Serving the Quantitative Finance Community
Why do futures/IMM FRA data show no convexity?
Perhaps some of you in traderland can explain a "puzzling" results in my research project.I obtained a small batch of high frequency data from Reuters (Dec 02 to March 03) - unfortunately this was all I could get - containing quotes on IMM FRAs out to the expiration of the eighth (short sterling) futures contract. When I compared the FRA midquotes to transaction prices on short sterling futures I found almost no difference. In fact FRA rates tended to exceed futures rates by 0.5 bp when I matched quotes/trades to within 60 seconds over the three-month sample period.While this was not surprising for the near FRA contracts, I had though that there would be evidence of a convexity adjustment in the longer-dated contracts. Rough back-of the-envelope calculations using the Hull-White and Ho-Lee model put the convexity adjustment for the eighth futures/FRA contract at 5-7 basis points, depending on assumptions.Hence I have the following questions:(i) Do sterling markets bother with pricing the convexity adjustent into FRAs?(ii) If convexity is priced into FRA rates, what are the popular models traders employ to estimate this adjustment?(iii) There is some recent US research that argues that the marking-to-market and collateralisation of swaps has reduced the convexity adjustment applied to swaps. How prevalent are such practices with FRAs in the UK? And have they had an impact on FRA pricing?Answers to any or all of these questions would be much appreciated!P.S. My IMM FRA quote data originated from a single dealer so I'm willing to concede that the data may be suspect.
Why do futures/IMM FRA data show no convexity?
I may be wrong but I would have expected the convexity correction for the 8th future to be ~ 0.8-1bp not the 5-7bp you quote?
Why do futures/IMM FRA data show no convexity?
"(ii) If convexity is priced into FRA rates, what are the popular models traders employ to estimate this adjustment?"On Euro$ futures, lots of traders use bloomberg I guess, I think their default is the Kirkos Novak 1997(Hull-White I guess, with no mean reversion Ho-Lee) .....also described in my 2nd edtion...however it is not so much about what model you use, it is more about how you use the model (know your weapon), many people only use models to fit market, prop traders tend to try to make up their own views on future fluctuations (I do not like vol term as it to often is linked to standard deviation that mainly is a statistical term, only catching up some of the relevant fluctuations)...second how can you capture that, most theory is here flawed... and few traders get it...(well more traders than academics have an idea about it)0.8 bp or 7 bp yes that is big difference, one of you forgot to divide or multiply by 10 ? Are both of you market makers? you both quote prices, I am in
Last edited by Collector on January 31st, 2007, 11:00 pm, edited 1 time in total.
Why do futures/IMM FRA data show no convexity?
Thanks for the lead on the model "Collector". What is the title of the book you are referring to? Looks like I should get hold of a copy.Re my convexity calculation using the Ho-Lee model:(i) for the eight futures, T1 = 2, T2=2.25(ii) basis point volatility (SD) of the 3 month spot rate over the 90 days preceding 19 December 2002 equals 1.75 (statistical approach, unfortunately)So convexity = 0.5 * 1.75^2 *2 *2.25 = 6.9 bpI'm not a market-maker so no chance of making any money out of me - book royalties are your only hope!