H2-2017: Where to invest?

H2-2017: Where to invest?

Key takeaway – The investment environment remains challenging. In their portfolios, investors should: i) privilege a multi-year horizon; ii) focus on capital preservation by adopting a defensive risk profile; and iii) accept lower expected returns in exchange for lower volatility. Wealth preservation requires conservative investment strategies, with a greater exposure to alternatives. The recommended strategic asset allocation (SAA) is depicted in Figure 1. In the liquid space, the portfolio should privilege bonds (25 percent) and stocks (20) over commodities (10) and cash (5). In the illiquid space, the allocation should be equally divided between real estate (RE, 20 percent) and private equity (PE, 20 percent). With respect to the SAA recommended in H1-2017, the portfolio’s risk profile has increased slightly, with lower allocations to Unites States (US) stocks and oil, to the advantage of—respectively—Eurozone (EZ) stocks and emerging markets (EM) corporate bonds. Recent data justify moderate risk-taking: in H1-2017, global growth mildly accelerated and inflation rose, while remaining below central banks (CBs) targets; most asset classes registered a positive performance, with EM stocks leading the gains. In H2-2017, these trends will continue, and global financial markets will be supported by liquidity. Yet, as market valuations are still disconnected from underlying macro fundamentals, over-valuation and political tensions could create instability, especially in the US, and bring about a correction. Over the next few years, “expected returns” are likely to be lower than in the pre-crisis period, and monetary policy divergence will create jitteriness: once liquidity starts to decline, bear markets—if not a full-fledged market crash —are likely.

Figure 1. SAA H2-2017 – Recommended portfolio

Note: As of June 20, 2017, the above portfolio returned 48.5 percent since inception (January 1st 2012), with a 6.9 percent annual average, a volatility of 1.8 percent, a Sharpe ratio of 1.19, and a Sortino of 1.77. Over the same period, a riskier, classic “60 percent stocks/40 percent bonds” portfolio would have delivered 47.8 percent, with a 6.8 percent annual average, a volatility of 1.9 percent, a Sharpe of 1.06 and a Sortino of 1.57.

Source: Authors’ elaboration, 2017.

1. SAA: defensive approach, lower expected returns, greater exposure to alternatives.

Suggested investment strategy: focus on capital preservation. The investment environment (Section 4) remains very challenging. Investors should privilege a multi-year horizon, adopt a defensive risk profile in their portfolios, and accept lower expected returns in exchange for lower volatility (i.e.: a lower standard deviation of the expected returns). Wealth preservation requires conservative, unconventional investment strategies. Unusual, less liquid portfolios—broadening exposure beyond conventional stocks and bonds, identifying opportunities in the illiquid space—are likely to perform better than conventional ones. As a result, only 60 percent of the assets should be liquid, and as much as 40 percent should be kept illiquid (Figure 1 and Table 1).

Going forward, “expected returns” are likely to be lower than in past years. With slow economic growth (Section 4), revenues and corporate earnings will stagnate. For most asset classes, returns will remain low[1]. Over the next few years, unless CB liquidity is withdrawn gradually, a rising disconnect between fragile fundamentals and elevated valuations will bring about a bear market (a drop of more than 20 percent from the 52-week high)—if not a full-fledged market crash (a more-than-40-percent plunge)[2]. In other words, fundamentals will eventually determine asset values.

In the liquid space, the allocation should privilege bonds (25 percent) and stocks (20) over commodities (10) and cash (5).

  • Stocks (20 percent) – Equities will continue to receive inflows, because: i) investors lack viable alternatives; ii) often face limits on cash holdings; and iii) will continue to ride liquidity-driven markets that keep achieving new highs. In developed markets (DMs), corporate cash will keep supporting share buybacks. The whole allocation should go to DM large caps with cash flow. In particular, dividend-paying blue chips are preferred—better if multinational brands with exposure to EM domestic demand. Within DMs, a reallocation from the US to the EU is justified by: a) lower prices of European stocks; and b) improving political and economic sentiment in the EZ (Section 3). As volatility risks remain elevated globally, small caps and EM stocks are better avoided.
  • Bonds (25 percent) – Held for capital preservation purposes, bonds should follow a “buy to hold” strategy, with profit-taking in case of yield-reducing risk-off episodes, unexpected liquidity-injections by CBs, and further reductions in long-term interest rates. In DMs, overbought sovereign should be avoided[3]; US yields will rise, but remain subdued – supported by the ongoing quest for financial safety: safe assets[4] are in demand and suffer from a chronic shortage. To prop-up yields, the allocation should privilege USD-denominated corporate debt: 20 percent to high quality, high-yield multinational brands—of which 15 percent in DMs and 5 percent in EMs (blue-chips only). An additional 5 percent should be allocated to EM sovereign bonds, only if USD-denominated, mostly investment grade and with above-inflation yields.
  • Commodities (10 percent) – The allocation should include energy (5 percent) and precious metals (5 percent) to capture the rise in oil and gold prices in the event of risk-off episodes and as an insurance against: a) currency debasement; b) sharp downturns; and c) periphery-to-core runs.
  • Cash (5 percent) – Cash is required as: a) seeding-capital to quickly seize opportunities; b) insurance against sharp downturns; and c) protection against negative rates. The 5 percent allocation should be held in money market funds, cash ETFs and hard cash. Given depreciation pressures on EM currencies, the US dollar (USD) and the Swiss Franc (CHF) should be favoured in DMs, and the Singapore Dollar (SGD) in EMs.

In the illiquid space, the allocation should be equally divided between RE (20 percent) and PE (20 percent). Years of aggressive monetary policy have inflated the prices of liquid assets, creating the risk of low returns and volatility, which in the portfolio are partially offset by selected RE and PE investments.

  • Real estate (20 percent) – As much as 15 percent invested in RE should be allocated in DMs – as a fixed-income play, based on rental income and should go to: a) undervalued “trophy” assets; and b) distressed properties in rapidly growing cities. The remaining 5 percent should be invested in EMs as a capital appreciation play based on fundamentals[5].
  • Private equity (20 percent) – As much as 15 percent should be allocated to DMs, invested in undervalued firms with positive cash-flow, able to produce non-replicable or luxury products in high demand in EMs. The remaining 5 percent should go to EMs: despite regulatory and currency risks, fundamentals are supportive.

Changes versus the SAA recommended in H1-2017: a slight “risk-on” tilt. Within a defensive positioning, there is room for moderate risk-taking. While the total allocation to stocks has remained unchanged (20 percent), the weight of EZ stocks has been increased (from 8 to 14 percent of the portfolio) via a reduction of US stocks (from 12 to 6 percent). Finally, exposure to oil has been reduced to the advantage of EM corporate bonds (Table 1).


Table 1. SAA H2-2017 – Recommended portfolio (with changes vs. SAA H1-2017)

*Note: Increased allocation to EZ stocks (from 8 to 14 percent) due to a decreased allocation to US stocks (from 12 to 6 percent).

2. Year-to-date (y-t-d) performance: mild growth and positive markets.

Over H1-2017, growth picked up in EMs … Between Q1-2016 and Q1-2017 (quarterly data, year-on-year), growth accelerated in most large economies. In DMs, GDP growth accelerated in the US (from 1.6 percent to 2.0), in the EZ (from 1.7 to 1.9) and in Japan (from 0.5 to 1.3). In EMs, growth rose in China (from 6.7 to 6.9 percent), Brazil (from -5.5 to -0.4), and Russia (from -0.4 to 0.5), while in India it decelerated (from 7.1 to 6.2). Over the same period, inflation accelerated in the US (from 1.1 to 2.5 percent) and in the EU (from 0.0 to 1.8). Conversely, in most EMs it declined: in China (from 2.1 to 1.4 percent), India (from 2.7 to 2.4), Russia (from 8.3 to 4.6), and Brazil (from 10.1 to 4.9). Prices rose in Turkey, from 8.6 to 10.2 percent.

… and most asset classes registered a positive performance, with EM stocks leading the gains. In USD terms, most asset classes have enjoyed a positive performance y-t-d[6]. In the equity space, EMs overperformed: the MSCI EM Index rose by 17.5 percent, more than the US S&P500 (9.6 percent) and the Eurostoxx 600 (14.1 percent, 8.2 in euro terms). In the fixed income space, EM high yield led increases, with a 5.7 percent price gain, while global corporate bonds rose by 4.2 percent. In Q2-2017, US Treasuries (UST) experienced strong inflows, with 10-year UST yields declining from 2.44 to 2.18. Commodities suffered a negative performance: the Bloomberg Commodity Index declined by 8.1 percent, driven mostly by the drop in oil prices (-19 percent in 2017).

3. 2017 outlook: mild growth acceleration, market performance still led by liquidity, further disconnect between fundamentals and valuations.

In H2-2017, growth and inflation will keep rising, but remain below trend. In several large DMs and EMs, growth will continue to rise. Inflation will remain subdued in DMs (2.4 percent in the US, 1.6 in the EZ, and 0.6 in Japan) and EMs (2.2 in China, and between 4 and 5 percent in India, Brazil, Russia, Mexico and Indonesia).

  • DMs: between 2016 and 2017, growth is expected to accelerate in the US (from 1.6 to 2.2 percent) and to remain unchanged in Japan (from 1.0 to 1.1) and in the EZ (at 1.7). In the US, the economy could lose momentum as President Trump’s electoral promises (namely, infrastructure spending, tax cuts and financial deregulation) are unlikely to materialize. The labour market remains strong, but a low participation rate and limited wage growth keeps inflation subdued. In the EZ, growth will be supported by a more benign political calendar and inflation will accelerate -but will remain below the 2.0 percent target of the European Central Bank (ECB), granting room for further monetary loosening. Owing to monetary policy divergence, the USD is likely to modestly appreciate against the EUR.
  • EMs: India, Russia, Brazil and Turkey will record faster growth than in 2016. Between 2016 and 2017, growth will accelerate in India (from 6.8 to 7.0 percent), Brazil (from -3.6 to 0.3) and Russia (from -0.2 to 1.1). In China, the deceleration will be mild, from 6.7 to 6.3 percent, but regulatory tightening has increased the risk of a hard-landing; however, the Chinese authorities are vigilant and will relax their stance were the situation to deteriorate. In the Middle East, the row between Qatar and the GCC will gradually be scaled down. However, if tensions reignite oil prices could rise significantly. The slow pace of policy normalization in the US will allow EM currencies to hold their ground against the USD.

In H2-2017, global markets will continue to be supported by liquidity, but political tensions could create instability.

  • Stocks: In H2-2017, in most US indices the risk of a correction (a fall of more than 10 percent from the 52-week high, over a single week) will increase due to the: a) elevated valuations: the forward P/E[7] is above historical averages[8] in the S&P500 (18.9), Dow Jones Industrial Average (18.9), Russell 2000 (19.8) and Nasdaq 100 (21.8); and b) rising likelihood of a destabilizing political event—e.g.: an impeachment process on President Trump. Given the size[9] of US stock market, some contagion is possible. European stocks enjoy lower valuations (forward P/E 5[10]) and stand to benefit from improving economic sentiment. With volatility near all-time lows (VIX[11] at 10.5), the price of hedges to protect investors against stock price declines reached historical lows[12].
  • Bonds: In DMs, bond prices are likely to stabilize at current levels. In the US, policy rates are expected to increase, but the strong demand for UST will inhibit a rise in yields. In the EZ, prices of European debt will be supported by the extension into 2018 of the ECB asset purchase program. In EMs, given a more favourable economic outlook, USD-denominated debt has the potential to perform better than DM debt. EM corporate bonds offer: a) higher yields—although below historical averages; and b) the potential for further price increases.
  • Commodities: oil prices are likely to remain in the range of USD 45 – 55 per barrel, as: a) the extension of the Opec agreement provides a floor to price declines; and b) the stable geopolitical environment makes strong increases improbable. Weak global demand and abundant inventories will keep the price of other commodities subdued. Price increases will only depend on supply shocks. Low global inflation and USD strength will remain headwinds for performance.

4. The context: low growth, high liquidity (but about to decline), volatility risk.

Since the 2008 GFC, global growth has remained stagnant, below potential. The global economy is: 1) in the middle of “a lost decade”, still suffering from the GFC long-lasting effects on credit markets, output, and employment; and 2) undergoing a long-term transformation, as: a) aging populations reduce consumption; b) higher aggregate savings depress interest rates; c) services—less capital-intensive than agriculture and manufacturing—keep rising as shares of global output, employment and value added; and, as a result: d) investment, productivity, demand and trade remain weak, adding to structural unemployment. Global growth decelerated from 4.2 percent (1998-2008 average) to 3.1 percent in 2016, and is expected it at about 3.5 percent (2017-2018 average).

Most CBs loosened monetary policy … To uphold growth after the GFC, CBs slashed policy rates and implemented quantitative easing (QE)[13] programs, pushing short- and long-term interest rates to historical-lows. Whenever the macroeconomic outlook deteriorated and every time the financial markets were stressed by risk-events (be it a US credit downgrade, an ‘earlier-than-expected’ US Federal Reserve (Fed) exit, the sovereign debt crisis in Europe, a sharp drop in oil prices, or China’s devaluation), CBs provided liquidity[14], encouraging risk-taking.

… but failed to lift growth … Cheap money created a culture of complacency: i) policy makers could afford delaying structural reforms; ii) entrepreneurs borrowed to re-finance their debts and back their own shares, not to invest in the real economy. In other words, abundant liquidly – putting creative destruction on hold – kept governments, banks and companies afloat – but brought about zombie-entities, which deliver low growth.

… and boosted market performance. In both DMs and EMs, CB liquidity repressed financial volatility – and boosted asset prices. Supported by monetary policy, financial markets around the world experienced sustained gains. Since March 2009: a) the market capitalization (a proxy of financial markets’ liquidity) of the main global indices almost-tripled: from USD 11.3tn to USD 27.0tn[15]; b) global stock markets rose 73.3 percent[16], reaching record-highs, and c) bond yields dropped to record-lows[17], while prices rose 30.1 percent, mostly in EMs[18]. In other words, as over-reliance on CBs buoyed financial markets, valuations got increasingly disconnected from underlying macro fundamentals.

Policy divergence will create jitteriness: once liquidity starts to decline, corrections are likely. Over the next few years, global CBs will maintain a highly accommodative stance: the upcoming tightening cycle will be well below historical norms and monetary policy normalization will take longer than expected. However, policy-divergence[19] will make CBs less able to act as volatility-repressors. Monetary policy has started to diverge in late 2016, driven by: 1) a very gradual normalization cycle in the Unites States (US)[20]; and 2) further loosening in the EZ and Japan, via additional QE programs and negative interest rates. Over-reliance on CBs can make monetary policy interventions ineffective, and CBs risk making policy mistakes, lose their credibility and create volatility.

We thank Mert Yildiz for comments and suggestions. All errors are ours.



[1] For most asset classes, returns will remain low because of: a) high valuations, compared to historical standards; and b) aging societies, insufficient investment, and weak productivity, which constrain potential growth and corporate earnings (Section 4).

[2]Market correction”: a fall of more than 10 percent from the 52-week high, over a single week. “Bear market”: a fall of more than 20 percent from the 52-week high, over 300 days. “Market crash”: a fall of more than 10 percent in just one day, or 40 percent from the 52-week high over 150 days.

[3] In DMs, yields on safe assets are likely to decline, reflecting expectations of a more gradual pace of monetary policy normalization, a higher global risk aversion and compressed term premia, but yields are already low or negative, and core DM bonds (UST and German Bunds) will benefit only from shocks.

[4] In bond markets, regulators – by forcing institutional investors to invest in triple-A assets, while the supply of these assets has declined by 50 percent – are pushing real interest rate even lower. More than 30 per cent of global government debt is trading at negative nominal yields. Yields on sovereign and corporate bonds, which pay a spread over government debt, have fallen in tandem. Conversely, in equity markets, historically-low real-interest-rates have pushed asset prices high and created asset bubbles, boosting investors’ returns via capital gains. However, as profits grow in line with the economy, investment income (the dividend yield) steadily declined. In 2017, the markets will benefit from almost 1.8 trillion of additional quantitative easing (QE), or 2.19 percent of the worldwide bond market (total debt outstanding), estimated at USD82.2 trillion. In particular, in 2017 the US Federal Reserve (Fed) will roll over (i.e.: no expansion of the balance sheet) maturing securities – purchased with its QE program – for $194bn. In the EZ, the European Central Bank (ECB) QE program will decline from 80bn per month (in January-March, for a total of €240bn) to €60bn per month (in the April-December period, for a total of €540bn) and amount to a yearly total of €780bn (USD816bn). In Japan, where in September 2016 the Bank of Japan (BoJ) abandoned pre-set amounts of monthly purchases and moved to price targeting, i.e. it will buy securities until the 10-year bond yield reaches around 0 percent, annual purchases are likely to remain unchanged from previous levels (¥60 to 70 trillion – USD512 to 615bn). In the UK the Bank of England (BoE) QE will likely amount to £130bn (USD160bn), disaggregated as follows: a) £10bn of Gilts; and b) £120bn of corporate bonds (over the 18 months between September 2016 and February 2018, the BoE will buy £180bn of corporate bonds).

[5] In EMs, RE has a lower regulatory volatility than PE.

[6] Data as of June 20, 2017.

[7] Data as of June 20, 2017. Price over expected earnings in the next 12 months.

[8]  Valuations (as of June 20, 2017) in the US are above historical averages (Dow Jones Industrial Average is trading at 18.2 in its forward estimate, 1.5 standard deviations above its 10-year average of 15.3), and above those in Europe (Eurostoxx 600, forward estimate of 15.5, 0.6 standard deviations above its 10-year average of 13.7). S&P 500 12-month forward P/E, currently at 18.8 multiples, is also above historical averages: 5-year (15.2), 10-year (14.4), 15-year (15.2), and 20-year (17.2). Price to book show a similar trend: Dow Jones trading (as of June 20, 2017) at 3.5 times its book value, while Eurostoxx 600 trades at 1.9 multiples.

[9] With 52.6 percent of the MSCI ACWI, the US stock market represents over half of the world’s market cap.

[10] As of May 31 2017.

[11] As of June 20, 2017.

[12] As of June 12 2017, the cost of the 97-93 per cent put spread on the S&P 500 (which requires the index to fall by between 3 and 7 per cent in a month to be profitable) allows a maximum profit of USD4 for contracts that cost USD0.16, or a 25 times return. Source: FT , 2017.

[13] CBs embarked in open market operations involving the purchase and sale of Treasury securities, a practice referred to as QE. Coupled with low interest rates, QE freed up capital and encouraged in risk appetite amid ultra-supportive monetary policy. As a guaranteed purchaser of assets, CBs gave confidence to bond investors. Bond-buying helped bring down governments’ borrowing costs. Weaker inflation lead to mounting calls for another round of QE. US: the Fed pioneered QE, an unconventional monetary stimulus through unsterilized purchases of Treasury and mortgage-backed securities (MBS). Between 2009 and 2014, the Fed undertook three rounds of QE. The third round was completed in October 2014, at which point the Fed’s balance sheet was USD4.5tn, five times its pre-crisis size. Overall, the Fed’s balance sheet expanded from USD2.2tn (December 2008) to USD4.5tn (December 2014). EZ: In January 2015, the ECB announced a EUR 1.1tn QE scheme under which it would buy EUR 60bn of EZ government bonds each month. Overall, ECB’s balance sheet expanded from USD2.4 (December 2008) to USD2.7tn (December 2014). Japan: in April 2013 the BoJ launched its QE program for purchase of government bonds and expanded it in October 2014, to increase its holdings at an annual pace of JPY 80tn. Overall, BoJ’s balance sheet expanded from USD 1.0 (December 2008) to USD2.5tn (December 2014). UK: Between March and November 2009, the BoE decided to purchase Pound Sterling (GBP) 200 billion of financial assets (mostly UK Government debt or ‘gilts’).  Since then the CB has decided on further purchases:  GBP 75bn in October 2011; GBP 50bn in February 2012 and GBP 50bn in July 2012, bringing total assets purchases to GBP 375 bn. Overall, BoE’s balance sheet expanded from USD 0.3 (December 2008) to USD 1.2tn (December 2014). Brazil: Banco Central Do Brasil (BCB) expanded its balance sheet from USD0.4 (December 2008) to USD0.8tn (December 2014). Russia: Central Bank of Russia (CBR)’s balance sheet expanded from USD15.8 (December 2008) to USD23.7bn (December 2014). India: Reserve Bank of India (RBI)’s balance sheet expanded from USD0.1 (December 2008) to USD0.2tn (December 2014). China: between January and Aug 2015, the PBoC induced further liquidity in the markets by cutting its bank reserve requirements, from 20 to 18 percent. Overall, PBOC’s balance sheet expanded from USD3.4 (December 2008) to USD5.5tn (December 2014). Across the world, money-printing brought about currency depreciation, or outright devaluation (via administrative decision – mostly in Asia so far). QE programs by the ECB and Bank of Japan (BoJ) have played a crucial role in depreciation of Euro (EUR) and Japanese Yen (JPY) respectively. Source: Trading Economics, 2015.

[14] The world’s main CBs have expanded their balance sheets significantly: in the US, from USD 2.2tn (December 2008) to 4.4tn (June 2017); in the EZ, from USD 2.4tn (December 2008) to USD 4.2tn (June 2017); in Japan, from USD 1.0tn (December 2008) to USD 5.0tn (January 2016); in the UK, from USD 0.3tn (December 2008) to USD 1.2tn (December 2014); in Switzerland, USD 0.2tn to USD 0.6tn (December 2015); in Brazil, from USD 0.4tn (December, 2008) to USD 0.7tn (November 2015); in Russia, from USD 15.8bn (December 2008) to USD 22.3bn (January 2016); in India, from USD 0.1tn (December 2008) to USD 0.2tn (December 2015); in China, from USD 3.4tn (December 2008) to USD 4.8tn (December 2015). Source: Trading Economics, 2016.

[15] Between March 1, 2009 and June 20, 2017, total market capitalization rose in the US (DJIA Index): from USD 2.2tn to USD 6.1tn; in the EZ (Eurostoxx 600): from USD 5.0 to USD 11.5tn; in Japan (Nikkei 225): from USD 1.6 to USD 3.2tn; in Brazil (IBOVESPA): from USD 0.35 to USD 0.7tn; in Russia (MICEX): from USD 0.2 to USD 0.5tn; in India (NIFTY): from USD 0.3 to USD 1.1tn; in China (Shanghai Composite): from USD 1.6 to USD 3.9tn. Source: Eikon; for Brazil, Bovespa, 2017.

[16] Growth took place mostly in DMs—US 125.2 percent, Japan 93.7 percent, EU 53.4 percent—but also in EMs—India 164.5 percent, Russia 80.1 percent, Brazil 46.2 percent, China 37.2 percent. Source: Thomson Reuters, as of May 8 2017. World: Dow Jones Global Index; US: Dow Jones Industrial Average (DJIA); EZ: Eurostoxx 600 ; Japan: Nikkei 225; Brazil: IBOVESPA; Russia: MICEX; India: Nifty; China: Shanghai Composite.

[17] Between March 2009 and June 2017, US treasuries (UST) declined from 2.65 to 2.21; 10-year German Bunds from 3.00 to –0.19%; 10-year Japanese bonds from 1.06 percent to -0.22 percent.

[18] In DMs bond prices rose: 25.4 percent in the US, 35.3 percent in Germany, and 17.9% in Japan. In EMs bond prices rose: 156.4 percent in Brazil, 83.9 percent in India, 49.3 percent in Russia, and 31.0 percent in China. Source for global prices: Bonds – World: JPM Global Aggregate Bond Index. Data as of May 8 2017.

[19] In mid-December 2015, the Fed decoupled from other CBs by embarking on an announced interest rate hiking cycle, with “four rate-increases in 2016, reaching a federal funds rate of 1.35 percent by 2017”. Instead, the Fed hiked only once in 2016, and twice in 2017 (March and June). The tightening cycle is proving slower than announced and well below historical norms: over the last 30 years, tightening cycles lasted, on average, 15 months, with an average increase of 2.3 percentage points in the policy rate. The current one has lasted for 19 months, with only one percentage point increase. Meanwhile, the European Central Bank (ECB) and Bank of Japan (BoJ) will continue to ease by combining quantitative easing (QE) with negative rates. In June 2014, the ECB became the first major central bank to venture below zero and since December 2015 it charges banks 0.3 per cent to hold their cash overnight. In January 2016, the ECB announced “unlimited” action to boost EZEZ inflation. In January 2016, BoJ adopted negative interest rates (- 0.1 per cent). In February 2016, Sweden’s central bank (Riksbank) cut its main repo rate to – 0.50 percent, down from – 0.35 percent. Until 2018, CB in Europe and Japan will continue their asset purchases programs.

[20] The Fed has started tightening, with: i) the first a rate increase in almost a decade (25 basis points – bps) in Q4-2016; ii) two 25bp hikes in H1-2017; iii) another likely in H2-2017; and iv) if macro data are strong enough, an announced reduction in the size of its balance sheet.