Model Assumptions are the large collection of explicitly stated or implicit premises, conventions, choices and other specifications on which any model is based.
Model assumptions are necessary to contain and simplify model development, or even make a model feasible. Assumptions can be numerical or conceptual in nature. Proper Model Documentation is the primary mechanism for controlling the possibly adverse influence of assumptions.
- Idealizations and simplifications of the underlying entity that is being modeled (e.g., market, counterparty)
- Choice of distributions among competing choices
- Choice of parameter fitting / calibration among competing choices
Issues and Challenges
- Undocumented or even unrecognised model assumptions