Average DSCR - think twice!

Ok so if you are in Project Finance you have probably calculated and analysed your fair share of average DSCR’s. If you take the average of a row of DSCR’s have a think if this is what you actually mean, its not wrong but you need to be aware of its limitations because it doesn’t work particuarly well when dealing with exotic cashflows or repayments.

There two ways to calculate the average DSCR:

Calculate the average of the period by period DSCR’s over the life of the loan
Divide the total CFADS by sum of total principal and interest

On the face of it there isn’t much difference but they can result in very different numerical outcomes, lets take a look at an example in the screenshot below where the loan life is from June 2010 to June 2013.

You can see that the sum of CFADS over the loan life is $267m and the sum of P+I is $172.94m, therefore the <DSCR> is $267,/$172.94m = 1.544x . If we calculate the average using the AVERAGE function (excluding zero values of course, best done using an {array} or n/a if you are old school!) then the result is 1.647x.

Capture of my screen showing dummy figures.

Capture of my screen showing the results of the two different methods.

Why are they different?

Well the average of the DSCR values over time treats all of the elements as equally important whereas method 2 weights each element by the relative importance of the sum of principal and interest in each period. This is best highlighted when we have extreme values such as the final repayment being very small compared to the CFADS in the period. This can lead to an enormously high ratio which is given equal importance in method 1 which distorts the average DSCR.

So what one is correct?

Both methods are what they say they are, the average DSCR, but in certain situations be aware that method 2 is probably more meaningful. Be aware of this - especially if you are trying to engineer a higher ratio !