Cause behind IRS curve reversal and structured bonds impact

Transcription

1 KOREA Chief Strategist (+822) Hyojin Kim Economist (+822) Hongcheol Moon Analyst (+822) Kyoungrok Lee Credit Analyst (+822) Yuna Park Analyst (+822) Sangwon Noh Analyst (+822) Ilhyuk Kim Analyst (+822) Kwangsook Hwang CFA, Senior Credit Analyst (+822) Cause behind IRS curve reversal and structured bonds impact Rate hike concerns and slowdown possibility: 3s5s, 2s5s, and 2s1s IRS spreads have begun to reverse. In turn, this is widening long term reversed swap spread (IRS KTB). Cause behind the reversal of IRS curve lies on the environment where the rate hike cycle is nearing its end while global economic slowdown fear is increasingly spreading. Furthermore, in tenors where extreme flattening occurs, reversing pace of the IRS curve tends to accelerate through structured bonds. In fact, the same phenomena were taking place in the ending phase of rate hike cycle between 7~ 8. Impact of CMS Spread Range Accrual: Issuers usually begin unwinding their previously hedged positions once forward IRS spread reaches around 1 bp, making them open to short gamma risk. This unwinding move widens reversed spread and again leads to increasing the unwinding demand (and quickening the widening pace), hence creating the vicious cycle. This hedging process ends up growing market volatility by (1) quickening the reversing pace of IRS curve and (2) widening the reversed swap spread in long term. Effect of the hedge to concentrate in 4 5y and 9 11y: Assuming the entire outstanding CMS (KRW 1.9 trillion) would be unwound, the Receive IRS effect would be the strongest in 4 5y and 9 11y tenors. Given the IRS curve was frequently reversed throughout all tenors near the end of rate hike cycles in the past, we still have to look out for Receive 1y IRS rush and widening possibility of reversed swap spread in 1y. IRS and KTB curves to continue flattening: It is quite natural to see the yield curve reverse near the end of rate hike cycle. 3s1s KTB spread will likely tighten to 3 bp or lower during Q3 and may even reverse in the first quarter next year. IRS curve too should invert frequently throughout all sectors at least for 6 months. Fixed In Fixed In Important disclosures are located on the back of this document. rund2you@dongbuhappy.com

2 Behind the IRS curve reversal Some tenors on IRS curve reverse in 32 months Term spread is rapidly narrowing. 3y1y KTB spread has tightened from 113 bp at the end of last year to 33 bp recently. The IRS curve was reversed in 3s5s, 2s5s, and 2s1s spreads, the first time in 32 months since its last reversal in December 8. Simultaneously, the reversed swap spread in long term is also widening due to fall in long term IRS rates. Widening in the reversed swap spread expands mark to market loss of bond swap position (Buy bond + Pay IRS). When issuers unwind these positions resulting in Sell bond + Receive IRS, rising bond yields, reversing IRS curve, and widening reversed swap spread will likely accelerate their moves. Risk is present that the entire bond market may fall into the vicious cycle. Worries over interest rate hike, slowdown possibility, and structured bonds There are a number of factors that caused the IRS curve reversal, both fundamental and technical. In fundamental perspective, there still remains some anxiety over a rate hike potential due to persisting inflation concerns while fears mount over a global recession. Current expectations are that, even if the BoK raises its policy rate, 1 2 will be the maximum number of actions it can take before the end of 11. The yield curve should flatten in such environments. On technical side of the spectrum, traders took loss cut of Pay IRS positions (previously betting on bond yields to rise). Also, demand arises for issuers to hedge their structured bonds by receiving IRS in medium and long term tenors. Most of these phenomena have already been witnessed in ending phases of rate hike cycles between 7 8. Speaking of structured bond issuances as one of the causes for IRS curve reversal, the more the bond (KTB) yield curve flattens, the stronger the demand for high yield structured bonds becomes. The reason is because coupons on these structured bonds are enhanced through sale of options. Structured bond issuers, in general, hedge their positions through receiving IRS. When maturities of these issues extend, demand for Receive IRS too rises. This also applies to Range Accrual structured bond, issuance of which has recently been increasing. Rise in CD rate following the rate hike possibility is another factor that would flatten the IRS curve. 2

3 Fig 1. 3s5s IRS spread reverses the first time in 32 mon. (bp) KTB IRS Source: Bond Pricing Agencies Average, ICAP and Dongbu Securities Fig 2. Reversed swap spread (IRS KTB) widens (bp) 1 2Y 3Y 5Y 1Y Source: Bond Pricing Agencies Average, ICAP and Dongbu Securities Impact of structured bond: CMS Spread Range Accrual Issuance and hedging structure When Constant Maturity Swap Spread Range Accrual (CMS) is issued, it comes with a bit of complex coupon payment system. Its coupon is determined through a predetermined formula: fixed rate*n/n, where n = number of days in a year when 5y CMS 3y CMS while N = total number of days in the same year (typically 365). Coupon payments are determined digitally, either zero or a fixed rate. The structured bond accrues interest on days when its spread is normal while no interest is paid on days the spread is reversed. Thus, the issuer is exposed to risk of loss when the IRS curve steepens. To hedge against the risk, CMS issuers Receive short term IRS + Pay long term IRS, which is considered an initial hedging stage. So, strong demand arises for the issuers (of 3s5s CMS maturing in 5 yrs) to hedge their risk by paying 1y IRS. Total amount of CMS that have been issued since 5 currently stands at KRW 2.6 trillion. KRW 1.9 trillion of that amount had all been issued in 6~ 7 period during which 3s5s IRS spread was close to zero. This was because many investors weren t expecting the IRS curve to reverse thinking it was just a means for exchanging CDs. There has hardly been any CMS issuance since 28, and yet approximately KRW 1.9 trillion won worth CMS bonds still remain to mature. On top of this, if we add structured deposits sold in the CMS type, the total amount would probably be more than double that amount. At the time of issue, there were mainly 3s5s CMS (1y mat) 3

4 in 6 7. Fig 3. Issuance Volume of CMS Spread Accrual (KRW 1mln) Callable CMS Spread Accrual CMS Spread Accrual 3,5 3, 2,5 2, 1,5 1, 5 Fig 4. Outstanding CMS (KRW 1mln) 8, 7, 6, 5, 4, 3, 2, 1, Y-1Y 5Y-1Y 5Y-2Y 5Y-3Y 3Y-1Y Source: Korea Asset Pricing (KAP) and Dongbu Securities Source: KAP and Dongbu Securities Short gamma: the reason for rapid unwinding Once CMS is issued, dynamic hedging comes into play depending on moves of IRS (CMS) spread. Given the coupon structure with either zero or a fixed rate digitally determined, the dynamic hedging on tenors where forward spread is normal (positive) wouldn t have too much impact on markets thanks to long gamma. Nonetheless, from the historical perspective, when forward spread reverses to about 1 bp or more, delta begins reducing sharply leading to a short gamma exposure. In other words, over hedged positions get unwound following reduction in delta, widening reversed spread. This again leads to unwinding demand and widening the reversed spread even more rapidly, hence becoming a vicious cycle. So far, majority of CMS structured bonds have been issued with the same exercise price structure (zero spread) but with different maturities and coupons. Thus, in essence, this type of hedging process of CMS structured bonds ends up increasing market volatility in terms of accelerating reversing pace of IRS curve and widening the long term swap spread. Impact of IRS Receive strong in 4-5y and 9-11y tenors From our past experiences, we know issuers usually begin unwinding their overhedged positions once forward IRS spread reaches around 1 bp. Let s think of 1s3s CMS maturing in 1yrs. When 2y (or below) forward 1s3s IRS spread reverses, effect of the hedge is strongest in 5y IRS tenor due to its unwinding demand (which comes 4

5 in the form of receiving that tenor); refer to Fig6. So, the unwinding effect comes out as Receive 5y IRS which is equivalent to addition of long end within reversed sector (2y forward) and long end of CMS tenor (3y). Fig5. shows sensitivity of Receive IRS that arises when assuming the entire outstanding CMS (KRW 1.9 trillion) would be unwound. We used a simple calculation method to examine its sensitivity throughout all tenors by adding termto maturity of outstanding CMS bonds and long term tenors. Currently, the unwinding effect is the strongest in 4 5y and 9 11y tenors. Recently, IRS curve reversal was particularly obvious in 5y (as the spread was reversed in 1s5s, 2s5s, and 3s5s). We believe this was partly because Receive IRS was concentrated mostly in 4 5y tenors, sensitivity of which tends to be high. Fig 5. Sensitivity of Receive IRS highest in 4-5y and 9-11y (KRW1mln) CMS' sensitivity on 'Receive IRS' 1, 2, 3, 4, 5, 6, 7, 8, (yrs.) Source: KAP and Dongbu Securities Fig 6. 1s3s IRS forward tenors currently under short- gamma influence (bp) Source: Bloomberg and Dongbu Securities Forward IRS 1s3s spread curve Short-term forward tenors now in the circle of shortgamma influence 3m 6m 9m 12m 15m 18m 21m 24m 3yr 4yr 5yr 6yr 7yr 8yr 9yr 1yr 12yr 15yr 2yr (Foward) Fig.7 exhibits outstanding amount of CMS structured bonds by term to mat. For example, 1y3y CMS with 21 months to maturity have KRW 21 billion in outstanding amount. Fig. 8 shows Forward IRS spread by maturity. Forward spread of 1y3y IRS in 21 months is presently at 8 bp, getting close to the 1bp threshold level. We believe it s now in the circle of short gamma influence. When we add term to mat (21 m) and long end CMS tenor (3y), the maturity comes to a little less than 5y. In easier terms, if unwinding were to take place due to 1y3y CMS (KRW 21 billion) getting into the circle of short gamma influence, it would be concentrated mostly on Receive 5y IRS. 5

6 Fig 7. CMS Issuance amount by term-to-mat (KRW 1 mln) Fig 8. IRS Forward Spread by maturity (Aug 3) (KRW 1 mln) 3Y-1Y 5Y-1Y 5Y-2Y 5Y-3Y 1Y-5Y Term-tomaturity Term-tomaturity 3Y-1Y 5Y-1Y 5Y-2Y 5Y-3Y 1Y-5Y 3m 5 9 3m m 5 6m m 2 9m m 4 12m m 15m m 2 18m m 2,1 2 21m m 24m yr 1 3yr yr 1 9 2, 5 4yr yr ,839 5yr yr 1, 1,8 5 6yr yr 1,5 7yr yr 8yr yr 9yr yr 2 1yr yr 12yr yr 15yr yr 2yr 계 3,8 3,2 2,543 7,239 2,2 Source: KAP and Dongbu Securities Source: Bloomberg and Dongbu Securities Assessing possibility for the reversed swap spread to widen It is interesting to see majority of outstanding amount (CMS) lie in 4 6y tenors. The CMS in theses tenors have 4 6y terms to maturity and long end CMS tenor is 5y (as in 1s5s, 2s5s, and 3s5s). Thus, if the CMS hedges unwind, Receive IRS will likely be concentrated in 9 11y and this is the reason that its sensitivity becomes high in these sectors. Fortunately, the forward spread (in 4~6yrs) for 1s5s, 2s5s, and 3s5s IRS yet remains positive. So far, we believe there hasn t been any unwinding ( Receive 9~11y IRS ) taken place just as yet. Going forward, should the forward spread nears 1 bp, investors must be cautious over the possibility that many issuers may rush for Receive 1y IRS. With widening reversed swap spread in 1y and bond swap positions loss cut, we could see 1y KTB yield relatively weakening. Despite the fact CMS bonds in 9 11y haven t been unwound yet, IRS rate fell the most in 1y amid the recent reversal of its curve. This is due to fundamental factors that would flatten the curve, not because of CMS. But then, it is still possible some of the 6

7 CMS issuers might have gone for Receive IRS in advance, as the short term forward tenors (in 2y or below) came under the influence of short gamma. They were probably expecting to see the long term forward spread soon reversing to around 1 bp. Another reason may also have come from their past experience: every time IRS curve was reversed, 1y tenor exhibited the highest volatility. Fig 9. IRS curve flattening and reversal Fig 1. Pay attention to 4~6y forward IRS 3s5s spread (bp) CD91d 1Y 2Y 7/28~8/3 Change in yields (L) Yields on Aug 3, 211 (R) 3Y 5Y 7Y 1Y 1Y 2Y 3Y 5Y 7Y 1Y (%) (bp) m 4-6y forward tenors still being long gamma Forward IRS 3y5y spread curve 6m 9m 12m 15m 18m 21m 24m 3yr 4yr 5yr 6yr 7yr 8yr 9yr 1yr 12yr 15yr 2yr MSB KTB IRS (Foward) Source: Bond Pricing Agencies Average, ICAP and Dongbu Securities Source: Bloomberg and Dongbu Securities Unlike our expectations, we haven t seen a severe fluctuation in the swap market even though recent reversal of IRS curve widened reversed swap spreads (IRS KTB) in long term sectors. As swap spreads in tenors from 1 to 5 yrs have repeatedly reversed and normalized, bond swap (Buy bonds + Pay IRS) perhaps might not have increased too much. Swap spreads in 7~1yrs, however, probably did experience more demand for Pay IRS than that in the past. For these swap spreads in the longterm (7 1y), duration sensitivity is quite high compared to that in short term and thus, it could create a significant market impact even with small volume of loss cuts. 7

8 Structured bonds and its influence: Power Spread Issuance and hedging structure IRS curve s recent reversal has taken place throughout 1s5s, 2s5s, and 3s5s, obviously around 5y. We believe power spread may have worked as one of the causes behind the reversal. Power spread has a cap and a floor predetermined on an initial indenture. Its coupon is determined by a specific formula; a fixed rate + leverage ratio *(3m CD 3m KTB). Floor is usually set zero for securing principal. Issuers hedge themselves by paying the same tenor IRS and buying treasury as soon as they issue the power spread. Hedging amount equals to face value*leverage ratio and therefore its market impact becomes magnified while reversed swap spread tightens. After issuance, the issuers then engage in dynamic hedging depending on moves of forward 3m CD 3m KTB spread. Power spread had been issued till August 28 and currently KRW 1.47 trillion remains outstanding. Merely 5 billion won of that amount is power spread with 5y or longer terms to maturity. 19 billion of the 5 billion won had initially been issued in 1y tenor and its leverage ratio is somewhere between times the equity, all of which are concentrated in 5y ( to maturity). Putting the ratio into a perspective, it s equivalent to KRW 2.54 trillion. Contribution to recent IRS curve reversal around 5y We have estimated a breakeven floor that would make the structured bond unprofitable. In essence, the threshold level would be at about 44 bp, generally somewhere between 41 and 48 bp. This level takes into account amount of issues and leverage ratio. Once the forward spread reaches close to the breakeven floor, gamma turns negative in some tenors and unwinding pace of over hedge accelerates. This, in turn, would lead to [ Receive IRS + Sell bond ], hence rapidly widening reversed swap spread. We also believe that the power spread with 5y to maturity recently being under short gamma influence might have helped the reversed swap spread (5y) widen. In fact, volatility of swap spread had grown whenever 5y forward 3m CD 3m KTB spread moved in and out of the 41 to 48 bp band in the past. Power spreads in other tenors 8

9 were excluded from this analysis since their maturities and thresholds are widely spread out. Fig 11. Issuance volume of Power Spread Fig 12. 5y Power Spread widened reversed swap spread (KRW 1mln) 3, Power Spread 2,5 2, 1,5 1, Source: KAP and Dongbu Securities (bp) Source: Dongbu Securities 5y Forward 3m CD - 3m MSB spread Breakeven floor: -41~-48bp Strategy IRS and KTB curves to continue flattening; 3s1s KTB spread to reverse in 1Q12 IRS curve in majority tenors had remained reversed between October 27 and September 28 during which the rate hike cycle neared its end. In the same period, the average 1s5s IRS spread was 23 bp. On the other hand, reversal of KTB yield curve had reversed only temporarily but frequently. It is quite natural to see the yield curve reverse in ending phase of a rate hike cycle. 3s1s KTB spread will likely tighten to 3 bp or lower during Q3 and may even reverse in the first quarter next year. IRS curve too should reverse frequently throughout all sectors at least for 6 months. Considering similar cases of other countries and Korea in the past, the yield curve after it had flattened or reversed remained where it had been at least for the next 6 months to 2 years. We can only see the curve steepening when a rate cut scenario or unexpected additional rate hike(s) becomes highly likely that would grow market expectations over either scenario. But, chances are low we will see any of these situations come true. 9

10 Structured bonds impact near the end of hiking cycle; stay alert on possibility of Receive 1y IRS -rush If the yield curve extremely flattens or reversed near the end of rate hike cycle, structured bonds influence becomes stronger in the market and accelerates its moving pace toward the same direction it was previously heading. Presently, unwinding effect (Receive IRS) of CMS bonds is mostly in 4 5y and 9 11y tenors. 9 11y sectors have not unwound yet. 4 6y forward IRS 1s5s, 2s5s, and 3s5s spreads still remain positive. Nonetheless, given that IRS curve was frequently reversed throughout all tenors near the end of rate hike cycle in the past, we still have to look out for widening possibility of reversed swap spread in 1y. Fig 13. Term spread in the ending phase of rate hike cycle (27.1~28.9) (bp) IRS 1Y-5Y 1Y-3Y 1Y-2Y 5Y-3Y 5Y-2Y 5Y-1Y 3Y-2Y 3Y-1Y 2Y-1Y Avg(A) Max Min (B) A-B KTB/MSB 1Y-5Y 1Y-3Y 1Y-2Y 5Y-3Y 5Y-2Y 5Y-1Y 3Y-2Y 3Y-1Y 2Y-1Y Avg(A) Max Min (B) A-B Source: Bond Pricing Agencies Average, ICAP and Dongbu Securities Fig 14. Swap spread (IRS-KTB) near the end of rate hike cycle (27.1~28.9) (bp) 1Y 2Y 3Y 5Y 7Y 1Y Avg(A) Max Min (B) A-B Source: Bond Pricing Agencies Average, ICAP and Dongbu Securities 1

11 FIXED INCOME BUSINESS TEAM RESEARCH Chief Strategist (+822) Cheongho Park Credit Analyst (+822) Minwook Park, KICPA Credit Analyst (+822) Eunjin Ko, CFA Quant Analyst (+822) Hongcheol Moon Analyst (+822) Yuna Park Analyst (+822) Sangwon, Noh Research Assistant (+822) Sam Moon Translator/RA (+822) Hwatak Chang Chief Economist (+822) Hyojin Kim Economist (+822) GLOBAL SALES Sanghee Han (+822) This report is a translation excerpt of the Korean version publicly released on August 4, 211. DONGBU SECURITIES CO., LTD. has not disclosed the material contained in this report to any institutional investor or third party prior to its publication. This report accurately reflects the personal views of the analyst(s) and was written without any undue external influence or intervention. No part of any of the analyst(s) compensation was, is, or will be directly or indirectly related to the specific recommendations or views expressed by the analyst(s) in this research. Disclosures This memorandum is based upon information available to the public. While reasonable care has been taken to ensure that the information contained herein is not untrue or misleading at the time of publication, Dongbu Securities Co., Ltd. ("DONGBU") makes no representation that it is accurate or complete. All opinions and estimates contained in this report constitute DONGBU s judgment as of the date of this report, are subject to change without notice and are provided in good faith but without legal responsibility. This report is not an offer to sell or a solicitation of an offer to buy any securities. Neither DONGBU nor any of its affiliates, nor any other person, accepts any liability whatsoever for any direct or consequential loss arising from any use of this report or the information contained herein. All materials in this report are the intellectual property of DONGBU. This report or any portion hereof may not be reproduced, distributed or published without the prior written consent of DONGBU. 11