what are the mechanics behind building the Euro strip.. Where do you derive the weights from.. Does it change.. where is the equation.. What does it represent.. So you say the "quick and dirty" implies risk associated with what.. "non parrallel shifts in the terms structures? or is it just Basis risk.. meaning they don't normalize to the same notional value at the tick level? thats what BPV is isn't it
CD, what I'm trying to say is, just like trading an 8 legged option spread, your execution costs is huge factor in the trade. It's not meant to discourage you, but rather explain why I have stated several times that these trades are "usually" better fit for a prop firm. Prop traders use all those legs and try to execute at favorable implied prices. They are looking for that 1/2 tick edge. Just as an option market maker who wants to make markets in less liquid options, many prop traders sit out on the less liquid parts of the curve and make a market. As Marty said, you certainly could take a longer macro view, but then you most likely want to structure the trade differently. I agree, learning as much as you can is a good thing. So many of my responses are not directed at you but are broader statements on the "off chance" that other readers might be curious about it.
Cool.. I understand that respectively.. What other structures express a view about credit risk... private lending against gov liquidity.. long credit freeze trade..
Spot on Mav. Marty is one of the people here I respect, and between CD's probing questions, Marty's responses and your illuminating contribution, this is a great thread. I've always been a little curious about Eurodollar trading, now I know I'll stick to other instruments more suited to a retail trader. Still, as CD said, this is fascinating stuff and well worth knowing.
Right, this is where it gets a little complicated... Basically, the only thing that measures like BPV (and DV01) tell you is the total interest rate risk on your position. The meaning of these measures is that they tell you what your expected PNL will be when/if the whole interest rate curve shifts by 1 basis point (up or down). So, in the specific case I have given, BPV tells you that for 1000 2y note futs you need a total of 1570(ish) Eurodollar contracts, but, by itself, it tells you nothing about how many of which. In order to get an idea of how much you need to buy of what, you need to do some reasonably repetitive work. Specifically, you need to "bucket" your risk by performing an iterative procedure where you "bump the curve" (alternatively, if you want a quick and dirty computational solution there's some linear algebra magic you can do, but let's forget about that for the mom). If you want me to go into a wee bit of gory detail on how you go about "bumping the curve", let me know.
Bumping the curve.... I think I get it... I don't know linear algebra... But getting the concept is more important... Its all about making it to where the rate curves are normalized so that a parallel shift doesn't incur pnl
Yes, parallel shifts are easy... The "bumping" is done by moving a single curve input (such as one specific Eurodollar contract) by 1bp up (and down), while keeping everything else unchanged, and then rebuilding the curve. You then compute the PNL on your position that resulted from the "bump", et voila. So it's like taking a partial derivative. Once you've done that, you're gonna end up w/something that people sometimes call the "risk ladder" or "bucket risk", which will tell you not only the total risk of your position, but also how risk is distributed along the curve
Ok.. Your rebuild the curve based on that fixed point in the curve (one specific Eurodollar contract) that has been bumped up 1 basis point.. Do you derive the changes in the rest of the curve by something relative to "delta" in options.. meaning you look at historical moves and come up with contract sensitivity to the front contract for each contract.. Yet instead of using the front contract , you use a point on the eurodollar curve.. I hope this makes sense..
No, no, that would be a very different kettle of fish indeed. You do nothing to the rest of the curve. So as a specific (and simplified) example, let's say the PV of your position with the original mkt curve is X. The original curve uses, say, 99.76 for the price of the Z3 contract. Move the price of the Z3 to 99.77 and don't touch anything else whatsoever. Build a new curve with the new price of Z3 and everything else unchanged. Suppose you revalue your position with the new curve and now the PV is Y. Y-X is the risk of your position in the Z3 "bucket". Et voila.