Credit risk foundations
Core concepts, lending lifecycle, ratings, and portfolio risk.
Learn
Learn credit risk
Build a clear path from credit-risk data and scorecards to IFRS 9 measurement, model validation, and ongoing monitoring.
Browse by topic
Credit scorecards
Build, scale, validate, and monitor application and behavioural scorecards.
Probability of default (PD)
Define, estimate, calibrate, and project default probabilities.
Loss given default (LGD)
Measure discounted losses after recovery, costs, and collateral.
Exposure at default (EAD)
Estimate drawn and undrawn exposure at the point of default.
IFRS 9 expected credit loss
Understand impairment stages, scenarios, and discounted expected losses.
Credit-risk data engineering
Build reliable point-in-time datasets, labels, joins, and controls.
Statistics and machine learning
Compare statistical and machine-learning approaches for risk modelling.
Validation and governance
Backtest, challenge, document, and govern credit-risk models.
Guide library
Methods, assumptions, formulas, worked examples, and primary references.
Behavioral Scorecards for Existing Customers: Build, Validate, and Monitor Monthly Risk Scores
A practical guide to behavioral scorecard design for existing accounts: point-in-time data, 12-month outcomes, PD calibration, validation, and monitoring.
Credit Portfolio Stress Testing: A Step-by-Step Scenario Walkthrough
Work through a synthetic credit portfolio stress test: build a coherent adverse scenario, calculate segment losses, explain the drivers, and monitor key assumptions. Distinguishes internal stress sensitivities from IFRS 9 ECL and regulatory capital.
Feature Engineering for Credit Risk: Behavioral Windows Without Leakage
Build point-in-time-safe credit-risk features from payment behavior, balances, delinquencies, categories, trends, and missing data. Includes a worked synthetic example, browser-safe Python, leakage controls, and a production monitoring checklist.
Competing Risks in Credit PD: Default, Prepayment, and Account Closure
Learn how default, prepayment, and account closure compete in credit PD modeling. Compare cause-specific hazards, cumulative incidence, Kaplan–Meier, and Fine–Gray with a worked example, monitoring guidance, and runnable synthetic-data Python.
Credit Scorecard Scaling: Convert Points to Odds and Probability of Default
Derive a points-to-double-odds scale from a chosen anchor, convert score points to good-to-bad odds and PD, work a synthetic example, and monitor whether the PD mapping remains calibrated.
Using Delinquency Roll Rates and Cure Rates to Monitor Credit Portfolios
A worked guide to delinquency transition matrices, forward roll rates, one-snapshot and sustained cures, account versus balance weighting, and multi-period projections—with clear denominators and synthetic examples.
How to Compare Two PD Models with a Paired Bootstrap
Compare discrimination from two PD models on the same matured validation cohort with a paired, stratified bootstrap confidence interval for the AUC difference.
IFRS 9 ECL Staging: A Step-by-Step Guide to 12-Month and Lifetime ECL
Understand IFRS 9’s three-stage general approach, assess significant increases in credit risk, and calculate 12-month versus lifetime ECL with a synthetic term-structure example.
PD Segmentation (3/3): Python Implementation and Validation Tests
Implement fixed PD bands in Python, calculate segment calibration and uncertainty, test boundaries and ordering, and validate the design with time-based holdouts.
PD Segmentation (2/3): Monitoring Calibration, Drift, and Migration
A practical PD-segment monitoring framework for assignment quality, population shifts, grade migration, observed-versus-expected defaults, uncertainty, and escalation.
PD Segmentation (1/3): Common Methods and How to Choose
A practical guide to product splits, score bands, expert rules, trees, model-based grades, and clustering—with a step-by-step process for choosing and testing PD segments.
Measuring Credit Portfolio Concentration: HHI, Exposure Shares, and Stress Tests
A practical, synthetic-data guide to measuring borrower, sector, geographic, and common-factor concentration with exposure shares, HHI, connected-group aggregation, and stress testing.
Handling Imbalanced Data in Credit Risk Models
A practical workflow for rare-default modelling: define the target, compare class weighting and resampling without leakage, evaluate precision and recall at useful thresholds, and recalibrate PDs on representative data.
A Step-by-Step Credit Risk Model Monitoring Framework
A practical monitoring workflow for PD, LGD, EAD, and IFRS 9 ECL: align populations and horizons, calculate core metrics, interpret uncertainty, and set review actions without relying on universal thresholds.
Lifetime EAD Backtesting: Methods, Metrics, and a Python Example
A practical guide to backtesting lifetime exposure-at-default forecasts: define mature cohorts, compare horizon-specific exposure, monitor calibration and error, and run deterministic Python with synthetic data.
How to Calculate Predictor Weights in Logistic Regression: Coefficients, Contributions, and Scorecard Points
Learn what logistic-regression coefficients mean, how to convert coefficient × sample standard deviation into percentage weights summing to 100%, and how coefficients become scorecard points using a worked synthetic PD example.
Default Rate, Observation Window, and Performance Period: Choosing 60 or 90 DPD
Define default-rate targets, distinguish observation and performance windows, compare 60- and 90-day delinquency labels, and monitor PD outcomes with synthetic examples.
EAD and CCF Estimation: From Facility History to Monitoring
A practical guide to exposure at default (EAD) and credit conversion factors (CCFs): define the observation window, build facility-level data, estimate drawdowns, handle unstable observations, validate predictions, and monitor performance. It distinguishes IFRS 9 expected drawdowns from regulatory-capital EAD.
LGD Recovery Cash Flows, Workout Costs, and Discounting
Learn which post-default recoveries and collection costs belong in LGD, how to discount them, and why IFRS 9, Basel IRB, and internal LGD measures can differ.
Reading PD Backtests When Defaults Are Rare
Use confidence intervals and an exact binomial tail check to interpret a one-year PD backtest when default counts are small. Learn the assumptions, calculate a synthetic example, and understand what a non-significant result does—and does not—show.
Estimating a One-Year Probability of Default
A practical guide to defining the default horizon, fitting a logistic PD model, calibrating its output, and validating estimates with matured outcomes.
What credit risk measures
Separate likelihood of default, loss severity, exposure, and portfolio impact.
Obligor, facility, and portfolio grain
Choose a consistent entity grain before building labels or aggregating exposures.
Credit lifecycle and risk appetite
Connect origination decisions with account management, collections, and portfolio limits.
Expected loss and uncertainty
Calculate the loss average, quantify dispersion, and see why equal expected losses can hide different outcomes.
A credit scorecard development blueprint
Design the sample, target, predictors, score scale, validation, and decision use as one workflow.
Binning, Weight of Evidence, and Information Value
Calculate bin-level evidence while guarding against unstable or empty bins.
Logistic regression and score scaling
Translate model log-odds into probabilities and a points scale with visible conventions.
Cut-offs, overrides, and reject inference
Choose operational thresholds with evidence and document what rejected applications conceal.
Default events and observation windows
Define the event, unit, and horizon before estimating probability of default.
One-year PD and lifetime PD
Keep the risk horizon and survival conditioning explicit when comparing PD measures.
PD calibration: point-in-time and through-the-cycle
Separate ranking from the calibration that assigns a probability scale to grades or scores.
Transition matrices and rating migration
Use rating migration to describe movement and derive horizon risk only under explicit assumptions.
Survival, hazard, and macro-conditioned PD term structures
Build period-by-period risk while accounting for surviving accounts and changing conditions.
Discounted net recoveries and workout LGD
Measure loss from cash-flow shortfalls, recovery costs, and timing after default.
Workout LGD and market LGD
Compare observed recovery cash flows with market-price proxies and understand their limits.
Collateral, haircuts, and downturn LGD
Translate security into plausible net cash recovery instead of treating collateral value as cash.
Cure, re-default, and recovery timing
Build transparent default episodes, discount workout cash flows, and monitor cure and repeat-default outcomes.
Drawn exposure, undrawn commitments, and CCF
Estimate how much additional committed credit may be used before default.
Revolving utilization and CCF estimation
Use account paths to learn how balances evolve between an observation date and default.
Product-level EAD forecasting
Forecast exposure paths using contractual terms, utilization behavior, and account exit patterns.
IFRS 9 EAD versus regulatory EAD
Avoid carrying a prudential conversion factor into accounting measurement without reconciliation.
IFRS 9 stages and ECL horizons
Connect impairment stage to 12-month or lifetime expected credit loss measurement.
Significant increase in credit risk
Assess the change in default risk since initial recognition, not merely the change in allowance.
Scenario-weighted, discounted ECL
Combine probability-weighted paths for PD, LGD, exposure, and cash shortfalls transparently.
Modifications, POCI assets, and ECL disclosures
Recognize special accounting paths where a standard model diagram is not enough.
Entity keys and source-system mapping
Create traceable obligor, facility, application, and collateral relationships before modeling.
Point-in-time snapshots and performance labels
Separate feature availability from future outcomes to prevent look-ahead leakage.
Joins, missingness, duplicates, and outliers
Treat data quality as a measurable property of the model population, not a one-time cleanup.
Time-based splits, leakage, and class imbalance
Validate on later periods and keep resampling or feature selection inside the training process.
Logistic regression, trees, and boosting
Choose model complexity based on use, sample size, data quality, and governance capacity.
Calibrating probability models
Re-map predictions to observed event rates on separate, representative data.
Survival analysis and censoring
Use event time when accounts have different follow-up or default risk changes over time.
Explainability, fairness, and challenger models
Assess model outputs and impacts across groups while preserving a clear purpose and review trail.
Discrimination: ROC-AUC, Gini, and KS
Measure ranking ability while keeping class balance, horizon, and threshold context visible.
Calibration backtesting and observed-to-expected
Compare probability forecasts with matured outcomes and explain sampling noise.
PSI, CSI, and drift monitoring
Compare population and feature distributions while disclosing bins, smoothing, and threshold limitations.
Portfolio outcomes and model governance
Bring migration, vintage, recovery, EAD, and ECL outcomes into a controlled model lifecycle.