160815 ED-FF term structureEver since I put out a few libor-spread-narrowing trades, I have been looking at the term structure of ED-FF spreads more carefully.  I thought the following was interesting.  I calculated a simple approximation of the ED vs FF spreads for each of the first five ED contracts.[1]  The two things that really stand out are:

  • There is a large drop-off between EDU and EDZ. As mentioned previously, there is a good chance libor-FF narrows after October 14 (some time reasonably soon thereafter).  I had suggested for the past few weeks that EDU-Z calendar spread could be much lower than the odds of the December Fed meeting.  You can see the markets pricing this in.  Currently, EDU-Z spread (two total meetings around Dec) is 4bps (or 3.25bps if you subtract a year-end turn) and just the Dec meeting is 7.75bps (in FFs).  On Friday, there was a 5bp difference in the ED-FF spreads priced in between EDU6 and EDZ6 (42.7bps for EDU6 vs 37.7bps for EDZ6, as per the table).  Even as recently as July 22, there was no difference in ED-FF spread between EDU6 and EDZ6.
  • There is NO drop-off in ED-FF spreads between EDZ6 thru EDU7. In fact, all these spreads (for each column in the yellow box) are almost exactly the same!  Some of those late 2017 FFs are not very liquid, so the fact that ED and FF managed to stay so aligned is not a coincidence.  On the one hand, my “most likely” scenario is that we get a massive drop-off in libor-FF spread by the end of the year.  The term structure of ED-FF being flat is consistent with this story.  But other possibilities include libor-FF narrowing slowly over time, as well as scenarios I think are less likely, like libor-FF widening by year-end or staying the same.  If we do see libor-FF narrow a lot by year-end, you would expect the ED-FF spreads in EDZ6 thru EDU7 to be mostly the same.[2]  I’m guessing the locals or some algo is trading the ED-FF this way.  But considering how wide libor still is, you would think that as time goes on, the odds of libor-FF staying this wide should decrease.  This is especially so, since the swap spread curve decreases over time, and eventually goes negative.  So having all the ED-FF spreads be a constant over time (and a LARGE constant at that), does not seem correct to me.  Hence the reason for Trade F31.  By EDZ7, I would think libor-FF should be back to some “normal” level.

You should develop your own view of the timing of any libor-FF normalization.  Admittedly, I don’t sit around watching short term money flows or anything that detailed.  We have negative information asymmetry – there are people who have a lot more information than we do.  However, that table to me does not seem like a likely equilibrium term structure for ED-FF.  My #1 scenario is that before year-end (EDZ6) is the point at which libor collapses.  So perhaps the flat term structure exhibited in EDZ6-FF thru EDU7-FF could be reasonable.  My #2 scenario is that it take a little longer for the markets to adjust, and we may go into EDH7 (and possibly longer).  So it seems like the ED-FF spreads should on the margin get narrower over time.

In any event, if the locals and algos are going to keep the flat term structure in ED-FF through the front of the reds, this makes it less unattractive to hold steepeners in EDs.  This is probably why the year flies near the front of the curve went back positive last week (despite the rally).  But just as a precaution, I still like swapping any steepener exposure in EDs for FFs.  And stay away from any derivation of the EDU6-Z6 spread.  That could be toxic waste.

[1] Basically I compare an ED contract with the average of the two FF contracts that are 1 and 2 months afterwards.  For example, I look at EDU6 vs the average of FFV6 and FFX6.  I could calculate this exactly, but this is close enough for an estimate – especially since the closes themselves can be noisy.

[2] You could argue for a small degree of positive or negative slope over time, which I don’t want to get into now.