What causes a gain or loss on a futures positsion when the IMM roll occurs?

if the spread between the front and the back contract changes you will make or lose money depending on your position. If you are long the front contract you are a seller of the roll - you speak in terms of the front contract - and will lose money if the spread widens (becomes more positive). As a rule of thumb, if the market sells off you would expect the roll to widen.

Put simply, on the IMM settlement date, the eurodollar contract settles to the 11:00am GMT offical BBA (British Bankers Association) LIBOR setting. This setting determines the final price of the contract, so if the 11am setting implies a price (1000-rate) which is higher than last night's closing price and you are long the contract then you make that bps differential.... oh and as a sad aside, the crooks (or chicago mafia) at the CME charge you a full trading commission for expiring any open contracts that are left open on IMM day, even though you did not physically trade them... ahhh and we all though extortion was on it's way out after gotti's funeral.....just a hunch but i suspect your question may be a bit more involved - are you refferring to the massive p/l swings that tend to happen in interest rate swap books on the day of the imm roll? if so i can explain this very comprehensively.... let me know

That does help - thanks. Yes I guess I am trying to understand why a gain or loss would result in this day. I don't trade or get involved in product accounting so my view is form a reporting aspect - hence why my questions in this forum ar always Billy Basic!Thanks

just think of the IMM roll date as a market trade - a market trade which occurs a a price based on that morning's LIBOR setting for the entire size of the current position....

posted by Johnself11:are you refferring to the massive p/l swings that tend to happen in interest rate swap books on the day of the imm roll? if so i can explain this very comprehensively.... let me know what p/l swings are you talking about? is this any artifact of (i) some methods of creating a libor curve (especially in the stub out to the 1st IMM date) (ii) (old) commonly used methods of hedging the stub risk?

generally stated, the p/l effects of an IMM roll on swap book come from the basic fact that after the front future rolls off, the curve suddenly becomes constructed with new, different intruments. the major p/l swings come from the two "transition" points in the curve. most swap curves consist of 3 intruments: cash deposit rates in the very front of the curve out the 3m point (3mL), then some number of convexity-adjusted eurodollar futures (usually no more than 20, or 5ys worth), and then par swap rates. the rule is generally that in an overlapping situaton, the prority of ordering is swaps, ed futures, then deposit rates, respectively.... so the day before the imm roll all the cash rates to 3m inputted into the curve are not being used, as the front future almost completely covers the first 3m, and so the curve ignores the cash rates. so the term structure of the first 3m of the curve just before imm roll is just the 3m rarte implied by the about-to-expire ed future...after the roll, the cash rate come back into use in the curve, as the next ed future is now 3m away... this means that all the short-dated forwards in the first 3m will change based on the change in curve construction.... so any swap book positions in these fwds will be revalued with the new curve structure.... the same thing happens at the futures/swaps transition point, as the futures strip is extended by 3m due to adding one more future, so all the fwd rates in that area of extension will change and create p/l effects.....

as I thought, the large changes in fwd rates you're talking about happens in books with archaic methods of marking the short end and IF the aim is market making / hedging & not short end prop trading, possibly the massive p/l swings occur because of lazy swap traders not hedging out the real 'future stub risk shape hiding behind' the 1st contract.I suspect this is all old stuff: most swap houses now have moved away from such book marking / management methods. I hope to be wronged though. edit: thx for the explanation. Possibly there are still banks using this method. That's why the 3s1s basis in the short end drastically changed signs this June IMM.

well you have a point in that using overlapped cash depos is pretty low-tech, but i assure you that it is still practiced by many dealers... even if a more sophiscticated "3mL is always repriced" model is used, the futures/swaps transition issue will always be present, and usuallly it is change in the forward rates in this region that generally have the largest p/l impact.... also i think you're right in your 1's/3's assertion - pretty sure you can count on one hand the number of dealers which actually have separate 1mL curves and 3mL curves....

surely this isn't a real p&l event as this money would have dripped in/out of the book over the previous 3 months as carry/rolldown?

yes you are absolutely corret - the p/l comes from the fact that your book has over/under-valued those fwd rates for the last 3 mos... certainly the money doesn't disappear into thin air!

though i will attest that when the roll p/l is negative (because you have had 3 months of false positive p/l) it still sucks to take the whole hit in one day!

You mentioned that some middle / back offices have this faulty system for valuing the book that generates erroneous short-dated fwds as one gets closer to the IMM roll (I'm saying middle / back office because I can't believe that the trader does not know what the 'true price' is). But in your latest post, are you implying that some swaps traders don't have a good estimate of the expected p/l on the next IMM expiry date because of the contract rolling off and that the 'hit' comes as a surprise? If so, have you actually seen this happen in houses of repute? Have you actually seen this happen in any of the firms you traded for? Because (and correct me if I'm wrong), using this system, a 1mth fra spanning the 1st libor futures contract would be mispriced a few days before the roll by as much as 7 - 10bp on some roll dates, no?

no you misunderstand my point.... traders know it's coming - they have to otherwise they will be mispricing swaps in the transition reasons left and right... my point is that it is psychologically crappy to bleed in small amounts of p/l over 3m and then give it all back in one day....traders know its coming but that doesn't make it any more pleasant...

another alternative sometimes employed is to reserve the overstated positive p&l every day to release against the "hit" which the contract expires.