Models generate the information on which a wide range of decisions are formed, from forecasting policy outcomes to estimating the financial feasibility of major infrastructure programmes. Errors can be very costly and highly embarrassing.
To aid those seeking to determine whether a model is robust and reasonable, this framework covers seven stages from reviewing the model concept to ensuring that the outputs produced by the model are robust, well communicated and used by decision-makers. It covers issues such as governance arrangements, and the quality of data and assumptions. It has the flexibility to be used for models of all levels of complexity and business risk, from relatively simple spreadsheets to detailed forecasts using specialist software. Its application can be tailored according to time and resource available, and the level of assurance needed to reach a judgement.
The framework is based on that used by the NAO when reviewing organisations’ models and builds on good practice guidance, including from HM Treasury and international standards.