by Christoph Schon, CFA, CIPM
As hedge funds pile up short Treasury future positions like it’s February 2020, certain parts of the financial markets are getting nervous that we could be in for a repetition of the liquidity crisis seen in early March that year. According to data compiled by the US Commodity Futures Trading Commission (CFTC), leveraged funds have accumulated more than $850 billion worth of short positions in US Treasury note futures, causing concern about the market implications if these trades need to be rapidly unwound in a March 2020-style ‘dash for cash’. Even though the current levels are way past the $660 million peak from three years ago, we believe that the number of outstanding derivatives contracts (open interest) could climb to even higher heights, as central bank rates are expected to stabilize over the next nine months. And if bond yields begin to decline steadily, it does not have to end in tears this time.
But let us start with the basi(c)s first.
What is a basis trade?
A basis trade aims to exploit a differential in price between a derivative and its underlying cash instrument. Part of this price difference reflects the fact that entering a futures position requires almost no initial cash outlay (apart from the initial margin), whereas buying the underlying security requires funds. If a trader can finance the purchase of the cash instrument through a repurchase agreement (repo) at a borrowing rate that is lower than the discount rate implied by the difference between the future and the cash price, buying the underlying and shorting the future can provide an attractive arbitrage opportunity.
This tends to be the case in the US Treasury bond market. Asset managers often prefer using futures to gain exposure when new funds arrive, instead of immediately buying the underlying cash securities. This creates a tiny price difference between the future contract and the so-called cheapest-to-deliver (CTD) bond[1]. As hedge funds have thus far been less regulated than other financial institutions (pre-Archegos collapse), such as banks and asset managers, they can often borrow money at cheaper rates and employ more leverage—especially when posting ultra-safe Treasury bonds as collateral—which makes those trades worthwhile for them, even for small differences in repo rates.
What are the risks?
The profitability of this kind of trade heavily depends on the trader’s ability to finance the purchase of the CTD at a lower cost than the implied repo rate (IRR) and on their ability to apply leverage. Overnight repo rates are often lower than those for a transaction that would match the exact term of the future expiry, so a hedge fund may decide to roll over the financing, which comes with the risk that repo rates could rise, making the trade more expensive and forcing the trader to post additional funds. Extra collateral may also be required to keep the short position in the future open, either if the divergence between the cash and future price gets so wide that changes in the former no longer cover the variation margin for the latter, or if the exchange demands a higher maintenance margin because of increased volatility.
Margin and collateral calls are, therefore, the biggest risks for this kind of highly leveraged trade, as they can force the trader to sell assets at fire sale prices in an already distressed market, in order to raise the required capital.
Causation or correlation?
Much has been made of the connection between these basis trades and the liquidity crisis observed in the US Treasury market at the onset of COVID in March 2020, but it is by no means an established fact that the former actually caused the latter. In a whitepaper from July 2020[2], the Office of Financial Research (OFR) finds that even though many hedge funds were forced to close their basis trades during that time, there was “no conclusive evidence that these sales in turn caused greater illiquidity in the Treasury market.”
What the study did conclude, however, was that this illiquidity in turn caused losses on the basis trade, which were possibly compounded by traders buying futures and selling cash treasury bonds in order to meet the greater margin requirements. Yet, the authors make it very explicit that they were “unaware of any hedge fund defaults associated with the basis trade” during that time.
How big is the problem?
As hedge funds are notoriously secretive about their holdings, it is difficult to estimate the exact total size of basis trades outstanding. The Federal Reserve notes in another recent article[3] on the topic that only around 10% of total hedge fund repo transactions are conducted through centrally cleared venues, which renders one leg of the trade virtually invisible. There is, however, a dataset available that can give us an indication of the other, exchange-traded half of the strategy.
The charts in Figure 1 show an extract from the weekly Traders in Financial Futures dataset compiled by the CFTC, which reports open interest in financial futures on American exchanges, broken down by type of institution, e.g. dealer, asset manager, and leveraged funds. For the purpose of this analysis, we show the long positions of asset managers (blue lines) and the short positions of leveraged funds (green lines) in the three most liquid Treasury note contracts. We have also added the dashed amber lines, which show the effective federal funds rate in the chart for the 2-year note contract and the 10-year US Treasury yield in the chart for the corresponding future, in order to highlight the close relationship between the level of rates and the open interest rate in these contracts.
Sources: CFTC Traders in Financial Futures data, U.S. Treasury, Board of Governors of the Federal Reserve System
As the charts illustrate, there clearly seems to be some interplay between the level of interest rates and the number of positions in futures contracts, which makes sense for both sides of the trade. Rising yields make bonds more attractive to investors, increasing the flows into fixed income funds. When asset managers receive those inflows, they may choose not to purchase the bonds directly but buy a future instead, which explains the rise in long positions. At the same time, hedge funds might decide to take advantage of rising interest rates by shorting bond futures—especially at the start of a central bank hiking cycle—which could explain the surge in short positions on the 2-year notes in both 2017/18 and since March 2022.
Enter the basis trade…
But if all of that is true, why then did the short open interest continue to climb once the Federal Reserve was finished raising rates at the end of 2018 (denoted by the first vertical dashed line in the top left chart of Figure 1)? This is where a basis trade comes in. Due to its high leverage and sensitivity to interest rates, the trade depends on a certain amount of stability in the level of those rates. This was the case in the first half of 2019, at least at the short end of the Treasury curve, which is why most of the increase in open interest occurred in the 2-year note, whereas the 10-year contract attracted no further short sellers, while long positions even declined in tandem with the corresponding yield.
It was only when the Federal Reserve started loosening monetary conditions again in July 2019 (the second vertical dashed line) that futures open interest rates also began to decline. Initially, the decrease was gradual, which would have given market participants enough time to unwind their positions in an orderly manner. However, the rapid, successive emergency cuts of 50 and 100 basis points on March 3 and 15, 2020, respectively, led to much bigger price swings (the third vertical line).
In such an environment, investors prefer more liquid instruments, such as futures and on-the-run Treasury bonds. However, the CTDs that the hedge funds would have held as the long legs of their basis trades would have been older, less liquid, off-the-run issues, so that their price appreciation may not have been sufficient to make up for the losses from the short futures position, especially when selling bonds in an emergency situation.
So, how likely are we to see a repetition of this scenario in the near future?
Where to from here?
When comparing the current environment with that of the previous Fed hiking cycle, one could argue that we are only just coming to the end of the first phase of rising interest rates, in which asset managers have built up long futures positions to satisfy the increasing investor demand for fixed income funds, while hedge funds have shorted futures to benefit from falling bond prices.
If we now assume—in line with market expectations—that the Federal Reserve will keep rates around the current level for the next nine months, the precedent from 2019 would indicate that open interest in Treasury futures will continue to rise, as fixed income fund managers enjoy ongoing demand for their products, while hedge funds scramble to pick up the nickels in front of the basis trade steamroller.
Eventually, though, market participants will start to anticipate the first rate cuts from the Federal Reserve, and bond yields will once again begin their descent. For example, in the 14 months between the start of 2019 and end of February 2020, the 2-year and the 10-year US Treasury steadily and gradually declined by around 150 basis points each. The asset manager open interest in Figure 1 indicates that this was accompanied by a reduced appetite for long futures from that investor group. The removal of that extra demand could have just been sufficient to make the basis trade less lucrative, thereby also decreasing the desire for short positions from the leveraged accounts.
Compare this to a 60-basis point drop in the 10-year yield between February 28 and March 9, 2020, a rebound of the same size over the next nine days, only to shed the same amount again over the remainder of the month. Pair this with 150 basis points of rate cuts from the Federal Reserve, and you will get the kind of volatility that will squeeze hedge funds out of their basis trade positions through margin and collateral calls.
Is this likely to happen again soon? Let us hope not…
[1] It may be worth noting that this type of basis trade focuses specifically on the price differential between the futures contract and its particular CTD, which both have the same maturity. It is therefore market neutral by its very nature and does not take or reflect any views on the overall direction of monetary policy, interest rates, or yield curve dynamics.
[2] https://www.financialresearch.gov/briefs/files/OFRBr_2020_01_Basis-Trades.pdf
