16 FAMILIES / FOUR LENDING CONTEXTS
The models behind
better-informed decisions.
Find the business question closest to yours. Each model page explains its purpose, where and how it can be used, the data it needs and the limits of its answer.
16 model families shown
Application default risk
Which new applications deserve a closer look?
Application probability of default (PD) estimates the chance that a borrower will meet a defined default condition within a specified period. It brings consistency to risk assessment at the point of application, while leaving eligibility, pricing and approval rules under your control.
MSME financial distress
Does the business have the financial capacity to support the loan?
A distress model examines financial strain in the borrowing business. Liquidity, leverage, operating performance and debt service provide different signals from repayment history. Distress and loan default are related concepts, but they are not interchangeable outcomes.
Behavioural risk
Which existing accounts are beginning to deteriorate?
Behavioural risk uses what an account has actually done: payment timeliness, balances, missed instalments and changes over time. It estimates a future arrears or default outcome from an agreed review date, rather than reusing an application score indefinitely.
Delinquency transitions
How might accounts move between arrears stages?
A transition model describes the likelihood of accounts moving from one repayment state to another over a defined interval. It makes the movement into and out of arrears visible, rather than treating the current overdue balance as a complete picture.
Cure & redefault
Which troubled accounts may recover—and stay recovered?
Cure estimates the likelihood of an account returning to an agreed performing condition. Redefault examines the risk of slipping back after that recovery. Looking at both helps avoid equating one catch-up payment with durable resolution.
Loss given default
If a loan defaults, how much might actually be lost?
Loss given default (LGD) measures economic loss after considering recoveries, their timing and the costs of obtaining them. Security can affect recovery, but a collateral valuation is not the same as cash recovered.
Exposure at default
How much could be at risk when default happens?
Exposure at default (EAD) estimates the amount outstanding if a borrower defaults. An amortising loan may shrink before default; a revolving facility may grow through additional drawings. A current balance alone cannot describe both.
Timing & prepayment
When might repayment, early closure or default occur?
Timing models describe when events may happen, not just whether they may happen. Default and prepayment affect different cash flows; an account that prepays cannot subsequently default on the closed facility in the same projected path.
Loan profitability
Is the proposed lending worth doing after costs and risk?
A profitability model brings expected receipts, funding costs, operating costs and credit losses into a consistent cash-flow view. It helps separate a high headline lending rate from a genuinely attractive expected contribution.
Credit-cycle adjustment
How could a different economic environment change risk?
Credit-cycle adjustment relates risk estimates to economic conditions. Point-in-time (PIT) risk reflects a particular environment; through-the-cycle (TTC) views aim to describe risk across a longer cycle. They answer different questions and need an explicit bridge.
Risk migration & SICR
How much has credit risk changed since origination?
Migration compares credit risk across time. Where relevant to the entity and accounting framework, significant increase in credit risk (SICR) assessments combine comparable risk measures with qualitative information and policy rules.
Expected credit loss
What loss estimate follows from the portfolio and scenarios?
Expected credit loss (ECL) combines the chance of default, loss severity, exposure and timing across explicit scenarios. It connects several models and accounting assumptions; it is not simply one score multiplied by today’s balance.
Portfolio loss & capital
How large could portfolio losses become when risks move together?
Portfolio loss analysis looks beyond average expected losses to the range of possible outcomes. Shared industries, locations, borrowers and economic drivers can cause losses to cluster, so diversification cannot be inferred from the number of loans alone.
Risk-adjusted return
Which opportunities earn an adequate return on the capital they use?
Risk-adjusted return relates expected contribution after credit cost to a clearly defined capital measure. It allows the business to compare opportunities that have different margins, loss exposure and capital needs.
Stress & growth planning
Can the business fund its growth and withstand a downside?
Stress and growth planning follows how lending cohorts, collections, losses, funding and capital interact over time. It tests the path to an outcome, not just whether a target loan-book size looks attractive at the end.
Model uncertainty
Would the decision change if the estimate were less certain?
Model uncertainty examines how results change when data samples, parameters, model choices or assumptions change. It helps a reviewer recognise when a precise-looking number is supported by limited evidence.