Learn / Regulation
What is a model inventory, and what should it hold?
A model inventory is one list of every model your institution uses, with who owns it, what it is used for and when it was last checked. RBI's draft model risk guidance says a model that is not on the list cannot be used. So the list has to be complete, and it has to be kept up to date.
Why does the model inventory matter?
Most institutions can name their main scorecards. Few can list every spreadsheet that sets a price, a limit or a provision, or every vendor score that feeds a decision. An inventory forces that list into one place. It also tells you which models are overdue for validation, who to call when one breaks, and what is affected when a data feed changes.
What should you put in a model inventory?
The draft names a minimum set. The last three rows are our suggestions, not RBI's wording.
| Field | What it tells you |
|---|---|
| Model name and status | Active, not in use, under development or retired. The draft wants all four kept. |
| Owner, builder, validator, approver | Named people, not teams. The draft requires validation to be independent of whoever built the model. |
| Risk tier | How important the model is, which sets how closely it is checked. |
| Intended use | The one thing the model is approved for. Using it elsewhere is a model risk. |
| Upstream and downstream links | The tables and feeds it reads from, and the decisions that use its output. |
| Findings from validation, monitoring and audit | Open issues in one place. |
| Vendor or in-house (suggested) | Vendor models need extra oversight and contract rights. |
| Last and next validation date (suggested) | Shows overdue models at a glance. |
| Retire date and keep-until date (suggested) | The draft says keep retired models for at least ten years. |
A worked example
A small NBFC with three models might have an inventory like this. The names and dates are made up for illustration.
| Model | Status | Tier | Owner | Validator | Next validation |
|---|---|---|---|---|---|
| PL application scorecard v3 | Active | High | Head of credit risk | External firm | Mar 2027 |
| Collections propensity model | Active | Medium | Collections head | Risk analyst, not the builder | Jun 2027 |
| Bureau vendor score (bought) | Active | High | Credit head | Not yet assigned | Overdue |
The last row shows what an inventory is for. A vendor score drives decisions, it is tiered high, and no one has been named to validate it. Without the list, that gap stays hidden.
What are the common gaps?
- Spreadsheets and vendor scores missing from the list.
- One person named as owner, builder, validator and approver.
- Dependencies left blank, so nobody knows which models a changed data feed will hit.
- Retired models deleted instead of kept on the list.
- A list kept in someone's inbox instead of a controlled file or table.
Try it on your own tables
The free tools give you the prompt, the checks and how to present the result. You enter table names only.
See the free toolsFrequently asked questions
What is a model inventory?
A single list of every model an institution uses, with owner, use, risk tier, dependencies and validation status.
Does a spreadsheet belong in the inventory?
If it drives a decision, yes. RBI's own example is a spreadsheet calculator used to set lending rates, customer margins or credit terms.
How long should retired models stay on the list?
The draft says at least ten years from the date it is retired. Longer if another law requires it.
Who should own the inventory?
The draft makes senior management responsible for keeping it accurate and up to date. Many teams give the job to the model risk or credit risk function.
Keep reading
RBI's draft model risk rules: what they mean for your credit models
RBI has drafted rules on how banks, NBFCs and credit bureaus manage the risk of their models. What is in the draft, what it means for scorecards and AI, and what to do now.
RegulationHow to tier your credit models by risk
Model tiering ranks each model by how much harm a wrong result could do. A simple two-factor method, a worked example for lenders, and the mistake that gets models tiered too low.
Data guideCredit-risk data guide: the tables and columns behind every tool
A plain-English guide to the loan, repayment, payment, bureau and application tables used in credit-risk work: what each column means, where to find it, and what goes wrong.