I am wondering what is the most uded method or the most accurate one to compute default probabilities out of bond prices.

Actually it seems there isn't any most used method. KMV or CreditMetrics provide some. It depends on what you're working on. Each of them have advanatges and bad points.

If you use your Bond Price to first calculate the Bond's Asset Swap spread (i.e. convert the current price to a floating spread over LIBOR) you can then calculate the Probability of Default (PD) as follows...PD = (1 - exp(- s T) ) / (1 - RR)where:s = The Asset Swap Spread for the BondT = The Maturity of the BondRR = Expected Recovery Rate

Actually, since you asked about "the most accurate" method I should add that when the Bond is far from Par, the Asset Swap Spreads tends to become increasingly unreliable as a valuation of the bond. Obviously, this effects your calculated PD but then again you still need to make a big assumption in chosing your Recovery Rate.

Hi all,ActuallY I am still confused. I want to calculate the default probabilities from bond prices. I have read the reduced form models in general and all I can understand so far is that the default probabilities are known and used to calculate the interest rates . Am I right? if not could anyone please choose any reduced form model and tell me how to calculate the default probabilities. If I understand one , I believe I will understand all.MB

Reduced form models assign a probability of default directly from accounting data, without use of bond or any security prices.If you know implied default probabilities and recovery rates, you can compute bond prices. If you know bond prices and implied recovery rates, you can compute implied default probabilities. These are just mathematical identities. If you know two, you can compute the third. This assumes you know risk-free interest rates.Reduced form models are typically used to estimate actual default probabilities. You can't compute any price from an actual probability.