Home Forums Main Forum Top-Down Curve Analysis

  • Author
    Posts
  • Curve Advisor
    Keymaster
    Post count: 612
    #3228 |

    Many times, when we look at a trade, we will look at the recent historicals or technicals to determine what our view of a particular trade is. However, it can also be useful to look at the curve as a whole, and develop an opinion of what is going on in the big picture, to give additional color to the our view on a smaller piece of the curve.

    I was having a discussion earlier today about the U8-U9-U0 1 year fly and how low it could go. The question was, could it get down to -5 (from the current -2 settle). Well, it could get to -5. But I thought it was very unlikely. And here is why:

    U6-U7 spread is 11.5.
    U0-U1 spread is 20.5.

    This means that the FOUR year flies in between have to equal -9bps (the value of the difference of the spreads).

    U6-U7 vs U0-U1 condor = U6-U7-U8 fly + U7-U8-U9 fly + U8-U9-U0 fly + U9-U0-U1 fly

    In the current environment, it is somewhat unlikely that U6-U7 spread noticeably lower – this is less than half a hike over an entire year. The Fed only has one ease left and that could very well occur on or before the Sept 2106 meeting (so not in the U6-U7 spread). Similarly, when there is a global yield grab, it’s difficult for U0-U1 spread to get wider. So if the spread between U6-U7 and U0-U1 is going to be loosely capped around -9bps (or hypothetically as wide as say -12bps), what kind of allocation can you make where the sum of:

    U6-U7-U8 fly + U7-U8-U9 fly + U8-U9-U0 fly + U9-U0-U1 fly = -9 bps, AND where U8-U9-U0 fly is taking up -5 of those -9bps? You would end up with a really funky fly curve, that is probably very exploitable. So if you want to make the claim that the fly can get to -5, you probably need for either: (1) U6-U7 to get noticeably lower, or (2) U0-U1 spread to get noticeably wider.

    I’m just not sure this is likely in the current environment. I’m not saying it’s not possible – just very unlikely.

You must be logged in to reply to this topic.