Skip to main content

Model Risk

Nowadays risk models are crucial in the analysis, valuation, measurement and management of risk. Financial institutions increasingly rely on mathematical and financial models to estimate possible scenarios and support business or trading decisions. Model risk has been growing as the sophistication increases alongside with greater computing processing power. Depending on the type and size of the firm, financial and economic models range from simple spreadsheet tools to complex statistical models running in distributed systems, applied to millions of transactions. Models are commonly used to identify and measure credit, operational, market and liquidity risks.

Independently of the level of sophistication of a given model, model usage reliance exposes a financial firm to model risk. Similarly as other types of risks, if materialised, model risk will typically involve the possibility of a financial loss or incorrect business decisions which might lead to financial losses or other adverse consequences such as reputational damage to the firm.

Model risk may arise from factors such as:

  • Inaccuracy or fundamental flaws in a the model design and development, including poor or informal updates to existing models. Examples of some of these flaws include errors in computer code and algorithmic errors
  • Inconsistent data, being it input by the model operator or historical data used by the model
  • Poor interpretation and misapplication of model results by the model users
  • Use of outdated models which no longer reflect current business assumptions

Investopedia defines Model Risk as:

A type of risk that occurs when a financial model used to measure a firm’s market risks or value transactions does not perform the tasks or capture the risks it was designed to.

Model risk is considered a subset of operational risk, as model risk mostly affects the firm that creates and uses the model. Traders or other investors who use the model may not completely understand its assumptions and limitations, which limits the usefulness and application of the model itself.


Our recent articles about Model Risk