Serving the Quantitative Finance Community
Salomon two factor
A question on rate modeling history;Bookstaber in 'demon of our design' talks of salomon brothers fixed arb desk using a two factor rate model. He also mention it was adjusted for fed rate decisions.Was this just a two factor hull white with r(0) fixed at the fed rate? If not, any published details?Just curious from a historical perspective.
Salomon two factor
slackerNot sure about the history. However in a two factor model, r(t) is "comprised of" two things, so r(t) = x(t) + y(t). If you say that one of the factors, say x(t) is Fed Funds or some other target, is a perfectly legitimate choice.
Salomon two factor
I think this is a reference to what is also referred to as the Gauss 2+ model. It is given a thorough treatment in the 3rd edition of the book by Tuckman and Serrat (perhaps not coincidentally, Bruce Tuckman spent time in the arb group). The short rate in the model has no quadratic variation, being just pulled (fairly rapidly) toward a target. The target is the sum of two correlated OU processes, so you end up with two factors (Brownian motions) and three state variables.
Salomon two factor
Thanks guys. Will have to take a look at Tuckman's book, not very familiar with it. The idea that they pinned a parameter directly to the Fed rate is interesting to me. I suppose in hindsight we now know two factor short rate models do not generate enough correlations, but its interesting to me how they spent months/millions trying to figure out why their relative value curve trades were bleeding since they were convinced their model was right.
Salomon two factor
The thing with micro RV trades is that you'll almost always be right, EVENTUALLY, but frequently, the interim losses are just too hard to bear...The Solly model is pretty useful for trading micro RV. As to big curve trades, it's as difficult to interpret as other simpler methods.