美國國債和英國國債期貨策略

1、May 14, 2020 02:44 PM GMTGlobal Futures Rolls Report | GlobalUS Treasury and Gilt Futures Rolls HYPERLINK mailto:Lorenzo.Testa Lorenzo.Testa+44 20 7677-0337contract rolls for the June to September cycle. We look at theAlina Zaytsevafundamental drivers of the roll, investor positioning, deliverySTRAT

2、EGIST HYPERLINK mailto:Alina.Zaytseva Alina.Zaytseva+44 20 7677-1120optionality, and relative value. Finally, we suggest optimalrolling strategies.TU Contract: Mildly Bearish: We expect the upcoming TU roll to be influenced by positioning and recent roll cycles. Aside from Feb 20 roll cycle, recent

3、trends have been characterized by cheapening into the First Notice Day.FV Contract: Mildly Bearish: We expect a mildly bearish roll as the long positioning of asset managers and recent roll cycles will likely influence the roll.TY Contract: Neutral: Futures positioning and current richness of the ro

4、ll will likely influence the upcoming TY roll cycle. Their effects could offset one another, so we remain neutral on the roll.UXY Contract: Neutral: We believe that the relative value of contract CTDs and positioning will likely influence the UXY roll. However, the expected impacts of these will lik

5、ely offset one another, so we remain neutral.US Contract: Bearish: A noticeable increase in long positioning of asset managers in the contract, combined with an already rich roll (based on model implied fair values) lend a bearish bias on the US roll.WN Contract: Mildly Bullish: Relative value lends

6、 a bullish bias to the roll, although positioning partially offsets this bullish bias. We remain mildly bullish overall.Gilt Contract: Neutral: While positioning provides a weak bearish signal, we see the roll as close to fair value and are neutral overall.Summary of our roll viewsDue to the nature

7、of the fixed income market, the issuers or bonds of the issuers recommended or discussed in this report may not be continuously followed. Accordingly, investors must regard this report as providing stand-alone analysis and should not expect continuing analysis or additional reports relating to such

8、issuers or bonds of the issuers.Morgan Stanley does and seeks to do business with companies covered in Morgan Stanley Research. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of Morgan Stanley Research. Investors should consider

9、 Morgan Stanley Research as only a single factor in making their investment decision.For analyst certification and other important disclosures, refer to the Disclosure Section, located at the end of this report.+= Analysts employed by non-U.S. affiliates are not registered with FINRA, may not be ass

10、ociated persons of the member and may not be subject to FINRA restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account.We present our views on US Treasury and Gilt futuresMORGAN STANLEY & CO. LLCKelcie GersonSTRATEGIST HYPERL

11、INK mailto:Kelcie.Gerson Kelcie.Gerson+1 212 761-3983Guneet Dhingra, CFASTRATEGIST HYPERLINK mailto:Guneet.Dhingra Guneet.Dhingra+1 212 761-1445MORGAN STANLEY & CO. INTERNATIONAL PLC+Lorenzo TestaSTRATEGISTContractPositioningRecent trendsOptionality/fair valueRelative valueCurveOverallTUMild bearish

12、Mild bearishMild bullishMild bullishNeutralMild bearishFVMild bearishMild bearishMild bullishMild bullishNeutralMild bearishTYMild bearishNeutralNeutralNeutralNeutralNeutralUXYMild bearishNeutralMild bullishMild bullishNeutralNeutralUSBearishNeutral Mild bearishNeutralNeutralBearishWNMild bearishNeu

13、tralBullishNeutralNeutralMild bullishGiltMild bearishNeutralNeutralNeutralNeutralNeutralTU Contract: Mildly BearishWe expect the upcoming TU roll to be mildly bearish, influenced primarily by the current positioning and recent roll cycles.Asset managers have maintained their long position in the con

14、tract, now representing 18.4% of open interest, while other reportables have also maintained slight longs. Overall, positioning seems mildly bearish for the roll. In most roll cycles, the roll cheapened significantly into the First Notice Day. The exception is the February 20 roll where the roll ric

15、hened substantially into the First Notice Day. However, we think the roll price action in February is not representative of typical roll patterns. As such, we believe the previous three roll cycles indicate that cheapening will likely occur into the First Notice Day, so we are mildly bearish on the

16、roll.Exhibit 1: Key metrics for the TU March - June rollFrontBackContractTUM0TUU0CTD2.375 Mar22s1.75 Jun22sMarket Value110-073110-120Model Implied Fair Value110-065110-101Implied repo (bp)17.9924.34Rich/Cheap0-0060-017Dv012.002.29Model Option Value0-000-0-000Repo to Delivery5.011.0RollDV01 Hedge*875

17、Vol Wtd Hedge*884Roll-0-045Curve Hedge1% 1s/2sDv01 Roll13-21+10 bp of Curve =0.1 TicksModel Implied Roll Fair Value-0-03+Optionality0-000AM Market Size33%AM Net Posn/ OI18%LF Market Size31%LF Net Posn/ OI-20%* Hedges represent number of back contracts per 1000 front contractsSource: Morgan Stanley R

18、esearch, CFTC, BloombergPositioning mildly bearish: Exhibit 2 shows the net positioning of various investor categories in the TU contract. The relatively price-insensitive other reportables has maintained a long position of approximately 4.9%. Asset managers have kept their long positions too at a l

19、evel close to 18.4% this net long positioning is flat relative to the last cycle and lower compared to the previous three cycles. The price-sensitive leveraged funds category has decreased its short position since the last roll cycle, to - 20.0%, maintaining an upward trajectory. Positioning seems m

20、ildly bearish for the roll. This roll cycle is progressing slightly below the standard rate (as shown in Exhibit 3).Exhibit 2: TU contract positioning from various investor categories around roll cyclesNet Posn/ Open Interest 50%40%30%20%10%0%-10%-20%-30%Exhibit 3: Roll progress compared to previous

21、 cyclesRoll Progress (%)Current Roll Progress: 3.7%100%90%80%70%60%50%40%30%20%10%0%-20-15-10-50510-40%-50%Jun-19Aug-19Nov-19Feb-20May-20Business Days to First Notice Day of Front ContractProgress Range of Four Prior Rolls Current Roll ProgressSource: Morgan Stanley ResearchRoll CycleLF Net Posn (%

22、of OI) AM Net Posn (% of OI)Other Net Posn (% of OI)Source: Morgan Stanley Research, CFTCOptionality and net basis mildly bullish: The model-implied optionality is zero for both contracts, as there is a very low probability of a switch of the CTD for either contract.When considering the relative ric

23、hness and cheapness of the futures contracts adjusted for optionality, the roll is a bit cheap given that the front contract is trading roughly in line with fair value and the back contract is rich by 2 3/8 ticks. Given this, the optionality and net basis of this roll seem mildly bullish for the rol

24、l. However, we must note that the back contract is less likely to normalize as quickly as a front contract would, given it has more time to maturity.Exhibit 4: Moves in roll from comparable positioning cyclesDV01 Adjusted Roll Change (32nds) 8.06.04.02.00.0-2.0-4.0-6.0-8.0-10.0-12.0Exhibit 5: Rich/c

25、heap spread between front and back CTDFront CTD RichBack CTD Richbp 12.010.08.06.04.02.00.0-2.0-4.0-17-12-7-238B-Days to First Notice Day of Front ContractMay-19Aug-19Nov-19Feb-20May-20Source: Morgan Stanley Research-6.0Nov 19 Dec 19 Jan 20 Feb 20 Mar 20 Apr 20 R/C Spread (Back CTD - Front CTD)Sourc

26、e: Morgan Stanley ResearchRecent roll cycles mildly bearish: Exhibit 4 shows the TU roll (DV01-adjusted) over the past four cycles. In most roll cycles, the roll cheapened significantly into the First Notice Day. The exception is the February 20 roll where the roll richened substantially into the Fi

27、rst Notice Day. The TU roll in February, as well as the rolls of other contracts, were significantly influenced by the cheapening (and later richening) of the back CTD relative to the front CTD (Exhibit 5). The back CTD cheapened in line with the underperformance of cash relative to futures as well

28、as the dislocation of term funding relative to overnight. Because of this, we think the roll price action in February is not representative of typical roll patterns. As such, we believe the previous three roll cycles indicate that cheapening will likely occur into the First Notice Day, so we are mil

29、dly bearish on the roll based on recent roll cycles.Relative value mildly bullish: The back CTD has richened substantially relative to the front CTD over the past two months after the front CTD richened aggressively, as shown in Exhibit 5. The substantial magnitude of the richening, however, could c

30、ause the roll to trade bullish as the richness of the back CTD fades. However, there may be structural reasons the back CTD is rich and it may take more time for the back CTD to normalize, so we think the relative value aspect of the roll is only mildly bullish.Macro curve view neutral: This cycle,

31、the correlation between the 1s2s curve and the spread of future-implied forward yields (back contract - front contract) over the past month has been minimal (R2=0.11) (Exhibit 6). Exhibit 7 highlights how curve moves have influenced the roll in previous cycles. Given that recent curve moves have not

32、 been met with similar changes in the spread between the CTDs, we do not think moves in the yield curve will influence this roll cycle strongly.Exhibit 6: Regression of the CTD spread on 1s2s curve spreadExhibit 7: 1s2s yield curve move since beginning of the cycleBack CTD - Front CTD (bp) 0.0-0.5-1

33、.0-1.5-2.0-2.5-3.0-3.5y = 0.06x - 2.82 R = 0.11Curve moves (bp) 2520151050-5-10-15-20-10010051015201s/2s CurveSource: Morgan Stanley ResearchB Days to First Notice Day of Front ContractAug-19Nov-19Feb-20May-20Source: Morgan Stanley ResearchFV Contract: Mildly BearishWe expect the upcoming FV roll to

34、 be influenced by positioning and recent roll cycles.Asset managers and other reportables have maintained their long positions over the past four cycles. The previous four roll cycles were characterized by modest cheapening throughout the cycle followed by richening into the First Notice Day. So we

35、anticipate that, from here until the First Notice Day, this roll cycle will experience slight bearish bias.Exhibit 8: Key metrics for the FV March - June rollFrontBackContractFVM0FVU0CTD1.25 Aug24s1.5 Nov24sMarket Value125-182125-136Model Implied Fair Value125-176125-116Implied repo (bp)13.5523.39Ri

36、ch/Cheap0-00+0-020Dv015.285.56Model Option Value0-0000-000Repo to Delivery5.011.0RollDV01 Hedge*949Vol Wtd Hedge*875Roll0-04+Curve Hedge-67% 3s/5sDv01 Roll6-18510 bp of Curve =-11.3 TicksModel Implied Roll Fair Value0-060Optionality0-000AM Market Size32%AM Net Posn/ OI31%LF Market Size27%LF Net Posn

37、/ OI-26%* Hedges represent number of back contracts per 1000 front contractsSource: Morgan Stanley Research, CFTC, BloombergPositioning mildly bearish: Asset managers open long position versus the total open interest in the contract has remained flat since the beginning of the last roll cycle at 31%

38、 (Exhibit 9). The net long position of relatively price-insensitive other reportables increased slightly, to 1.8%. Leveraged funds short positions are similar to the last roll cycle, at around 26% short. Overall, we remain mildly bearish on positioning.Exhibit 9: FV contract positioning from various

39、 investor categories around roll cyclesNet Posn/ Open Interest 40%30%20%10%0%-10%-20%-30%-40%Exhibit 10: DV01-adjusted roll for current cycle against roll for the past four cyclesDV01 Adjusted Roll Change (32nds) 10.05.00.0-5.0-10.0-15.0-20.0Jun-19Aug-19Nov-19Feb-20May-20-17-12-7-238B-Days to First

40、Notice Day of Front ContractRoll CycleLF Net Posn (% of OI) May-19Aug-19Nov-19AM Net Posn (% of OI)Other Net Posn (% of OI) Feb-20May-20Source: Morgan Stanley Research, CFTCSource: Morgan Stanley ResearchRecent roll cycles mildly bearish: Exhibit 10 shows the FV roll (DV01-adjusted) over the past fo

41、ur cycles. The previous four roll cycles were characterized by modest cheapening throughout the cycle followed by richening into the First Notice Day. So we anticipate that, from here until the First Notice Day, this roll cycle will experience slight bearish bias. As in TU, the February roll was unc

42、haracteristic of usual roll developments due to an extreme richening (and then cheapening) of the front CTD relative to the back CTD (Exhibit 11) and dislocation of term funding leading to net basis spread moves.Relative value mildly bullish: Exhibit 12 shows that both the back and front FV contract

43、s CTDs stand cheap relative to the curve spline. We note that the contracts CTDs were slightly richer one month ago, so they may richen a bit and remain structurally rich through the cycle. Both CTDs are similar to where they were a month ago in relation to the spline, which suggests there are struc

44、tural reasons for the richness/cheapness.Exhibit 11: FV contract CTD rich/cheap spread over the last six monthsFront CTD RichBack CTD RichbpExhibit 12: CTD rich/cheap on the splineRich vs. SplineRich/ cheap (bp) 0.50.0-0.5-1.0-1.5-2.0Jul-24Sep-24Oct-24Nov-24Dec-24-2.5Cheap vs. SplineJan-25Feb-25All

45、Bonds1m AgoCTD FrontCTD BackSource: Morgan Stanley ResearchSource: Morgan Stanley ResearchOptionality and net basis mildly bullish: There is virtually no possibility of a shift in the CTD for the front or the back FV contract, rendering the model-implied optionality zero for both contracts. Based on

46、 our model-implied fair value, we see that the front contract and back contract are trading at fair value and 2 ticks rich, respectively (Exhibit 8). Given that the front contract is cheaper than the back contract, the roll is currently cheap, which could lead to richening as we approach the First N

47、otice Day.Macro curve view neutral: In the last month, the correlation between the 3s5s curve and the spread of future-implied forward yields (back contract - front contract) has been null (Exhibit 13). Given no change in CTDs, curve exposure is only due to moves in repo curves between the front and

48、 back tenors. We dont expect a large change in the repo curve during this roll cycle.Exhibit 13: Regression of the CTD spread on 3s5s curve spreadExhibit 14: 3s5s yield curve moves since the beginning of the cycley= -0.14x + 2.85R = 0.05Back CTD - Front CTD (bp) 332211Curve Moves (bps) 1086420-2-4-6

49、Current Curve: 12.7 bp0-20-15-10-50510-1B-Days to First Notice Day of Front Contract05101520 May-19Aug-19Nov-19Feb-20May-203s5s Curve (bp)Source: Morgan Stanley ResearchSource: Morgan Stanley ResearchTY Contract: NeutralFutures positioning and current richness of the roll will likely influence the u

50、pcoming TY roll cycle. Overall, we remain neutral on the roll.Though asset managers drove the move lower over the last few roll cycles, their net long positioning has decreased since the November 2019 cycle. The CTD on the back contract is the 2.375 May27s, which is 2.2 ticks rich on the spline. The

51、 CTD on the front contract is the 2.25 Feb27s, which is at fair value based on our model.That both CTDs are similar to where they were a month ago in relation to the spline suggests there are structural reasons for the richness/cheapness; we dont believe they will impact the rolls significantly.Last

52、ly, according to our modeling, there is virtually no probability that the new 7y bond to be auctioned later this month would become the back contracts CTD unless the 7y auction clears with a coupon of 2.125% or greater.Exhibit 15: Key metrics for the TY March - June rollFrontBackContractTYM0TYU0CTD2

53、.25 Feb27s2.375 May27sMarket Value138-30+138-22+Model Implied Fair Value138-30+138-200Implied repo (bp)5.0625.30Rich/Cheap0-0000-02+Dv018.819.10Model Option Value0-000-0-000Repo to Delivery5.011.0RollDV01 Hedge*967Vol Wtd Hedge*1020Roll0-080Curve Hedge-31% 5s/10sDv01 Roll4-24610 bp of Curve =-8.8 Ti

54、cksModel Implied Roll Fair Value0-10+Optionality0-000AM Market Size34%AM Net Posn/ OI15%LF Market Size23%LF Net Posn/ OI-18%* Hedges represent number of back contracts per 1000 front contractsSource: Morgan Stanley Research, CFTC, BloombergPositioning mildly bearish: Looking at the net positioning o

55、f various investor categories in the TY contract, asset managers currently hold a decent 14.6% long net position versus the total open interest in the contract (Exhibit 16), a decrease from the previous cycle. The relatively price-insensitive other reportables category remains slightly net short at

56、-0.5%. This net short positioning offsets the bearish bias we would otherwise have from looking at the net long positioning of asset managers. Levered funds, on the other hand, have decreased their net short positioning to -18.0%.In our previous roll cycle analyses, we compared the current positioni

57、ng in the contract to similar positioning going into previous roll cycles. The closest match to the current positioning was around the last cycle (Feb 2020), where the roll cheapened slightly atthe onset of the roll cycle, but richened substantially on the last few days into the First Notice Day (Ex

58、hibit 18). Consistent with TU and FV, the richening was due to extreme cheapening of the back CTD relative to the front CTD due to increased term funding costs and illiquidity in the UST market. Because of these unusual circumstances, we dont think positioning led to those moves in February, so we r

59、emain mildly bearish on this roll cycle from a positioning standpoint.Exhibit 16: TY contract positioning from various investor categories around roll cyclesNet Posn/ Open Interest 40%30%20%10%0%-10%-20%-30%-40%Exhibit 17: DV01-adjusted roll for current cycle against roll for the past four cyclesDV0

60、1 Adjusted Roll Change (32nds) 12.010.08.06.04.02.00.0-2.0-4.0-17-12-7-238Jun-19Aug-19Nov-19Feb-20May-20B-Days to First Notice Day of Front ContractRoll CycleLF Net Posn (% of OI) AM Net Posn (% of OI)Other Net Posn (% of OI)Source: Morgan Stanley Research, CFTCMay-19Aug-19Nov-19Feb-20May-20Source: