Chart 17: SXEP calls would have generated a higher average payoff and Chart 18: Selling the USO 25d put (~8.7 strike as of 5-Aug) to buy the more frequent positive returns vs. USO calls (when sized for an upfront SXEP 42d call (~280 strike as of 5-Aug) for ~0 provides early upside cost equal to the current price of the USO 3M 25d put) participation and a ~12% downside buffer on the short USO put leg 300 13 Avg Payoff Freq of >0 F 60% . Long SXEP call strike ° 2.1x SXEP 3M 25d calls: 2.9% 27% 290 (unlevered implementation) N 42 50% 1.3x USO 3M 25d calls: 1.4% 10% 280 —S o—— — ss “p 970 \ ‘iN } V | Vea 40% Vip pe 260 AY 10 9 . | | 30% 250 ) ? | " 9 20% 240 7) = = = = ° y d Short USO put strike|_ 9 10% 230 220 7 0% SaSCGe eR RE SES ees 08 = 09 10 140 M12—~CO18 "1415 16 & & & @ @ g g £2285 5 BSBs8 ———SXEP Call Payoff === USO Call Payoff —— SXEP — USO (rhs) Source: BofA Merrill Lynch Global Research. Data from 1-Jan-16 tO 5-Aug-1 6. Source: BofA Merrill Lynch Global Research. Data from 17-Jan-16 to 5-Aug-16. Backtesting is Backtesting is hypothetical in nature and reflects application of the screen prior to its introduction. It hypothetical in nature and reflects application of the screen prior to its introduction. It is not is not intended to be indicative of future performance. intended to be indicative of future performance. Chart 19: Number of bonds purchased by the ECB in their CSPP Chart 20: Oil drawdowns have recently coincided with $ weakening programme by sector . . "mt Bonds Purchased ti“‘é~S™S**SSS!SC 70% Negative correlation would have helped 140 % of Eliaibl = 55% dampen potential USD losses on the 490 m% of Eligible — 50% short USO put leg 100 iii 45% 0 80 a," + w | 250 5 60 = 35% 30% i] i] 2 P| we 8 1 1 =" ° & “> © dS Oo a2 Sw ASOD 19% -10% FF SF PF & GO & £@ For GF Ke oy" & = -30% as & .S RS OS < * SY & Ss ' ' ' ' ' ' ' ' ' ' < AS RS eS o7 608 = 09 10 11 12 13 ‘4 15 ‘6 oy SF of e ——USO vs. EURUSD correlation* Conditional on USO down Source: BofA Merrill