Have you ever seen that online ad, where they show you a bunch of fruit in a series of math problems, and you have to figure out what the value of each fruit is?  That’s exactly what I thought of when I looked at the Fed Fund futures.  The market price of the FF future would be the numbers on the right and you have to figure out what the factors being priced are (the fruit).  You use the values of the fruit from one level to work on the next.

I had been trying to figure out why some of the FF futures looked “fully priced” for a 25bp hike when clearly there are some risk factors that should cause the June meeting not to be fully priced.  The front FF futures just seem very well offered all the time.  So I thought I would take a closer look at the math behind each of the FF futures, so that I can make a determination as to what factors I agree or disagree with:

  • FFK8 (market 98.2975/98.30). This is fairly straightforward.  2975s have been actively trading for over a week.  The FFER has been 1.70 for the first 10 days of May, implying a FFK8 price of 98.30.  If the rest of the month kept pricing at 1.70, and month-end trades 1 bp below (or 1.69), we should get a final FFK8 price of 98.30032 (30 days of 1.70 and 1 day of 1.69).  Since there is no other meeting this month, the only way that FFK8 could be trading so actively below 98.30 is the possibility of a move to 1.71 in the FFER some time later this month.  If you think the move higher in FFER occurs Monday May 14, FFK8 would be 98.29467 (17 days of 1.71 and 1 day of 1.70 for month-end).  You can figure out the break-even for the market price of 98.2975 as being a 1.71 FFER 9 days before month-end (with the 9th/month-end day being 1.70, which would be 1bp below the previous days).  I see no flows, so I have no competitive advantage in trying to guess when this would happen.  However, this is what the markets are pricing in.  Another interpretation of FFK8 is to say that the markets think there is a 50% probability of the rate being 1.71 and staying there (except month-end) by the end of the month (assuming the FFER this past Friday is not 1.71 and assuming a uniform probability distribution of the day 1.71 could occur).  I will discuss the reservations I have later in the essay – I just want to discuss what is priced into the FFs currently.
  • FFM8 (market 98.155/98.16). From FFK8, we have that the markets are pricing in a 50% chance of a 1.71 fix by the end of the month.  Let’s assume the base rate (the rate other than month-ends) for June is then 1.705.  That means if there is no move, FFM8 should settle 98.29533 (assuming month end is 1bp lower at 1.695).  The Fed meeting is on June 13th, so the new rate would theoretically be 1.955 (25 bps higher) as of June 14th.  But if we again assume that the FFER will go up yet another bp around the time of the Fed meeting (as before the previous two hikes, that would get us to a FFM8 price of 98.14700.  Of course, the markets aren’t going to be pricing in a full 25bps, when there is no chance of a 50bp move.  From the ED options markets, we can see that there is about a 2bp discount (for example EDM8 97.75 call is 1.5 and the 97.875 call is 0.5) from a full 25bps being priced.  Leading up to the Fed meeting, I expect there to be only 24 to 24.5 bps priced in (i.e. no fully priced in).  So if we use 23.1bps of the June meeting currently being priced, with the FFER expected to increase 1 bp three days prior to the meeting (fraction thereof), we get an EDM8 price of 98.15827, which is in line with the FFM8 pricing.  Note that we are currently assuming almost ONE AND A HALF additional “1bp” increases in the FFER by the end of June.  If we didn’t have these “1bp” increases, FFM8 price would be FULLY PRICED at 98.15867 (current market).  We’ll use 0.924bps of another “1bp” increase in the FFER as being priced (since the June meeting is 23.1/25 priced) in the calculation for FFN8…
  • FFN8 (market 98.055/98.06). There is nothing going on in July, so using 23.1bps of a June hike, plus 0.5bps of a 1bp increase in the FFER in May, plus another 0.924bp increase in the FFER in June gets us to 98.05612 (current market price).  Note that if there were no additional “1bp” increases in the FFER, FFN8 would be 98.05032.  We would be a half bp from being fully priced in FFN8.
  • FFQ8 (market 98.04/98.045). The pricing of FFQ8 is wholly dependent on what you think of the odds of an August hike.  This should be roughly 1bp or so, and I suppose the markets could always add the chance of another “1bp” increase.  As with the previous meetings, if there were no “1bp” increases and no August hike (after a June hike), FFQ8 would be 98.05032.

As you may have ascertained, the crux of the FF pricings depend on what you think the odds of the “1bp” increases in the FFER are.  Keep in mind that there was ONE “1bp” increase around the time of the Dec 2017 hike (from 1.16 to 1.42) and TWO “1bp” increases around the time of the March 2018 hike (from 1.42 to 1.69), with a further “1bp” increase in mid-late April (to 1.70).  I can buy the argument for one 1bp increase.  Maybe.  But weren’t the rest of the “1bp” increases from a rising Libor, increased T-bill issuance and generally tighter financial conditions?

Libor has been coming off.  Doesn’t it stand to reason that pressures on the FFER could also come off?  The only exception would be if the “1bp” increases in the FFER was somehow due to the unwinding of the balance sheet.  It’s possible.  But I think the most likely explanation is the rise in short term rates overall (especially Libor).  And that is normalizing, with much further normalization priced into the curve.  The second most likely explanation of the rise in the FFER is the increased T-bill issuance.  And now that the Treasury has received tax payments, they do not have to issue as much.

If the markets want to price in 2+bps of a “1bp” FFER increase at every meeting, we will blow through the high end of the Fed’s target range at the end in 2-3 more hikes.  In other words, the FFER is at 1.70 currently on a 1.5-1.75 target range.  Surely the Fed will react in some way if the FFER was above 2.5 on a 2.25-2.5 target range?  The Fed has tools to control both the upper range and the lower range.  So there could be a cap to how much the FFERs could rise, relative to the target range.

Here’s something almost no one is probably considering… what if the FFER actually starts DECLINING “1bp” rather than increasing “1bp”?  That last jump from 1.69 to 1.70 on April 20 looked like it occurred on low volume, when Libor-FF was near its widest.  If volumes recover, could that put downward pressure on rates?  I’m not a market-maker, so I don’t know.  Now that ED-FF is pricing Libor-FF to narrow over 25bps from the highs, couldn’t we take that last 1bp increase in the FFER back?  How about the 1bp increase in early April 3?  Couldn’t we take that back too?  I’m not saying multiple “1bp” declines in the FFER are likely.  I’m just saying the markets don’t seem to be factoring this in.  I thought earlier in the year that any excess FFER rises (above the 25bp hikes) would decline as Libor-FF narrowed as the year went on.  I’m somewhat shocked that the FFER is priced to continue to increase – especially as we get close to the Fed’s upper bound on the target rate.

Isn’t it also possible that the month-end ratess could start trading more than 1bp below the regular monthly rate again?  Prior to the last two months, the month-ends were trading 7-8bps below the regular monthly rate.  Couldn’t March and April month-ends (where the FFER only declined 1bp) have also been affected by rising Libor and increasing T-bill issuance?  Having the month-end FFER go back to “normal” would mean an additional quarter bp of value.

These factors make a “cheap” crisis punt on the front FFs that much more attractive.  Let’s look at how the FFER trades for clues.  I don’t have the benefit of seeing flows, but I would think that as the year goes on and the markets normalize, the FFER should also normalize.  And if the FFER starts to get too unruly, the Fed may step in to put it back in line.