VAR(DSCR) is a useful measure for comparing projects.

By Nick Crawley on May 28 2009

I recently had an interesting discussion on my earlier article on how a little more science could make the DSCR a more powerful metric. In summary two project finance professionals debated if running scenario analysis isn’t more meaningful. My comment to this is below and I think articulates my earlier point a little better by putting the ‘variance’ of the DSCR out there as a measure of the ‘base case’ rather than of the sensitivity of the model to movements in underlying assumptions.

Posted on LinkedIn 28th May 2009

Hi guys, thank you for your comments - I completely understand and agree with your points of view.

Let me refine my comment a little more by observing that sensitivity / scenario analysis is a way of measuring the stress on an output when a variable changes, that is valid and well established. My thought is more focused on communicating the stress on an output (say the DSCR) when NOTHING changes….this is the Base Case and is independent of views on probability distributions of inputs, scenario analysis etc etc and is purely another metric for describing the DSCR string.

For example, these two projects, which on the face of it appear to be quite similar in DSCR terms

Project A <DSCR>=1.65x, DSCR(min) = 1.65x
Project B <DSCR>=1.65x, DSCR(min) = 1.65x

Can be shown to be quite difficult when the variance is calculated (just as simply as the average).

Project A <DSCR>=1.65x, DSCR(min) = 1.65x, DSCR (VAR) = 1.7%
Project B <DSCR>=1.65x, DSCR(min) = 1.65x, DSCR (VAR) = 9.0%

Of course, in a full analysis scenarios are performed and from time to time Monte Carlo routines are run but at the end of the day if you can communicate how ‘unsettled’ a Base Case is based on one additional number….isn’t that a good thing?