Financial toolbox/Credit risk utilities/Transprob function/SnapsperYear

For a single client this is the information on dates an ratings 30.06.12 AAA 31.08.12 AAA 30.09.12 AA 31.10.12 AA 30.11.12 AA 31.12.12 A 31.01.13 A 28.02.13 A 31.03.13 BBB 30.04.13 BB 31.05.13 CCC 30.06.13 D

Transprob function, cohort method, snaps 6, I get the following transition probabilities (from 30.06.12-30.06.13) AAA to AAA 1.5625, AAA to AA 1.5625, AAA to A 15.625, AAA to BB 12.5, AAA to D 68.75. AA to A 3.125, AA to BB 3.125, AA to D 93.75, etc

Can someone please explain: How is 1.5625, 15.625, 12.5, 3.125 actually calculated (at least on one example)? If Snaps per year=12, the result is

AAA to AAA 0.771, AAA to AA 4.624, AAA to A 12.717, AAA to BBB 5.299, AAA to BB 6.503, AAA to CCC 7.804, AAA to D 62.282, etc.

How is for example 0.771 calculated? Or 12.717?

Duration method:

AAA to AAA 1.872, AAA to AA 7.447, AAA to A 14.438, AAA to BBB 5.631, AAA to BB 6.651, AAA to CCC 7.099, AAA to D 56.864, etc

How is e.g. 7.447 (AAA to AA) calculated or 7.099 (AAA to CCC)? Best regards Aleksandar

Answer 1

[Shashank Prasanna]

The Algorithm section in the transprob doc page tells you about how they go about calculating it:

http://www.mathworks.com/help/finance/transprob.html

If you want further info you will have to open up the code:

>> edit transprob

And you can see the exact implementation there.

[Aleksandar]

Thank you for your answer. I went through the material. If I understood correctly, transition probability should be (based on the provided example on the link) totals.totalsMat / totals.totalsVec (e.g. IG to IG should be 4721/4808)?