TBAC Basis Trade Presentation
TBAC Basis Trade Presentation
TBAC Basis Trade Presentation
“According to Treasury futures positioning data, asset manager long positions and leveraged fund
short positions have increased significantly. Please discuss the factors that could be driving this
dynamic. What are the reasons for asset managers to prefer the Treasury futures market over the
cash market for their duration needs? What type of activity does the leveraged short positioning
reflect; for example, to what extent could this be cash-futures basis trading activity? What are
important factors to consider when monitoring changes in cash-futures basis positions?"
1
Summary
➢ In this presentation, we examine the large increases in asset manager longs and
leveraged fund shorts in Treasury futures and propose potential structural factors that
may be contributing to the persistent increases over time.
➢ We discuss specific reasons why asset managers use Treasury futures and find that
their strategic asset allocation has resulted in lower Treasury allocations and a potential
structural need for Treasury futures.
➢ We also discuss how leveraged funds use Treasury futures, and what types of levered
strategies might result in consistently opposing positions relative to asset managers.
➢ We conclude that relative-value trading strategies are the most likely driver of structural
leveraged fund shorts in Treasury futures and propose some relevant relationships for
Treasury to monitor on the Treasury futures basis trade.
➢ We suggest that a mutually beneficial relationship exists between asset managers and
leveraged fund positioning in Treasury futures.
2
Review of CFTC Futures Data
3
Overview of CFTC Positioning Data
➢ The CFTC Traders in Financial Futures (TFF) report separates large traders in the
financial markets into four categories(1):
4
(1) Information sourced from CFTC Traders in Financial Futures (TFF) Explanatory Notes
Overview of CFTC Positioning Data
Figure 1: Treasury Futures Open Interest (Face Amount) vs. Figure 2: Treasury Futures Open Interest (Face Amount)
Treasury Coupons Outstanding (ex. SOMA) % of Treasury Coupons Outstanding (ex. SOMA)
2.5 20 35%
30%
2.0 16
1.0 8 15%
10%
0.5 Futures Open Interest (Left 4
Axis, $Tr) 5% Reduced Leverage /
Contained Range Post GFC
0.0 0 0%
06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
(1) Data on Treasury futures is sourced from the CFTC Traders in Financial Futures report as of Dec 2023 5
(2) Data on Treasury securities is sourced from US Treasury and NY Fed as of Dec 2023
Overview of CFTC Positioning Data
➢ Recently there have been large increases in both asset manager and leveraged fund positions in
Treasury futures.
➢ Since November 2021, asset managers (AM) have increased their net position from $100bn to over
$700bn, while leveraged funds (LF) have moved from flat to short almost $800bn. (1)
➢ When normalized by the size of the Treasury market (2) in Figure 4, recent trends look similar. One
interesting thing to note is the size of AM and LF positions is similar to 2019 peak (not larger).
➢ In the 2006-2007 period, AM and LF relative positions were flipped. This change could indicate a
structural evolution in the ecosystem and how different investors behave.
Figure 3: Net Treasury Futures By Investor Type Figure 4: Net Treasury Futures by Investor Type
($Bn Face) (% Treasury Coupons ex. SOMA)
1,000 6%
800 Asset
4% Managers
600 Asset
Managers
400
2% Other
200 Other
0 0%
-200 Dealer
-2% Dealer
-400
-600 Leveraged
Funds -4% Leveraged
-800 Funds
-1,000 -6%
06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
(1) Data on Treasury futures is sourced from the CFTC Traders in Financial Futures report as of Dec 2023 6
Overview of CFTC Positioning Data
➢ Most of the net position in futures is concentrated in shorter duration contracts (TU, FV, and
TY) likely due to larger amounts of Treasuries outstanding in these maturity buckets.
➢ Currently, these three contracts represent 80% of the net asset manager notional position and
55% of the net TY equivalent position.
➢ For purposes of the analysis we have focused on notional exposure, as opposed to DV01
exposure, because it is most consistent with quantities of repo/funding needs.
• That said, there are likely important things to consider around DV01 exposure, since it would more
accurately reflect interest rate risk concentration in longer duration securities.
0 0 0
7
(1) Data on Treasury futures is sourced from the CFTC Traders in Financial Futures report as of Dec 2023
Overview of CFTC Positioning Data
➢ We also show data for US and WN contracts (note the axis change) as well as all contracts in
TY equivalent terms as of the end of 2023.
➢ While the contours are not all the same, the broad trends recently (expansion of AM longs and
LF shorts) are consistent across all Treasury futures contracts.
Figure 8: Net US Position Figure 9: Net WN Position Figure 10: Asset Manager
($Bn Face) ($Bn Face) Current Face Amount & TY
Equivalent Breakdown
80 80
800
60 60 Asset
Asset
Managers 700 WN WN
Managers
40 40 US
600 UXY
20 20 US
500 TY
UXY
0 0
400
TY
FV
-20 -20 79%
300
Leveraged FV 55%
-40 -40
Funds Leveraged
200
8
(1) Data on Treasury futures is sourced from the CFTC Traders in Financial Futures report as of Dec 2023
Asset Manager Discussion
9
How do Asset Managers Use Futures?
➢ Active asset managers likely use Treasury futures for the following reasons:
1. Could simplify execution and operations (negates need for cash / repo trades).
2. Eliminates reporting repo interest expense.
3. Guidelines might not allow repo leverage (mostly separate accounts).
4. Likely more leverage flexibility compared to 40-Act repo limits.
10
How do Asset Managers Use Futures?
➢ The CFTC data shows that in recent years, asset managers have had a net long bias in Treasury futures,
even with large swings in Treasury yields (10y yield of 0.5% in ‘21 to 5.0% in ’23). Why would this be?
➢ First, what types of positions might cause asset managers to be net long Treasury futures?
Figure 11: Asset Manager Net Futures Figure 12: Asset Manager Empirical
vs. 10Y Yield Duration Using Three Factor
Regression(1)
1,000 6.0
Asset Managers Net 7.0
800 Index
Futures ($bn Face, 5.0 6.0 Fund
Left Axis) 5.0
600
4.0
4.0
400 Average of 20
3.0 3.0
200 Average “Active” Large Active Core
2.0 Duration Plus Managers
2.0 1.0
0
0.0
-200 1.0
10Y Yield -1.0
-400 (Right Axis) 0.0 -2.0
06 08 10 12 14 16 18 20 22 24 17 18 19 20 21 22 23
11
(1) Sourced from Bloomberg and the presenter’s firm’s calculations. Fund returns are regressed on US 10Y, DXY, and IG Corporate OAS. Data as of 01/17/2024
Asset Managers: Product Mix Tends to Favor Futures
Relative to Treasury Repo
➢ Below, we show in Table 1 a broad overview of different asset management products
with associated regulatory requirements for repo and futures usage.
➢ We focus on active fixed-income, since passive strategies, which largely track cash
indices, are probably less likely to use Treasury futures.
➢ Based on industry estimates(1), we conclude that most of CFTC asset manager data
represents mutual funds and actively-managed separate accounts.
Table 1: Asset Management Products Regulatory Requirements for Repo and Futures Usage
Active ETF SEC / 1940 Act Prospectus/SAI Yes Yes(2), 50% Yes
(1) When looking at eVestment data for several large asset managers, we found that nearly 90% of AUM was in mutual funds and separate accounts as of Q3 2023
(2) 40-Act limit depends on the mutual fund’s approach to repo under SEC rule 18f-4. Prior to 2020 introduction of 18f-4, repo was typically limited to 50%. 12
Asset Managers: Product Mix Tends to Favor Futures
Relative to Treasury Repo
➢ Mutual funds are regulated by the Investment Company Act of 1940 (“40-Act”)
➢ While established in 1940, various rules and amendments to the 40-Act have been adopted over time.
➢ For many years prior to 2020, the applicable rules created incentives for mutual funds to favor derivatives like futures
over repo in certain cases, including limitations on the size of repo borrowing.
➢ Rule 18f-4 adopted in 2020 established a new option for mutual funds to treat repo like other derivatives (which are
subject to a “VaR test”), though some funds may not have updated their documentation or practices to utilize it. Given
the change was relatively recent, its potential impacts warrant further investigation.
➢ If not using this new option, mutual funds must limit their repo borrowing to 50% of NAV to meet the applicable asset
coverage requirements.
➢ In our view, many mutual funds are still limiting the size of their repo borrowing and achieving leverage through futures.
➢ Separate accounts are typically governed by the client investment management agreement (IMA)
and associated investment guidelines.
➢ Partial but representative data(1) suggests that most institutional accounts do not permit repo.
➢ Futures therefore are in many cases the only way to obtain leverage in a separate account.
13
(1) Sourced from presenter firm’s data and may not correlate with aggregate industry level
Asset Managers: 40-Act Mutual Fund Investment Expense
Accounting Also Favors Futures
➢ While 40-Act mutual funds can use repurchase agreements, the “interest expense” associated with
repos must be reported separately as an investment expense.
➢ In 2018, in accordance with FINRA Rule 2210(d)(5), some fund analysis providers started including
“interest expense” in their databases of total expense ratios which include management fees and
administrative fixed-costs.
➢ This change had the effect of making providers who prioritize Treasury repo over futures look more
expensive, even though the two products are similar economically and repo is often less expensive.
➢ FINRA issued clarification D.1.2. in 2019, and while efforts have been made to exclude interest
expense from published expense ratios(1), the reality is that if a fund has high interest expense, it
raises questions, even if interest expense generally associated with portfolio structures or leverage
that are expected to add value to the overall portfolio.
Distribution Fees
Legal Fees
Interest Expense
14
(1) One Expense Ratio to Rule Them All | Morningstar
Asset Managers: Largest Fixed-Income Categories
➢ The largest fixed-income mutual fund category is intermediate core / core plus at 39% of total
assets. Many funds are benchmarked to the Bloomberg U.S. Aggregate Index (“the Agg”),
which is a fairly good, but not complete, representation of the U.S bond market.
➢ Recently flows into short-duration products have also been substantial, where many funds are
benchmarked to 1-3Y Government/Credit and 1-3Y Corporate indices.
➢ We go into some detail on the composition of these indices in the next slide.
Table 3: Fixed-Income Mutual Fund Sector Breakdown (1) Figure 13: Cumulative Fixed-Income Strategy Flows
(2006-2023, $bn) (1)
Total Net Assets % Fixed-Income
Morningstar Sub-Category
($Bn) Category 800
Intermediate Core-Plus Bond 681 25% Core + Core Plus +
Intermediate Core Bond 381 14%
700
Multisector
Short-Term Bond 339 12% 600
Multisector Bond 267 10% All Other
High Yield Bond 244 9% 500
Ultrashort Bond 142 5%
Nontraditional Bond 109 4%
400
Inflation-Protected Bond 85 3% 300
Intermediate Government 82 3%
Corporate Bond 75 3% 200
Short
Bank Loan 75 3%
100 Duration
Global Bond-USD Hedged 56 2%
Global Bond 46 2% 0
Emerging Markets Bond 43 2%
Other 109 4% -100
Total 2,734 100% 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
15
(1) Sourced from Morningstar as of 12/29/2023
Asset Managers: Benchmark Indices
➢ Presently, the Agg is about 42% Treasuries, and in total, about 80% government risk (1).
It is an investment-grade only, market value weighted index. The 1-3Y Gov/Credit index
is 66% Treasuries and over 70% government risk.
➢ Notably, the main explanatory market risk factor for the Agg is the level of Treasury
yields, which generally account for over 75% of the modeled risk of the index.
Table 4: Bloomberg Aggregate Index (2) Table 5: Bloomberg 1-3Y Government Credit (2)
(1) The term “government risk” refers to Treasury debt, Agency debt, and Agency MBS and is highlighted in red. 16
(2) Sourced from Bloomberg as of 12/29/2023. Treasury weights exclude SOMA holdings.
Asset Managers: Structural Credit Overweights
➢ When managing against the Bloomberg Aggregate Index, active managers will often
allocate to higher yielding spread assets to achieve an income advantage and a more
diversified portfolio given the heavy interest rate sensitivity of the benchmark.
➢ Spread sector investment decisions are made in assets of varying durations and
maturities, and are often thought of separately from interest rate investment decisions.
➢ Although asset managers have different investment approaches, we believe it’s common
to separate decisions made on interest rate duration and credit spread duration.
➢ Futures allow asset managers to make credit allocation decisions relatively seamlessly,
without impacting interest rate risk exposures, but can introduce basis risk between the
futures allocation and the Treasury allocation in the index.
➢ While the process is more precise with Treasuries + repo, it is more cumbersome to
implement operationally.
➢ It’s likely that structural overweight positions in credit products could result in
persistently higher allocations to Treasury futures amongst asset managers.
17
Asset Managers: Structural Credit Overweights
When allocating out of Treasuries into credit sectors, we highlight two broad transaction types. We focus
on the second type and the implications for how asset managers hedge duration and curve.
➢ Matched substitution: The manager buys a credit and can sell a Treasury of similar duration.
➢ Results in a reduction of the Treasury allocation and an increase in credit exposure. Minimal need for futures or repo.
➢ Mismatched substitution: The manager buys a credit and must sell a Treasury of different duration.
➢ To control for duration and curve risk, the manager will need to use futures or repo and reverse repo(1).
➢ This happens for several reasons, to name a few:
(1) Manager buys a floating rate credit asset with zero duration.
(2) Manager runs out of similar duration Treasuries to sell from the portfolio.
(3) Manager buys a non-USD credit asset.
Figure 14: Bloomberg Agg Index Table 6: Ways to Hedge Duration and Curve Risk With
Treasury Allocation by Maturity Bucket(2) Mismatched Treasury/Credit Substitution
16.0%
14.4%
14.0% Type Transactions Involved Cost/Benefit
12.0% Perfect Hedge with (1) Buy credit - Most operationally complex
10.4% Repo and Borrow (2) Short perfect Treasury hedge - Generates interest expense
10.0% (3) Borrow perfect Treasury
(4) Repo another Treasury to fund credit
8.0% Partial Hedge with (1) Buy credit - Some operational complexity
Repo Only (2) Short Treasury futures hedge - Generates interest expense
6.0% 4.7% 4.9% (3) Repo another Treasury to fund credit - Introduces some curve risk
3.6% 3.6%
4.0%
Futures Hedge (1) Buy credit - Less operationally complex
(2) Sell Treasury - Introduces basis and curve risk
2.0% (3) Execute futures curve trade
0.0%
1-2y 3-5y 5-7y 7-10y 10-20y 20y+
(1) Note, we assume portfolio managers will typically use the cheapest form of repo with lowest haircut (Treasuries) 18
(2) Sourced from Bloomberg as of 12/29/2023.
Asset Managers: Structural Credit Overweights
➢ A separate complication, which is obscure but meaningful, is that portfolio managers will
sometimes treat credit duration differently than risk-free duration. This matters more for
lower quality credit products, where spread volatility dwarfs interest rate volatility. Bonds
that trade on dollar price, like lower rated high yield, exhibit lower “empirical duration.”(1)
➢ Therefore, an allocation out of Treasuries into lower quality credit, would typically require
a manager to buy back more duration in futures, compared to higher quality credit.
(1) Empirical duration can be thought of as a statistical duration as opposed to a cash flow duration
19
(2) Sourced from Bloomberg as well as the presenter’s firm’s calculations as of 01/10/2024
Asset Managers: Structural Credit Overweights
Four examples(1) where a mismatched substitution results in a net long position in futures:
Duration
Trade Security Notional % Cash % Wtd Avg Life Yield OAS
Contrib
Sell Treasury -5.0% -5.0% (0.31) 7.9 4.10%
Buy Treasury Futures 4.8% 0.31
Buy CLO 5.0% 5.0% 0.00 2.0 8.00% 2.32%
Net Treasury Futures 4.8%
Net Change to Fund(2) 0.0% 0.00 0.12% 0.12%
Duration
Trade Security Notional % Cash % Wtd Avg Life Yield OAS
Contrib
Sell Treasury -5.0% -5.0% (0.31) 7.9 4.1
Buy Treasury Futures 4.8% 0.31
Buy Short-Dated ABS 5.0% 5.0% 0.06 1.2 6.7 2.14%
Sell TU Futures Hedge -3.1% (0.06)
Net Treasury Futures 1.7%
Net Change to Fund(2) 0.0% 0.00 0.11% 0.11%
Four examples(1) where a mismatched substitution results in a net long position in futures:
Notional Duration
Trade Security Cash % Wtd Avg Life Yield OAS
% Contrib
Sell Treasury -5.0% -5.0% (0.31) 7.9 4.1
Buy Treasury Futures 4.8% 0.31
Buy EUR IG Credit 5.0% 5.0% 0.22 7.2 3.9 1.40%
Sell 5Y Bobl Futures Hedge -5.0% (0.22)
Sell EUR/USD Fwd -5.0%
Net Treasury Futures 4.8%
Net Change to Fund(2) 0.0% 0.00 0.07% 0.07%
➢ Stylized example shows what an active manager’s allocation could often look like (Figure 17)
compared to an index (Figure 16). In this example, the actively managed portfolio has a 20% out-of-
index credit allocation and a higher yield than the index but maintains the same duration and yield
curve exposure.
➢ To accomplish this, the active positions are funded by selling Treasuries, the net duration and curve
are hedged with Treasury futures (resulting in a net long futures exposure and positive net leverage).
➢ Alternatively, the manager could have repo’d Treasuries to fund the allocations which would have
generated interest expense.
Figure 16: Bloomberg Agg Index Figure 17: Hypothetical Active Manager Figure 18: Recent Growth in Floating
Securitized Assets(2)
0% 20% 40% 60%
-20% 0% 20% 40%
Treasuries 42% 10%
Treasuries 22%
Treasury Futures
Treasury Futures 14% Compared to the AGG ex.
Government Related 5%
Government Related 5%
9% Treasury Market Value
Corporates 25%
Corporates 25%
Agency MBS 29% Agency MBS 29%
8%
Floating Rate Credit Floating Rate Credit 5%
High Yield High Yield
7%
5%
Securitized Products Securitized Products 5%
Non-US Credit 6%
Non-US Credit 5% Compared to the
Non-US Rate Futures Non-US Rate Futures -4% AGG Market Value
5%
4%
Aggregate Index Active Manager
Yield Duration Avg. OAS Yield Duration Avg. OAS 3%
4.52 6.17 0.38 5.36 6.17 1.07 14 15 16 17 18 19 20 21 22 23
(1) Sourced from Bloomberg as of 12/29/2023 and the presenter’s firm’s calculations
22
(2) Sourced from Bank of America as of Dec 2023
Asset Managers: Structural Credit Overweights
Looking deeper at index and industry data, two observations are worth investigating further:
(1) There appears to be a positive relationship between the Treasury allocation in the Aggregate Index(1),
and the CFTC asset manager net futures position (as a % of the Treasury market), shown in Figure 19
➢ It’s not obvious why this relationship exists. One possible explanation is that as Treasury allocations have risen over time, asset
managers chose to increase their overweights to higher returning credit sectors relative to Treasuries.
(2) Using eVestment data(2) and looking at rough sector positioning for 14 large core and core plus active
management platforms ($900bn AUM), there does appear to be a correlation between the AUM
weighted active credit sector exposure and the CFTC asset manger net futures position
➢ This relationship lends some support to the hypothesis that Treasury futures are used to fund active credit sector positions
➢ It does not prove the point definitively, however. More investigation into the types of credit assets owned is required.
Figure 19: Treasury Index Weight vs. Figure 20: eVestment Active Non-Govt Allocation
Asset Manager Net Futures Position % of Tsy Mkt AUM Weighted Avg of 14 Large Asset Managers
45% 6.0% 30% Active Non- 800
Government
Treasury Weight in Exposure
28% CFTC Asset 700
Aggregate Index 4.0% Manager Net (%NAV, Left
40% (%,Left Axis) Axis)
26% Futures ($Bn 600
Face, Right
2.0%
24% Axis) 500
35%
0.0% 22% 400
30% CFTC Asset Manager 20% 300
Net Futures Position -2.0%
(%Tsy Mkt ex. SOMA, 18% 200
25% Right Axis) -4.0%
16% 100
(1) Sourced from Bloomberg, CFTC, and the presenter’s firm’s calculations as of 12/29/2023 23
(2) Sourced from eVestment as of Q3 2023
Illustration of Futures Replication Costs
➢ In Figure 21 we show the cumulative performance costs from replicating the Bloomberg
Treasury Index with the 6 available Treasury futures contracts.
➢ The methodology assumes a key-rate duration hedged portfolio and subtracts the cash
return using the overnight SOFR rate.
➢ Importantly, the return shown here is not the same as a pure basis trade (long CTD,
short Treasury future), it includes the performance of the whole Treasury market.
➢ From 2018-2023, futures replication had an average annual cost of 0.45%. Most fixed-
income spread products have hurdled that carry cost (Figure 23).
Figure 21: Cost of Replicating Figure 22: Trough to Peak Figure 23: Credit Spreads vs.
Bloomberg Treasury Index with Changes in Futures Replication Cost of Tsy Futures (2018-2023
0.5% Futures(1) 0.8% Performance(1) Avg)
0.0% 0.7%
0.6% HY Credit 4.03%
-0.5%
0.5%
-1.0% EM External Debt 3.68%
0.4%
-1.5% 0.3% IG Credit 1.21%
-2.0% Average Annual 0.2%
Return = -0.45%
-2.5% 0.1% Agency MBS 1.00%
0.0%
-3.0%
Oct-23
Apr-18
Oct-18
Apr-19
Oct-19
Apr-20
Oct-20
Apr-21
Oct-21
Apr-22
Oct-22
Apr-23
Futures Replication 0.45%
Apr-18
Apr-23
Apr-19
Apr-20
Apr-21
Apr-22
Oct-21
Oct-18
Oct-19
Oct-20
Oct-22
Oct-23
25
How Do Leveraged Funds Use Treasury Futures?
➢ Leveraged funds typically use futures for some of the same reasons asset managers do:
➢ One important difference, however, is that leveraged funds employ relative-value strategies with
greater size relative to their assets under management (more leverage).
➢ These relative-value strategies seem likely to reside mostly within the multimanager community, and
potentially within some single strategy relative-value or quantitative hedge funds.
➢ As shown in Figure 24, because of strong performance, multimanager strategies have grown faster
than the rest of the industry. With this may have come a growth in dedicated relative value AUM.
26
(1) Sourced from Goldman Sachs Research as of Dec 2022
How Do Leveraged Funds Use Treasury Futures?
➢ While asset managers have shown a net long bias in futures, the opposite is true of leveraged funds,
which have showed a persistent net short bias, despite large swings in markets.
➢ Similar to the discussion about asset managers, we think it’s unlikely that directional positions are
generating a structural bias to futures positioning recently.
➢ But large hedge funds and CTA’s are likely to have influenced shorter term positions at various times.
➢ As has been discussed extensively by regulators and market participants, it seems most likely to us
that the persistent futures short by leveraged funds reflects Treasury basis relative-value positions.
➢ As we showed on the prior slides, there is a persistent risk premium to be harvested in Treasury
futures basis. It appears that asset managers are willing to pay this premium to invest in higher
yielding products, while hedge funds earn the premium.
➢ We can estimate the returns to the basis trade using the data on the cost of futures replication:
➢ We assume a leverage ratio of 20x and show cumulative performance (excess of cash return).
➢ Anecdotally, 20x appears to be a good approximation of leverage typically used in these trades. Repo and futures margins
do permit higher leverage, however. 20x leverage could generate 9-10% annualized excess returns using this framework.
➢ March 2020 would have more than wiped out equity capital in such a strategy, explaining the stress during this period. To
the extent these strategies live inside of multimanagers, other capital may have been injected during this period.
Figure 26: Cumulative Excess Returns Figure 27: Peak to Trough Drawdown
Treasury Index vs. Futures Using 20x Leverage Multiple
60.0% 0.0%
-2.0%
50.0%
20X
-4.0%
Leverage
40.0%
-6.0% "Equity"
Capital
30.0% -8.0%
-10.0%
20.0%
-12.0%
Peak to
10.0% Trough
-14.0%
Drawdown
0.0% -16.0%
2018 2019 2020 2021 2022 2023 2018 2019 2020 2021 2022 2023
➢ The performance we show is an approximation and not that of a pure basis trade (future vs. CTD).
➢ That said, it’s likely that hedge funds who implement the basis trade do so in ways that involve using
Treasuries other than CTDs, with numerous other real-world complexities to consider, including:
➢ Type of repo funding (term v. daily)
➢ CTD switch option (recall volatility in USZ3 in Fall 2023)
➢ Spot vs. deferred CTD analysis
➢ Futures roll timing, optimal delivery and wildcard option
➢ Barriers to entry given balance sheet scarcity/repo capacity
➢ Usage of cheaper off-the runs
➢ In addition to expensive implied repo, looking below, we see that futures CTDs can often trade with
negative spreads to Treasury fitted curve.
➢ Similar to “on-the-run” Treasuries, it’s likely that CTDs trade this way due to a liquidity premium vs. off-the-runs.
➢ Recalling the period of March 2020, dealers reported(1) large liquidations in off-the-run sectors from the relative-value
community, suggesting that basis implementation involves more varied expressions.
(2) 10Y
(4) 20Y
(6) 30Y
(8) OTR
(10)
CTD
(12)
0 5 10 15 20 25 30
Maturity (Yrs) 29
(1) See page 56 of this report https://home.treasury.gov/system/files/221/CombinedChargesforArchivesQ22021.pdf
Leveraged Funds: Estimating Size of the Basis Trade
➢ Recent research from the Fed (1) suggests that the basis trade has grown in size lately, similar to
levels seen in 2018-2019. They cite the futures data, but also show some evidence from the
sponsored repo market and collateral posting data, shown below:
➢ Another piece of research(2) released in September of 2023 quantifies the basis trade at $550bn at the end of 2022.
➢ The authors used data from form PF, and the size seems fairly consistent with size of futures exposure at the time.
➢ Bottom line, estimating size of relative-value AUM is hard especially as a large portion of capital pursuing these strategies
may be embedded in multimanager hedge-funds, where little public data on sub-strategy allocations is available.
30
(1) https://www.federalreserve.gov/econres/notes/feds-notes/recent-developments-in-hedge-funds-treasury-futures-and-repo-positions-20230830.html
(2) https://www.federalreserve.gov/econres/notes/feds-notes/hedge-fund-treasury-exposures-repo-and-margining-20230908.html
Leveraged Funds: Monitoring the Basis Trade
Item Notes
Treasury Futures Open Interest Relative to the Size of Treasury Market • Disproportionate growth could indicate excessive
systemic leverage.
Size of Cleared Repo Market • Will be more valuable when all Treasury repo is
cleared in 2026.
Treasury-Futures Basis Valuations • Viewed in the context of changes to open interest,
valuations could be an important indicator.
• For example, if cash richens vs. futures and open
interest is growing, suggests position crowding.
Dealer Inventories of Treasury Securities • Supplement with real time flow color in off-the-run
trading activity.
Term structure of Repo Funding Levels • Monitor overnight vs. term repo spreads as an
indicator of funding pressures.
Changes in Margin Requirements • Monitor CME exchange margins and repo.
Concentration of Funding Sources in Bilateral Repo Market • Overreliance on single institutions could create
counterparty risk.
31
Potential Interactions Between Asset
Managers & Levered Fund Positions
32
Potential Interactions to Consider
➢ Amidst record Treasury issuance, and procyclical fiscal stimulus, we conclude that a symbiotic
relationship has emerged between asset managers and hedge funds.
• Large procyclical deficits have driven Treasury weights higher in bond indices at a time when the economy is
relatively healthy and active bond managers prefer credit relative to Treasuries.
• This cyclical preference for credit, alongside the structural asset allocation of active bond managers, may be
resulting in lower cash Treasury demand and increased use of Treasury futures.
• This preference for futures vs. cash improves valuations and the investment opportunity for RV hedge funds.
➢ While the system seems stable to us now, it’s worth thinking through potential risks that might arise,
as asset managers have moved down in liquidity while hedge funds have increased leverage.
• A reduction in credit risk by asset managers would naturally facilitate unwinding the basis trade by hedge funds.
• However, if asset managers experience rapid outflows while market liquidity is impaired, these outflows could
initially be funded through Treasury sales. This type of behavior could pressure RV basis positions, and it’s possible
this dynamic was at play during March 2020 when everyone in the market scrambled for liquidity. We see below a
high tail risk correlation between Treasury-futures basis performance and 1-3y Corporate spreads. This period
likely also simply reflected extreme liquidity preference which would pressure secured vs. unsecured funding
spreads.
Figure 29: Trailing 1Y US Deficit (%GDP) vs. Figure 30: Short Maturity Credit Spreads
1Y Average Unemployment Rate vs. Proxied Basis Trade Drawdowns
5
U3 (Inverted, Left Axis) -3 4.25 12.0%
1-3Y Corporate
3.75 OAS (Left Axis)
0 -5 10.0%
3.25 Tsy Index Excess
-7 Return 15X Leverage 8.0%
-5 2.75
Drawdown from
-9 2.25 Peak 6.0%
-10
1.75
-11 4.0%
1.25
-15 Deficit (%GDP, Right
-13 2.0%
Axis) 0.75
-20 -15 0.25 0.0%
90 92 94 96 98 00 02 04 06 08 10 12 14 16 18 20 22 24 2018 2019 2020 2021 2022 2023
33
(1) Sourced from Bloomberg and the presenter’s firm’s calculations as of Dec 2023
Conclusions and Areas of Further Research
34