Mike GParticipantFebruary 16, 2017 at 8:15 pmPost count: 20
Just wanted to say the information you have provided here in this forum is very educational. Thanks so much.
How do you interpret the prices of the Butterflies and Doubleflies? I am having a hard time relating the fly and d-fly prices to the the fundamentals (fed rate hike path, economic data releases, etc) Can you elaborate more about this?
Curve AdvisorKeymasterFebruary 17, 2017 at 7:46 pmPost count: 612
There are a lot of factors affecting the shape of the curve. By far the largest driver is the future path of Fed rate moves (assuming we are talking about US interest rates). This is not to say there aren’t other factors, including positioning, general supply/demand, risk premiums, inflation expectations, convexity, discounting, to list some other factors. But all other things being equal (and they may not be), you would expect interest rates to eventually converge to some function of the Fed rate path. Typically for Eurodollar futures, this means settling to some function of the Funds funds over a 3 month period, plus a spread for 3mo libor.
You can consider a calendar spread (say EDZ7-EDZ8 spread) as the amount of hikes you would expect in the period covered between the EDZ7 and EDZ8 contracts (note that this will be roughly 1/2 the Jan 2018 meeting, all the rest of the 2018 meetings and 1/2 of the Jan 2019 meeting). So if we assume all the other factors are somewhat negligible, a EDZ7-Z8 spread of 50bps would mean there were 2 hikes priced in. Similarly, a Z8-Z9 spread of 35bps would mean there were 1.4 hikes priced in. When you look at the fly (which would be 15bps), you are comparing the relative number of hikes between the two periods. This may or may not help you visualize the curve.
Another way to think about the fly height is as a relative measure of the landing probabilities of the end of the hiking cycle. For example, if it was somehow known to everyone that the Fed was going to hike 2x a year, and do their last hike in Dec 2018, you would expect the EDZ7-Z8-Z9 fly to be 50. Well, we don’t know with certainty where that last hike will be, so there will be some curvature around the most likely candidates of the last hike. So you can think of the fly curve as a function of the probability distribution of the last hike. I discuss this a little bit in my first CME article. Note that depending on the pace and length of hikes the curve shape can look differently.
Note that this is a gross simplification so that you can visualize a major driver of the curve. But I think it helps you visualize what is going on.
Mike GParticipantFebruary 19, 2017 at 9:08 amPost count: 20
Thanks so much for getting back to me! I am slowly getting a feel for what the prices are implying!
I’ve attached 3,6,9, and 12 month turn adjusted fly curves as of 2017-2-17. I have 2 more questions.
1). From the images, I find that 3M and 6M flies are rather choppy. Am I incorrectly plotting this? The article you linked shows rather smooth flies curves. Is the current regime rather different? How do your interpret 3 and 6M flies currently? More specifically, are there any information from them?
2). What I’m a bit still stuck on is the following: a) when fly prices are positive, b) when fly prices are negative, c) when fly prices are increasing and d) when fly prices are decreasing. Can you elaborate on this more?
Attachments:You must be logged in to view attached files.
Curve AdvisorKeymasterFebruary 20, 2017 at 9:04 amPost count: 612
1) well the fly curves are “jagged” now partly because there are larger-than-usual wagers being placed on particular meetings. In particular the Dec meetimgs, because the last (and only) two hikes have been in Dec. So the “turns” you back out may look unusually large. However, the 6 mo flies starting with “H” and “U” have no turns so they should be cleaner. There is a lot more mark to market noise (as a percentage if the fly level) on the narrower flies, so some more choppiness should be expected in 3 and 6 mo flies. The date on your charts are incorrect – let me look at the previous day when I get back in the office.
2) Just to (over)simplify, when there are more hikes (and possibly a smaller ease tail) in the front spread than the back spread of a fly, the fly will be higher and vice versa. When flies increase, the hikes in the earlier spread leg will be perceived to be larger and vice versa. Just watch how the 6-12 month flies move on various types of economic data and news releases. In general, stronger growth data will cause more curvature further out. Stronger inflation data will cause more curvature closer to the front of the curve. QE (even abroad) will tend to affect the curve further out (area most affected by demand).
Curve AdvisorKeymasterFebruary 20, 2017 at 5:48 pmPost count: 612
I took a look and it does look like it’s a turn-adjustment “issue”. There is nothing wrong with the H and U flies (since there is no turn). If you just look at those, you will probably see the shape resembles something closer to the shape of the year flies. I mentioned somewhere that two out of the only two hikes the Fed has had has been in the Dec meeting. This is going to cause the Dec pricings to be higher than the other meetings. You probably want to make some adjustment for this in your turn calculation. Because currently, you are accounting for all of this type of Dec pricings as “turn”. One thing you can do is to set a cap on the year-end turn. I am currently using 1.25bps.
I suppose this is a probably a good take-away. Now that the meetings are more “gradual” and more probable, you can expect the markets to take positions on certain meetings. For example, June has also been an attractive candidate for some time. So you can’t just “smooth” the curve because the historicals looks good.
Take a look at what happened to the U7 3mo double fly earlier this year… People are going to start taking bets on specific meetings.
Attachments:You must be logged in to view attached files.
NickParticipantApril 5, 2018 at 9:48 amPost count: 8
I assume this is the best thread for this question – although different from the previous…
Re-reading the first CME paper you wrote, in the section Example of Nonlinearity, Note 5 you state you implied the rate from ED12. How do you do this? I assume you mean (forward) LIBOR, otherwise you would just look at Fed Funds futures and wouldnt need to imply anything from ED??
…or am I missing something obvious?…
Curve AdvisorKeymasterApril 24, 2018 at 3:33 pmPost count: 612
Apologies for the delay in replying. When I referred to the “rate implied by ED12”, I just meant
rate = 100 – ED12
But the “ED12” I used was a constant maturity generic 12th quarterly ED contract. Constant maturity contracts ensure that you are comparing apples to apples in terms of the maturity that each ED contract is expressing. For example, ED1 can have as few as 1 day to settlement to as many as 98 days (which could be the equivalent to and ED2) to settlement. So when looking at historicals, you want to make sure you are looking at EDs that are the same days from maturity.
You must be logged in to reply to this topic.