LOIS Maximus

Since this is my first post, hello! In the words most commonly used during the reign of Mike Francesa on WFAN, long time first time. I've agreed to do some posts since MM hung up the shingle a couple weeks ago, and am thrilled to be getting the chance at it. Please feel free to reach out to me at the following addresss with any questions, comments, concerns, or otherwise: cackalackcap@gmail.com.

Today we'll be returning to the wonderful world of STIRs, and specifically the now-falling LOIS spread. This has been dealt with a number of times lately by Jurofalco ("An Exercise In Connecting The Dots") and Shawn ("The Anatomy of a LIBOR Panic" and "Why Is LIBOR Moving Higher"). I'll assume given that background readers are basically familiar with the story.

Since peaking on April 6th at 59.6 bps, the 3m LIBOR-OIS spread has slipped below 55 bps.  That date is a nice, convenient trading week following the close of Q1, a date we think is significant. In the prior recent LOIS blowout, spreads peaked out just before an October 14th 2016 deadline for new fund rules which forced huge rotation out of prime funds and into government backed funds.

This time, it's not a question of money market funds. Instead, as the prior posts detailed, the catalyst has been less clear but there's broad consensus that tax reform has played a huge role. Included in the year-end tax cut for US corporates was a provision that incentivized repatriation of offshore cash, specifically 15.5% on all offshore income rather than the previous 35% rate when repatriated; a lighter 8% levy applies for illiquid assets but the 19.5% lower rate for cash/cash equivalents is the key thing here. As a result of that repatriation, there have been a wave of announcements around what companies will do with that cash and other benefits from the tax cut. 

That all sounds boffo for USD! Unfortunately, it's not that simple. Offshore assets were offshore on paper only. For instance, see the tale of Braeburn, Apple's Reno, Nevada-based credit mutual fund corporate treasury operation. Profits earned abroad weren't kept in local currency, but instead sold into USD and held in accounts invested in USD. So buying USD on a repatriation thesis didn't work out so well. But the investment of those USD assets is where LIBOR comes in.

It might be helpful to look at this visually. The chart below shows the share of US corporate assets sitting in liquid low risk assets (time/savings deposits, checkable deposits, money market funds, UST, Agencies, munis, mutual funds, and foreign deposits). Since the start of the current phase of US corporate-lead globalization, the share has nearly doubled. Short term interest rate assets only (time/savings deposits, checkable deposits, foreign deposits, and money market funds) have seen similar increases.

In effect, the US tax system has created an incentive for big chunks of cash to sit "offshore" in low risk investments. At the same time, the offshore market has lapped up that dollar funding. Now, in a one-off shock, that dollar funding (or at least, a large chunk of it) has headed out the door. Eventually, it will work its way back into the financial system, but probably not in the same way. With Q1 now in the books, and the biggest marginal flow of the stock of offshore dollar assets finished flowing out, rates are now reversing. That's visible in both the LOIS spread (offshore dollar deposits) and the CP-OIS spread.

There's one more piece to this story. FTAV has done a good job covering Credit Suisse's Zoltan Poszar's theory that specific provisions of tax reform are resulting in a shift in behavior for international banks' New York branches. This theory suggests those branches will shift from borrowing parent company cash to funding with commercial paper, and that repatriated cash will then be lent out via parent companies in the FX swap market (or, alternatively, those parent companies will need less USD borrowing via FX swaps because they now have USD funding back). Under this theory, commercial paper issuance by foreign branches should be rising, while the amount due to foreign offices from those branches should be falling. Instead, on a net basis, the amount due is going up.

There's an argument about gross flows here to be made, but for now, keeping things as simple as possible, if tax changes are forcing US branches of foreign banks to reduce funding from the home office and replace it with CP, it seems unlikely that net liabilities to foreign offices would be rising sharply. The second leg also doesn't make much sense when taking a look at the data. Foreign commercial paper outstanding is up a paltry $13.5bn since the end of Q3, hardly the stuff of massive funding market shifts!

So, takeaways. The argument we're making here is that LOIS's spike isn't going to lead to a permanent plateau. Similar to money market reform, it was a one-off event that the market had to digest. Changes in US tax law haven't created a permanent spread between LIBOR and Fed funds, but did create a lot of selling pressure for a while. Barring new marginal widening pressures like a large spike in credit spreads more broadly, LOIS should contract considerably from here. 

The easiest trade to would be to receive LIBOR versus OIS swaps, or own ED contracts versus forward OIS. If you're not able to avail yourself of the OIS market, an easier to execute version would be to own ED futures versus Fed Funds futures. As shown in the chart below, these are respectable proxies for spot LOIS, though of course they're not exactly the same thing. Assuming (perhaps too optimistically) a very rapid unwind of the LOIS spike, the June Eurodollar contract (M8, which has a valuation date of June 18th) should settle close to 25 bps tighter versus the average of the July (N8) and August (Q8) Fed Funds futures. A less aggressive approach would be to receive the September Eurodollar (U8) versus October (V8)/November (X8) Fed Funds futures. That gives more time for the spread to tighten, but the spread is already closer to its pre-blow out levels of ~20 bps. Because EDs trade with a DV01 of $12.50/tick and Fed Funds trade $20.835/tick, actual execution would be long 4 EDM8 versus short 1 each FFN8 and FFQ8 (or the later maturity, outlined previously). The best case for either version would be a move to 20-25 bps, with stops of 50 bps (EDM8) or 42 bps (EDU8) give-or-take. If held right to "maturity", the whole thing should be sold in the days before EDM8 or U8 cash-settle, to save a load of headaches.

Astute readers will note that we're sort of cheating here; Eurodollars mature to the value of LIBOR over a 3 month period, but we're only using Fed Funds futures for 2 months. Unfortunately, the value dates, maturity, and mechanics of the two different futures contracts don't work out exactly, so barring some might specific hedge ratios this is just the most efficient way we can see to put on a LOIS tightener using generic futures. Do get in touch if you're aware of an alternative!

Edit: An astute reader has pointed out that to get a perfect hedge ratio of $125 DV01 per leg (technically $125 ED and $125.01 FF) you would need to buy 10 EDs for every 3 of each FF (6 total), versus the 4:(1+1) ratio advocated above. That is correct, and if you want to add the extra size, it's probably the best way to go about it.