1、|TRADE IDEA 25/06/2019 1 FOCUS|EM 24 June 2019 EM:Look at systemic factors and see the future KEY MESSAGES Our PCA monitor shows that the rally in EM local currency rates and sovereign credit has been driven almost entirely by systemic forces.US corporate credit,2 year US real rates,and the meltdown
2、 of the US and European 5y and 10y rates have been the most important drivers,our models show.An important observation is that systemic dynamics prevail.Ninety percent of the weekly variability of EM local currency rates and sovereign credit is explained by the first two common latent factors.Althou
3、gh the sources of variability may change among countries,the results show that these assets move in a synchronised manner.As expected,the country with the most idiosyncratic component has been Turkey.We expect structurally lower real interest rates in the US to bode well for EM.Ceteris paribus,we co
4、ntinue to see low systemic risk for the asset class going forward.We reinforce our bullish call for emerging markets.Please refer to important information and at the end of this report EM STRATEGY Gabriel Gersztein,Global Head of Emerging Markets Strategy|Banco BNP Paribas Brasil Principal component
5、 for EM how it works We analyse the relationships between a set of emerging market(EM)sovereign credit and external common factors by using Principal Component Analysis(PCA).This approach helps us to categorise the factors that have been driving the variability in EM IR and sovereign credit.Forward-
6、looking scenarios can then be constructed for these drivers to assess the theoretical market impact of sudden changes in their trends.We use a two-step procedure.The first step is to extract the two most important common latent factors from these data series.In the second,we calculate the co-varianc
7、e between(a)these factors and(b)potential explanatory variables.Our approach uses a 35-weekly rolling window to calculate the principal components from a basket of EM sovereign credit.This basket includes the 5y CDS of the most relevant and liquid EMs:Brazil,Mexico,Colombia,South Africa,Turkey,Indon
8、esia,Russia,Chile,Philippines,Malaysia and Argentina.For interest rates,the basket includes the 5y local currency swap rate from some of the most relevant EM countries:Brazil,Mexico,Colombia,South Africa,Turkey,Indonesia,South Korea,Thailand and Malaysia.We have not included China and India,due the
9、high idiosyncrasies of these markets the idea was to use systemic common factors and not domestic ones.Main strategies in place Structural weak US dollar view expressed through long CZK and PLN against USD and long BRL,HUF,CLP,COP against EUR(total allocation of USD85mn).Flattening in Brazil(DI Jan-
10、22sJuly20s and Jan-21sJan-25s;total allocation:USD35,000 DV01)and in Mexico TIIE5y10s with an allocation of USD25,000 DV01).Receiving 10y rates in Czech Republic and 2y rates in Mexico using local currency swaps(total allocation:USD25,000 DV01).Tactical paying local currency rates in both Colombia(5
11、y IBR)and Chile(1y CLPxCAM)with a total allocation of USD22,000 DV01.Sell protection in Russia and Turkey via 5y CDS(total allocation:USD40mn and USD13mn respectively)Selected thematic publications EM:Fed insurance,EM endurance EM Flows:Slight reversal in May EM Strategy:Understanding the transmissi
12、on channels on local yields Link to our piece from Q3 2018:EM Strategy:The worst is over|FOCUS 25/06/2019 2 The process we use to estimate the market factors provides the data series in which selected local assets co-vary,but it does not indicate what is effectively driving these factors.The results
13、 of step two are shown in Tables 1-2:US Corporate credit),2-year and 5y5y US real rates,and loosening financial conditions are the explanatory variables with the highest co-variance with the first common factor.Compared to Q4 2018,there has been a substantial jump in the explanatory power in all of
14、them(Table 1).As expected,countries such as Mexico,Peru,Indonesia,Malaysia,and Colombia show a substantial co-variance with US corporate credit and US real rates(Table 2).Lastly,it is worth mentioning the increase in the explanatory power of BNPP political risk index for Turkey.Systemic risk is cons
15、iderable:89.6%of the weekly variability of EM sovereign credit has been explained by the first two common latent factors(Table 3).Although the sources of variability may change among countries,the results show that these assets generally tend to move in a synchronised manner.In our PCA publication o
16、f 15 April 2019,we reported that the credits with the most idiosyncratic component have been Argentina(systemic factors explained only 6.7%of the weekly variability of 5y CDS)and Chile(which continued to decouple from the rest of the countries and performed more like a developed country).PCA suggest
17、s that systemic factors dominate returns in both countries this time around(Table 3).EM sovereign credit EM STRATEGY Tables 1-3 and Fig.1-3:Ranking of 25 variables with the highest co-variance to the 1st and 2nd common latent factors Gabriel Gersztein,Global Head of Emerging Markets Strategy|Banco B
18、NP Paribas Brasil Sources:Bloomberg,BNP Paribas;as of 24 June 2019 R2 with 1st market factorR2 with 2nd market factor Explanatory powerCurrent28-Dec-18Current28-Dec-181st Market2nd MarketSumAverage(5 first variables)82.8%Average(5 first variables)49.0%Factor(a)Factor(b)(a)+(b)US 2y real rates89.5%24
19、.2%BNPP Political risk idx60.0%2.4%Brazil66.6%22.2%88.9%US Corporate IG(CDX)86.4%7.5%BNP Global Risk Premium58.6%29.9%Turkey 24.9%74.9%99.8%Financial Stress82.2%1.5%USDCNY44.1%3.3%Mexico87.9%0.0%87.9%Global Financial Conditions78.5%6.4%US 12m Euro hedged Xccy basis(3m)43.4%46.4%Colombia83.4%0.1%83.6
20、%US 10s30s77.7%8.7%UST Implied Volatility39.1%65.5%South Africa71.5%21.8%93.4%US 5y5y real rate75.8%4.8%Industrial metals30.6%3.2%Indonesia87.0%5.5%92.4%EUR Swap rate 5Y72.1%3.2%US 2y Euro hedged Xccy basis(3m)24.8%70.1%Russia80.0%1.0%81.0%US Money Market conditions70.6%56.5%US 2y Yen hedged Xccy ba
21、sis(3m)24.1%51.4%Peru90.4%0.0%90.4%US Corporate HY(CDX)68.5%1.3%US Eco Surprises21.8%9.8%Malaysia97.4%1.8%99.2%US 5y swap rate64.7%2.0%US 5y Euro hedged Xccy basis(3m)17.0%60.0%Chile77.1%0.8%77.9%Global slope 2s10s64.6%26.1%Euro Funding 3m16.2%8.5%Philippines97.7%0.6%98.3%US 10y swap rate63.9%1.5%EU
22、RUSD13.3%28.8%Argentina49.3%32.6%81.8%US 1y1y forward63.4%4.4%US 1y1y forward6.1%2.9%Average76.1%13.5%89.6%US 5y Euro hedged Xccy basis(3m)57.3%8.9%US 10y swap rate5.1%0.7%US 2y Euro hedged Xccy basis(3m)48.6%0.1%US 5y swap rate4.7%0.2%US Eco Surprises46.0%19.9%US 5y5y real rate4.7%28.4%EURUSD45.4%3
23、5.1%US 2y breakeven4.6%33.5%US 2y Yen hedged Xccy basis(3m)39.7%13.7%EUR Swap rate 5Y4.5%2.6%US 2y breakeven38.4%10.9%Global Financial Conditions4.3%64.9%USDCNY30.6%56.2%US Corporate HY(CDX)3.7%66.1%US 5y5y break even29.4%0.5%US Corporate IG(CDX)3.6%63.4%Crude Oil29.0%0.9%Crude Oil2.1%38.0%US 12m Eu
24、ro hedged Xccy basis(3m)16.9%2.1%US 10s30s1.9%56.8%BNPP Political risk idx7.1%24.0%US Money Market conditions0.7%10.9%BNP Global Risk Premium6.6%9.1%US BAA 10y spread0.5%53.8%2552672792913030.60.40.30.10.10.20.40.5Oct-18Dec-18Feb-19Apr-191st Market componentUS Corp IG(rhs;inverse)050100150200-0.6-0.
25、4-0.3-0.10.10.20.40.5Oct-18Dec-18Feb-19Apr-191st Market componentUS 2y real rates(rhs)0255075100-0.6-0.4-0.3-0.10.10.20.40.5Oct-18Dec-18Feb-19Apr-191st Market componentUS 5y5y real rates(rhs)|FOCUS 25/06/2019 3 According to our calculations,the long-term impact of US corporate credit on different EM
26、 USD sovereign and quasi-sovereign names is relatively low,yet the results show that EMs are not insensitive to negative shocks from US corporates in the short-term.Any credit event or negative news flows from US corporate credit has the potential to impact EM sovereign credit premium(mostly in spec
27、ulative grade names).Likewise,the recent rally in US corporate credit has boded well for EM sovereign credit(Fig.4).The financial transmission mechanism Negative and low interest rates have occurred on two occasions since the 2008-09 global financial crisis,periods that coincided with strong flows a
28、nd the outperformance of EM fixed income assets.A constant question regarding exchange rate determination(and US dollar flows)is whether to use nominal or real rate differentials?At a theoretical level,real rates are what matter,in our view,as we assume that a fundamental external factor for emergin
29、g markets is the lending of US dollars outside the US,which is directly linked to the domestic level of real interest rates.Investors can look through the loss of purchasing power created by inflation,so exchange rates tend to hover around real rate,not nominal,differentials.In addition,real rates m
30、atter most when explaining currency moves(and financial flows)in the medium and long-term.In terms of the use of short-or long-term interest rates,the Federal Reserve and the IMF suggest focusing on longer-term rates now.According to a study by the Bank of International Settlements,central banks in
31、most emerging markets have limited ability to influence the long-term interest rate in their own currencies.The direct influence of changes in their policy rate relative to that of the Federal Reserve is small.A rise in the federal funds rate(other interest rates being constant)has a much smaller di
32、rect effect than often assumed.A 100bp increase in the federal funds rate adds directly only 7bp to long-term rates overseas(BIS working paper:Low long-term interest rates as a global phenomenon;Peter Hrdahl,Jhuvesh Sobrun and Philip Turner;August 2016).On average,since 2005 a 100bp rise in the US 1
33、0-year yield is associated with an 80bp rise in the yields in other bond markets swamping the effects of changes in short-term rates.EM sovereign credit EM STRATEGY Fig.4:EM hard currency yield Gabriel Gersztein,Global Head of Emerging Markets Strategy|Banco BNP Paribas Brasil Sources:Bloomberg,BNP
34、Paribas;as of 24 June 2019 6.516.244.065.092.84.35.87.38.810.320112012201320142015201620172018EM USD unhedged|LatamUS Corporate HYEM USD unhedged|AsiaEM USD unhedged|EMEAHard Currency I RegionSources:Bloomberg,BNP Paribas;as of 24 June 2019 US long-term yields can be broken down into two elements:(i
35、)the average of expected future short-term rates;and(ii)a term premium.The term premium seems to matter more for international correlations than the average of expected future short rates,and this is particularly true for emerging markets.As Fig.5 shows,there has been a strong co-movement between lo
36、ng-term US real interest rates and EM sovereign credit(benchmarked by the EMBIG yield unhedged).Fig.5:US real rates and EM credit Sources:Bloomberg,Macrobond,and BNP Paribas;as of 24 June 2019 4.55.36.16.97.7-50050100150Jul-12Nov-13Mar-15Jul-16Nov-17Mar-19LT US real rates(in bp)EMBIG 10y yield blend
37、ed(rhs)Taper tantrumOpec/RMB devaluationTrump;UST steepening|FOCUS 25/06/2019 4 EM local currency rates EM STRATEGY Gabriel Gersztein,Global Head of Emerging Markets Strategy|Banco BNP Paribas Brasil Sources:BNP Paribas;as of 24 June 2019 Our principal component analysis suggests that the meltdown o
38、f US and EUR interest rates has been the most important driver of local currency rates in Asia,CEEMEA and Latin America.Systemic risk is high:Over the last months,the first common latent factor has explained almost 80%of the variability of EM local currency rates.US and EUR 5y swap rates,US 1y1y for
39、ward rate and 5y5y real rates are among the factors with the highest explanatory power(Table 5).While contagion from the commercial channel and trade tensions via the strengthening of the US dollar poses a risk,global liquidity and US financial conditions remain at supportive levels for local curren
40、cy rates.Positions in EM were mostly in hard currency and/or FX hedged.In a nutshell,there was a widespread concern as to whether it was too early to take on local currency risk.We have been in the longest cycle of strong US dollar since 1970s and EMs have shown resilience to it.Due to the key role
41、that the USD plays as both invoice and funding currency,a weaker greenback should benefit EM through increasing cross-border bank lending and higher commodity prices in USD terms.Thus,in an environment of depressed nominal and real rates in DMs,commodity currencies would have a considerable margin f
42、or catch up.We reinforce our scenario of weaker US dollar going forward.Trends are easier to identify when they are already in place and markets,sometimes,wait for a facilitator to adjust to new scenarios.In this case,the Fed was the long-waited catalyst,we believe.We expect structurally lower real
43、interest rates in the US and Europe to bode well for EM local currency rates.Ceteris paribus,we continue to see low systemic risk for the asset class going forward,supporting our bullish scenario.Tables 5-7:Ranking of 20 variables with the highest co-variance to the 1st,2nd and 3rd common latent fac
44、tors Table 4:1st and 2nd common latent factors for EM 5y local currency rates Sources:BNP Paribas;as of 24 June 2019 R2 with 1st componentR2 with 2nd componentCurrent28-DecCurrent28-DecAverage(5 first variables)96.2%Average(5 first variables)29.9%EUR Swap rate 5Y98.1%4.7%Crude Oil41.3%46.0%US 5y swa
45、p rate96.9%15.7%US BAA 10y spread40.4%60.1%US 1y1y forward96.3%26.5%Commodity Prices38.0%63.0%EU Swap 2s5s95.3%1.8%BNP Global Risk Premium16.3%34.6%US 5y5y real rate94.3%0.0%US 5y breakeven13.4%74.6%St Louis Fed Financial Stress Index49.3%49.0%US Money Market conditions13.4%5.3%US Money Market condi
46、tions44.7%81.3%US Eco Surprises12.2%43.9%US Eco Surprises44.0%0.9%US Economic Uncertainty Policy11.9%0.0%Global Financial Conditions43.0%12.6%St Louis Fed Financial Stress Index8.6%39.7%US Swap 2s5s35.0%0.4%BNP Global Risk5.9%12.6%Atlanta Fed GDP31.5%16.5%US dollar TWI5.1%92.7%US BAA 10y spread22.9%
47、22.4%Atlanta Fed GDP3.5%65.7%Renminbi nominal TWI19.5%16.9%Renminbi real TWI2.7%57.9%BNP Global Risk19.5%7.6%Renminbi nominal TWI2.1%54.8%Renminbi real TWI16.9%17.4%Global Financial Conditions2.0%73.1%US Economic Uncertainty Policy15.9%4.7%UST Implied Volatility1.7%37.2%Proxy of Global Demand11.2%9.
48、5%US Swap 2s5s1.4%76.3%USDCNY9.4%19.0%USDCNY0.8%63.4%US 5y breakeven9.1%6.6%Proxy of Global Demand0.7%65.4%UST Implied Volatility5.3%15.0%US 5y swap rate0.6%0.1%Explanatory power(R2)1st Market2nd MarketSumFactor(a)Factor(b)(a)+(b)Brazil 5Y80.9%10.0%90.9%Mexico 5Y84.9%0.7%85.6%Colombia 5Y90.1%5.0%95.
49、1%South Africa 5Y89.8%2.8%92.6%Turkey 5Y20.9%78.1%99.0%Indonesia 5Y91.8%2.7%94.5%Korea 5Y90.3%1.8%92.1%Thailand 5Y73.1%9.2%82.3%Malaysia 5Y94.9%1.4%96.3%Average79.6%12.4%92.1%Legal Notice This document has been written by our Strategist and Economist teams within the BNP Paribas group of companies(c
50、ollectively“BNPP”);it does not purport to be an exhaustive analysis,and may be subject to conflicts of interest resulting from their interaction with sales and trading which could affect the objectivity of this report.This document is non-independent research for the purpose of the UK Financial Cond