* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Probability of Default for Microfinance Institutions
Survey
Document related concepts
Modified Dietz method wikipedia , lookup
Systemic risk wikipedia , lookup
Financialization wikipedia , lookup
Pensions crisis wikipedia , lookup
Yield spread premium wikipedia , lookup
Peer-to-peer lending wikipedia , lookup
Financial economics wikipedia , lookup
Moral hazard wikipedia , lookup
Securitization wikipedia , lookup
Continuous-repayment mortgage wikipedia , lookup
Credit rationing wikipedia , lookup
Syndicated loan wikipedia , lookup
Harry Markowitz wikipedia , lookup
Student loan wikipedia , lookup
Modern portfolio theory wikipedia , lookup
Transcript
Probability of Default for Microfinance Institutions SPTF Annual Meeting - Senegal May 2014 2 Moody’s and SPI4 Moody’s has been a supporter of the SPTF’s efforts in obtaining and reporting information on the social performance of MFIs » SPTF was represented on the Moody’s task force in the development of our Social Performance Assessment (SPA) » Moody’s has served on various working groups on social standards for the industry » Moody’s is committed to obtaining information on social performance issues that are standardized in order to report information and model risks of poor social performance » The Moody’s research on PD in microfinance is consistent with the data being collected through the SPI4 » Moody’s is interested in working with Cerise and the SPI4 to continue to model social data and look at its impact both on clients and the credit risk of MFIs Probability of Default Modeling 2 Agenda Modeling probability of default (PD) Social performance and default probability Probability of Default Modeling 3 1 Overview Probability of Default Modeling 4 Probability of Default by Moody’s Grade Importance of Calculating PD Pricing loans Investor return Portfolio risk Probability of Default Modeling 5 6 Developing a Model Deciding on Number of Factors for Scorecard Marginal Contribution to Accuracy Pos Helps Accuracy Recommended Range 0 For model building purposes, we may want to have more factors initially, with understanding that some will be discarded Hurts Accuracy Neg 4 8 12 16 20 n Number of Factors in Scorecard Probability of Default Modeling 6 2 Data Preparation Probability of Default Modeling 7 Overview of Data Preparation Data preparation involves collection of the required data, and deciding sources and systems to extract data. It also involves cleansing the data by removing financial statements that do not satisfy the following criteria: » Ratio checks: running the dataset through a series of data cleansing rules » Default definition: consistent definition of default has to be determined to properly classify the obligors of the underlying data into defaulters and non-defaulters » Determine the default horizon: determining a time window to classify the financial statements into defaults and non-defaults Above criteria ensure that the data contains information of all obligors and the information is consistent with the business segment for which the model is being built . Probability of Default Modeling 8 Defining Default Methodology for tagging financial statements as default » If financial statements were less than 3 month before default event then these statements were removed from the model development » If 2 statements were available from 4 to 21 months before default event then statement closer to default event was kept and tagged as default and other statement was dropped » If a defaulted obligor had a statement that was more than 21 months before default event then the statement was tagged as non-default Probability of Default Modeling 9 Basic Checks All statements were passed through a series of filtering criteria » Total Assets <=0 » Total Liabilities < =0 » Total Revenue <=0 » Total assets do not match to the sum of total liabilities and total equity reserves (a threshold of 2% was used) » Total Current Assets < 0 » Total Non Current Assets < 0 » Depreciation and Amortization < 0 » Total Operating Expenses < 0 » Total Long Term Liabilities < 0 » Cash and Equivalents < 0 FINAL DATA SAMPLE (Before Basic Checks) Total Statements: 868 Unique MFIs: 293 Defaults: 16 (1.84%) Basic Checks1 34 (3.9%) statements dropped FINAL SAMPLE FOR MODEL DEVELOPMENT Total Statements: 834 Unique MFIs: 292 Defaults: 16 (1.92%) 1. Refer appendix 11 for details of basic check analysis Probability of Default Modeling 10 3 Candidate Quantitative Factors: Single Factor Analysis Probability of Default Modeling 11 The available data yields 46 potential factors for single factor analysis Different sources were considered to come up with a list of candidate factors for model development » Microfinance Handbook by Joanna Ledgerwood » Microfinance Consensus Guidelines Published by CGAP/The World Bank Group, September 2003 Category Sustainability/Profitability Factor Name GrossMargin (Total_Revenue - Financial_Costs) / Total_Revenue OperatingMargin (Total_Revenue - Financial_Costs - Loan_Loss_Provision - Operating_Expense)/Total_Revenue ROE (Total_Revenue - Financial_Costs - Loan_Loss_Provision - Operating_Expense)/(Total_AssetsTotal_Liabs) ROA (Total_Revenue - Financial_Costs - Loan_Loss_Provision - Operating_Expense)/Total_Assets Operational_self_sufficiency Total_Revenue/(Financial_Costs + Loan_Loss_Provision + Operating_Expense) InterestCoverage CashtoLiabs RevenuetoTotalAsts Total_RevenueGrowth Total_Revenue/Interest and fee expense on all funding liabilities (v3210 ) Cash & Cash Equivalents – Audited (v1110)/Total_Liabs (Total_Revenue - Financial_Costs - Loan_Loss_Provision Operating_Expense)/Gross_Loan_Portfolio (Total_Revenue + Non_Operating_Income - Financial_Costs - Loan_Loss_Provision Operating_Expense - Non_Operating_Expense)/Gross_Loan_Portfolio Current_Assets/Current_Liabs Interest and fee expense on all funding liabilities (v3210 )/Gross_Loan_Portfolio Total_Liabs/(Total_Assets-Total_Liabs) Total_Liabs/Total_Assets Total_Liabs/(Total_Revenue - Financial_Costs - Loan_Loss_Provision - Operating_Expense + Depreciation and Amortization(v3530)) Total_Revenue/Total_Assets (Total_Revenue-Total_Revenue_Prev)/Total_Revenue_Prev GrossPortfolioGrowth (Gross_Loan_Portfolio-Gross_Loan_Portfolio_Prev)/Gross_Loan_Portfolio_Prev LoanPortfolio_CPIAdj Total_Assets_CPIAdj Avg_outstanding_loansize (229.601/CPI_INDEX)*Gross_Loan_Portfolio (229.601/CPI_INDEX)*Total_Assets (229.601/CPI_INDEX)*Gross_Loan_Portfolio/nb outstanding loans (v8040) Yield_on_Loan_Portfolio Gross_Yield_on_Loan_Portfolio Asset/Liability Management CurrentRatio Funding_expense_ratio LiabtoNetWorth LiabtoAssets LiabtoEBITDA Growth Size Calculation Probability of Default Modeling 12 The available data yields 46 potential factors for single factor analysis (cont’d) Category Factor Name Loan_officer_productivity Calculation number of active borrowers (v8050)/ number of loan officers (v8010) Personnel_productivity Branch_Productivity Type_Of_Loans number of active borrowers (v8050)/ Number of employees (v8020) number of active borrowers (v8050)/ Number of branches (v8030) (229.601/CPI_INDEX)*(Total_Revenue + Non_Operating_Income - Financial_Costs Loan_Loss_Provision - Operating_Expense - Non_Operating_Expense)/number of loan officers (v8010) (229.601/CPI_INDEX)*(Total_Revenue + Non_Operating_Income - Financial_Costs Loan_Loss_Provision - Operating_Expense - Non_Operating_Expense)/ Number of employees (v8020) (229.601/CPI_INDEX)*(Total_Revenue + Non_Operating_Income - Financial_Costs Loan_Loss_Provision - Operating_Expense - Non_Operating_Expense)/Number of branches (v8030) Number of loans outstanding(v8040)/number of active borrowers (v8050) Operating_Expense/Gross_Loan_Portfolio Financial_Costs/Gross_Loan_Portfolio (229.601/CPI_INDEX)*Operating_Expense/number of active borrowers (v8050) (229.601/CPI_INDEX)*Gross_Loan_Portfolio/number of loan officers (v8010) Portfolio at risk above 30 days (v7030)/Gross_Loan_Portfolio Of which portfolio at risk above 180 days (v7100)/Gross_Loan_Portfolio On-time portfolio (v7010)/Gross_Loan_Portfolio Write offs (v7140)/Gross_Loan_Portfolio Loan loss reserve – Audited (v1220)/ Portfolio at risk above 30 days (v7030) Loan loss reserve – Audited (v1220)/Gross_Loan_Portfolio Portfolio in arrears (v7130)/Gross_Loan_Portfolio reprogrammed and refinanced loans (v7115)/Gross_Loan_Portfolio mean(v8174,v8184,v7914,v7924,v7934,v7944) sum(Urban clients - volume of portfolio (v8410), Semi-Urban clients - volume of portfolio (v8420),0)/Gross_Loan_Portfolio Female clients - volume of portfolio (v8320)/Gross_Loan_Portfolio Financial revenue from investments – Audited (v3120)/Total_Revenue sum(Self-help groups (v8250), Solidarity groups (v8260), Communal banks loans/Self-help groups – volume (v8270))/Gross_Loan_Portfolio 6-nmiss(v8110,v8120,v8130,v8140,v8150,v8160) Loans_to_Ind_Types 10-nmiss(v8510,v8520,v8530,v8540,v8542,v8544,v8546,v8548,v8549,v8550) PBT_per_loan_officer Efficiency/Productivity PBT_per_employee PBT_per_branch Portfolio quality loans_per_borrower Operating_expense_ratio Financial_Expense_ratio Cost_per_borrower Avg_portfolio_per_credit_officer PAR_30_Ratio PAR_180_Ratio OnTime_Portfolio Writeoff_Ratio Risk_coverage_ratio LoanLossReserve_Ratio Arrears_rate Pct_Refinanced Avg_maturity_of_loans Pct_Urban_Clients_Volume Others Pct_Female_Clients_Volume Pct_Revenue_From_Investments Pct_Group_Loans Probability of Default Modeling 13 In general, factors are evaluated on the following set of criteria » Position Analysis: There must be enough observations. Observations where many values are missing typically indicate that the information is difficult to obtain. This information should therefore not be included in the final model » Factors must be intuitive. Experienced credit analysts should be familiar with the factor and its relationship with credit risk given the credit culture in which they operate » Factors must be consistent with expectations. Factor behaviour should be consistent with business judgment and any deviations in expectations should be easily explained » Factors must be powerful. The ultimate list of factors incorporated into the model should exhibit a high degree of discriminatory power on the basis of credit risk Probability of Default Modeling 14 Single Factor Analysis Performance: 21 factors recommended for further exploration in MFA Category Sustainability/ Profitability Asset/Liability Management Growth Size Factor Name AR* Default Rate Relationship Missing Recommend % ation GrossMargin 36% Good 2% OperatingMargin -13% Counterintuitive 2% ROE -5% Counterintuitive 3% ROA -7% Counterintuitive 3% Operational_self_sufficiency -11% Counterintuitive 2% InterestCoverage 37% Good 2% Yield_on_Loan_Portfolio -5% Counterintuitive 2% Gross_Yield_on_Loan_Portfolio -9% Counterintuitive 2% CurrentRatio -28% Counterintuitive 2% Funding_expense_ratio 39% Strong 1% Financial_Expense_ratio 46% Strong 1% LiabtoNetWorth 12% Good 2% LiabtoAssets 13% Good 2% LiabtoEBITDA -7% Counterintuitive 2% CashtoLiabs 19% Good 0% Total_RevenueGrowth 39% GrossPortfolioGrowth 38% LoanPortfolio_CPIAdj -13% Counterintuitive 0% Total_Assets_CPIAdj -14% Counterintuitive 2% 4% Weak 5% Avg_outstanding_loansize Comments High correlation with LiabtoAssets High correlation with LiabtoAssets High missing % High missing % Used as a proxy for Income level of the borrowers *AR = Accuracy Ratio Probability of Default Modeling 15 Single Factor Analysis Performance : 21 factors recommended for further exploration in MFA (cont’d) Category Efficiency/ Productivity Factor Name AR* Default Rate Relationship Loan_officer_productivity 23% Good 5% Personnel_productivity 27% Good 5% Branch_Productivity 18% Good 6% PBT_per_loan_officer -8% Counterintuitive 6% PBT_per_employee -17% Counterintuitive 6% PBT_per_branch 3% Moderate 7% RevenuetoTotalAsts 12% Moderate 2% Operating_expense_ratio 28% Good 0% Cost_per_borrower 19% Good 5% Avg_portfolio_per_credit_officer 6% Good 4% PAR_30_Ratio -8% Counterintuitive 4% PAR_180_Ratio -32% Counterintuitive 8% 1% Good 4% Writeoff_Ratio 8% Moderate 7% Risk_coverage_ratio 11% Moderate 6% LoanLossReserve_Ratio -1% Moderate 2% Arrears_rate -2% Weak 9% OnTime_Portfolio Portfolio Quality Others Missing Recommend % ation Pct_Refinanced 14% Avg_maturity_of_loans 23% loans_per_borrower 32% Strong 6% Pct_Urban_Clients_Volume 23% Good 0% Pct_Female_Clients_Volume 29% Good 5% Pct_Revenue_From_Investments -1% Counterintuitive Pct_Group_Loans 1% 20% Type_Of_Loans 3% Moderate 0% Loans_to_Ind_Types 10% Good 0% Comments High missing % High missing % Used as a proxy for Debt to Income ratio of borrowers High missing % Low diversity of responses and very low accuracy ratio Used as a proxy for portfolio diversity Probability of Default Modeling 16 PAR 30 Ratio Key statistics: Relative Entropy 0.96, Accuracy Ratio -8% Frequencies and Default Rates for PAR_30_Ratio 0.025 to 0.05 0.01 to 0.025 0 to 0.01 3.0% 200 2.5% 2.0% 150 1.5% 100 1.0% 50 0.5% 0 0.0% 1 0.75 % Default 0.05 to High 250 Default Rate Frequency missing CAP Curve of PAR_30_Ratio 0.5 0.25 0 0 Answer 0.25 0.5 0.75 1 % Population » This factor performs inadequately with no discriminatory power » Counterintuitive relationship between the responses and the default rate Probability of Default Modeling 17 PAR 180 Ratio Key statistics: Relative Entropy 0.96, Accuracy Ratio -32% Frequencies and Default Rates for PAR_180_Ratio 0.003 to 0.012 0 to 0 7.5% 7.0% 6.5% 6.0% 5.5% 5.0% 4.5% 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 200 Frequency >0 to 0.003 150 100 50 0 1 0.75 % Default 0.012 to High Default Rate missing 250 CAP Curve of PAR_180_Ratio 0.5 0.25 0 0 Answer 0.25 0.5 0.75 1 % Population » Counterintuitive relationship between the responses and the default rate Probability of Default Modeling 18 Avg_outstanding_loansize ? Key statistics: Relative Entropy 0.95, Accuracy Ratio 4% Frequencies and Default Rates for Avg_outstanding_loansize 500 to 1500 2500 to 4000 4000 to High 1 4.0% 3.5% 200 Frequency 1500 to 2500 3.0% 2.5% 150 2.0% 100 1.5% 1.0% 50 0.75 % Default < 500 Default Rate missing 250 CAP Curve of Avg_outstanding_loansize 0.5 0.25 0.5% 0 0.0% 0 0 Answer 0.25 0.5 0.75 1 % Population » This factor performs inadequately with low discriminatory power » Weak relationship between the responses and the default rate i.e. higher the score lower the default rate Probability of Default Modeling 19 4 Candidate Quantitative Factors: Multi Factor Analysis Probability of Default Modeling 20 Starting with 21 Candidate Factors from SFA Section Sustainability/ Profitability Asset/Liability Management Size Efficiency/ Productivity Portfolio Quality GrossMargin InterestCoverage 36% 37% Default Rate Relationship Good Good Financial_Expense_ratio 46% Strong LiabtoAssets CashtoLiabs 13% 19% Good Good Avg_outstanding_loansize 4% Weak Loan_officer_productivity Personnel_productivity Branch_Productivity PBT_per_branch RevenuetoTotalAsts Operating_expense_ratio Cost_per_borrower 23% 27% 18% 3% 12% 28% 19% Good Good Good Moderate Moderate Good Good Avg_portfolio_per_credit_officer 6% Good OnTime_Portfolio Writeoff_Ratio Risk_coverage_ratio loans_per_borrower 1% 8% 11% 32% Good Moderate Moderate Strong Pct_Urban_Clients_Volume 23% Good Pct_Female_Clients_Volume 29% Good Loans_to_Ind_Types 10% Good Factor Name AR Comments Used as a proxy for Income level of the borrowers Used as a proxy for Debt to Income ratio of borrowers Others Used as a proxy for portfolio diversity » As number of defaults are very low i.e. 16, we kept all the factors with positive accuracy ratio for MFA » Return ratios e.g. ROA and ROE are not present in the candidate factors list because MFIs typically operate on low return and higher base i.e. large assets Probability of Default Modeling 21 Pct_Female_Clients_Volume Key statistics: Relative Entropy 0.88, Accuracy Ratio 29% Frequencies and Default Rates for Pct_Female_Clients_Volume 5.0% 4.5% 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 500 Frequency 0.35 to High 400 300 200 100 0 1 0.75 % Default 0 to 0.35 Default Rate missing 600 CAP Curve of Pct_Female_Clients_Volume 0.5 0.25 0 0 Answer 0.25 0.5 0.75 1 % Population » This factor performs adequately with moderate discriminatory power » Good relationship between the responses and the default rate i.e. higher the score lower the default rate Probability of Default Modeling 22 1% 4% -6% -3% -12% -1% -15% -7% Funding_expense_ratio 34% 73% 13% 100% 3% 3% -32% 26% 8% 5% 6% 15% -7% -9% 75% LiabtoNetWorth 48% 18% 5% 3% 100% 97% 34% -32% -3% 4% -7% -18% 26% -7% 25% 13% 30% LiabtoAssets 48% 17% 6% 3% 6% -4% -15% 24% -7% 26% 13% 30% RevenuetoTotalAsts 50% 13% -2% -32% 34% 34% 100% -41% Avg_outstanding_loansize -27% 5% -2% 26% -32% -30% -41% 100% -8% -14% -16% 14% -19% 3% 7% Loan_officer_productivity 11% 10% 6% 8% -3% -1% 3% -8% 100% 47% 22% 18% -6% 2% 10% 23% -4% -6% Personnel_productivity 22% 13% 1% 5% 4% 6% 20% -14% 47% 100% 26% 18% 2% 6% 9% 35% 7% -7% Branch_Productivity 9% 8% 4% 6% -7% -4% 12% -16% 22% 26% 100% 18% -13% -2% 11% 17% 3% 1% PBT_per_branch 1% 13% -6% 15% -18% -15% 0% 7% loans_per_borrower 21% 5% -3% -7% 3% 12% 33% 14% 8% Operating_expense_ratio -10% -19% -12% -9% 0% 11% 4% 0% Financial_Expense_ratio 58% 53% -1% -5% 100% 5% -7% Cost_per_borrower 27% 5% -15% -7% 5% 100% 24% portfolio_per_credit_officer 29% -9% -7% -30% 30% 30% 46% -48% -4% OnTime_Portfolio 1% -11% -9% 97% 100% 34% -30% -1% 20% 12% 0% 22% 14% 18% 18% 18% 100% -15% 5% 2% -13% -15% 100% -10% -7% -7% -8% 3% 2% 6% -2% 5% 75% 25% 26% -15% 7% 10% 9% 11% 7% -10% 100% -5% 3% 13% 13% 33% -29% 23% 35% 17% 22% 12% 1% 0% -5% -17% -6% Writeoff_Ratio -10% -7% -10% 12% -5% -5% -24% -1% Risk_coverage_ratio 11% 14% Pct_Urban_Clients_Volume 0% Pct_Female_Clients_Volume Loans_to_Ind_Types -2% -9% -11% -7% 15% -4% 2% -3% 14% 15% -9% -10% -7% 13% -7% 7% 12% 14% 11% -8% -1% 1% -5% -5% -7% 14% -4% 0% -5% -5% -6% 15% -4% -5% -24% 4% -2% 20% -8% -29% -48% -17% -1% 0% 15% -41% -10% -2% 3% -1% 7% -9% 7% -3% 13% -10% -10% 10% 3% 17% -1% 22% -17% 11% 16% 26% 5% 2% -8% 2% -19% 5% -1% 2% -2% -1% 6% 7% 16% 16% 3% 5% -3% -7% -30% -4% -4% 7% -3% 16% -13% 26% -9% 3% -17% 33% 11% -7% 24% 100% 16% -4% 1% 1% -7% 1% 11% 14% 4% 7% 7% 16% 100% 43% 43% -13% 9% 9% 8% 0% 16% -3% -4% 43% 100% 14% -9% -8% -3% 2% 2% 16% 16% 1% 43% 14% 100% 1% -9% -10% 16% 14% -5% -5% 4% 0% 3% 7% 10% 26% -7% -6% -2% 15% -1% -3% 3% 15% 2% -7% -8% 14% 15% 20% -41% 7% 13% 17% 2% 5% -4% -3% 7% -1% -4% -8% -10% -4% -10% -1% -8% -1% 5% 0% 0% 7% 15% 13% 11% -4% 5% -8% -15% 33% 46% 26% 24% 22% -19% -6% -4% -7% 3% -19% 53% Loans_to_Ind_Types 6% Pct_Female_Clients_Volu me 5% -2% -10% 11% Pct_Urban_Clients_Volum e 13% -2% 1% Writeoff_Ratio 8% 6% OnTime_Portfolio 10% 13% 5% Cost_per_borrower 5% 10% 27% 100% 13% 21% -10% 58% 27% 29% Risk_coverage_ratio portfolio_per_credit_officer Financial_Expense_ratio Operating_expense_ratio 62% 100% 27% 73% 18% 17% 13% CashtoLiabs loans_per_borrower PBT_per_branch InterestCoverage LiabtoAssets 1% LiabtoNetWorth 9% CashtoLiabs 100% 62% 10% 34% 48% 48% 50% -27% 11% 22% InterestCoverage GrossMargin GrossMargin Branch_Productivity Personnel_productivity Loan_officer_productivity Avg_outstanding_loansize RevenuetoTotalAsts Funding_expense_ratio Candidate Factor Correlation Matrix -13% 20% 1% 5% 3% -13% -13% -13% -9% 1% 100% -2% -4% -1% 5% 26% 20% 9% -8% 1% -2% 100% 2% 6% -3% -9% 9% -3% 5% -4% -19% -2% 1% 2% 100% Probability of Default Modeling 23 Logistic Regression Models Model Number of Factors Significance Level1 AR2 Model 1 5 P Value <= 0.05 69.5% Model 2 8 P Value <= 0.1 77.8% Model 3 6 P Value <= 0.1 73.4% Best model after dropping Pct_Urban_Clients_Volume Model 4 4 P Value <= 0.05 65.5% Best model after dropping Avg_outstanding_loansize Model 5 5 P Value <= 0.1 69.4% Best model after dropping Avg_outstanding_loansize Comments » Due to low number of defaults we also considered models with 90% significance level of estimated coefficients » Pct_Urban_Clients_Volume represents percentage of urban and semi-urban borrowers of an MFI’s portfolio. Though this factors comes significant at 90% significance but we recommend not to include this factor in the model because MFIs typically have semi-urban and rural borrowers. Model should not penalize an MFI for having large base of rural clients » Avg_outstanding_loansize was used as a proxy for income level of borrowers of MFIs. But given low accuracy ratio of this factor we also considered models after dropping this factor which resulted in a drop of 6% in AR for model 4 and 11% for model 5 compared to model 1 and model 2 respectively 1. For estimated coefficients and p value refer appendix 1 2. AR = Accuracy Ratio Probability of Default Modeling 24 Beta Model – Factor Weights Section Sustainability/Profitability Asset/Liability Management Size Efficiency/ Productivity Portfolio Quality Others Factor Name GrossMargin InterestCoverage Financial_Expense_ratio LiabtoAssets CashtoLiabs Avg_outstanding_loansize Loan_officer_productivity Personnel_productivity Branch_Productivity PBT_per_branch RevenuetoTotalAsts Operating_expense_ratio Cost_per_borrower Avg_portfolio_per_credit_officer OnTime_Portfolio Writeoff_Ratio Risk_coverage_ratio loans_per_borrower Pct_Urban_Clients_Volume Pct_Female_Clients_Volume Loans_to_Ind_Types Number of Factors Model AR Factor AR 36% 37% 46% 13% 19% 4% 23% 27% 18% 3% 12% 28% 19% 6% 1% 8% 11% 32% 23% 29% 10% Model 1 Model 2 Model 3 Model 4 Model 5 22.4% 14.0% 18.2% 32.1% 26.7% 15.5% 7.9% 13.7% 11.6% 14.8% 8.6% 17.8% 14.0% 16.0% 25.1% 21.2% 17.6% 15.2% 19.5% 26.7% 14.4% 7.8% 19.5% 24.2% 23.4% 18.2% 14.6% 19.3% 5 69.5% 8 77.8% 6 73.4% 4 65.5% 5 69.4% » All models do not give any weight to sustainability/profitability and portfolio quality factors Probability of Default Modeling 25 5 Candidate Social Factors Probability of Default Modeling 26 New Data Preparation Qualitative (SPA Data) Quantitative (non SPA Data) Total Statements: Unique MFIs: Defaults: 731 249 16 Quantitative model prepared as before. Data for ‘Total Revenue Growth’ and ‘Gross Portfolio Growth’ updated for missing values Merging two datasets Total Statements: Unique MFIs: Defaults: (1.98%) 506 161 10 Total Statements : Unique MFIs: Defaults: 167 167 10 6 MFI dropped due to no exact match with quant data Total Statements : Unique MFIs: Defaults: 161 161 10 Qualitative Model was prepared on this data Remove statements from the quantitative data where MFI’s are not common to SPA (Qualitative) data 225 (30.8%) statements dropped Combined Model has been estimated on this data 1. Quantitative Models have been estimated on 731 records and 16 defaults 2. Qualitative Models for have been estimated on 161 records and 10 defaults 3. The combined model uses 506 records and 10 defaults Probability of Default Modeling 27 Candidate Social Factors Candidate social factors were based on availability of reliable data. Data sourced from the MIX and analyzed with Moody’s SPA Variable ProbChiSq Pricing Transparency Practices 0.463 AR 6% Disclosure of components of pricing 0.383 9% Manner of communication of pricing 0.106 16% Debt Collection Practices 0.059 27% Specific debt collection policies 0.218 17% Definition of acceptable and unacceptable collection practices 0.218 Voluntarily adopted consumer protection standards 0.060 Range of Products offered 0.159 24% Policies included in Code of Ethics 0.351 15% Written policies on hiring women 0.111 18% Corruption Score 0.098 19% 17% 27% Low AR Probability of chance occurrence is high Probability of Default Modeling 28 29 Rejected Social Variables Pricing Transparency Code of Ethics Greater than 0.9 20% 15% 10% 5% 0% Answer 80 70 60 50 40 30 20 10 0 Less than equal to 0.2 0.2 to 0.6 0.6 to 0.9 Greater than 0.9 15% Default Rate 0.5 to 0.9 Frequencies and Default Rates for Policies included in Code of Ethics Frequency 100 90 80 70 60 50 40 30 20 10 0 Less than equal to 0.5 Default Rate Frequency Frequencies and Default Rates for Pricing Transparency Practices 10% 5% 0% Answer Probability of Default Modeling 29 30 Accepted Social Variables Range of Products Offered Debt Collection Practices Frequencies and Default Rates for Range of Products offered 0.1 to 0.45 0.45 to 0.9 Greater than 0.9 20% 10% 5% 0% Answer CAP Curve of Debt Collection Practices Frequency Default Rate 15% 80 70 60 50 40 30 20 10 0 Less than equal to 0.2 0.2 to 0.4 0.4 to 0.6 0.6 to 0.8 Greater than 0.8 10% 5% Default Rate Less than equal to 0.1 100 90 80 70 60 50 40 30 20 10 0 0% Answer CAP Curve of Range of Products offered 1 1 0.75 0.75 % Default % Default Frequency Frequencies and Default Rates for Debt Collection Practices 0.5 0.25 0.5 0.25 0 0 0 0.25 0.5 % Population 0.75 1 0 0.25 0.5 % Population 0.75 1 Probability of Default Modeling 30 31 Combined Model Combining the Quantitative and Qualitative factors give an AR of 79.0% Section Quantitative Score Qualitative Score Section Weight Factor Cash to Liabilities Loans per borrower 64% Operating expense ratio Financial Expense ratio Percent Female Clients Volume Debt Collection Practices 35.6% Range of Products offered Factor Weight Final Weight 13.77% 8.9% 16.48% 10.6% 22.62% 14.6% 26.19% 16.9% 20.94% 13.5% 38.9% 13.9% 61.1% 21.8% Probability of Default Modeling 31 6 Structural Component Probability of Default Modeling 32 Qualitative factors are not necessarily judgmental, but cannot be empirically confirmed by the data Franchise Operating Environment Systems » Market position and sustainability » Macroeconomic stability » Market size and geographic diversification » Regulatory strength » Board independence and governance » Legal system and corruption » Financial reporting and transparency » Asset concentration and earnings diversification » Audit process » Strength of credit scoring and risk management » Access to alternative funding sources Probability of Default Modeling 33 7 Impact of Probability of Default on Recovery from Social Events Probability of Default Modeling 34 35 Impact of Probability of Default on Recovery from Social Events* From 2008-2010, a series of repayment crises associated with social disruptions struck some of the world’s oldest and most advanced microfinance markets, such as India and Nicaragua. Tens of thousands of borrowers defaulted, institutions buckled, apparent development gains were reversed, and outside investors suffered severe losses, dampening confidence in what had looked like a fast-growing and resilient market. The waves of default involved social phenomena on the ground, sudden changes of attitude associated with cultural shifts, and political movements. Scholars and practitioners identified a number of these “social default” repayment crises in the Andhra Pradesh region of India, in Nicaragua, Bosnia-Herzegovina, Morocco, Pakistan, Kazakhstan, and others. They have sought to understand these events and the kinds of practices and institutional arrangements that led to them. *Based on research by Columbia university SIPA Program Probability of Default Modeling 35 36 Issues » How does a repayment crisis caused by a social event affect the financial performance of MFIs? » Do strong social performance practices mitigate the severity of those financial effects? Probability of Default Modeling 36 37 Database Social performance data were obtained from MIX and then cleaned and sorted by Moody’s according to the SPA. Probability of Default Modeling 37 38 Relationship Between PD and SPA Grade No direct relationship between PD and SPA Probability of Default Modeling 38 39 Relationship Between PD and SPA Grade MFIs that scored best in social performance tended to show the most improvement in PD following a social event. Human resources was the social area that was most indicative of PD recovery Probability of Default Modeling 39 © 2012 Moody’s Analytics, Inc. and/or its licensors and affiliates (collectively, “MOODY’S”). All rights reserved. ALL INFORMATION CONTAINED HEREIN IS PROTECTED BY COPYRIGHT LAW AND NONE OF SUCH INFORMATION MAY BE COPIED OR OTHERWISE REPRODUCED, REPACKAGED, FURTHER TRANSMITTED, TRANSFERRED, DISSEMINATED, REDISTRIBUTED OR RESOLD, OR STORED FOR SUBSEQUENT USE FOR ANY SUCH PURPOSE, IN WHOLE OR IN PART, IN ANY FORM OR MANNER OR BY ANY MEANS WHATSOEVER, BY ANY PERSON WITHOUT MOODY’S PRIOR WRITTEN CONSENT. All information contained herein is obtained by MOODY’S from sources believed by it to be accurate and reliable. Because of the possibility of human or mechanical error as well as other factors, however, all information contained herein is provided “AS IS” without warranty of any kind. Under no circumstances shall MOODY’S have any liability to any person or entity for (a) any loss or damage in whole or in part caused by, resulting from, or relating to, any error (negligent or otherwise) or other circumstance or contingency within or outside the control of MOODY’S or any of its directors, officers, employees or agents in connection with the procurement, collection, compilation, analysis, interpretation, communication, publication or delivery of any such information, or (b) any direct, indirect, special, consequential, compensatory or incidental damages whatsoever (including without limitation, lost profits), even if MOODY’S is advised in advance of the possibility of such damages, resulting from the use of or inability to use, any such information. The credit ratings, financial reporting analysis, projections, and other observations, if any, constituting part of the information contained herein are, and must be construed solely as, statements of opinion and not statements of fact or recommendations to purchase, sell or hold any securities. NO WARRANTY, EXPRESS OR IMPLIED, AS TO THE ACCURACY, TIMELINESS, COMPLETENESS, MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OF ANY SUCH RATING OR OTHER OPINION OR INFORMATION IS GIVEN OR MADE BY MOODY’S IN ANY FORM OR MANNER WHATSOEVER. Each rating or other opinion must be weighed solely as one factor in any investment decision made by or on behalf of any user of the information contained herein, and each such user must accordingly make its own study and evaluation of each security and of each issuer and guarantor of, and each provider of credit support for, each security that it may consider purchasing, holding, or selling. Probability of Default Modeling 40