November 19th, 2007, 1:38 pm
You can use the same math one would use if instead one had interest rate futures data instead of FRA data. Ignoring liquidity concerns, you should in principle get better results using FRAs because there is no need for a convexity correction. Look up any of the many articles on building swap zero curves (e.g. Uri Ron's Bank of Canada paper) and just drop in the FRA rates instead of the implied futures rates.