There's principal, I just write the first period coupon, remaining parts is the regular

<t>a bond coupon will be paid at the end of each year, if today is the middle of the year, then the price should be1. P=0.5Coupon*Survival*RiskFreeDF2. p=Coupon*Survival*RiskFreeDF-0.5CouponWhich one is correct?I know Dirty price-accrue(0.5Coupon)=clean price, but the difference between 1and 2 is no...

If real measure cash flow is given, what kind of research we can do? Thanks!

nobody can help me?Since All the paper talk about synthetic CDO, Iam wondering how to dicsount cash CDO (waterfall)? Real measure or risk neutral measure?

Is any one can help me to find some materials which are related to "cash CDO tranch pricing"? Thanks a lot!

Is any one can help me to find some materials which are related to "cash CDO tranch pricing"? Thanks a lot!

<t>I was asked to build a model to calculate tranch present value1. You can assume a simple process for the cash flow(underlying is corporate bond)2. Then discount to current time(maybe I need to consider interest rate, default rate)But I feel I have no idea to discount cash flow, real measure, or r...

<t>I was asked to build a model to calculate tranch present value1. You can assume a simple process for the cash flow(underlying is corporate bond)2. Then discount to current time(maybe I need to consider interest rate, default rate)But I feel I have no idea to discount cash flow, real measure, or r...

Please help on this question:You pick up 3 numbers out of {1,2,…,10}, what’s the probability of the sum to be divisible by 3?Thanks a lot!

<t>You are right. But I still confuse.I need forcasting quartely CMT 10y from next period to future 10 years (so 40 numbers)And I already get the parameters.So if we let current cmt10y is r0, how to get r1?r1=r0+E(dr(0)); r2=r1+E(dr(1))..... Use this way, r1..r40 nearly flat.My question is this meth...

Sorry, some typo. I take mean directly, and forecasting 10y weekly data, all of them are very flat.

<t>I use MLE with UKF and historical 10 year full term structure data. I don't think any wrong with estimation.Problem is when I forcasting, can I calculate mean value directly E[r(t)]=r(t-1)+E[dr(t)] or should I simulate 1000 value, then take the mean?If tried first method, it doesn't work. Future ...

<t>I use CIR to forecasting.After finish estimation, we get a group parameter {K,theta,sigma,lamda}then for forcasting, you should take the mean value for next period. r(t)=r(t-1)+E(dt)=r(t-1)+k(theta-r(t-1))dt=k*theta*dt+(1-kdt)r(t-1)so first term is cconstant, and basically is very flat, like falt...

macca:Any chance to send me a copy of CDO model? My email is daerbao@hotmail.comThanks a lot!

Would you please explain it a little bit detail.I have another question, for the CIR estimation the parameters, which kind of method we can use? It seems MLE is not so easy. Or the result is not so good. ANy other suggesion.