Test your financial model auditors!

By Nick Crawley on June 16 2009

The best model audit is not an Audit

On most project finance transactions the financial model needs to be audited. If blue chip organisations are involved this is an unavoidable measure, in the equity markets where standards of due diligence are generally lower (no offence!) this happens less often. After you have been in this market long enough you build up repository of memories of occasions when everybody, especially the auditors, would prefer not to perform an audit - these are unfortunately the ones when an audit is really needed. We have all been there and boy wasn’t it painful, frustrating and expensive!

Having worked on hundreds of these in various capacities from model author, lender, adviser, and facilitator and for many years Model Auditor my view is that the process of rebuilding a model identifies more issues than just an audit of a bad model. This is not because those teams are not competent, on the contrary they (generally) really know their stuff, but the fact is that working through each line of often badly prepared code, with the crutch of ‘spreadsheet auditing software’ and the need to document all queries and then enduring the open-ended process of the resulting iterations is a really inefficient and therefore an expensive exercise to go through to ‘tick the box’.

Shadow rebuilding…a nice idea but isn’t a stand alone solution

The alternative is to throw the original model up on one screen, fire up a new spreadsheet and rebuild, challenge and reconcile all lines but whilst ensuring transparency and flexibility results in a far superior interrogation of the code and on the basis the original model is of the usual market standard (pretty poor) then the resulting model will add a lot more value to the overall process and the management team that have to live with it. I propose that this can be done quicker than auditing a bad model!

However, here is the rub, this alternative process doesn’t result in an Opinion Letter….which for the vast majority of cases is the sole reason for getting the model audited - rightly or wrongly. I caveat this because there is at least one Model Audit firm whose preferred approach is to perform a shadow rebuild and then sign off the original model, what the ?!

Test your model auditors

So next time you engage a Model Auditor have a think about what you really want to achieve, more often than not the financial cost of rebuilding and auditing the resulting model is less than or equal to the cost of auditing the original model with the exception that with the later process you are still left with a model that is probably going to fall over when actuals need to be incorporated or covenants reported….

If you really have to get the original model audited then consider embedding three or more intentional errors into the model upon submission of the initial version to the model auditors, ranging from an easy one to an error they will only find in a scenario if they fully understand the project (not just the model) which is the true test of how well a model audit firm actually understand. Go on see how well they do and what about pegging their fees to finding these errors?!