《全球外汇展望：美元牛市压力测试.docx》由会员分享，可在线阅读，更多相关《全球外汇展望：美元牛市压力测试.docx（52页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、Stress testing the USD bullWe outline three possible paths that may lead to USD weakness In our view these are possibilities, not likelihoods. .however they could quickly start to dominate the FX marketsEvery forecast needs a stress test every so often. We have been bullish on the USD since April 20
2、18, a time when the consensus was resolutely looking for USD weakness. Since then, the USD has rallied between 6-10% depending on which basket of currencies you choose to measure it against. Throughout that time, the consensus mistakenly retained its conviction that the USD would weaken and even now
3、, it looks for the USD to depreciate by 5% by the end of next year. The past year has seen the “street” move in our direction but curiously stick to their bearish USD view. The consensus has been wrong and has responded to the dilemma of what to do next by simply shifting their previously forecast m
4、ove to happen from the new, and lower, starting point. The consensus in the last year seems to be to take whatever the spot level is at the time and add six big figures. Despite the decline in the EUR-USD the thinking of the consensus refuses to budge. The consensus seems to think uwhy change a losi
5、ng formula” and is happy to stick with the forecasting formula of “spot + sixn.Source: Bloomberg, HSBCWe face a different dilemma. The markets have played out much as we expected which means a number of our year-end 2019 forecasts are now in close proximity - EUR-USD at 1.10, USD-JPY at 105, AUD-USD
6、 at 0.68 for example. A number of the key reasons for our USD positivity - a relatively stronger cyclical story, higher yields than others in G10, safe haven status, Fed policy flexibility and a lack of alternatives - remain in play. But to retain confidence in our bullish USD when the spot market h
7、as already moved so close to our forecasts, we need to stress test our logic. It is time to ask what might prompt us to change our bullish view on the USD?3) A weak USD policy including FX interventionThe impact of FX intervention on gold is tricky. On the face of it a weak USD policy is outright po
8、sitive for gold. An important question is how other monetary authorities would respond - JPY and CHF are two safe haven currencies, which also have history of intervention. If there were tit-for-tat intervention elsewhere then gold would benefit more broadly - indeed the precious metal has already o
9、utperformed the major “reserve currencies in recent months (Chart 13). A weak USD policy would aid gold, as bullion stands out as a safe haven asset when the worlds reserve currency may lose ground. Much would depend on the levels of uncertainty and volatility generated by a weak USD policy and what
10、 part it would play in trade relations. It is possible that should the world descend into competitive 1930s style devaluation then when faced with greater uncertainty and a new precedent of direct intervention by the Fed, then investors could flock into gold in greater levels. According to Bloomberg
11、 data, gold rallied from around USD20/oz in late 1932 to above USD34/oz in 1934 - a move of 67%. Applying a similar adjustment to todays prices would suggest gold pushing towards USD2500/oz.What you missed over summerA warm summer for the USDWith the holiday over, we give a quick recap of some of th
12、e key events that happened over the last few months, and their impact on the USD, EUR, and GBP. Summer for the USD really kicked off with dovish commentary from the Fed on 19 June, which initially saw the USD end June as the worst performer in G10. This Fed dovishness culminated in a 25bp rate cut o
13、n 31 July - the first rate cut since 2008. However, with this cut largely expected, and strong economic data releases coming out during the month of July (see chart 1), the USD fought its way to the top and ended July as the strongest G10 performer - a mirror image of June.In August it was all about
14、 trade tensions, as the US started the month by imposing additional 10% tariffs on USD300bn of Chinese imports on 1 August, China imposed its own retaliatory tariffs towards the end of the month. Overall, the DXY gained 0.7% All prices from this section are taken from Bloombergthis summer, a period
15、from 1 June to 10 September.1. The US activity surprise index saw big moves following Fed statementsSource: HSBC, BloombergEurozone: A two-year low for the EURThe EUR fell 1.1% against the USD this summer. Activity data releases for the Eurozone repeatedly fell below consensus over the same period,
16、illustrated by chart 2.The EUR strengthened in June, despite the ECB keeping its policy rate unchanged and extending forward guidance that this would be the cause until mid-2020. July proved not so warm for the EUR, as it fell 2.6% against the USD. Much of the downward pressure came as the market di
17、gested a disappointing bag of data releases, helping to fuel cyclical concerns about the Eurozone. In late August the EUR slipped as President Trump tweeted that the EUR was dropping like crazy and harming US exporters, increasing concerns that the US would turn more of its trade war attention towar
18、ds the Eurozone. EUR-USD reached a two-year low of 1.0926 on 3 September.UK: The Brexit saga continuesGBP depreciated 2.1% against the USD over the summer period, as Brexit headlines and domestic politics continued to drive much of the GBP movement. Activity data was lower than expected throughout J
19、une, but from July onwards started to beat consensus estimates. As a result the UK activity surprise index is little changed over the summer (chart 3).June saw Boris Johnson surge ahead in the race to become the next PM, after Theresa May stepped down from her position on 7 June. GBP was one of the
20、weakest G10 currency performers in the same month. GBP-USD fell 4.2% in July, as the expected likelihood of a no dear Brexit continued to weigh on GBP. Boris Johnson became PM on 24 July. August saw GBP driven by more Brexit developments, although it ultimately ended the month flat against the USD.
21、A two-year low of 1.1959 was reached by GBP-USD on 3 September.2. Eurozone activity underwhelmed.3. . while UK activity redeemed itselfEurozone activity surprise indexEurozone activity surprise indexSource: HSBC, BloombergUK activity surprise indexJun-19 Jun-19 Jul-19 Aug-19 Aug-19Source: HSBC, Bloo
22、mbergii Mar zuiy - un bep zuiycca-atcG6 4 2.2,4.6E o.6so-O-O-O-OAE-au 山 sroo工roEsnpu- CJSnRVZ osrvMo:工 Qsrvonq QW0.03 CJSrvBN- CJSMtrB sCDwnb Esel 工山 s三nb山工M山乏山 s.=nbD0Sx an山 pao d_0。工 osn s三nbeo-cEq。七07 sasnbdon山 slDb(Dog-oEd asn，NX 工 po_s en=n.2J64 (asn二qap 江山 (Sep工山 AO ;_FdL2J AOL sn upal。d-8 0-
23、osn&8 o- dm。AOL xn:pal。d-8 o- an 山JPY - The true safe havenWe have identified that we may be moving back into a RORO world, our recent multi-asset research also suggests the JPY is still one of only three Iraditionaf safe havens, alongside gold and developed market government bonds. This means that
24、if we go risk-off in the future the JPY could see strong appreciation. A caveat to this is that previously Japans Ministry of Finance, through the Bo J, has intervened in the FX market to prevent the JPY from appreciating excessively. However, we would argue that this is no longer a free option, as
25、it may provoke political resistance from the US administration through tariffs. In this case, USD-JPY could fall with less support from the Bo J.The return of ROROCross-asset correlations over the past six months have shown a distinctly Risk On - Risk Off profile, as shown by chart 1. In this RORO w
26、orld, assets lose many of their idiosyncratic drivers, and become broadly split into the two groups of risk on and risk off. During normal times1, these correlations are not so evident, so when they rear their head, it can undermine the diversification of a portfolio if not carefully considered (see
27、: The divers币cation illusion, 18 July 2019/It seems this return of RORO behaviour has somewhat coincided with the rise in global trade tensions. But regardless of what has driven this rise in cross asset correlations the underlying question for investors is: what can you rely on for portfolio protec
28、tion during a market sell-off?We identify the JPY as being a reliable and traditional safe-haven asset. With the backdrop of continued geopolitical and trade tensions, a RORO correlation structure, and the JPY retaining safe haven status, it may not take much for another wave of risk-aversion to sen
29、d USD-JPY lower.1. Cross-asset correlations over the past 6 monthsJapan 10Y EUR IG corp credit German 10Y UK 10Y GBP IG corp credit USD IG corp credit US 10Y | JPY-USP Precious Metals HUF-USD EUR-USD GBP-USD Italy 10Y EM debt (local) EM debt (USD) Agriculture Livestock MXN-USD Pacific equities Europ
30、e equities North America equities USD HY corp credit EUR HY corp credit EM Asia equities EM E&ME equities EM LatAm equities BRL-USD INR-USDCAD-USD AUD-USD KRW-USD ZAR-USD Industrial Metals EnergySource: Bloomberg, Refinitiv Datastream, HSBCNote: All correlations are calculated using local-currency t
31、otal return indices. The assets are ordered such that those with similar correlations to other assets are together.What can stop JPY strength?Previously when there was strong demand for the JPY, the Ministry of Finance has intervened in the FX market through the Bo J, to prevent its currency from ap
32、preciating too much, as illustrated in chart 2. However this may have changed. We recently argued the reaction function of the BoJ may now be influenced by the threat of tariff responses from the US, as the Trump administration seeks to take action against central banks purposefully weakening their
33、currencies against the USD (see: Tariffonomics & FX, 16 July 2019). If the Japanese authorities were to intervene, they may be met with increased US tariffs, which might render the intervention somewhat impotent. With more pressure on the BoJ to leave the JPY in the hands of the market, a “risk off
34、move could see USD-JPY fall even further and with less support than it has previously received from the BoJ.2. A timeline of BoJ FX interventionsSource: Bloomberg, Japan MoF, HSBCData Science UpdateDuring September we published five pieces using Machine Learning techniques. The research has been app
35、lied to several asset classes and has covered a wide range of applications: from short term tactical trading strategies to long term strategic analysis of market cycles. Below we give a short summary of these pieces.Strategic Analysis of Market CyclesIn Global Credit Strategy: 2020 and beyond、5 Sept
36、ember 2019, we apply a model called XGBoost to over 70 years of data in order to predict credit bear markets.This model currently assigns an 84% probability to a credit bear market in one years time.This model has had good out of sample performance and correctly predicted five of last the six credit
37、 bear markets with only one major false positive.The lead time of the model on successful predictions has been between 7 and 16 months.In Detecting cycles , 4 September 2019, we use a clustering algorithm known as K-Means to analyse the equity market cycle. The model works by grouping time periods t
38、ogether.This analysis suggests that at this time we are in a late cycle stage for equity markets.The equity market has been edging closer to the bear market cluster since 2018.More encouragingly, this cycle analysis can provide useful signals which investors can use to switch between cyclical and de
39、fensive sectors.Short term tactical trading signalsIn CEEMEA Quant Strategies, 4 September 2019 we train logistic regression and XGBoost models to predict whether CEE rates will fall over the following month.Out of sample performance has been high with model accuracy ranging from 61% to 70%.Performa
40、nce has been particularly good for CZK - this fits with anecdotal evidence that model trading activity is particularly important for CZK rates. The CZK rates model turned to give a pay signal on 19 August.This piece pairs well with the Trading EUR duration, 19 August 2019 note published earlier in t
41、he summer which used a similar approach for EUR rates.Natural Language ProcessingNatural Language Processing (NLP) is the subfield of Machine Learning techniques which can be applied to text.In The climate is chanqing for airlines、10 September 2019, we track discussions of climate change on airlines
42、 earnings calls. Climate is suddenly a hot topic for airlines - there has been a surge in discussions of climate change during 2019. This is a much hotter topic for European airlines than for US airlines. Discussion of climate issues by US airlines has historically focussed more on fuel efficiency.I
43、n Data Matters: Uncertainty W Risk, 4 September 2019, we again turn to our database of earnings call transcripts to help us understand why markets are not more volatile in the face of a raft of political and economic uncertainties.In case you missed itEarlier in the year we published two Machine Lea
44、rning primers: Machine Learning 101, 11 March 2019Machine Learning 102, 1 May 2019We believe that Machine Learning techniques are becoming increasingly important for financial market data analysis. These two primers should help our readers gain a more intuitive understanding of how these techniques
45、work.Mark McDonaldHead of Data Science and AnalyticsHSBC Bank plcmark.mcdonaldhsbcib +44 20 7991 5966Song Jin Lee, CFAEuropean Credit StrategistHSBC Bank plcsongjin.leehsbc +44 20 7991 5259Piran PhippenCEEMEA Rates StrategistHSBC Bank plcpiran.c.phippenhsbc +44 20 7991 5693Shiva JoonData ScientistHS
46、BC Bank plc shiva.joonhsbcib +44 20 7991 1356Alastair Pinder, CFAGlobal Equity StrategistHSBC Securities (USA) Inc. alastair.pinderus.hsbc +1 212 525 4131Subhrajit Banerjee, CFA Fixed Income StrategistHSBC Bank plcsubhrajit.banerjeehsbcib +44 20 7991 6851Ashim PaunCo-Head, ESG Research; ClimateChang
47、e StrategistHSBC Bank plcashim.pai.jnh$bcibcom+44 20 7992 3591Andrew Lobbenberg*AnalystHSBC Bank plcandrew.lobbenberghsbcib +44 20 7991 6816Employed by a non-US affiliate of HSBC Securities (USA) Inc, and is not registered/ qualified pursuant to FINRA regulationsG10 at a glanceUSD-CADUSD-CAD05 06 07
48、 08 09 10 11 12 13 14 15 16 17 18 19Source: HSBC, Bloomberg05 06 07 08 09 10 11 12 13 14 15 16 17 18 19Source: HSBC, Bloomberg05 06 07 08 09 10 11 12 13 14 15 16 17 18 19Source: HSBC, BloombergCanada: Risks are buildingWe see the CAD weakening over the coming year. This outlook is determined by a blend of the generally better than expected domestic outlook, and a challenging external environment With the market no longer aggressively priced for BoC easing, we believe the balance of risks tilts towards currency