad”
Loan Size to Turnover # Good #Bad Total Bad Rate
< 10% 1,000 0 1,000 0.0%11 - 25% 970 30 1,000 3.0%26 - 50% 930 70 1,000 7.0%>50% 900 100 1,000 10.0%
TOTAL 3,800 200 4,000 5.0%
Points037
10
Models Rank Applicants by Risk
Full Model Cross Tabulation: Risk = “Bad Rate”
More Points Indicate Higher Risk
4,000 Loans, 200 are “Bad
Score Range Goods Bads Total Bad Rate0 - 14 200 0 200 0.0%15 - 28 397 3 400 0.8%29 - 42 792 8 800 1.0%43 - 57 1,179 21 1,200 1.8%58 - 71 745 55 800 6.9%72 - 85 348 52 400 13.0%86 - 100 139 61 200 30.5%
TOTAL 3,800 200 4,000 5.0%
20
Application Credit Scorecards and Basel 2
Basel 2 has spurred financial institutions to develop internal ratings based models to calculate expected loss based on predictions of its components:
Expected Loss = Probability of default (PD) x Loss given default (LGD) x Exposure at Default (EAD)
The Basel 2 accord does not specify any one methodology for the development of credit scorecard, but several overriding principles.
Thank you
Questions?