How to build a binomial tree model for bond valuation?

see the topic. I got several questions here, 1. in the book, CFA metioned a binomial tree model need be build before value a bond, ie, interest rate has 60% chance going up and 40% chance going down… my question here is how to predict the percentage? 2.after getting the percentage, which forward rate should be employeed as the benchmark rate? 3. for a specific bond, we shouldn’t just use the benchmark rate. Some spread should be added in, how to value the spread is my third question? I hope I can get some advise here, I have Bloomberg access, so if you find any clue in Bllomberg, it will be perfect! Thanks

I advise you to learn how to spell the word ‘advice’.

i remember it is in the option part, you can have a look there. you are reading Fixed income securities parts, arent you?

  1. It is actually not important. What is important is matching the mean and variance of the process you are trying to match. Ultimately your binomial tree is supposed to approximate a normal distribution (or a Gaussian process or something) and that is characterized by its mean and variance not what the steps look like. Obviously, depending on your up/down percentages you will need to pick up/down steps carefully. 2) It depends on what you are trying to do and it is a huge morass of a problem. 3) Again a monster of a problem. Theoretically, you can get lots of market estimates of credit spreads from CDS spreads but there are lots of other possibilities.

“It depends on what you are trying to do and it is a huge morass of a problem.” Morass:Something that hinders, engulfs, or overwhelms:Eg. a morass of detail. My $0.02 : I learnt this section inside out last year…and they ended up asking the most arb questions on fixed income one can imagine. Not to say they wont ask it this year… but what it seems is this test is a test in obscurity-ok-enough moroseness for one day! happy studying!