
Hello everyone and welcome back! This week’s Chartbook will focus mainly on 10 year Treasuries and some technical factors affecting the markets for these securities. Since the start of 2021, there has been significant borrowing, repo market imbalances, and a notable shift in futures positioning around the 10 year maturity. To understand how these parts could all fit together, let’s start by looking at some NY Fed securities lending statistics.
Treasury Securities Lending
Alongside its many other activities in the Treasury and Treasury repo markets, the NY Fed provides the primary dealers with a securities lending facility (see details here). Through this program, primary dealers can use available Treasuries as collateral with the NY Fed to borrow Treasuries that the Fed holds in its own portfolio. The next day, the transaction is reversed and the Treasuries are returned to their original portfolios. Since primary dealers must pay a minimum 5 basis point annualized rate on trades through the facility, they are discouraged from borrowing securities unless they have a valid commercial incentive to do so. One such incentive may be high demand for a specific security from other borrowers. If a Treasury is in high demand in the repo markets and repo rates on it are significantly below general repo rates (called a special), primary dealers can intermediate between this demand and the Fed’s supply by re-lending a security borrowed from the Fed. With this background out of the way, let’s check out the statistics for the 10 year note maturing Nov 15, 2030.
The blue and black bars on this chart show the volume of bids to borrow this security accepted and rejected by the NY Fed over the course of this year. The red marks show the weighted average rate at which dealers borrowed from the Fed on each day. We see that for the first six weeks of the year, roughly $8B was borrowed every day in this security, which was at that moment the on-the-run 10 year maturity. However, the red lines show that in mid-January there was a significant spike in borrow rates and a number of rejected bids (which occur when dealers bid in excess of the Fed’s explicit or discretionary limits). There was also a smaller rise in rates in early-February, followed by borrowing volumes dropping off in late-February as a new 10 year note was auctioned and this security became off-the-run.
The sharp rise in rates to 1.8% in mid-January indicates that primary dealers were experiencing high demand for this specific note, likely from short sellers willing to pay a high financing rate to borrow the on-the-run 10 year issue. If short sellers are taking a directional position, high annualized rates like this are somewhat sustainable, since the directional moves can make up for the cost of funding if gains are realized quickly. However, for arbitrageurs who are looking to go short a Treasury and long a related instrument to arbitrage a spread, high funding costs like this are prohibitive since they would erode much of the available profit. Now let’s look at the 10 year note maturing on Feb 15, 2031 which replaced this one as the on-the-run 10 year maturity.
In this chart we see the same view of the statistics for the February note, which first became available for borrowing in the second half of that month. We see that in the first weeks of March the demand and borrow rate for these notes was very high, with rejected bids showing dealers bidding near their size limits and rates approaching the important threshold of 3%. At a rate above 3% it becomes questionable to lend US Treasuries since this is the Fail Penalty charged at most repo venues for failing to deliver a security to close out a trade. If it becomes cheaper for shorts to fail to close out their trades and pay the Fail Penalty versus paying the contractual rate, repo lending of securities is problematic. In the first two weeks of March the on-the-run 10 year approached this territory, but borrowing has cooled off since then after a smaller peak in early April. Now let’s look at some more indirect signs of this activity that show up in broader rates markets and bank financial statements.
In this chart we see the average overnight rates for cash in the DTCC Delivery vs Payment Service (details here), the DTCC General Collateral Service (details here) and general collateral tri-party repos traded through BNY Mellon (details here), in white, blue, and red respectively. The Delivery vs Payment Service (DVP) is special since it is the only one of these rates that is not based on general collateral. A DVP repo through DTCC must be settled with a specific security and is thus a way for traders looking to borrow Treasuries to source the specific issue they are interested in. We see from the chart that the DVP rate plunged well below the general collateral rates briefly in mid-January and more dramatically in the first half of March, corresponding to the periods where demand for on-the-run 10 year notes was highest at the NY Fed. This is likely because short sellers were borrowing these notes in significant enough size in the DVP repo service to drag the average rate down (as a high borrow rate for Treasuries corresponds to a negative rate to borrow cash on the other side of the transaction). Finally, now that all of the major banks have released their Q1 details, let’s look at which ones were participating most actively in lending their Treasury portfolios as well.
This chart is an update of one presented in Chartbook #6, showing average quarterly securities lending activity in the repo market for four major US banks. As we can see all four of these stepped up their securities lending volumes in Q1, with JPM and GS showing the largest increases. This is a logical move from the banks as they hold Treasuries in size and have access to the Fed to borrow more if need be. When there is demand to borrow these securities in the broader market, they can make a low-risk profit by lending out this portfolio in repo.
Now let’s consider another source of demand for Treasuries to borrow: arbitrageurs. These players are more rate sensitive, as they will not borrow if available rates are too high, but there are a few signs that they may also active in the background.
Treasury Futures Basis
To understand how arbitrageurs fit into the picture, let’s consider a simple Treasury futures strategy. There are many futures products available on exchanges that are settled for ‘physical’ Treasuries at maturity. Logically, when the maturity date arrives, these futures should trade at exactly the price of the underlying bonds or notes. However, prior to maturity there might be slight deviations from this pricing due to various market imbalances. In this hypothetical strategy, an arbitrageur looks at whether bonds or futures are trading at a higher price, and goes long the cheaper instrument and short the more expensive one. Over time as the contract approaches maturity, the arbitrageur makes a low-risk profit as the prices converge, though there is some risk as they may first diverge more before converging (see OFR paper here on how some of these trades went wrong in March 2020). This type of strategy is called a futures basis trade and is popular among some funds that take such positions with leverage to generate yield. To start, let’s look at levered fund positioning in some of the futures contracts tied to 10 year Treasuries.
This chart shows the net positioning of levered funds in the CME 10 year note futures (contract code TY, white line) and the CME Ultra 10 year note futures (contract code UXY, yellow line). We see that levered funds recently went net long the TY contract for the first time since 2017, and have built up the largest reported net long position in the series for the more recently introduced UXY futures. The difference between these two contracts is the underlying basket of Treasuries that can be delivered at settlement. For the ‘regular’ TY contract, any Treasury with between 6.5 and 10 years to maturity is acceptable for delivery, while for the ‘Ultra’ UXY contract only notes originally issued as 10 years with at least 9.5 years remaining are acceptable. This restricts the UXY contract to only two securities suitable for delivery (since 10 year notes are issued quarterly) while the TY contract has a large range of options. Though both contracts use a conversion factor to normalize deliverable issues of different coupons and maturities, this calculation assumes a locally flat yield curve at a constant 6%, so in practice there will always be a specific security that trades at the most favorable market price for delivery (called the cheapest-to-deliver or CTD). In the TY contract, recent market conditions favor the shorter end of range, so the contract trades more in line with the 7 year maturity than the 10 year. Since the UXY contract only has the on-the-run 10 year and the 10 year immediately prior to it, these considerations are less relevant there.
The resulting market behavior is shown here in the implied yields of the TY (blue) and UXY (white) contracts. When overlaid with the generic 10 year (red) and 7 year (green) yields, we see that the UXY implied yield closely tracks the 10 year while TY tracks the 7 year. Though UXY is in theory a more ‘accurate’ 10 year note futures product, it is an evolution of the original TY product which remains more heavily traded, as shown by the open interest in the bottom panel.
Now let’s return to the point of view of an arbitrageur. If you wanted to take advantage of a small price dislocation in the TY or UXY futures relative to the underlying Treasuries, you would have to find the cheapest-to-deliver notes and take a long/short position. One way to think of this is as an implied repo trade where the arbitrageur buys the notes outright for cash and sells a futures contract to deliver them into some time later. Once the term of the ‘repo’ is over, the arbitrageur unwinds the position by delivering the notes and picks up a profit equal to the implied repo rate over the term of the trade.
This chart shows the implied repo rate for the June 2021 TY contract relative to its cheapest-to-deliver Treasury (a 2028 maturity note). A positive implied repo rate indicates that an arbitrageur may profit by being long the Treasury outright and short the futures, provided that the cost of leveraging the position does not erode too much of the spread. From the chart we see that while the implied repo rate was fairly high in the last quarter of 2020, it collapsed in December and went slightly negative in late-March. This means that currently arbitrageurs could profit slightly in the other direction (going short an outright Treasury and long the TY futures) provided funding costs were very low. Now let’s turn to the UXY contract which more directly affects the current 10 year note.
Here we see the implied repo rate for the June 2021 UXY contract against its cheapest-to-deliver Treasury (the on-the-run 10 year). While implied repo rates in UXY were also high towards the end of last year, they have fallen much more significantly and are now more than 1% negative. This means there is significant incentive for arbitrageurs to be long the UXY futures and short 10 year Treasuries outright, provided they can borrow the notes to sell short. Since the implied repo rate here is in the neighborhood of 1%, arbitrageurs would not drive borrow rates to around 3% as speculative shorts would, but given cheaper supply could possibly contribute to a significant short base. Given that we see a large long position by levered funds in the UXY futures, it seems reasonable to think that this is indeed happening and there is an offsetting short position in 10 year Treasuries.
That is all for this week! As always thanks for reading if you made it all the way through and hope to see you back next week for another Chartbook :)
Cheers,
DC