Shadow model: Difference between revisions

From ACT Wiki
Jump to navigationJump to search
imported>Doug Williamson
No edit summary
imported>Doug Williamson
(Expand definition.)
 
Line 1: Line 1:
''Financial modelling - audit and assurance.''
''Financial modelling - audit and assurance - spreadsheet risk.''


Shadow models are a tool to obtain understanding and assurance about the correct functioning of a financial model under review.
Shadow models are a tool to obtain understanding and assurance about the correct functioning of a financial model under review.


The shadow model reproduces the logic of the model under review.
Shadow modelling is about gaining assurance that a model is functioning correctly and appears to be error free.


The expectation is that, taking the same inputs as the model under review, the shadow model will produce the same outputs, using the same logic.
We build a shadow model that aims to reproduce the logic of the model under review.  If our shadow model produces a different result from the original model, either there is an error in one or both of the models, or something important has not been properly understood.
 
The expectation is that, taking the same inputs as the model under review, our shadow model will produce the same outputs, using the same logic.


Any differences are investigated, corrected for, or explained.
Any differences are investigated, corrected for, or explained.

Latest revision as of 23:10, 16 June 2021

Financial modelling - audit and assurance - spreadsheet risk.

Shadow models are a tool to obtain understanding and assurance about the correct functioning of a financial model under review.

Shadow modelling is about gaining assurance that a model is functioning correctly and appears to be error free.

We build a shadow model that aims to reproduce the logic of the model under review. If our shadow model produces a different result from the original model, either there is an error in one or both of the models, or something important has not been properly understood.

The expectation is that, taking the same inputs as the model under review, our shadow model will produce the same outputs, using the same logic.

Any differences are investigated, corrected for, or explained.


Shadow modelling of this kind is a very robust way to gain a high degree of assurance and understanding about a financial model.


See also