BCA Indicators/Model

The performance of European stocks relative to the U.S. has been dismal in the post-Lehman period. However, the Eurozone economy is performing impressively, profit growth is accelerating and margins are rising. This points to a period of outperformance for Eurozone stocks, at least in local currency terms. Standard valuation measures based on index data suggest that Eurozone stocks are cheap to the U.S. Nonetheless, the European market almost always trades at a discount, due to persistent lackluster profit performance. In Part II of our series on valuation, we approach the issue from a bottom-up perspective, utilizing the powerful analytics provided by BCA's exciting new Equity Trading Strategy (ETS) platform. The ETS software allows us to compare U.S. and European companies on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach avoids the problems of index construction. Investors can be confident that they will make money on a 12-month horizon by taking a position when the new bottom-up indicator reaches +/-1 standard deviations over- or under-valued, although technical information should be taken on board to sharpen the timing. The +/-2 sigma level gives clear buy/sell signals irrespective of fundamental or technical factors. Valuation alone does not justify overweight Eurozone positions at the moment, although we like the market for other reasons. The bottom-up valuation indicator will not replace our top-down version that is based on index data, but rather will be considered together when evaluating relative value. Total returns in the European equity market have bounced relative to the U.S. since 2016 in both local-currency and common currency terms (Chart II-1). However, this has offset only a tiny fraction of the dismal underperformance since 2007. In local currencies, the relative EMU/U.S. total return index is still close to its lowest level since the late 1970s. Compared with the pre-Lehman peak, the U.S. total return index is more than 96% higher according to Datastream data, while the Eurozone total return index is only now getting back to the previous high-water mark when expressed in U.S. dollars (Chart II-2). Chart II-1EMU Stocks Lag Massively... EMU Stocks Lag Massively... EMU Stocks Lag Massively... Chart II-2...Due To Depressed Earnings ...Due To Depressed Earnings ...Due To Depressed Earnings The yawning return gap between the two equity markets was almost entirely due to earnings as market multiples have moved largely in sync. Earnings-per-share (EPS) generated by U.S. companies now exceed the pre-Lehman peak by about 19%. In contrast, earnings produced by their Eurozone peers are a whopping 48% below their peak (common currency). This reflects both a slower recovery in sales-per-share growth and lower profit margins. Operating margins in Europe have been on the upswing for a year, but are still depressed by pre-Lehman standards. Margin outperformance in the U.S. is not a sector weighting story; in only 2 of 10 sectors do European operating margins exceed the U.S. The return-on-equity data tell a similar story. Nonetheless, a turning point may be at hand. Chart II-3Europe Trades At A Discount Europe Trades At A Discount Europe Trades At A Discount The Eurozone economy has been performing well, especially on a per-capita basis, and forward-looking indicators suggest that growth will remain above-trend for at least the next few quarters. U.S. profit margins have also been (temporarily) rising, but the Eurozone economy has more room to grow because there is still slack in the labor market. There is also more room for margins to rise in the Eurozone corporate sector than is the case in the U.S., where the profit cycle is further advanced. Traditional measures of value based on the MSCI indexes suggest that European stocks are on the cheap side. But are they really that cheap? Based on index data, Eurozone stocks trade at a hefty discount across most of the main valuation measures (Chart II-3). This is the case even for normalized measures such as price-to-book (P/B). However, Eurozone stocks have almost always traded at a discount. There are many possible explanations as to why there is a persistent valuation gap between these two markets, including differences in accounting standards, discount rates and sector weights. The wider use of stock buybacks in the U.S. also favors American stock valuations relative to Europe. But most important are historical differences in underlying corporate fundamentals. U.S. companies on the whole were significantly more profitable even before the Great Financial Crisis (Chart II-3). U.S. companies also tend to have lower leverage and higher interest coverage. Better profitability metrics in the U.S. are not solely an artifact of sector weighting either. RoE and operating margins are lower in Europe even applying U.S. sector weights to the European market.1 Why corporate Europe has been a perennial profit under-achiever is beyond the scope of this paper. U.S. companies reaped most of the benefit from productivity gains over the past 25 years, with the result that the capital share of income soared while the labor share collapsed. European companies were less successful in squeezing down labor costs. Measuring Value In the first part of our two-part Special Report on valuation, published in July 2016, we took a top-down approach to determine whether Eurozone stocks are cheap versus the U.S. after adjusting for different sector weights and persistent differences in the underlying profit fundamentals. A regression approach that factored in various profitability measures performed reasonably well, but the top-down "mechanical" approach that relied on a 5-year moving average provided the most profitable buy/sell signals historically. We approach the issue from a bottom-up perspective in Part II of our series, utilizing the powerful analytics provided by BCA's exciting new Equity Trading Strategy (ETS) platform. The software allows us to compare U.S. and European companies on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach avoids the problems of index construction when trying to gauge valuation across countries. The web-based platform uses over 24 quantitative factors to rank approximately 10,000 individual stocks in 23 countries, allowing clients to find stocks with winning characteristics at the global level. Users can rank and score individual equities to support a broad set of investment strategies and apply macro and sector views to single-name investments. The ETS approach has an impressive track record. Historically, the top-decile of stocks ranked using the "BCA Score" methodology have outperformed stocks in the bottom decile by over 25% a year.2 The BCA Score includes all 24 factors when ranking stocks, but we are interested in developing a valuation metric that provides valued added on its own and is at least as good as the top-down index-based measure developed in Part I. The five valuation measures in the ETS database are trailing P/E, forward P/E, price-to-book, price-to-sales and price-to-cash flow. We combine all of the Eurozone and U.S. companies that have total assets of greater than $1 billion into one dataset. The ETS platform then ranks the stocks from best to worst on a daily basis (i.e. cheapest to most expensive), using an equally-weighted average of the five valuation measures. The average score for U.S. stocks is subtracted from the average score for European stocks, and then divided by the standard deviation of the series. This provides a valuation metric that fluctuates roughly between +/- 2 standard deviations. Chart II-4 presents the resulting bottom-up indicator, along with our previously-published top-down valuation measure. A high reading indicates that European stocks are cheap to the U.S., while it is the opposite for low readings. Chart II-4Eurozone Equity Relative Valuation Indicators Eurozone Equity Relative Valuation Indicators Eurozone Equity Relative Valuation Indicators The underlying bottom-up data extend back to 2000. However, the bursting of the tech bubble in the early 2000's causes major shifts in relative valuation among sectors and between the U.S. and Eurozone that skew the indicator when constructed using the entire data set. We obtain a cleaner indicator when using only the data from 2005. As with any valuation indicator, it is only useful when it reaches extremes. We calculated the historical track record for a trading rule that is based on critical levels of over- and under-valuation. For example, we calculated the (local currency) excess returns over 3, 6, 12 and 24-month horizon generated by (1) overweighting European stocks when that market was one and two standard deviations cheap versus the U.S. market, and (2) overweighting the U.S. when the European market was one and two standard deviations expensive (Table II-1). Table II-1Value Indicator: Trading Rule Returns And Batting Average August 2017 August 2017 The trading rule returns were best when the indicator reached two standard deviations cheap or expensive, providing average returns of almost 11 percent over 12 months. The trading rule returns when the indicator reached +/-1 standard deviation were not as good, but still more than 3% on 12- and 24-month horizons. Table II-1 also presents the trading rule's batting average. That is, the number of positive excess returns generated by the trading rule as a percent of the total number of signals. The batting average ranged from 50% on a 3-month horizon to 68% over 24 months when buy/sell signals are triggered at +/- 1 standard deviation. The batting average is much higher (80-100%) using +/- 2 standard deviations as a trigger point, although there were only five months over the entire sample when the indicator reached this level. The charts and tables in the Appendix present the results of the same analysis at the sector level. The results are equally as good as the aggregate valuation indicator, with a couple of exceptions. European stocks are cheap to the U.S. in the Energy, Financials, and Utilities sectors, while U.S. stocks offer better value in Consumer Discretionary, Consumer Staples, Health Care, Industrials and Technology. Materials, Real Estate, and Telecommunications are close to equally valued. Sharpening The Buy/Sell Signals We then augmented the valuation analysis by adding information on company fundamentals, such as EPS growth and profit margins among others. The ETS software ranked the companies after equally-weighting the valuation and fundamental factors. However, this approach yielded poor results in terms of the trading rule. This is because, for example, when European stocks reach undervalued levels relative to the U.S., it is usually because the European earnings fundamentals have underperformed those of the U.S. companies. Thus, favorable value is offset by poor fundamentals, muddying the message provided by valuation alone. In contrast, adding some information from the technical factors in the ETS model does add value, at least when using +/-1 standard deviations as the trigger point for trades (Chart II-5). Excess returns to the trading rule rise significantly when the medium-term momentum and long-term mean reversion factors are included in the valuation indicator (Table II-2). The batting average also improves. Chart II-5Indicators: Value And Value With Technical Information Indicators: Value And Value With Technical Information Indicators: Value And Value With Technical Information Table II-2Value And Technical Indicator: Trading Rule Returns And Batting Average August 2017 August 2017 Adding technical information does not improve the trading rule performance when +/-2 sigma is used as the trigger point. Investment Conclusions Our new ETS platform provides investors with a unique way of picking stocks by combining top-down macro themes with company-specific information. It also allows us to develop valuation tools that avoid some of the pitfalls of index data by comparing stocks on a head-to-head basis. Historical analysis using a trading rule demonstrates that the new bottom-up valuation indicator provides real value to investors. We would normally evaluate its track record using stretching analysis, where we use only the historical information available at each point in time when determining relative value. However, the relatively short history of the available data precludes this test because we need at least a few cycles to best gauge the underlying volatility in the data. Still, investors can be fairly confident that they will make money on a 12-month horizon by taking a position when the bottom-up indicator reaches +/-1 sigma over- or under-valued, although technical information should be taken on board to sharpen the timing. The +/-2 sigma level gives clear buy/sell signals irrespective of the fundamental or technical factors. The bottom-up valuation indicator will not replace our top-down version that is based on index data, but rather will be considered together when evaluating relative value. At the moment, the top-down version proposes that European stocks are somewhat cheap to the U.S., while the bottom-up indicator points to slight overvaluation. Considering the two together suggests that valuation is close enough to fair value that investors cannot make the decision on value alone. Valuation indicators need to be near extremes to be informative. Our global equity strategists recommend overweighting Eurozone stocks versus the U.S. at the moment, although not because of valuation. Rather, the Eurozone economy and corporate earnings have more room to grow because of lingering labor market slack. This also means that the ECB can keep rates glued to the zero bound for at least the next 18 months while the Fed hikes, which will place upward pressure on the dollar and downward pressure on the euro. Mark McClellan Senior Vice President The Bank Credit Analyst Appendix: Trading Rule Returns By Sector Chart II-6, Chart II-7, Chart II-8, Chart II-9, Chart II-10, Chart II-11, Chart II-12, Chart II-13, Chart II-14, Chart II-15, Chart II-16. Chart II-6Consumer Discretionary Consumer Discretionary Consumer Discretionary Chart II-7Consumer Staples Consumer Staples Consumer Staples Chart II-8Energy Energy Energy Chart II-9Financials Financials Financials Chart II-10Health Care Health Care Health Care Chart II-11Industrials Industrials Industrials Chart II-12Materials Materials Materials Chart II-13Real Estate Real Estate Real Estate Chart II-14Utilities Utilities Utilities Chart II-15Technology Technology Technology Chart II-16Telecommunication Telecommunication Telecommunication 1 Please see The Bank Credit Analyst Special Report, "Are Eurozone Stocks Really That Cheap?" July 2016, available at bca.bcaresearch.com. 2 Please see Equity Trading Strategy Special Report, "Introducing ETS: A Top Down Approach to Bottom-Up Stock Picking," December 2, 2015, available at ets.bcaresearch.com.

Highlights Major central banks outside the U.S. have fired a warning shot across the bow of global bond markets by signaling that "emergency" levels of monetary accommodation are no longer required. Pipeline inflation pressures have yet to show up at the consumer price level outside of the U.K. Most central bankers argue that temporary factors are to blame, but longer-lasting forces could be at work. There are numerous examples of deflationary pressure driven by waves of innovation, cost cutting and changing business models. However, this is not confirmed in the productivity data. Productivity is dismally low and we do not believe it is due to mismeasurement. The Phillips curve is not dead. We expect that inflation will firm by enough to allow central banks to continue scaling back monetary stimulus. The real fed funds rate is not far from the neutral short-term rate, but it is still well below the Fed's estimate of the long-run neutral rate. Market expectations for the Fed are far too complacent; keep duration short. The failure to repeal Obamacare could actually increase the motivation of Republicans to move forward on tax cuts. Expansionary fiscal policy would make life more difficult for the FOMC, given that unemployment is on course to reach the lowest level since 2000. This would force the Fed to act more aggressively, possibly triggering a recession in 2019. The peak Fed/ECB policy divergence is not behind us, implying that recent dollar weakness will reverse. However, the next dollar upleg has been delayed. Fading market hopes for U.S. fiscal stimulus this year have not weighed on equities, in part because of a solid earnings backdrop. Global EPS growth continues to accelerate in line with the recovery in industrial production. In the U.S., results so far suggest that Q2 will see another quarter of margin expansion. Overall earnings growth should peak above our 20% target later this year. It will be tougher sledding in the equity market once profit growth peaks in the U.S. because of poor valuation. Expect to downgrade stocks in the first half of 2018. Corporate bonds are also benefiting from the robust profit backdrop. Balance sheet health continues to deteriorate, but the spark is missing for a sustained corporate bond spread widening. Feature Chart I-1Sell-Off In Global Bond Markets ##br##Triggered By Central Bank Talk Sell-Off In Global Bond Markets Triggered By Central Bank Talk Sell-Off In Global Bond Markets Triggered By Central Bank Talk Major central banks outside the U.S. fired a warning shot across the bow of global bond markets by signaling a recalibration of monetary policy at the ECB's Forum on Central Banking in late June (Chart I-1). The heads of the Bank of England (BoE), Bank of Canada (BoC) and Swedish Riksbank all took a less dovish tone, warning that the diminished threat of deflation has reduced the need for ultra-stimulative policies. The BoC quickly followed up in July with a rate hike and a warning of more to come. The central bank now expects the economy to reach full employment and hit the inflation target by mid-2018, much earlier than previously expected. The Riksbank also backed away from its easing bias at its most recent policy meeting. The ECB's shift in stance was evident even before its Forum meeting, when President Draghi gave a glowing description of the underlying strength of the Euro Area economy. The labor market is about two percentage points closer to full employment than the U.S. was just before the infamous 2013 Taper Tantrum.1 European core inflation is admittedly below target today, but so was the U.S. rate leading up to the 2013 Tantrum. We have not forgotten about Europe's structural problems or the inherent contradictions of the single currency. Banks are still laden with bad debt (although the recapitalization of Italian banks has gone well so far). Nonetheless, from a cyclical economic standpoint, solid momentum this year will allow Draghi to scale back the ECB's ultra-accommodative monetary stance by tapering its asset purchase program early in 2018. The message that "emergency" levels of monetary accommodation are no longer needed is confirmed by our Central Bank (CB) Monitors, which measure pressure on central bankers to raise or lower interest rates (Chart I-2). The Monitors became less useful when rates hit the zero bound and quantitative easing was the only game in town, but they are becoming relevant again as more policymakers consider their exit strategy. All of our CB Monitors are currently in "tighter policy required" territory except for Japan and the Eurozone (although even those are close to the zero line). The Monitors have been rising due to both their growth and underlying inflation components. Another tick higher in PMI's for the advanced economies in July underscored that the rebound in industrial production is continuing (Chart I-3). Our short-term forecasting models, which include both hard and soft data, point to stronger growth in the major countries in the second half of 2017 (Chart I-4). Chart I-2Most In The "Tighter Policy Required" Zone Most In The "Tighter Policy Required" Zone Most In The "Tighter Policy Required" Zone Chart I-3Industrial Production Recovery Is Intact Industrial Production Recovery Is Intact Industrial Production Recovery Is Intact On the inflation side, our pipeline indicators have all signaled a modest building of underlying inflation pressure over the past year (although they have softened recently in the U.S. and Eurozone; Chart I-5). In terms of the components of these indicators, rising core producer price inflation has been partly offset by slower gains in unit labor costs in some economies. Chart I-4Our Short-Term Growth Models Are Bullish Our Short-Term Growth Models Are Bullish Our Short-Term Growth Models Are Bullish Chart I-5Some Rise In Pipeline Inflation Pressure Some Rise In Pipeline Inflation Pressure Some Rise In Pipeline Inflation Pressure These pipeline pressures have yet to show up at the consumer level. Most central bankers argue that temporary special factors are to blame, but many investors are wondering if longer-lasting forces are at work. There are numerous examples of deflationary pressure driven by waves of innovation, cost cutting and changing business models. Amazon, Uber, robotics and shale oil production are just a few examples. If this is the main story, then the inability for central banks to reach their inflation targets is a "good thing" because it reflects the adaptation of game-changing new technology. There is no doubt that important strides are being made in certain areas where new technologies are clearly driving prices down. The problem is that, at the macro level, it is not showing up in the productivity data. Productivity is dismally low across the major countries and we do not believe it is simply due to mismeasurement. A Special Report from BCA's Global Investment Strategy2 service makes a convincing case that mismeasurement is not behind the low productivity figures. In fact, it appears that productivity is over-estimated in some industries. It is also important to keep in mind that technological change is nothing new. There is a vigorous debate in academic circles on whether today's new technologies are anywhere near as positive as previous ones like indoor plumbing, electricity, the internal combustion engine and the internet. We are wowed by today's new gizmos, but they are not as transformative as previous innovations. While productivity is surging in some high-profile firms, studies show that there is a long tail of low-productivity companies that drag down the average. A full discussion is beyond the scope of this report and more research needs to be done, but we are not of the view that technology and productivity preclude rising inflation. We expect that inflation will firm by enough to allow central banks to continue scaling back monetary stimulus in the coming months and quarters. Did Yellen Turn Dovish? As with other central banks, the consensus among Fed policymakers is willing to "look through" low inflation for now. Yellen's Congressional testimony did not deviate from that view, although investors interpreted her remarks as dovish. The financial press focused on her statement that "...the policy rate is not far from neutral." However, this was followed up by the statement that "...because we also anticipate that the factors that are currently holding down the neutral rate will diminish somewhat over time, additional gradual rate hikes are likely to be appropriate over the next few years to sustain the economic expansion and return inflation to our 2 percent goal." Chart I-6Bond Market Does Not Believe The Fed Bond Market Does Not Believe The Fed Bond Market Does Not Believe The Fed The Fed believes there are two neutral interest rates: short-term and long-term. Yellen argued that the actual policy rate is currently close to the short-term neutral level, which is depressed by economic headwinds. However, Yellen and others have made the case that the short-term neutral rate is trending up as headwinds diminish, and will converge with the long-term neutral rate over time. The Fed's Summary of Economic Projections reveals what the FOMC thinks is the neutral long-term real fed funds rate; the median forecast calls for a nominal fed funds rate of 2.9% at the end of 2019 and 3% in the longer run. Incorporating a 2% inflation target, we can infer that the Fed anticipates a real neutral rate of 1% in the longer run. The Fed is likely tracking the real neutral fed funds rate using an estimate created by Laubach and Williams (LW).3 Chart I-6 shows this estimate of the neutral rate, called R-star, alongside the real federal funds rate that is calculated using 12-month trailing core PCE. The resulting real fed funds rate has risen sharply during the past seven months due to both three Fed rate hikes and a decline in inflation. If the Fed lifts rates once more this year and core inflation stays put, then the real fed funds rate would end 2017 close to zero, only 42 bps below neutral. However, it's more likely that the Fed will need to see inflation rebound before it delivers another rate hike. In a scenario where core inflation rises to 1.9% and the Fed lifts rates once more, then the real fed funds rate would actually decline between now and the end of the year. The implication is that the real fed funds rate is not far from R-star, but the nominal rate will have to rise a long way before the real rate reaches the Fed's estimate of the long-term neutral rate. Investors simply don't believe Fed policymakers. According to the bond market, the real fed funds rate will not shift into positive territory until 2021 (see real forward OIS line in Chart I-6). We think this is far too complacent. U.S. Health Care Reform: RIP The speed at which short-term rates converge with the long-run neutral rate will depend importantly on the path of fiscal policy. The Republicans' failure to pass their health care legislation is leading the investors to doubt the prospect for (stimulative) tax cuts. This may be premature. Ironically, the failure to jettison Obamacare may turn out to be a blessing in disguise for President Trump and the Republican Party. According to the Congressional Budget Office, the proposed legislation would have caused 22 million fewer Americans to have health insurance in 2026 compared with the status quo. The Senate bill would have also led to substantial cuts to Medicaid relative to existing law, as well as deep cuts to insurance subsidies for many poor and middle-class families. Many of these voters came out in support of Trump last year. The failure to repeal Obamacare could actually increase the motivation of Republicans to move forward on tax cuts anyway. The chances for broad tax reform have certainly diminished, since that will be just as difficult to get passed as healthcare reform. The GOP also wanted to use the roughly $200 billion in savings from healthcare reform to fund reduced tax rates. However, tax cuts are something that all Republicans can easily agree too, and they will need to show a legislative victory ahead of next year's mid-term elections. The difficulty will be how to pay for these cuts. We expect them to be "fully funded" in the sense that there will be offsetting spending cuts, but these will be back-loaded toward the end of the 10-year budget window, whereas the tax cuts will be front-loaded. This would generate a modest amount of fiscal stimulus over the next few years. Sub-4% U.S. Unemployment Rate Followed By Recession? Chart I-7Inside The Fed's Forecasts Inside The Fed's Forecasts Inside The Fed's Forecasts Expansionary fiscal policy would make life more difficult for the FOMC, which may have already fallen behind the curve. The unemployment rate is below the Fed's estimate of the full employment level, and it will continue to erode unless productivity picks up soon. We backed out the productivity growth rate implied by the Fed's latest Summary of Economic Projections, given its assumption that real GDP growth will be roughly 2% over the next couple of years and that the unemployment rate will stabilize near the current level. This combination implies that productivity growth will accelerate from the average rate observed so far in this expansion (0.7%) to about 1%, which is consistent with monthly payrolls of 135,000 assuming real GDP growth of 2% (Chart I-7). If we instead assume that productivity does not accelerate (and real GDP growth is 2%), then payrolls must jump to 160,000 and the unemployment rate would fall below 4% next year. The implication is that the unemployment rate is likely to soon reach levels not seen since 2000, which would force the FOMC to tighten more aggressively. The Fed would hope for a soft landing as it tries to nudge the unemployment rate higher, but the more likely result is a recession in 2019. For this year, we expect the Fed to begin balance sheet runoff in the autumn, followed by a rate hike in December. The latter hinges importantly on at least a modest rise in core PCE inflation in the coming months. A rebound in oil prices would help the Fed reach its inflation goal, even though energy prices affect the headline by more than the core rate. Saudi Energy Minister Khalid al-Falih indicated at a recent press conference in St. Petersburg that no changes are presently needed to the production deal under which OPEC and non-OPEC producers pledged to remove 1.8mn b/d from the market. The Saudi energy minister's remarks leave open the possibility of deeper cuts later this year if global inventories do not draw fast enough, or for the cuts to be extended beyond March 2018 if officials are not satisfied with progress on the storage front. We still believe they are capable of meeting this goal, despite rising shale production. Chart I-8Forecast Of Oil Inventories Forecast Of Oil Inventories Forecast Of Oil Inventories Our commodity strategists expect OECD oil inventories to reach their five-year average level by year-end or early 2018 Q1 (Chart I-8). In the absence of additional cuts, the five-year average level of OECD inventories will be higher than we estimated earlier this year, indicating that our expectation for the overall inventory drawdown later this year has been trimmed. Still, our oil strategists believe the inventory drawdowns will be sufficient to push WTI above the mid-$50s by year-end. If this forecast pans out, rising oil prices will push up headline inflation and inflation expectations in the major advanced economies. The bottom line is that the backdrop has turned bond-bearish now that central bankers in the advanced economies are in the process of scaling back the easier monetary policy that followed the deflationary 2014/15 oil shock. Duration should be kept short within global fixed income portfolios. In terms of country allocation, our global fixed income strategists have downgraded the Eurozone government bond market to underweight, joining the Treasury allocation, in light of the pending ECB tapering announcement that could place more upward pressure on yields. This was offset by upgrading Japan to maximum overweight. Max Policy Divergence Has Not Been Reached Chart I-9Europe Has A Lower Neutral Rate Europe Has A Lower Neutral Rate Europe Has A Lower Neutral Rate The change in tone by central bankers outside the U.S. has weighted heavily on the U.S. dollar. The Canadian dollar and the Euro have been particularly strong. Investors have apparently decided that the peak Fed/ECB policy divergence is now behind us. We do not agree. The ECB may be tapering, but rate hikes are a long way off because there remains a substantial amount of economic slack in the Eurozone. Laubach and Williams estimate R-star in the Eurozone to be close to zero, which is 50 basis points below the U.S. neutral rate (Chart I-9). The difference is related to slower potential growth and greater unemployment. Labor market slack across the euro area as a whole is still 3.2 percentage points higher than in 2008, and 6.7 points higher outside of Germany. The current real short-term rate is about -1%. We expect U.S. R-star to rise in absolute terms and relative to the neutral rate in the Eurozone because the U.S. is further advanced in the economic expansion. As Fed rate hike expectations ratchet up in the coming months, interest rate differentials versus Europe will widen in favor of the dollar. It is the same story for the dollar/yen rate because the Bank of Japan is a long way from raising or abandoning its 10-year bond yield peg. Japanese core inflation has fallen back to zero and medium-to-long-term inflation expectations have dipped so far this year. The annual shunto wage negotiations this summer produced little in the way of salary hikes. The major exception to our "strong dollar" call is the Canadian loonie, which we expect to appreciate versus the greenback. We also like the Aussie dollar, provided that the Chinese economy continues to hold up as we expect. Stocks Get A Free Pass For Now Chart I-10Global EPS And Industrial Production Global EPS And Industrial Production Global EPS And Industrial Production Fading market hopes for U.S. fiscal stimulus have weighed on both U.S. Treasury yields and the dollar, but the equity market has taken the news in stride. Are equity investors simply in denial? We do not think so. The equity market appears to have been given a "free pass" for now because earnings have been supportive. The combination of robust earnings growth, steady real GDP growth of around 2%, and low bond yields has been bullish for stocks so far in this expansion. At the global level, EPS growth continues to accelerate in line with the recovery in industrial production, which is a good proxy for top line growth (Chart I-10). Orders and production for capital goods in the major advanced economies have been particularly strong in recent months. The global operating margin flattened off last month according to IBES data, although margins continued to firm in the U.S. and Europe (Chart I-11). The profit acceleration is widespread across these three economies in the Basic Materials and Consumer Discretionary sectors. Industrials, Energy, Health Care and Consumer Staples are also performing well in most cases. Telecom is the weak spot. Our sector profit diffusion indexes paint an upbeat picture for the near term (Chart I-12). Chart I-11Operating Margins On The Rise Operating Margins On The Rise Operating Margins On The Rise Chart I-12Earnings Diffusion Indexes Are Bullish Earnings Diffusion Indexes Are Bullish Earnings Diffusion Indexes Are Bullish In the U.S., the second quarter earnings season is off to a good start. Results so far suggest that Q2 will see another quarter of margin expansion. We believe that U.S. margins are in a secular decline, but they are in the midst of a counter-trend rally that will last for the rest of this year. Using blended results for the second quarter, trailing S&P 500 EPS growth hit 18½% on a 4-quarter moving total basis (Chart I-13). The acceleration in earnings is impressive even after excluding the Energy sector. We projected early this year that EPS growth would peak at around 20%4 by year end, but it appears that earnings will overshoot that level. Chart I-13Robust EPS Growth Even Without Energy Robust EPS Growth Even Without Energy Robust EPS Growth Even Without Energy It will be tougher sledding in the equity market once profit growth peaks in the U.S. because of poor valuation. We are expecting to scale back our overweight equity recommendation sometime in the first half of 2018, although the global rally could be extended by constructive earnings data in Europe and Japan. The earnings recovery in both economies is behind the U.S., such that peak growth will come later in 2018. There is also more room for margins to expand in Europe than in the U.S. The relative earnings cycle is one of the reasons why we continue to favor Eurozone and Japanese stocks to the U.S. in local currency terms. Japanese stocks are also cheap to the U.S. based on our top-down valuation indicator (Chart I-14). European stocks are not far from fair value relative to the U.S., after adjusting for the fact that Europe trades structurally on the cheap side. The message from our top-down valuation indicator for European stocks is confirmed when using the bottom-up information contained in the new BCA Equity Trading Strategy platform. The Special Report beginning on page 20 describes a bottom-up valuation measure that we will use in conjunction with our top-down (index-based) measures. Corporate Bonds: Kindling And Sparks Healthy EPS growth momentum is also constructive for corporate bonds, although overall balance sheet health continues to erode in the U.S. The release of the U.S. Flow of Funds data allows us to update BCA's Corporate Health Monitor (CHM) for the first quarter (Chart I-15). The level of the CHM moved slightly deeper into "deteriorating health territory." Chart I-14Top-Down Relative Equity Valuation Top-Down Relative Equity Valuation Top-Down Relative Equity Valuation Chart I-15Deteriorating Since 2015, But... Deteriorating Since 2015, But... Deteriorating Since 2015, But... The Monitor has been a reliable indicator for the trend in corporate bond spreads over the years, calling almost all major turning points in advance. However, spreads have trended tighter over the past year even as the CHM began to signal deteriorating health in early 2015. Why the divergence? The CHM is only one of three key items on our checklist to underweight corporate bonds versus Treasurys. The other two are tight Fed policy (i.e. real interest rates that are above the neutral level) and the direction of bank lending standards for C&I loans. On its own, balance sheet deterioration only provides the kindling for a spread blowout. It also requires a spark. Investors do not worry about high leverage or a profit margin squeeze, for example, until the outlook for defaults sours. The latter occurs once inflation starts to rise and the Fed actively targets slower growth via higher interest rates. Banks see trouble on the horizon and respond by tightening lending standards, thereby restricting the flow of credit to the business sector. Defaults start to ramp up, buttressing banks' bias to curtail lending in a self-reinforcing negative feedback loop. The three items on the checklist normally occurred at roughly the same time in previous cycles because a deteriorating CHM is typically a late-cycle phenomenon. But this has been a very different cycle. High stock prices and rock-bottom bond yields have encouraged the corporate sector to leverage up and repurchase stock. At the same time, the subpar, stretched-out recovery has meant that it has taken longer than usual for the economy to reach full employment. It will be some time before U.S. short-term interest rates reach restrictive territory. As for banks, they tightened lending standards a little in 2015/16 due to the collapse of energy prices, but this has since reversed. The implication is that, while corporate health has deteriorated, we do not have the spark for a sustained corporate bond spread widening. Indeed, Moody's expects that the 12-month default rate will trend lower over the next year, which is consistent with constructive trends in corporate lending standards, industrial production and job cut announcements (all good indicators for defaults). Chart I-16 presents a valuation metric that adjusts the HY OAS for 12-month trailing default losses (i.e. it is an ex-post measure). In the forecast period, we hold today's OAS constant, but the 12-month default losses are a shifting blend of historical losses and Moody's forecast. The endpoint suggests that the market is offering about 200 basis points of default-adjusted excess yield over the Treasury curve for the next 12 months. This is roughly in line with the mid-point of the historical data. In the past, a default-adjusted spread of around 200 basis points provided positive 12-month excess returns to high-yield bonds 74% of the time, with an average return of 82 basis points. It is also a positive sign for corporate bonds that the net transfer to shareholders, in the form of buybacks, dividends and M&A activity, eased in the fourth quarter 2016 and the first quarter of 2017 (Chart I-17). Ratings migration has also improved (i.e. moderating net downgrades), especially for shareholder-friendly rating action, which is a better indicator for corporate spreads. The diminished appetite to "return cash to shareholders" may not last long, but for now it supports our overweight in both investment- and speculative-grade bonds versus Treasurys. That said, excess returns are likely to be limited to the carry given little room for spread compression. Chart I-16Still Some Value In ##br##High-Yield Corporates Still Some Value In High-Yield Corporates Still Some Value In High-Yield Corporates Chart I-17Net Transfers To Shareholders ##br##Eased In Past Two Quarters Net Transfers To Shareholders Eased In Past Two Quarters Net Transfers To Shareholders Eased In Past Two Quarters Within balanced portfolios, we recommend favoring equities to high-yield at this stage of the cycle. Value is not good enough in HY relative to stocks to expect any sustained period of outperformance in the former, assuming that the bull market in risk assets continues. Investment Conclusions A key change in the global financial landscape over the past month is a signal from central banks that they see the need for policy recalibration. Policymakers view sub-target inflation as temporary, and some are concerned that low interest rates could contribute to the formation of financial market bubbles. The bond market remains skeptical, given persistent inflation undershoots and growing anecdotal evidence that new technologies are very deflationary. It would be extremely bullish for stocks if these new technologies were indeed boosting the supply side of the economy at a faster pace than the official data suggest. Robust advances in output-per-worker would allow profits to grow quickly, and would provide the economy more breathing space before hitting inflationary capacity limits (keeping the bond vigilantes at bay). We acknowledge that there are important technological breakthroughs being made, but we do not see any evidence that this is occurring on a widespread basis sufficient to "move the dial" in terms of overall productivity growth. Indeed, the stagnation of middle class personal income is consistent with a poor productivity backdrop. Chart I-18 highlights that "creative destruction" is in a long-term bear market. Chart I-18Less Creative Destruction Less Creative Destruction Less Creative Destruction That said, the equity market is benefiting from the mini-cycle in corporate profits, which are still recovering from the earnings recession in 2015/early 2016. We expect the recovery to be complete by early 2018, which will set the stage for a substantial slowdown in EPS growth next year. It won't be a disaster, absent a recession, but demanding valuations suggest that the market could struggle to make headway through next year. We expect to trim exposure sometime in the first half of 2018. To time the exit, we will watch for a roll-over in the growth rate of S&P 500 EPS on a 4-quarter moving total basis. Investors should look for a peak in industrial production growth as a warnings sign for profits. We are also watching for a contraction in excess money, which we define as M2 divided by nominal GDP. Finally, a rise in core PCE inflation to 2% would be a signal that the Fed is about to ramp up interest rates. For now, remain overweight equities relative to bonds and cash. Favor equities to high yield, but within fixed-income portfolios, overweight investment- and speculative-grade corporates versus Treasurys. We are comfortable with our pro-risk recommendations and our below-benchmark duration stance. Unfortunately, that can't be said of our bullish U.S. dollar and oil price house views. Both are controversial calls among our strategists. As for oil, supply and demand are finely balanced and our positive view hinges importantly on OPEC agreeing to more production cuts. The obvious risk is that these cuts do not materialize. The dollar call has gone against us as the latest signs of improving global growth momentum have admittedly been outside the U.S. Meanwhile, the U.S. is stuck in a political morass, which delays the prospect of fiscal stimulus. This is not to say that U.S. growth will slow. Rather, the growth acceleration may fall short of the high expectations following last November's election. We continue to believe that the market is too complacent on the pace of Fed rate hikes in the coming quarters. An upward adjustment in rate expectations should push the dollar higher on a trade-weighted basis, as outlined above. Nonetheless, this shift will require higher U.S. inflation, the timing of which is highly uncertain. We remain dollar bulls on a 12-month horizon, but we are stepping aside and calling for a trading range in the next three months. Mark McClellan Senior Vice President The Bank Credit Analyst July 27, 2017 Next Report: August 31, 2017 1 Please see Global Fixed Income Strategy Weekly Report, "Central Banks Are Now Playing Catch-Up," dated July 4, 2017, available at gfis.bcaresearch.com 2 Please see Global Investment Strategy Special Report, "Weak Productivity Growth: Don't Blame The Statisticians," dated March 25, 2016, available at gis.bcaresearch.com 3 Kathryn Holston, Thomas Laubach, and John C. Williams "Measuring The Natural Rates Of Interest: International Trends And Determinants," Federal Reserve Bank of San Francisco, Working Paper 2016-11 (December 2016). 4 Calculated as a year-over-year growth rate of a 4-quarter moving total of S&P data. II. The BCA ETS Trading Platform Approach To Valuing Eurozone Stocks The performance of European stocks relative to the U.S. has been dismal in the post-Lehman period. However, the Eurozone economy is performing impressively, profit growth is accelerating and margins are rising. This points to a period of outperformance for Eurozone stocks, at least in local currency terms. Standard valuation measures based on index data suggest that Eurozone stocks are cheap to the U.S. Nonetheless, the European market almost always trades at a discount, due to persistent lackluster profit performance. In Part II of our series on valuation, we approach the issue from a bottom-up perspective, utilizing the powerful analytics provided by BCA's exciting new Equity Trading Strategy (ETS) platform. The ETS software allows us to compare U.S. and European companies on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach avoids the problems of index construction. Investors can be confident that they will make money on a 12-month horizon by taking a position when the new bottom-up indicator reaches +/-1 standard deviations over- or under-valued, although technical information should be taken on board to sharpen the timing. The +/-2 sigma level gives clear buy/sell signals irrespective of fundamental or technical factors. Valuation alone does not justify overweight Eurozone positions at the moment, although we like the market for other reasons. The bottom-up valuation indicator will not replace our top-down version that is based on index data, but rather will be considered together when evaluating relative value. Total returns in the European equity market have bounced relative to the U.S. since 2016 in both local-currency and common currency terms (Chart II-1). However, this has offset only a tiny fraction of the dismal underperformance since 2007. In local currencies, the relative EMU/U.S. total return index is still close to its lowest level since the late 1970s. Compared with the pre-Lehman peak, the U.S. total return index is more than 96% higher according to Datastream data, while the Eurozone total return index is only now getting back to the previous high-water mark when expressed in U.S. dollars (Chart II-2). Chart II-1EMU Stocks Lag Massively... EMU Stocks Lag Massively... EMU Stocks Lag Massively... Chart II-2...Due To Depressed Earnings ...Due To Depressed Earnings ...Due To Depressed Earnings The yawning return gap between the two equity markets was almost entirely due to earnings as market multiples have moved largely in sync. Earnings-per-share (EPS) generated by U.S. companies now exceed the pre-Lehman peak by about 19%. In contrast, earnings produced by their Eurozone peers are a whopping 48% below their peak (common currency). This reflects both a slower recovery in sales-per-share growth and lower profit margins. Operating margins in Europe have been on the upswing for a year, but are still depressed by pre-Lehman standards. Margin outperformance in the U.S. is not a sector weighting story; in only 2 of 10 sectors do European operating margins exceed the U.S. The return-on-equity data tell a similar story. Nonetheless, a turning point may be at hand. Chart II-3Europe Trades At A Discount Europe Trades At A Discount Europe Trades At A Discount The Eurozone economy has been performing well, especially on a per-capita basis, and forward-looking indicators suggest that growth will remain above-trend for at least the next few quarters. U.S. profit margins have also been (temporarily) rising, but the Eurozone economy has more room to grow because there is still slack in the labor market. There is also more room for margins to rise in the Eurozone corporate sector than is the case in the U.S., where the profit cycle is further advanced. Traditional measures of value based on the MSCI indexes suggest that European stocks are on the cheap side. But are they really that cheap? Based on index data, Eurozone stocks trade at a hefty discount across most of the main valuation measures (Chart II-3). This is the case even for normalized measures such as price-to-book (P/B). However, Eurozone stocks have almost always traded at a discount. There are many possible explanations as to why there is a persistent valuation gap between these two markets, including differences in accounting standards, discount rates and sector weights. The wider use of stock buybacks in the U.S. also favors American stock valuations relative to Europe. But most important are historical differences in underlying corporate fundamentals. U.S. companies on the whole were significantly more profitable even before the Great Financial Crisis (Chart II-3). U.S. companies also tend to have lower leverage and higher interest coverage. Better profitability metrics in the U.S. are not solely an artifact of sector weighting either. RoE and operating margins are lower in Europe even applying U.S. sector weights to the European market.1 Why corporate Europe has been a perennial profit under-achiever is beyond the scope of this paper. U.S. companies reaped most of the benefit from productivity gains over the past 25 years, with the result that the capital share of income soared while the labor share collapsed. European companies were less successful in squeezing down labor costs. Measuring Value In the first part of our two-part Special Report on valuation, published in July 2016, we took a top-down approach to determine whether Eurozone stocks are cheap versus the U.S. after adjusting for different sector weights and persistent differences in the underlying profit fundamentals. A regression approach that factored in various profitability measures performed reasonably well, but the top-down "mechanical" approach that relied on a 5-year moving average provided the most profitable buy/sell signals historically. We approach the issue from a bottom-up perspective in Part II of our series, utilizing the powerful analytics provided by BCA's exciting new Equity Trading Strategy (ETS) platform. The software allows us to compare U.S. and European companies on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach avoids the problems of index construction when trying to gauge valuation across countries. The web-based platform uses over 24 quantitative factors to rank approximately 10,000 individual stocks in 23 countries, allowing clients to find stocks with winning characteristics at the global level. Users can rank and score individual equities to support a broad set of investment strategies and apply macro and sector views to single-name investments. The ETS approach has an impressive track record. Historically, the top-decile of stocks ranked using the "BCA Score" methodology have outperformed stocks in the bottom decile by over 25% a year.2 The BCA Score includes all 24 factors when ranking stocks, but we are interested in developing a valuation metric that provides valued added on its own and is at least as good as the top-down index-based measure developed in Part I. The five valuation measures in the ETS database are trailing P/E, forward P/E, price-to-book, price-to-sales and price-to-cash flow. We combine all of the Eurozone and U.S. companies that have total assets of greater than $1 billion into one dataset. The ETS platform then ranks the stocks from best to worst on a daily basis (i.e. cheapest to most expensive), using an equally-weighted average of the five valuation measures. The average score for U.S. stocks is subtracted from the average score for European stocks, and then divided by the standard deviation of the series. This provides a valuation metric that fluctuates roughly between +/- 2 standard deviations. Chart II-4 presents the resulting bottom-up indicator, along with our previously-published top-down valuation measure. A high reading indicates that European stocks are cheap to the U.S., while it is the opposite for low readings. Chart II-4Eurozone Equity Relative Valuation Indicators Eurozone Equity Relative Valuation Indicators Eurozone Equity Relative Valuation Indicators The underlying bottom-up data extend back to 2000. However, the bursting of the tech bubble in the early 2000's causes major shifts in relative valuation among sectors and between the U.S. and Eurozone that skew the indicator when constructed using the entire data set. We obtain a cleaner indicator when using only the data from 2005. As with any valuation indicator, it is only useful when it reaches extremes. We calculated the historical track record for a trading rule that is based on critical levels of over- and under-valuation. For example, we calculated the (local currency) excess returns over 3, 6, 12 and 24-month horizon generated by (1) overweighting European stocks when that market was one and two standard deviations cheap versus the U.S. market, and (2) overweighting the U.S. when the European market was one and two standard deviations expensive (Table II-1). Table II-1Value Indicator: Trading Rule Returns And Batting Average August 2017 August 2017 The trading rule returns were best when the indicator reached two standard deviations cheap or expensive, providing average returns of almost 11 percent over 12 months. The trading rule returns when the indicator reached +/-1 standard deviation were not as good, but still more than 3% on 12- and 24-month horizons. Table II-1 also presents the trading rule's batting average. That is, the number of positive excess returns generated by the trading rule as a percent of the total number of signals. The batting average ranged from 50% on a 3-month horizon to 68% over 24 months when buy/sell signals are triggered at +/- 1 standard deviation. The batting average is much higher (80-100%) using +/- 2 standard deviations as a trigger point, although there were only five months over the entire sample when the indicator reached this level. The charts and tables in the Appendix present the results of the same analysis at the sector level. The results are equally as good as the aggregate valuation indicator, with a couple of exceptions. European stocks are cheap to the U.S. in the Energy, Financials, and Utilities sectors, while U.S. stocks offer better value in Consumer Discretionary, Consumer Staples, Health Care, Industrials and Technology. Materials, Real Estate, and Telecommunications are close to equally valued. Sharpening The Buy/Sell Signals We then augmented the valuation analysis by adding information on company fundamentals, such as EPS growth and profit margins among others. The ETS software ranked the companies after equally-weighting the valuation and fundamental factors. However, this approach yielded poor results in terms of the trading rule. This is because, for example, when European stocks reach undervalued levels relative to the U.S., it is usually because the European earnings fundamentals have underperformed those of the U.S. companies. Thus, favorable value is offset by poor fundamentals, muddying the message provided by valuation alone. In contrast, adding some information from the technical factors in the ETS model does add value, at least when using +/-1 standard deviations as the trigger point for trades (Chart II-5). Excess returns to the trading rule rise significantly when the medium-term momentum and long-term mean reversion factors are included in the valuation indicator (Table II-2). The batting average also improves. Chart II-5Indicators: Value And Value With Technical Information Indicators: Value And Value With Technical Information Indicators: Value And Value With Technical Information Table II-2Value And Technical Indicator: Trading Rule Returns And Batting Average August 2017 August 2017 Adding technical information does not improve the trading rule performance when +/-2 sigma is used as the trigger point. Investment Conclusions Our new ETS platform provides investors with a unique way of picking stocks by combining top-down macro themes with company-specific information. It also allows us to develop valuation tools that avoid some of the pitfalls of index data by comparing stocks on a head-to-head basis. Historical analysis using a trading rule demonstrates that the new bottom-up valuation indicator provides real value to investors. We would normally evaluate its track record using stretching analysis, where we use only the historical information available at each point in time when determining relative value. However, the relatively short history of the available data precludes this test because we need at least a few cycles to best gauge the underlying volatility in the data. Still, investors can be fairly confident that they will make money on a 12-month horizon by taking a position when the bottom-up indicator reaches +/-1 sigma over- or under-valued, although technical information should be taken on board to sharpen the timing. The +/-2 sigma level gives clear buy/sell signals irrespective of the fundamental or technical factors. The bottom-up valuation indicator will not replace our top-down version that is based on index data, but rather will be considered together when evaluating relative value. At the moment, the top-down version proposes that European stocks are somewhat cheap to the U.S., while the bottom-up indicator points to slight overvaluation. Considering the two together suggests that valuation is close enough to fair value that investors cannot make the decision on value alone. Valuation indicators need to be near extremes to be informative. Our global equity strategists recommend overweighting Eurozone stocks versus the U.S. at the moment, although not because of valuation. Rather, the Eurozone economy and corporate earnings have more room to grow because of lingering labor market slack. This also means that the ECB can keep rates glued to the zero bound for at least the next 18 months while the Fed hikes, which will place upward pressure on the dollar and downward pressure on the euro. Mark McClellan Senior Vice President The Bank Credit Analyst Appendix: Trading Rule Returns By Sector Chart II-6, Chart II-7, Chart II-8, Chart II-9, Chart II-10, Chart II-11, Chart II-12, Chart II-13, Chart II-14, Chart II-15, Chart II-16. Chart II-6Consumer Discretionary Consumer Discretionary Consumer Discretionary Chart II-7Consumer Staples Consumer Staples Consumer Staples Chart II-8Energy Energy Energy Chart II-9Financials Financials Financials Chart II-10Health Care Health Care Health Care Chart II-11Industrials Industrials Industrials Chart II-12Materials Materials Materials Chart II-13Real Estate Real Estate Real Estate Chart II-14Utilities Utilities Utilities Chart II-15Technology Technology Technology Chart II-16Telecommunication Telecommunication Telecommunication 1 Please see The Bank Credit Analyst Special Report, "Are Eurozone Stocks Really That Cheap?" July 2016, available at bca.bcaresearch.com. 2 Please see Equity Trading Strategy Special Report, "Introducing ETS: A Top Down Approach to Bottom-Up Stock Picking," December 2, 2015, available at ets.bcaresearch.com. III. Indicators And Reference Charts Stocks continue to outperform bonds against a constructive backdrop of improving global economic prospects and accelerating EPS growth, while low inflation is expected to keep central banks from tightening quickly. Our main equity and asset allocation indicators remain bullish for risk, with a few exceptions. Our new Revealed Preference Indicator (RPI) jumped back to a 100% equity weighting in July. We introduced the RPI in last month's Special Report. Quite simply, it combines the idea of market momentum with valuation and policy measures. It provides a powerful bullish signal if positive market momentum lines up with constructive signals from the policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. Our Willingness-to-Pay (WTP) indicators are also bullish on stocks for the U.S., Europe and Japan. These indicators track flows, and thus provide information on what investors are actually doing, as opposed to sentiment indexes that track how investors are feeling. Investors often say they are bullish but remain conservative in their asset allocation. The U.S. WTP remains bullish, but has topped out, suggesting that flows into the U.S. market are beginning to moderate. In contrast, the WTP indicators for both the Eurozone and Japan are rising from a low level. This suggests that a rotation into these equity markets is underway, although it has not yet shown up in terms of equity market outperformance versus the U.S. On the negative side, our Monetary Indicator last month fell a little further below the zero line and our composite Technical Indicator appears to be rolling over; the latter generates a 'sell' signal when it drops below its 9-month moving average. Value is stretched, but our Valuation Indicator has not yet reached the +1 standard deviation level that indicates clear over-valuation. As highlighted in the Overview section, the U.S. and global earnings backdrop continues to support equity markets. Forward earnings estimates are in a steep uptrend, and the recent surge in the net revisions ratio and the earnings surprise index suggests that EPS growth will remain impressive for the remainder of the year. Bond valuation is largely unchanged from last month, sitting very close to fair value. We still believe that fair value is rising as economic headwinds fade. However, much depends on our forecast that core inflation in the major countries will grind higher in the coming months. Central banks stand ready to "remove the punchbowl" if they get the green light from inflation. The dollar's downdraft in July reduced some of its overvaluation based on purchasing power parity measures. The dollar appears less overvalued based on other measures. Our composite Technical Indicator has fallen hard, but has not reached oversold levels. This suggests that the dollar has more downside before it finds a bottom. EQUITIES: Chart III-1U.S. Equity Indicators U.S. Equity Indicators U.S. Equity Indicators Chart III-2Willingness To Pay For Risk Willingness To Pay For Risk Willingness To Pay For Risk Chart III-3U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators Chart III-4Revealed Preference Indicator Revealed Preference Indicator Revealed Preference Indicator Chart III-5U.S. Stock Market Valuation U.S. Stock Market Valuation U.S. Stock Market Valuation Chart III-6U.S. Earnings U.S. Earnings U.S. Earnings Chart III-7Global Stock Market And Earnings: ##br##Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance Chart III-8Global Stock Market And Earnings: ##br##Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance FIXED INCOME: Chart III-9U.S. Treasurys And Valuations U.S. Treasurys and Valuations U.S. Treasurys and Valuations Chart III-10U.S. Treasury Indicators U.S. Treasury Indicators U.S. Treasury Indicators Chart III-11Selected U.S. Bond Yields Selected U.S. Bond Yields Selected U.S. Bond Yields Chart III-1210-Year Treasury Yield Components 10-Year Treasury Yield Components 10-Year Treasury Yield Components Chart III-13U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor Chart III-14Global Bonds: Developed Markets Global Bonds: Developed Markets Global Bonds: Developed Markets Chart III-15Global Bonds: Emerging Markets Global Bonds: Emerging Markets Global Bonds: Emerging Markets CURRENCIES: Chart III-16U.S. Dollar And PPP U.S. Dollar And PPP U.S. Dollar And PPP Chart III-17U.S. Dollar And Indicator U.S. Dollar And Indicator U.S. Dollar And Indicator Chart III-18U.S. Dollar Fundamentals U.S. Dollar Fundamentals U.S. Dollar Fundamentals Chart III-19Japanese Yen Technicals Japanese Yen Technicals Japanese Yen Technicals Chart III-21Euro/Yen Technicals Euro/Yen Technicals Euro/Yen Technicals Chart III-20Euro Technicals Euro Technicals Euro Technicals Chart III-22Euro/Pound Technicals Euro/Pound Technicals Euro/Pound Technicals COMMODITIES: Chart III-23Broad Commodity Indicators Broad Commodity Indicators Broad Commodity Indicators Chart III-24Commodity Prices Commodity Prices Commodity Prices Chart III-25Commodity Prices Commodity Prices Commodity Prices Chart III-26Commodity Sentiment Commodity Sentiment Commodity Sentiment Chart III-27Speculative Positioning Speculative Positioning Speculative Positioning ECONOMY: Chart III-28U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop Chart III-29U.S. Macro Snapshot U.S. Macro Snapshot U.S. Macro Snapshot Chart III-30U.S. Growth Outlook U.S. Growth Outlook U.S. Growth Outlook Chart III-31U.S. Cyclical Spending U.S. Cyclical Spending U.S. Cyclical Spending Chart III-32U.S. Labor Market U.S. Labor Market U.S. Labor Market Chart III-33U.S. Consumption U.S. Consumption U.S. Consumption Chart III-34U.S. Housing U.S. Housing U.S. Housing Chart III-35U.S. Debt And Deleveraging U.S. Debt And Deleveraging U.S. Debt And Deleveraging Chart III-36U.S. Financial Conditions U.S. Financial Conditions U.S. Financial Conditions Chart III-37Global Economic Snapshot: Europe Global Economic Snapshot: Europe Global Economic Snapshot: Europe Chart III-38Global Economic Snapshot: China Global Economic Snapshot: China Global Economic Snapshot: China

Highlights Trading The Yield Curve: Butterfly trades, going long or short a bullet versus a barbell, offer exposure to changes in the slope of the yield curve while remaining insulated from small parallel curve shifts. This will always be true provided that the cash allocation to the two bonds in the barbell is chosen to make the dollar-duration of the barbell equal to that of the bullet. Yield Curve Models: Using a model of the butterfly spread versus the slope, we can calculate how much curve steepening or flattening is being discounted by the current yield curve. Our strategy is to only implement a steepening/flattening butterfly trade if we think the curve will steepen/flatten by more than what is currently priced in. Empirical Performance: Incorporating the reading from our butterfly spread model improves on the performance of a purely macro-based yield curve trading strategy, both in theory and empirically. A purely mechanical trading rule based on our model also displays encouraging results over time. Feature One of the mandates of this publication is to take a view on the slope of the yield curve. Typically, we implement these views by recommending butterfly trades. A butterfly trade consists of two legs: A Barbell. Defined as a weighted combination of the two bonds that bound the yield curve segment you want to trade. For example, to take a view on the 2/10 slope, the barbell leg of the trade would be a weighted combination of the 2-year and 10-year notes. A Bullet. Defined as a bond that sits near the middle of the yield curve segment you want to trade. For example, the 5-year note would be a good choice for the bullet leg of a trade designed to profit from shifts in the 2/10 slope. A butterfly trade is defined as going long either the bullet or barbell while simultaneously shorting the other. This provides exposure to the slope of the curve because bullets tend to outperform barbells when the yield curve steepens and vice-versa (Chart 1 on page 1). Chart 1Gain Curve Exposure Through Butterfly Trades Gain Curve Exposure Through Butterfly Trades Gain Curve Exposure Through Butterfly Trades In this Special Report, we explain why butterfly trades are the best way to gain exposure to changes in the slope of the yield curve. We also explain how we think about the trade-off between our macro-informed view of whether the yield curve will steepen or flatten and how much steepening/flattening is already discounted in the market. To determine what is discounted in the market we rely on fair value models of the butterfly spread, which are also described in this report.1 Note: In the remainder of this report we focus exclusively on the 2/10 slope of the curve and the 2/5/10 butterfly spread, although the logic of butterfly trades applies to any yield curve segment. We will explore different yield curve segments in future reports. The Mechanics Of Butterfly Trades The first choice that must be made when implementing a butterfly trade is how to weight the two bonds used in the barbell. The chosen weighting scheme depends on what sort of curve movement you want to profit from. For our purpose, which is to gain exposure to changes in the slope of the yield curve while remaining insulated from parallel shifts, we adopt a dollar duration (DV01)2 weighting scheme. In this weighting scheme, the barbell weights are set so that the DV01 of the bullet leg of the trade matches the DV01 of the barbell. Table 1 presents an illustrated example of how this works. Table 1Butterfly Trade Performance Illustrated Bullets, Barbells And Butterflies Bullets, Barbells And Butterflies The top half of Table 1 shows an example based on hypothetical bonds derived from the Federal Reserve's par coupon constant-maturity yield curve. By definition, each of these hypothetical bonds trades at par ($100) and we use that fact along with the par coupon yield to calculate each bond's duration. After calculating the DV01 for each hypothetical bond by multiplying its duration by its price and dividing by 104, we can calculate that placing 40% of the barbell's cash in the 10-year note and 60% in the 2-year note leads to identical DV01's in both the bullet and barbell. Shocking The Yield Curve Identical DV01's in each leg of the trade means that if we go long one leg and short the other, our butterfly trade is immune to small parallel shifts in the yield curve. This is shown in the sixth column of Table 1, where we see that a +1 basis point parallel shift in the curve results in a loss of $0.0475 in both the bullet and barbell. It should be noted that this immunization from parallel curve shifts only works for small changes in yields. This is because while we have matched the DV01 between each leg of the trade, we have not matched the convexity. In this weighting scheme the barbell will always have a greater convexity than the bullet and will outperform in the event of a large parallel curve shift (in either direction). However, large parallel curve shifts are quite rare in practice. Usually, big yield moves are associated with either a steepening or a flattening of the curve. As such, convexity differences are only a minor consideration when we recommend butterfly trades. While the DV01 of each leg of the trade is the same, within the barbell itself there is a mismatch between the 2-year and 10-year notes. The fifth column of Table 1 shows that the weighted DV01 contribution to the barbell is $0.0117 from the 2-year note and $0.0357 from the 10-year note. The greater "weighted DV01" means that the barbell is more sensitive to changes in the 10-year yield than to changes in the 2-year yield. It is this mismatch that gives the butterfly trade exposure to the slope of the curve. For example, column 7 of Table 1 presents a scenario where the curve steepens by a small amount. Specifically, the 10-year yield rises 1 bp, the 2-year yield falls 1 bp and the 5-year yield remains flat. In this scenario, the losses in the 10-year note more than offset the gains in the 2-year note, causing the barbell to underperform the bullet. The opposite scenario is presented in column 8, which shows that the barbell outperforms the bullet when the curve flattens. The bottom half of Table 1 replicates the same analysis using the current on-the-run 2-year, 5-year and 10-year notes instead of hypothetical par bonds. It shows that the same logic and methodology apply in both cases. Bottom Line: Butterfly trades, going long or short a bullet versus a barbell, offer exposure to changes in the slope of the yield curve while remaining insulated from small parallel curve shifts. This will always be true provided that the cash allocation to the two bonds in the barbell is chosen to make the dollar-duration of the barbell equal to that of the bullet. Modeling The Butterfly Spread Often, it is not sufficient to just know whether the curve will steepen or flatten and then put on the appropriate butterfly trade. In an efficient market the butterfly spread (defined in this report as the yield on the bullet minus the yield on the DV01-matched barbell) should adjust to expected changes in the slope of the curve so that no excess profits can be earned. We see evidence for this in the bottom panel of Chart 1 on page 1. Here, the 2/5/10 butterfly spread widens as the 2/10 slope steepens and vice-versa. The logic of this relationship depends on mean reversion. As the curve steepens investors start to discount a greater probability of curve flattening in the future. This means that investors will also demand greater compensation to enter steepener trades (long bullet, short barbell) as the curve steepens. We can take advantage of this positive relationship between the slope of the curve and the butterfly spread by creating a fair value model (Chart 2). The model is simply a regression of the 2/5/10 butterfly spread on the 2/10 Treasury slope. Chart 22/5/10 Butterfly Spread Fair Value Model 2/5/10 Butterfly Spread Fair Value Model 2/5/10 Butterfly Spread Fair Value Model We tested the model using many different time intervals and settled on a regression coefficient of 0.14. As shown in Chart 3, the coefficient has been reasonably close to 0.14 for most of its history, with the exception of the period immediately following the financial crisis when the fed funds rate was pinned at the zero-lower-bound. The zero-lower-bound caused the relationship between the butterfly spread and the slope to weaken dramatically, but it began to re-assert itself once the Fed started to lift rates at the end of 2015. At present, the coefficient from a 3-year trailing regression is 0.17. Chart 3Choosing The Right Beta Choosing The Right Beta Choosing The Right Beta What's Priced Into The Curve? One obvious application of our fair value model is that we can identify periods when the butterfly spread is too high or too low relative to the slope of the curve. Put differently, when the butterfly spread's deviation from fair value is above zero, the bullet looks attractive relative to the barbell. When the deviation from fair value is below zero, the barbell looks attractive compared to the bullet. However, if we make a few simplifying assumptions, we can express the model's deviation from fair value in a more helpful way. If we assume that: The butterfly spread will revert to its fair value during the next 6 months During this time period returns to the bullet and barbell legs of the trade will be equal3 Then we can calculate how much the slope of the curve must change to satisfy both conditions. In other words, we can answer the question of what change in the slope is being discounted by today's butterfly spread. Chart 4How Our Models Add Value How Our Models Add Value How Our Models Add Value The third panel of Chart 2 shows the change in the 2/10 slope that is currently being discounted by the butterfly spread. The bottom panel shows the level of the slope that is implied by the model compared to the actual 2/10 slope. A recent example of why it's important to consider what is priced into the curve is shown in Chart 4. Last December 20,4 we recommended entering a butterfly trade that is long the 5-year bullet and short the 2/10 barbell, a trade designed to profit from curve steepening. Since then, however, the 2/10 slope has flattened 44 bps. Despite the curve flattening, our recommended trade is 21 bps in the money. The reason is that, according to our model, on December 20 the butterfly spread was discounting a whopping 49 bps of flattening during the next 6 months. Significantly more flattening than what actually occurred. We continue to recommend this trade going forward, even though the curve is now already priced for 6 bps of 2/10 steepening during the next six months. This means that we will need the yield curve to steepen more than 6 bps for our trade to outperform. We continue to see this as the most likely outcome.5 Bottom Line: Using a model of the butterfly spread versus the slope, we can calculate how much curve steepening or flattening is being discounted by the current yield curve. Our strategy is to only implement a steepening/flattening butterfly trade if we think the curve will steepen/flatten by more than what is currently priced in. Empirical Testing Charts 5 and 6 illustrate the importance of relying on both a macro call about the slope of the curve and the reading from our butterfly spread model. In Chart 5 we plot 6-month excess returns in the 5-year bullet over the 2/10 barbell versus the 6-month change in the 2/10 slope. While we see a reasonably strong positive correlation, there are still many periods of steepening when the bullet underperforms and many periods of flattening when the barbell underperforms. Chart 5Performance Of A Bullet Over Barbell Strategy Vs. ##br##The Actual Change In The 2/10 Nominal Treasury Slope Bullets, Barbells And Butterflies Bullets, Barbells And Butterflies Chart 6Performance Of A Bullet Over Barbell Strategy Vs. The Difference Between ##br##Actual And Discounted Change In The 2/10 Nominal Treasury Slope Bullets, Barbells And Butterflies Bullets, Barbells And Butterflies Chart 6 plots the same 6-month excess return, but this time against the difference between the actual change in the 2/10 slope and what was priced-in according to our model. Here we observe a much stronger correlation and fewer examples of the butterfly trade not performing as expected. Going one step further, Table 2 shows the results of implementing butterfly trades over 6-month horizons assuming perfect knowledge of how the yield curve will move. The first row shows that, during our sample period, a long 5-year bullet, short 2/10 barbell trade produced positive returns 71% of the time when the 2/10 slope steepened, for an average un-levered 6-month return of 34 bps. Similarly, long 2/10 barbell, short 5-year bullet trades produced positive returns 71% of the time when the 2/10 slope flattened, for an average un-levered 6-month return of 24 bps. Table 2Performance Of Butterfly Trades Over 6-Month Horizons ##br##Assuming Perfect Knowledge Of Curve Movements (1976-Present) Bullets, Barbells And Butterflies Bullets, Barbells And Butterflies The bottom two rows of Table 2 show that the performance of these trades improves when we also incorporate the reading from our model, only putting on trades when the steepening or flattening is greater than what was initially priced in. In fact, incorporating the output from our butterfly spread model led to 128 instances when we would have reversed the trade that would have been implemented if all we knew was which direction the slope would move. Out of those 128 instances, 60% of the time the change led to a better trade. Cumulatively, incorporating the reading from the model produced an extra return of more than 11% throughout our entire sample. Can We Just Follow The Model? This begs the question of whether we can create a mechanical trading rule based purely on the output from our butterfly spread model that will produce positive results. To test this we first look at excess returns in the 5-year bullet over the 2/10 barbell in 6-month periods following different readings from our model (Table 3). Table 3Performance Of Butterfly Trades Over 6-Month ##br##Horizons Based Only On Our Model (1976-Present) Bullets, Barbells And Butterflies Bullets, Barbells And Butterflies We find that bullets outperform barbells in more than 70% of 6-month periods when the 5-year bullet appears more than 5 bps undervalued. Similarly, barbells outperform in 56% of 6-month periods when the 2/10 barbell is more than 5 bps undervalued. Second, we created a trading rule where every month you invest either: Chart 7A Model-Driven Curve Trading Strategy A Model-Driven Curve Trading Strategy A Model-Driven Curve Trading Strategy 100% in the 5-year bullet, if the bullet appears more than 5 bps cheap on our model. 50% in the 5-year bullet and 50% in the 2/10 barbell, if the bullet is between 5 bps expensive and 5 bps cheap compared to the barbell. 100% in the 2/10 barbell, if the barbell appears more than 5 bps cheap on our model. The cumulative results from this model since 1980 relative to a curve-neutral benchmark that is always invested 50% in the bullet and 50% in the barbell are shown in Chart 7. We observe a clear outperformance over time, with relatively few periods of sustained losses. Bottom Line: Incorporating the reading from our butterfly spread model improves on the performance of a purely macro-based yield curve trading strategy, both in theory and empirically. A purely mechanical trading rule based on our model also displays encouraging results over time. Going forward we will consider both the output from our butterfly spread model and our macro view of the yield curve when recommending butterfly trades. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com 1 These models were first introduced in a Global Fixed Income Strategy Special Report from February 1, 2002. Please contact your sales representative to request a copy. 2 DV01 is the dollar value of a basis point. It measures the dollar change in the price of a given bond assuming a one basis point change in its yield. It is calculated as the bond's duration times its price, divided by 104. 3 A 6-month time period was arbitrarily chosen to line up with our preferred investment horizon. We also need to assume how much of the discounted shift in the yield curve occurs at the long-end relative to the short-end. We assume that half the change in slope occurs at each maturity, but the results are not very sensitive to changing this assumption. 4 Please see U.S. Bond Strategy Special Report, "Seven Fixed Income Themes For 2017", dated December 20, 2016, available at usbs.bcaresearch.com 5 For further details on our macro outlook for the yield curve please see U.S. Bond Strategy Weekly Report, "The Yield Curve On A Cyclical Horizon", dated March 21, 2017, available at usbs.bcaresearch.com

Highlights BCA's Central Bank Monitors support the case for less stimulus. Yellen's "dovish" testimony does not change our Fed call. The BCA Beige Book Monitor and related indicators support our view on the economy and Fed. Maximum central bank policy divergence has not been reached. Too early to predict Trump's replacement for Yellen. Now that economic surprise index has bottomed, risk assets can outperform as the metric mean reverts. Some wage measures are accelerating as the economy approaches full employment. Feature Chart 1Sell-Off In Global Bond Markets##BR##Triggered By Central Bank Talk Sell-Off In Global Bond Markets Triggered By Central Bank Talk Sell-Off In Global Bond Markets Triggered By Central Bank Talk Global bond investors were shocked in June when central bankers announced at the ECB's Forum on Central Banking what appeared to be a global recalibration of monetary policy. Until that time, investors had been lulled into a false sense of security that growth headwinds would prevent the Fed from hiking by more than once a year and keep the other major central banks on hold "indefinitely." The heads of the Bank of England (BoE), the Bank of Canada (BoC) and the Riksbank all took a less dovish tone, as they signaled less need for ultra-stimulative policies because the threat of deflation had diminished. Together with some better-than-expected U.S. economic data, this shift in tone led to a sharp sell-off in global bond markets (Chart 1). The BoC followed up last week by kicking off a prolonged tightening cycle. The central bank now expects the Canadian economy to reach full employment and hit the BoC's inflation targets by mid-2018, which is much earlier than expected. The global bond mini-rout actually began before the ECB Forum, when the ECB President gave a very upbeat description of the underlying strength of the Euro Area economy. BCA's Global Fixed Income Strategy service highlights that the Euro Area is about two percentage points closer to full employment than the U.S. was just before the infamous 2013 Taper Tantrum.1 European core inflation is admittedly below target today, but so was the U.S. rate leading up to the 2013 Tantrum. Draghi's comments confirm that the ECB will announce this fall that a further tapering of its asset purchase program will take place in early 2018. The message that "emergency" levels of monetary accommodation are no longer needed is confirmed by our Central Bank Monitors (CB), which measure pressure on central bankers to raise or lower interest rates (Chart 2). The Monitors became less useful when rates hit the zero bound and quantitative easing became popular, but the measures are relevant again. All of our CB Monitors are in "tighter policy required" territory except for Japan (although even that one appears on the verge of breaking above the critical zero line). The Monitors have been rising due more to their growth than their inflation components. Bond investors may be startled by the ECB's posture because inflation remains well below target in all the major economies except the U.K. What is most worrying is the recent deceleration in U.S. inflation, where the economy is very close to or at full employment. Almost all of the major central banks point to temporary factors that will soon fade, which would allow inflation to escalate toward the target. Our Aggregate Inflation Indicators have all signaled a modest building of underlying inflation pressure over the past year (although they have softened recently in the U.S. and Eurozone; Chart 3). In terms of the components of these indicators, rising core producer price inflation has been partly offset by slower gains in unit labor costs in some economies. Chart 2All In The "Tighter Policy Required" Zone All In The "Tighter Policy Required" Zone All In The "Tighter Policy Required" Zone Chart 3BCA Aggregate Inflation Indicators BCA Aggregate Inflation Indicators BCA Aggregate Inflation Indicators These and other indicators support our view that core consumer price inflation will grind higher in the coming months in most of the advanced economies, including the U.S. Admittedly, all models and indicators have been poor predictors of inflation in this recovery. Nonetheless, historical relationships might begin to re-establish now that capacity utilization is rising and labor market slack has moderated significantly. Did Yellen Turn Dovish? June's FOMC minutes indicated that the consensus among Fed policymakers is willing to "look through" low inflation and maintain the current timetable on rate hikes. Yellen's Congressional testimony last week did not deviate from that view, although investors interpreted her remarks as dovish. The financial press focused on her statement that "...the policy rate is not far from neutral." However, this was followed up by the statement that "...because we also anticipate that the factors that are currently holding down the neutral rate will diminish somewhat over time, additional gradual rate hikes are likely to be appropriate over the next few years to sustain the economic expansion and return inflation to our 2 percent goal." The Fed asserts there are two neutral interest rates: short-term and long-term. Yellen argued that the actual policy rate is close to the short-term level, which is depressed by economic headwinds. However, Yellen and others have made the case that the short-term neutral rate is trending up as headwinds diminish, and will converge with the long-term neutral rate over time. The Fed Chair is at risk of confusing investors by discussing the concept of two neutral rates, although this may have been to head off demands by some Congressional lawmakers that the Fed should follow a mechanical policy rule when setting policy (such as the Taylor Rule). Nonetheless, the important point is that Yellen is not saying that the actual policy rate is close to the peak for the cycle. Yellen's testimony has not altered our Fed call for this year: balance sheet runoff beginning in the fall, followed by a rate hike in December. The latter hinges importantly on at least a modest rise in core PCE inflation in the coming months. We expect more rate hikes in 2018/19 than are discounted in the bond market. That said, the soft June CPI data challenges our view that inflation will move higher in the second half. The bottom line is that the backdrop has turned decidedly bond-bearish now that central bankers in the advanced economies are in the process of scaling back the easier monetary policy that followed the deflationary 2014/15 oil shock. Global bond yields have already taken a step up in recent weeks, but they will have to rise further to catch up with the solid pace of global growth and diminishing economic slack. Duration should be kept short. The Beige Book: Another Inflation Anomaly The Beige Book released on July 12 supports the Fed's base case outlook for the economy and inflation. It also keeps the Fed on track to begin to trim its balance sheet in September and boost rates by another 25 basis points in December. Our quantitative approach2 to the qualitative data in the Beige Book points to an acceleration in GDP and inflation, less business unease from a rising U.S. dollar, and ongoing improvement in real estate, both commercial and residential (Chart 4). Chart 4Beige Book Monitors Support Fed's Outlook##BR##On Economy And Inflation Beige Book Monitors Support Fed's Outlook On Economy And Inflation Beige Book Monitors Support Fed's Outlook On Economy And Inflation At 62%, the BCA Beige Book Monitor remained near its cycle highs in July, providing more confirmation that the economy rebounded in Q2 after a desultory Q1. The July 12 Beige Book covered the period from late May through June 30. Based on the Beige Book, the dollar should not be much of an issue in Q2 earnings season. The greenback seems to have faded as a concern for small businesses and bankers, which is in sharp contrast to 2015 and early 2016 when mentions of a strong dollar in the Beige Book surged. The Q2 earnings reporting season will provide corporate managements with another forum to express their views of currency impact on their operations. Business uncertainty over government policy (fiscal, regulatory and health) remained elevated in the most recent Beige Book (not shown). The implication is that the business community is mindful of the lack of progress by Washington policymakers on Trump's agenda. Our analysis of the Beige Book also shows that real estate was still stout as Q2 ended. This implies that both residential and commercial real estate, the former a source of strength in Q1, will add to growth again in Q2. Moreover, the latest reading on the BCA Real Estate Monitor further widened the gap between the BCA Beige Book Real Estate Monitor and the relative performance of REITS to the S&P 500. Nonetheless, BCA's U.S. Equity Strategy service recently downgraded REITS to neutral,3 citing our expectation of higher Treasury yields, modest rent growth, some cracks in CRE credit quality, and tightening standards for bank lending in the CRE marketplace. Echoing the market's disagreement with the Fed on inflation, the big disconnect in the Beige Book showed up in the number of inflation words. Inflation words hit a new peak in July, in sharp contrast with the recent soft readings on CPI and PCE. In the past, increased references to inflation have led measured inflation by a few months, suggesting that the CPI and core PCE may soon turn up. Bottom Line: The Beige Book backs the Fed's assertion that the economy will expand around 2% this year and inflation will move higher in the coming months, supporting a gradual removal of policy accommodation. Uncertainty in Washington is distressing, but worries over the dollar seem to be fading. Max Policy Divergence Has Not Been Reached What about the dollar? Tighter Fed policy is dollar-bullish on its own, but some of the major central banks are also starting to remove the monetary punchbowl as well. Recent dollar action suggests that investors have decided that the peak Fed/ECB policy divergence is now behind us. We do not agree. The ECB may be tapering, but raising interest rates is a long way off because there is still a lot of economic slack in the Eurozone. In contrast, the Fed is increasingly concerned that allowing the unemployment rate to fall further below its estimate of full-employment risks too large an overshoot of the 2% target. We still believe that market pricing for the fed funds rate is too benign. As Fed rate hike expectations ratchet up in the coming months, interest rate differentials versus Europe will widen in favor of the dollar. It is the same story for the dollar/yen rate. The major exception is the Canadian dollar, which we expect to appreciate versus the greenback. Does Gary Cohn Have What It Takes? A key wildcard in the financial outlook is the Fed Chair's replacement. Yellen's term as Chair will end in February 2018 and the markets have not yet shown any concerns about her potential replacement. The current frontrunner is Gary Cohn, the Chairman of President Trump's National Economic Committee; his appointment would conform to some historical precedents but violate others. Our March 6 Weekly Report4 provides a list of potential Fed appointees and also provides some background on the potential for the Fed to become more politicized under Trump. Since the late 1970s, Presidents have selected the Fed Chair based on their trust relationship with a candidate. Arthur Burns (Chair from 1970-1978) was the head of President Eisenhower's Council of Economic Advisors (CEA) and was a special counselor to President Nixon. William Miller (1978-1979) worked for the presidential campaigns of Hubert Humphrey and Jimmy Carter. Alan Greenspan (1987-2006) served as the Chair of President Reagan's Social Security Commission in the early 1980s, was the Chair of President Ford's CEA and advised President Nixon's campaign in 1968. Ben Bernanke (2006-2014) was George W. Bush's chief economist in 2005 and 2006 before Bush chose him to lead the Fed. Janet Yellen (2014-present) was Chair of Bill Clinton's CEA in the late 1990s, when she worked with many of Obama's economic team members. Paul Volcker (1979-1987) was the lone exception to this rule; he worked for Nixon, but not Carter, before becoming Fed Chair (Table 1). Table 1Characteristics Of Fed Chairs Since 1970 Global Monetary Policy Recalibration Global Monetary Policy Recalibration Cohn does not have any experience as a central banker, but that does not preclude him from holding the position. Volcker, Bernanke and Yellen, all held posts in the Federal Reserve System before their appointments as Chair. However, Miller was an outside director for the Boston Fed, and Burns and Greenspan had no prior experience at the monetary authority. Party identification is one area where Gary Cohn would stand out. Since at least 1970, the party affiliation of a new Fed Chair has matched that of the President. However, Presidents have crossed party lines to reappoint sitting Fed Chairmen to additional terms. Volker, Greenspan and Bernanke were reappointed to lead the Fed by Presidents from opposing political parties. The timing of Trump's announcement on Yellen's replacement may be critical. In the summer of 2013, names were already being floated by the Obama White House (and mainly rejected) by markets, before he finally settled on Yellen. The official announcement came in early October 2013. In August 2009, President Obama reappointed Bernanke for a second four-year term. Bernanke was initially nominated to be Fed Chair by George W. Bush in October of 2005. If the appointment comes in October and the nominee is perceived to be hawkish, the risk is that markets may begin to price in the regime change sometime in the next few months. Our U.S. Bond Strategy service argued in a recent report5 that rate hike expectations may already be ramping up, while the data on the economy and inflation begin to beat expectations again. Bottom Line: It is too early for the markets to be concerned about the next Fed Chair and their policies. The names mentioned in the summer may not be the ones offered the job in the fall. Surprise Index Finally Bottomed Out The June employment report marked a turning point for the Citigroup surprise index, following an extended period of disappointment that depressed the dollar and bond yields. The June reports on CPI and retail sales were disappointing, but June industrial production exceeded expectations. What does this mean for relative asset returns? After 86 days, expectations moved low enough to allow economic reality to begin to run ahead. It took as few as 8 business days (in 2009) and as many as 164 (2015) for the surprise index to return to the zero line, an average of 52 days (Chart 5). Chart 5Risk Assets Tend To Outperform As Economic Surprise Index Rebounds Risk Assets Tend To Outperform As Economic Surprise Index Rebounds Risk Assets Tend To Outperform As Economic Surprise Index Rebounds Mean-reversions in the surprise index following troughs have generally been good for risk assets in this recovery (Table 2). We have identified 11 periods since late 2009 when the surprise index bottomed out and then moved up toward zero. In 8 of those episodes, the total return on stocks was higher than 10-year Treasuries. Equities beat Treasuries by an average of 286 bps across all 11 periods, with a median outperformance of 400 basis points. Table 2U.S. Financial Market Performance As Economic Surprise Index Rises Global Monetary Policy Recalibration Global Monetary Policy Recalibration The total return on investment-grade corporate debt outperformed Treasuries in 6 of 11 episodes. In those six instances, investment grade credit outperformed on average by 132 bps. Nonetheless over all 11 episodes, the excess return was 0%. In contrast, high-yield bonds beat Treasuries in 7 of the 11 periods, with a median outperformance of 188 basis points. Similarly, small caps beat large caps 72% of the time as the economic surprise index moved back toward the zero line. The median outperformance of small over large in all 11 periods was 124 basis points. The performance of commodities was mixed as economic surprises climbed. Gold rose in 6 of the 11 times, but fell in 5. Oil prices posted increases in only 5 of the 11, but the median return for oil after economic surprise bottomed was -2.7%. Bottom Line: Economic expectations that ramped up post-election have now declined and allowed the economic surprise index to trough. The implication for investors is that risk assets tend to outperform as the economic surprise index moves back to zero. This supports our tactical views of stocks over bonds, small over large caps, and credit over Treasury. What's Up With Wages? The June jobs report released in early July6 only added to the market's fears that the Phillips Curve is dead because wage growth softened even as the labor market tightened. Unfortunately, no Fed officials including Yellen have addressed the topic in depth recently. The market does not believe the Fed when it says that the tighter labor market is pushing up wages. We see it another way. Chart 6 shows that wage inflation has accelerated since mid-to-late 2012, but some measures of wages have made more progress than others. Chart 7 and Chart 8 reinforce that, setting aside the rollover in average hourly earnings (AHE), wage inflation is accelerating, albeit modestly. Chart 6Plenty Of Signs That##BR##Wages Are Accelerating Plenty Of Signs That Wages Are Accelerating Plenty Of Signs That Wages Are Accelerating Chart 7Compositional Effects Do Not##BR##Explain Recent Rollover Compositional Effects Do Not Explain Recent Rollover Compositional Effects Do Not Explain Recent Rollover Chart 8Acceleration In Hours Worked Should##BR##Lead To Faster Wage Growth Acceleration In Hours Worked Should Lead To Faster Wage Growth Acceleration In Hours Worked Should Lead To Faster Wage Growth The Employment Cost Index (ECI) excluding bonuses (Chart 6, panel 1) is our favorite measure of labor compensation. It has accelerated steadily since 2010. It adjusts for compositional changes in the labor market (unlike the average hourly earnings measure) and is the broadest and most comprehensive wage metric. Its drawbacks are that it is released with a long lag. For example, the Q2 ECI data will not be released until the end of July. The AHE data is already available for June and Q2. On the other hand, unit labor costs (ULC) (panel 2) have stagnated for the past five years. Data starts in 1947, so it has the most history of any of the wage measures. However, it is even more delayed than the ECI: it is released five weeks after the end of the quarter. Moreover, these data are subject to revisions and tend to be more volatile than other wages measures, which makes it difficult to identify a change in trend. Productivity, which is used to construct ULC, is also very difficult to estimate. A recent BIS report7 notes that there is evidence that the relationship between ULC and labor market slack has diminished over time, but that ULC is a better measure of inflationary pressures than AHE. Median usual weekly earnings (panel 3) have also accelerated. This is not a pure wage measure; it combines hourly pay and hours worked and, therefore, is a good proxy for incomes. Income growth has picked up the pace, providing a solid underpinning for consumer spending. Panel 4 shows compensation per hour worked. It, too, has stalled and is subject to the same strengths and weakness as ULC because it is part of the quarterly Productivity and Costs report. This metric has run near 2% with no trend. Finally, average hourly earnings (panel 5) have sped up since 2012, but rolled over in late 2016. This wage gauge gets most of the market's attention although it is only one of many measures that the Fed watches. AHE is a timely data set, released alongside monthly payroll numbers. It includes average earnings of private non-farm production and non-supervisory positions. The major disadvantage of this measure is that hourly wage earners represent only about 58% of workers and do not account for trends in salaried jobs. Earnings do not include bonus pay or employee benefits. The data are available beginning only in 2006. In Chart 7, we created an "equally-weighted" AHE measure to adjust for shifts in the composition of the labor market, but we found that the recent deceleration is not linked to compositional effects. Since wage growth bottomed out in late 2012, the compositional shifts slightly lowered wage inflation on average, but the growth rates today are roughly the same. The Atlanta Fed wage tracker (not shown) is in a distinct uptrend. The Tracker has the advantage that it is not biased by compositional shifts. Chart 8 shows our update to a study by the Kansas City Fed8 that found only a few industries (mostly in the goods-producing sector of the economy) have accounted for most of the rise in wages, notably manufacturing, construction and wholesale trade. Financial services, retail, professional and business services, and leisure and hospitality - all service sector industries - were the laggards. The report shows that although earnings growth has fallen behind in service-oriented industries since 2015, hours worked have seen faster growth than in the goods-producing sector. We concur with the author that labor demand was strong in the past few years in areas that have not experienced much wage growth. As the labor market continues to tighten, wages in these industries may accelerate, but patience will be required. Bottom Line: The various measures of wage inflation provide a mixed picture. Taken as a group, however, we believe that wage growth has indeed accelerated as the labor market has tightened. The acceleration has admittedly been modest, but it is only recently that unemployment reached a full employment level. The real test for the Phillips curve will be in the coming quarters as the economy moves further into "excess labor demand" territory. John Canally, CFA, Senior Vice President U.S. Investment Strategy johnc@bcaresearch.com Mark McClellan, Senior Vice President The Bank Credit Analyst markm@bcaresearch.com 1 Please see BCA's Global Fixed Income Strategy Weekly Report, "Central Banks Are Now Playing Catch-Up", dated July 4, 2017. Available at gfis.bcaresearch.com. 2 Please see U.S. Investment Strategy Weekly Report "The Great Debate Continues", dated April 17, 2017. Available at usis.bcaresearch.com. 3 Please see U.S. Equity Strategy Weekly Report "SPX 3000?, dated July 10, 2017. Available at uses.bcaresearch.com. 4 Please see U.S. Investment Strategy Weekly Report "Trump And The Fed", dated March 6, 2017. Available at usis.bcaresearch.com. 5 Please see BCA's U.S. Bond Strategy Weekly Report "Summer Snapback", dated July 11, 2017. Available at usbs.bcaresearch.com 6 Please see U.S. Investment Strategy Weekly Report "Sizing up the Second Half", dated July 10, 2017. Available at usis.bcaresearch.com. 7 Monetary policy: inching towards normalization", Bank for International Settlements (BIS), 25 June 2017. 8 Wage Leaders and Laggards: Decomposing The Growth In Average Hourly Earnings", Willem Van Zandweghe, Federal Reserve Bank of Kansas City, February 15, 2017.

Highlights Portfolio Strategy The energy bear market is drawing to a close. Lift exposure to above benchmark. Firming refining operating conditions, at the margin, suggest that it no longer pays to underweight this energy sub-group. Book gains and lift to neutral. Softening profit fundamentals are weighing on the real estate outlook. Trim REITs to neutral. Recent Changes S&P Energy - Lift to overweight. S&P Oil & Gas Refining & Marketing - Upgrade to neutral, lock in profits of 14.3%. S&P Real Estate - Trim to neutral and remove from high-conviction overweight list. Table 1 SPX 3,000? SPX 3,000? Feature Equities wrestled to hold on to gains last week, fighting a spike in geopolitical tensions, synchronized global central bank hawkish rhetoric and sector rotation. Investors continued to shed tech stocks in favor of financials, pushing our recently initiated long financials/short tech pair trade return near double digits. Our view remains that a rotational correction is the dominant market theme. Nevertheless, on the eve of earnings season, profits will soon take center stage and serve as a catalyst for the overshoot equity phase to resume. Our sense is that before the end of the business cycle, there are high odds that the S&P 500 will hit the 3,000 mark. That does not mean it will be a straight line advance from here. Garden variety 5-10% corrections are all but inevitable. Rather, our point is that before the next recession hits late in the decade, the SPX can attain 3,000. How did we come up with this figure? We derived the S&P 500's peak cycle value using three different methods: Dividend Discount Model (DDM) Forward P/E and EPS growth sensitivity analysis Equity Risk Premium (ERP) Table 2SPX Dividend Discount Model SPX 3,000? SPX 3,000? Table 2 shows our DDM on the S&P 500. It assumes healthy dividend growth in 2017 and 2018. Our expectation of a 2019 recession drives a steep decline in dividends that year, followed by a slow climb in 2020 and 2021, in line with the 2009-2011 experience (Chart 1). 2022 is our terminal year when dividend growth settles at 6.6%, close to the long-term average. Our discount rate assumes a 3.2% 10-year Treasury yield and a 5% equity risk premium (the past decade's average, Chart 2). This discount rate mirrors the historical average corporate junk bond yield. This valuation model delivers an S&P 500 value of 2904. Chart 1Joined At##br## The Hip Joined At The Hip Joined At The Hip Chart 2FX10 ERP And The Economy##br## Are Inversely Correlated ERP And The Economy Are Inversely Correlated ERP And The Economy Are Inversely Correlated Alternatively, we examine the S&P 500's sensitivity to EPS growth rates and forward valuation multiples. If we use the street's 160.8 (or 10.6% implied CAGR) S&P 500 2019 EPS estimate and assign the current 12-month forward multiple as a starting point, Table 3 shows an S&P 500 value of just under 3,000. Downside risks look limited. Using this EPS forecast, even a 2-turn multiple contraction results in the S&P 500 appreciating 10% from here. Table 3SPX EPS & Multiple Sensitivity SPX 3,000? SPX 3,000? Lastly, a conservative ERP analysis reveals that SPX 3,000 is a realistic peak cycle estimate. Our assumptions include: a 200 bps ERP, a 3.2% 10-year Treasury yield and 160.8 SPX EPS. These assumptions result in an S&P 500 value of slightly over 3,000. How do we justify a decline in the ERP from its current level of 338 bps to our assumed 200 bps? G10 central banks are no longer putting out GFC-related fires; in fact, a slew of them are briskly turning from dovish to hawkish following the Fed's lead (Chart 3). As a result, a sustained decline in the ERP should follow as interest rates rise. Chart 3G10 Central Banks Map SPX 3,000? SPX 3,000? Chart 4Negative Correlation Is Re-Established Negative Correlation Is Re-Established Negative Correlation Is Re-Established The bottom panel of Chart 2 drives this point home. Since the history of SPX forward EPS data, the year-over-year change in the ERP has been almost perfectly inversely correlated with the ISM manufacturing index, i.e. an improving economy is synonymous with a receding ERP and vice versa. Lastly, keep in mind that a 200 bps ERP is still significantly higher than the 80 bps mean ERP that prevailed in the 1998-2007 decade (middle panel, Chart 2). The depreciating greenback is another source of support for our SPX 3,000 view. The yearlong positive correlation between the U.S. dollar and commodities has likely come to an end and the three plus decade inverse correlation has been re-established (Chart 4). As the cycle matures and enters its late stages, commodities and resource-related equities tend to pick up steam as profits rebound. Even energy stocks may catch a bid. Buy Energy Stocks... Energy equities are down roughly 20% year-to-date versus the broad market. In fact, the energy sector has broken down to a level last seen in 2004, when oil traded near $30/bbl (Chart 5). The three main culprits have been rising U.S. shale oil production, inventory accumulation, and investor doubts about whether all nations will comply with OPEC's mandated production cuts. While going overweight the energy space has been a "widow maker" trade recently, we are now tempted to take a punt on the S&P energy sector from the long side. There are tentative signs that this relative performance bear phase is drawing to a close. Three main drivers support our modestly sanguine view of energy stocks. First, as we mentioned above, the inverse correlation between the U.S. dollar and the commodity complex has been re-established after a one-year hiatus. Synchronized global growth suggests that a corresponding tightening interest rate cycle is brewing (Chart 3). Thus, there are high odds that a number of G10 central banks will hike rates later this summer or early this fall, now that the Fed has paved the path.1 As long as the greenback drifts lower, even energy stocks should catch a bid (Chart 6). Chart 5Crude Oil... Crude Oil... Crude Oil... Chart 6...And The Dollar Say Buy Energy Stocks ...And The Dollar Say Buy Energy Stocks ...And The Dollar Say Buy Energy Stocks Second, on the domestic operating front, the steepest drilling upcycle in recent memory is showing signs of fatigue. Baker Hughes reported the first weekly decline in 24 weeks in the oil rig count for the week ending June 30th. At least a modest deceleration in shale oil production is likely. Encouragingly, Cushing crude oil inventories are contracting on a year-over-year basis and OECD oil stocks appear poised to contract in late autumn/early winter (Chart 7). Predicting OPEC's compliance is tricky. However, BCA's Commodity & Energy Strategy service believes that little to no cheating will occur and in a worst case scenario Saudi Arabia will step in and curtail production were Libya and/or Iraq to pump oil above quota. Finally, our S&P energy sector Valuation Indicator has gravitated back to the neutral zone. Technicals are also washed out with our Technical Indicator breaching one standard deviation below its historical mean, a level that typically heralds a reversal (Chart 8). Recent anecdotes that the sell-side is throwing in the towel on their bullish oil forecasts for the remainder of the year are also contrarily positive. Chart 7Improving Supply Dynamics Improving Supply Dynamics Improving Supply Dynamics Chart 8S&P Energy Unloved And Fairly Valued Unloved And Fairly Valued Unloved And Fairly Valued Our newly introduced S&P energy sector relative EPS model encapsulates this cautiously optimistic industry backdrop (Chart 9). Simultaneously, the budding recovery in our S&P energy Cyclical Macro Indicator also signals that profits should best those of the overall market (second panel, Chart 8), giving us comfort to lift the S&P energy sector to a modest overweight position. ... As Refiners Are No Longer Cracking Under Pressure We are executing the upgrade to overweight in the broad energy sector via booking gains of 14.3% since inception in the S&P oil & gas refining & marketing sub-group and lifting exposure to neutral from underweight. It no longer pays to remain bearish on the pure play downstream energy business. Back in late September 2015, when we turned negative on refiners, we were anticipating a cyclical earnings downturn on the back of a refined product glut in this low margin / high volume industry. Fast forward to 2017 and that bearish profit view has materialized as relative EPS have fallen by roughly 60 percentage points from the most recent peak, and have only lately managed to stabilize (Chart 10). Chart 9EPS Model Waves Green Flag EPS Model Waves Green Flag EPS Model Waves Green Flag Chart 10Refining Profit Contraction Is Over Refining Profit Contraction Is Over Refining Profit Contraction Is Over If relative EPS have indeed troughed, then relative performance should soon find a bottom. Relative profit fortunes move with the ebb and flow of gasoline consumption. The latter is on the cusp of expanding for the first time since last November, heralding the same for relative profitability (bottom panel, Chart 10). Industry shipments tell a similar story. After recently bottoming at levels similar to those reached during the GFC, refinery shipments have staged a mini V-shaped recovery (top panel, Chart 11). Crack spreads have not collapsed to razor thin levels as the nearly eliminated Brent/WTI spread would suggest, but have remained resilient in the high-teens per barrel (third panel, Chart 11). Three forces are likely in play. First, not only is domestic gasoline demand underpinning refining margins, but petroleum products are also finding their way into foreign markets with net exports running at over 3 million bbl/day (bottom panel, Chart 11). Second, the U.S. dollar selloff since mid-December is making U.S. refined products more competitive in global markets. Finally, crude oil inventories are nearly 40% higher than gasoline inventories. Lower industry feedstocks represent a boost to refining margins (third, Chart 11). Nevertheless, we refrain from turning outright bullish on refiners. Refinery production hit all-time highs recently, refinery runs are climbing steadily and utilization rates are running hot north of 90%. Tack on, historically high refined products inventories and rising industry capacity growth and the profit backdrop darkens (Chart 12). Chart 11Three Positives... Three Positives... Three Positives... Chart 12...But Do Not Get Carried Away ...But Do Not Get Carried Away ...But Do Not Get Carried Away Netting it out, we expect a balanced refining profit outlook in the coming quarters. Bottom Line: Upgrade the S&P oil & gas refining & marketing index (PSX, VLO, TSO, MPC) to neutral and lock in profits of 14.3%. This also pushes the S&P energy index to an above benchmark allocation. Downgrade REITs We are making space for the energy sector upgrade to overweight via trimming the niche S&P real estate sector to neutral and concurrently removing it from the high-conviction overweight list. REITs have marked time year-to-date, but recently operating conditions have downshifted a notch. Three key drivers argue for lightening up exposure on this newly formed S&P GICS1 sector. First, REITs have been unable to materially benefit from the 50bps fall in the 10-year Treasury yield from the mid-December peak to the mid-June trough. As the economy recovers from the first quarter lull, Treasury yields will resume their advance. This is a net negative for the fixed income proxy real estate sector (top panel, Chart 13). Second, real estate occupancy rates have crested and generationally high supply additions in the apartment space are all but certain to push vacancies higher still. The implication is that rental inflation will remain under intense downward pressure (Chart 13). Worrisomely, credit quality in select commercial real estate (CRE) segments is deteriorating at the margin. The bottom panel of Chart 13 shows that retail and office delinquency rates have taken a turn for the worse, and represent a yellow flag. Finally, according to the Fed's latest Senior Loan Officer Survey, bankers are less willing to extend CRE credit. In fact, if one excludes the GFC spike, the tightening in CRE lending standards is near the two previous recessionary highs. If banks continue to close the credit taps, CRE prices will suffer a setback (Chart 14). Chart 13Time To Move To the Sidelines Time To Move To the Sidelines Time To Move To the Sidelines Chart 14Conflicting Signals Conflicting Signals Conflicting Signals Chart 15 puts the CRE price appreciation in historical perspective. Currently, CRE prices are on track to climb to two standard deviations above the long-term trend. Such exuberance is a cause for concern as it has historically marked the beginning of a corrective phase in CRE prices. Nevertheless, there are some positive offsets that prevent us from throwing in the towel in the S&P real estate sector. The tight labor market and accelerating industrial production explain the reacceleration in our REITs Demand Indicator, while the recent selloff in the bond market is a modest offset. If CRE appetite remains upbeat, this in turn suggests that CRE prices have a bit more room to run before reaching a cyclical peak (bottom panel, Chart 14). In addition, compelling relative valuations and washed out technicals argue against becoming overly bearish on REITs (Chart 16), as some of the bad news is already reflected in relative share prices. Chart 15An Historical Perspective An Historical Perspective An Historical Perspective Chart 16Positive Offsets Positive Offsets Positive Offsets Bottom Line: Trim the S&P real estate sector to neutral and remove it from the high-conviction overweight list. Anastasios Avgeriou, Vice President U.S. Equity Strategy & Global Alpha Sector Strategy anastasios@bcaresearch.com 1 Please see the June 30th, 2017 Foreign Exchange Strategy Service Special Report titled "Who Hikes Next?", available at www.bcaresearch.com Current Recommendations Current Trades Size And Style Views Favor small over large caps and stay neutral growth over value.

GAA DM Equity Country Allocation Model Update The GAA DM Equity Country Allocation model is updated as of June 30th, 2017. The model has continued to reduce its allocation to the U.S. and now for the first time the U.S. allocation is slightly below benchmark. Within the non-U.S. universe of the 11 countries, the model has also made significant adjustments to shift weights to Italy (now the largest overweight) and Australia (now for the first time it is overweight by the model), while underweight Germany and France. These adjustments are mainly due to liquidity and technical indicators, as shown in Table 1. As shown in Table 2 and Charts 1, 2 and 3, the overall model underperformed its benchmark by 14 bps in June, largely due to the underperformance of Level 2 model where the overweight of the Euro area v.s. the underweight of Australia and Canada hurt the performance. Since going live, the overall model has outperformed its benchmark by 139 bps. Table 1Model Allocation Vs. Benchmark Weights GAA Model Updates GAA Model Updates Table 2Performance (Total Returns In USD) GAA Model Updates GAA Model Updates Chart 1GAA DM Model Vs. MSCI World GAA DM Model Vs. MSCI World GAA DM Model Vs. MSCI World Chart 2GAA U.S. Vs. Non U.S. Model (Level 1) GAA U.S. Vs. Non U.S. Model (Level1) GAA U.S. Vs. Non U.S. Model (Level1) Please see also on the website http://gaa.bcaresearch.com/trades/allocation_performance. For more details on the models, please see the January 29th, 2016 Special Report "Global Equity Allocation: Introducing the Developed Markets Country Allocation Model". http://gaa.bcaresearch.com/articles/view_report/18850. GAA Equity Sector Selection Model The GAA Equity Sector Selection Model (Chart 4) is updated as of June 30, 2017. Chart 3GAA Non U.S. Model (Level 2) GAA Non U.S. Model (Level 2) GAA Non U.S. Model (Level 2) Chart 4Overall Model Performance Overall Model Performance Overall Model Performance Table 3Allocations GAA Model Updates GAA Model Updates Table 4Performance Since Going Live GAA Model Updates GAA Model Updates The model continues to overweight cyclicals versus defensives. Additionally, the model has turned overweight financials and underweight consumer discretionary on the back of momentum. This overweight in financials is now in line with our global sector recommendations published yesterday. For more details on the model, please see the Special Report "Introducing The GAA Equity Sector Selection Model," July 27, 2016 available at https://gaa.bcaresearch.com. Xiaoli Tang, Associate Vice President xiaoli@bcaresearch.com Aditya Kurian, Research Analyst adityak@bcaresearch.com

Our new Revealed Preference Indicator (RPI) is the latest installment in our ongoing research into trading rules that can augment our top-down macro approach to asset allocation. The RPI borrows from Paul Samuelson's "revealed preference" theory of consumer behavior to market behavior. It combines the idea of market momentum with valuation and the monetary policy backdrop. A trading rule for the stock/bond allocation based on the RPI outperforms traditional technical, monetary, and valuation indicators. It provides a powerful bullish signal if positive equity market momentum lines up with positive signals from policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. This model adds value on its own, but we feel that it is best used in conjunction with other indicators designed to improve performance around major market turning points. Future research will experiment with combining the RPI with other indicators to further enhance performance. In the meantime, we will present the RPI's signals each month in Section III of the monthly publication. As with all indicators and models, however, the RPI is only one input to our decision process. We base our asset allocation decision on a combination of indicators, macro themes, detailed data analysis and judgment. In 1938, economist Paul Samuelson published a paper entitled, "A Note on the Pure Theory of Consumer's Behavior," in which he outlined an alternative to the well-known economic principle, utility theory. He dubbed his work "revealed preference theory."1 His goal was to redefine utility - a measure of consumer satisfaction with a good or service - by observing behavior. He posited that when consumers reveal their preferences by buying one item rather than another, they reveal the way in which they maximize happiness or satisfaction. For instance, one can measure preferences via experiments in which a subject is given $200 and the choice between two brands of shoes at different prices. Repeating the exercise at different levels of income and relative prices generates "preference axioms." Samuelson's theory has many more layers of complexity, but this Special Report focuses on modelling investors' preferences through observed behavior. Borrowing from Samuelson's reasoning, we developed a methodology to identify investors' axioms of preference for equities and bonds at different levels of incomes and prices. Then we compared investors' real actions with those anticipated by our methodology. This allowed us to generalize our findings and analyze the effects on a portfolio of equities and bonds. The main finding of our statistical exercise is that asset allocators can profit from understanding how short-term moves are linked to the market's revealed preferences at different times during the economic cycle. We then use the results to construct an indicator and a trading rule that not only outperforms a buy-and-hold strategy by a wide margin, it outperforms other traditional trading rules as well. Building A Revealed Preference Model Our primary objective in constructing a revealed preference indicator (RPI) is to understand: (1) how market preferences shape the behavior of investors; and (2) how they ultimately affect future returns. To do so, we broke down our analysis into three key areas: Part I identifies market preferences for different levels of income and price in the economy. Part II defines a general investment strategy that utilizes historic preferences and short-term market movements as a market timing tool. Part III optimizes the RPI and compares its historical track record with a buy-and-hold asset allocation and trading rules based on other indicators. Part I - Developing The Framework The first step in building the RPI is to establish the proper control variables. We limited our basket of investable "goods" to U.S. equities and 10-year Treasurys. We also need a variable that is analogous to the income measure that Samuelson used in his study. For the choice facing investors who are deciding between buying two financial assets, we believe that a measure of market "liquidity" is more appropriate than income. By this we do not mean the ease by which financial assets can be bought and sold. Rather, it is "funding liquidity", or how easily it is to borrow to invest. BCA often uses the four phases of the Fed cycle, as interest rates fluctuate around the equilibrium level, as a measure of funding liquidity (Chart II-1):2 Chart II-1Fed Funds Rate As A Proxy For Income Fed Funds Rate As A Proxy For Income Fed Funds Rate As A Proxy For Income Phase I = Policy is accommodative but the fed funds rate is rising. Phase II = Policy is tight and the Fed is still tightening. Phase III = Policy is tight but the Fed is cutting rates. Phase IV = Policy is easy and the Fed is cutting rates. The rationale for using the fed funds-rate cycle as a proxy for income is that, when interest rates are below equilibrium, monetary conditions are accommodative. Leverage is easy to obtain because there is plenty of liquidity (income) to fund investments. When conditions are tight, funding liquidity is relatively scarce. To measure relative prices, we first divided the S&P 500 price index by its 12-month moving average and second, we took the inverse of the 10-year Treasury yield divided by 12-month moving average.3 We then used the ratio of these two deviations-from-trend to construct a relative price measure (Chart II-2). This ratio provides a single measure of how expensive stocks and bonds are, not only to each other, but to their own history as well. We then grouped the relative price data into four sets of percentiles, or buckets, shown in Table II-1. Stocks are expensive and bonds cheap at the top of the table, while the reverse is true at the bottom. Chart II-2Constructing A Single Price Measure For Equities And Bonds Constructing A Single Price Measure For Equities And Bonds Constructing A Single Price Measure For Equities And Bonds Table II-1Distribution Of Relative Price July 2017 July 2017 Table II-2A presents the average historical monthly percent change in stock prices for each combination of the four relative price and liquidity buckets over the entire dataset. In the fourth phase of the Fed cycle (when monetary conditions are easy and the Fed is still cutting interest rates), and when relative prices are in the first bucket (i.e. stocks are expensive), the average stock price increase during the month was slightly above 1% percent. Table II-2B provides the same breakdown for the average change in bond yields (shown in basis points, not returns). Tables II-2A and II-2B are recalculated at each point in time - meaning that we used an expanding sample to calculate the price buckets, and updated the results for the ensuing price or yield movements as new data are added. That way, we completely avoid the advantage of hindsight. To simplify our methodology, we coded the results to end up with the stock and bond returns for the 16 different combinations of Fed and relative price buckets. Table II-3 uses the results from Tables II-2A and II-2B in the last period of history as an example. The "Liquidity" and "Price" columns indicate the bucket (e.g. price in bucket 1 and liquidity in bucket 1). The "Stocks" and "Bonds" columns are coded as "1" if the asset appreciated during the month given the indicated liquidity/price bucket, and a "0" if it depreciated that month. Table II-2AEquity Market Reactions At Given Levels Of Price And Liquidity July 2017 July 2017 Table II-2BTreasury Market Reactions At Given Levels Of Price And Liquidity July 2017 July 2017 Chart II-3Revealing What Investors Prefer July 2017 July 2017 Part II - Habits Create Expectations It is important to keep in mind that the objective of our revealed preference model is not to use the revealed market preferences as forecasts but rather to examine what happens when investors decide to follow or ignore them. Our hypothesis in building this model is that, when investors go against their historical preferences, the result should be interpreted as short-term noise. It is only when preferences and (subsequent) short-term market moves are aligned that we should heed the signal and invest accordingly. Table II-3 can be thought of as the market's revealed preference. Again, keep in mind that we allowed revealed preferences to change over time by recalculating it under our stretching-sample approach. The following steps detail how we used investor preferences to create a trading rule that verifies our hypothesis empirically: Step 1 - Expected vs. Actual: The first step is to examine how actual equity prices and bond yields behaved relative to their expected trajectory. We created two variables - one for equities and one for bonds. If revealed preference last month (t-1) suggested that the asset's return should be positive in the subsequent month (t), and it indeed turned out to be positive in period t, then we coded month t as "1." If both the revealed preference and the actual outcome were negative, we coded it as "-1." If they did not match, the code is "0" (in other words, the market did not follow the typical historical revealed preference). Thus we have two time series, one for bonds and one for stocks, which are made up of 1s, -1s and zeros. Step 2 - Bullish, Bearish, and Neutral: We combined the coded series for stocks and bonds to encompass the nine possible outcomes in our model (i.e. both bonds and stocks can have a value in any month of 1, 0 or -1, providing 9 different combinations). Table II-4 presents the nine outcomes along with the asset allocation that would have maximized investor returns based on our historical analysis. For example, investors were paid to be overweight equities when equities and bonds have a code of "1" and "-1," respectively (top row in Table II-4). In other words, stocks tended to outperform bonds when revealed preferences from the month before predicted rising stock prices and rising bond yields, and these predictions were confirmed. Table II-4Understanding The Signals From Preferences July 2017 July 2017 If revealed preference is not confirmed for both bonds and stocks, then it is best for investors to stand aside with a benchmark allocation. Step 3 - "If It Don't Make Dollars, It Don't Make Sense": To test whether our theory would add strategic value, we computed a trading rule to see how well it performed against a benchmark portfolio of 50% equities and 50% Treasurys. The trading rule was computed as follows: when the revealed preference for equities is positive (at time t-1) and this signal is confirmed in t, then in t+1 we allocate 100% to the S&P 500 and 0% to Treasurys. When the revealed equity preference signal is correctly bearish, we removed all exposure to equities and allocated 100% to Treasurys. When the signal was neutral, we kept a benchmark allocation of 50% equities and 50% Treasurys. Chart II-3 shows that this trading rule outperforms the benchmark, confirming our initial hypothesis - one should fade the short-term movements when investors go against their preferences, and only follow the signals when those movements align with historical preferences. History shows that investors tend to underperform in terms of the stock/bond allocation when they deviate from their revealed preference. Chart II-3Correctly Gauging How Investors Behave Pays Off Correctly Gauging How Investors Behave Pays Off Correctly Gauging How Investors Behave Pays Off Part III - Validating The Results One drawback is that this trading rule would require frequent portfolio allocation changes every month, as shown in Chart II-4. As such, we constructed a smoothed version by imposing the rule that asset allocation is unchanged unless the model provides a new signal for two months in a row (Chart II-5).4 These restrictions not only dramatically reduced the frequency of the asset allocation adjustments, but it also augmented historical cumulative excess returns (Chart II-6). Chart II-4Revealed Preference Indicator Is Inherently Volatile Revealed Preference Indicator Is Inherently Volatile Revealed Preference Indicator Is Inherently Volatile Chart II-5Removing Some Of The Noise Removing Some Of The Noise Removing Some Of The Noise Any new indicator of course must be able to outperform a buy-and-hold strategy to be useful but it is also interesting to see how its performance ranks compared to a set of random portfolios. This way, we can identify if the indicator truly provides additional information. Random portfolios are generated using a monthly allocations of 100% or 0% to equities, with the remainder in Treasurys. Chart II-7 shows the performance of the smoothed indicator versus a set of 1,000 randomly generated portfolios. Chart II-6Once Smoothed, The RPI Truly Shines Once Smoothed, The RPI Truly Shines Once Smoothed, The RPI Truly Shines Chart II-7The RPI Adds A Significant Amount Of Information The RPI Adds A Significant Amount Of Information The RPI Adds A Significant Amount Of Information We compared the indicator's trading rule to simple moving averages or BCA's other indicators. We also wanted to ensure that the RPI adds value beyond investing based strictly on the four phases of the liquidity cycle or based on relative value alone. We therefore compared the track record of the RPI trading rule to rules that are based on: (1) the deviation of the S&P 500 from its 12-month moving trend; (2) BCA's monetary conditions indicator; (3) BCA's valuation indicator; (4) BCA's technical indicator; (5) the four phases of the Fed cycle; and (6) the relative price index. Charts II-8A and II-8B highlights that RPI indeed impressively dominates the other trading rules. The one exception is that, during the Great Recession, the model's performance fell to roughly match the performance of a S&P 500 technical trading rule. Chart II-8AThe RPI Outperforms The Sum Of Its Parts... July 2017 July 2017 Chart II-8B...As Well As Other Indicators July 2017 July 2017 Part IV - Conclusions The RPI is the latest installment in our ongoing research into trading rules that can augment our top-down macro approach to asset allocation. Quite simply, it combines the idea of market momentum with valuation and policy measures. It provides a powerful bullish signal if positive market momentum lines up with constructive signals from the policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. This model adds value on its own, but we feel that it will best be used in conjunction with other indicators designed to improve performance around major market turning points. Future research will experiment with combining the RPI with other indicators to further enhance performance. In the meantime, we will present the RPI's signals each month in Section III of the monthly publication. As with all indicators and models, however, the RPI is only one input to our decision process. We base our asset allocation decision on a combination of indicators, macro themes, detailed data analysis and judgment. The indicator's current reading for stocks versus bonds, at benchmark, is more conservative than our official recommendation. The benchmark reading reflects the fact that equities are overvalued and that investors have deviated from their preferences in their past two quarters. David Boucher Associate Vice President Quantitative Strategist 1 For more information, please see P. A. Samuelson, "A Note on the Pure Theory of Consumer's Behavior," Economica 5:17 (1938), pp. 61-71. 2 Please see U.S. Investment Strategy Special Report "Stocks And The Fed Funds Rate Cycle," dated December 23, 2013, available at usis.bcaresearch.com 3 We tested a few other measures, most notably the stock-to-bond total return ratio (measured by comparing each asset's total returns), but the chosen measures provided the best and most robust results. 4 We conducted a statistical exercise to validate and optimize the allocations in Table 4 to provide a smoother performance.

Highlights A whiff of global deflation shook-up financial markets in June, driven by melting oil prices and a startling May U.S. CPI report. Nonetheless, we have not changed our recommended asset allocation. Bond markets have over-discounted the impact of the commodity price weakness, especially with regard to Fed policy and long-term inflation expectations in the major countries. We do not see the selling pressure in the commodity pits as a harbinger of slower global growth. Above-trend growth in the U.S. is likely in the second half of the year, along with continuing robust activity at the global level. Oil prices should rebound, based on our view that consumption will outstrip production in the second half of the year; the surprise will be how strong oil prices are in the coming months. The FOMC appears more determined than in the past to stick with the current policy normalization timetable. Unemployment will edge further below the full-employment level if the FOMC does not slow the pace of job creation. We believe that the labor market is tight enough to gradually push up inflation. Together with a rebound in the commodity pits, this means that the recent bond rally will reverse. Soft U.S. CPI readings are a challenge to our view. The Fed will delay the next rate hike into next year if core inflation does not move up in the next few months. The equity market is vulnerable to unforeseen shocks given stretched valuation. Nonetheless, none of the main indicators that have provided leading information in the past are warning of an equity bear market. The profit backdrop remains constructive. Our base case is that stocks beat bonds and cash for the remainder of 2017. We expect to trim exposure to equities next year, but the evolution of a number of indicators will influence the timing. The same is true for corporate bonds. The dollar's bull phase has one more upleg left. Japanese, European and U.K. equities will outperform the U.S. in local currency terms. Feature A whiff of global deflation shook-up financial markets in June, driven by melting oil prices and a startling May U.S. CPI report. Investors quickly concluded that the Fed will have to proceed even more slowly in terms of its policy normalization plan which, in turn, sent the dollar and global bond yields sharply lower. Equity indexes held up because of the dollar and bond yield "relief valves". Stocks are also benefiting from the continuing rebound in corporate earnings growth in the major economies. Nonetheless, the commodity pullback and soft U.S. inflation data are a challenge to our reflation theme, which includes a final upleg in the U.S. dollar and a negative view on bond prices. We believe that markets have over-discounted the impact of the commodity price weakness, especially with regard to Fed policy and long-term inflation expectations in the major countries. Above-trend growth in the U.S. is likely in the second half of the year, along with continuing robust activity at the global level. We also think that the FOMC is more determined than in the past to stick with the current policy normalization timetable. The bottom line is that we are not changing our recommended asset allocation based on June's market action. We remain overweight stocks and corporate bonds relative to government bonds and cash. We are also short duration and long the dollar. A key risk to our asset allocation relates to our contrarily bullish view on oil prices. Oil Drove The Bond Rally... The decline in long-term bond yields since March reflected in large part a drop in inflation expectations (Chart I-1). BCA's fixed-income strategists point out that the slump in long-term inflation expectations has been widespread across the major countries, irrespective of whether actual inflation is trending up or down.1 Core inflation has moved lower in the U.S., Japan, Canada and (slightly) in the Eurozone, but has increased in Australia and the U.K. In terms of diffusion indexes, which often lead core inflation, they are falling in the U.S., Japan and Canada, but are rising in the U.K., the Eurozone and Australia (Chart I-2). Chart I-1 Inflation Expectations Drive Bond Rally Inflation Expectations Drive Bond Rally Inflation Expectations Drive Bond Rally Chart I-2Diverging Inflation Trends Diverging Inflation Trends Diverging Inflation Trends Given all these diverging signals within the national inflation data, it is odd that there has been such a uniform decline in inflation expectations across the major bond markets. That leads us to look to the commodity price decline as the main driver of the downshift in expectations. Short-term moves in oil prices should not affect long-term inflation expectations, but in practice the correlation has been strong since the plunge in oil prices beginning in 2014. Weaker oil and other commodity prices have also fed investor concerns that global growth is waning. We see little evidence of any slowdown in global growth, although some leading indicators have softened. Key monthly data such as industrial production, retail sales and capital goods orders reveal an acceleration in growth for the advanced economies as a group (Chart I-3). There has also been a general upgrading of the consensus growth forecast for the major countries and for the world in both 2017 and 2018 (Chart I-4). This is unlike previous years, when growth forecasts started the year high, only to be slashed as the year progressed. Chart I-3No Slowdown In Advanced Economies No Slowdown In Advanced Economies No Slowdown In Advanced Economies Chart I-4Growth Expectations Revised Up Growth Expectations Revised Up Growth Expectations Revised Up ...But Watch Out For A Reversal The implication is that we do not see the selling pressure in the commodity pits as a harbinger of slower global growth. Nonetheless, the mini oil meltdown in June went against our medium-term bullish view. In a recent report,2 our Energy Sector Strategy team noted that investors are confused about conflicting supply signals in oil markets. Traders do not yet see the physical shortage that the IEA/EIA/OPEC and BCA's top-down supply & demand analyses argue will prevail in the coming months. Chart I-5Falling Inventories To Drive Oil Rebound Falling Inventories To Drive Oil Rebound Falling Inventories To Drive Oil Rebound The investment community is being overly pessimistic in our view. The coalition led by the Saudi Arabia and Russia will have removed 1.4 MMB/d of production on average from the market between January 2017 and end-March 2018, versus peak production in November of last year. This will be diluted somewhat by the Libyan and U.S. production gains, but the increased production will not be sufficient to counter the OPEC/Russia cuts entirely. We expect global production to increase by only 0.7 MMB/d in 2017, an estimate that includes rapid increases in U.S. shale output. Meanwhile, we expect consumption to grow by 1.5 MMB/d, implying that oil inventories will fall over the remainder of this year. If history is any guide, this will lead to a rebound in oil prices (Chart I-5). It will be quite a shock to markets if crude reaches $60/bbl by December as we expect. As for base metals, it appears that the correction is largely related to reduced speculative demand rather than weak global and/or Chinese demand. It is true that the Chinese economy has slipped a notch according to some measures, such as housing starts and M2 growth. Nonetheless, the government remains cognizant of the risks of tightening policy too aggressively, especially with the National Party Congress slated for this autumn. The PBoC injected 250 billion yuan into the financial system in June and fiscal policy has been eased. Real-time measures of industrial activity such as railway freight traffic, excavator sales, and electricity production remain upbeat. Retail sales continue to expand at a healthy clip. Export growth is accelerating thanks to a weaker currency and stronger global activity. Given that many investors remain concerned about a hard landing in China, the bar for positive surprises is comfortably low. If China can clear this bar, as we expect it will, it will be good news for the commodity currencies and other commodity plays. A rebound in base metal and, especially, oil prices would boost global inflation expectations and bond yields, especially since inflation expectations have fallen too far relative to underlying non-energy inflation pressures. This forecast also applies to the U.S. bond market, although there was more to the soft May CPI report than oil prices. Is The Fed's Inflation Target Credible? Investors are questioning whether the Fed has the ability to reach its inflation goals. Is it possible that the U.S. is following Japan's roadmap where even an over-heated labor market is insufficient to generate any meaningful inflation? We argued above that the moderation in inflation expectations in the major markets was mostly related to the decline in commodity prices. However, in the U.S., it also reflected a fairly widespread pullback in CPI inflation this year. This is contrary to Fed Chair Yellen's assertion that most of it reflects special factors such as wireless telecommunications prices. The deceleration in inflation began around the start of the year. The three-month rate of change of the headline index fell by more than five percentage points between January and May, of which energy accounts for 3.3 percentage points. The deceleration in the core rate was a less severe, but still substantial at 2.8 percentage points. Table I-1 presents the components of the CPI that made the largest contribution to the deceleration in core inflation. Motor vehicles, owners' equivalent rent, apparel, recreation, wireless telecom and medical care services accounted for 1.2 percentage points as a group. However, many other sectors contributed in a small way to the overall deceleration of core inflation in the first five months of the year. Table I-1Key Drivers Of U.S. Core Inflation Deceleration In 2017 July 2017 July 2017 Some special factors were at play. The moderation in rent inflation likely reflects the bottoming of the vacancy rate. Discounting in the auto sector is not a surprise given weak sales. Wireless prices can be viewed as a special case as well. Nonetheless, the breadth and suddenness of the deceleration in core inflation across such diverse sectors, some unrelated to labor markets, commodity prices, the dollar or on-line shopping, is worrying. The disinflation this year in the Fed's preferred measure, the PCE price index, is not as extended but the data are published almost a month behind the CPI data. A diffusion index made up of the components of the PCE index is still in positive territory, unlike the CPI's diffusion index (Chart I-6). Nonetheless, the CPI data suggest that core PCE inflation will edge lower when the May data are released at the end of June. There has also been a moderation in some of the wage inflation data, such as average hourly earnings (Chart I-7). The slowdown has been fairly widespread across manufacturing and services. However, the soft patch already appears to be over; 3-month rates of change have firmed almost across the board (retail is a major exception). There is no slowdown evident at all in the better-constructed Employment Cost Index (ECI) as of the first quarter (Chart I-8). The ECI is adjusted to avoid compositional effects that can distort the aggregate index. The related diffusion indexes also remain constructive. Chart I-6PCE Inflation Rate To Follow CPI Lower PCE Inflation Rate To Follow CPI Lower PCE Inflation Rate To Follow CPI Lower Chart I-7AHE SoftPatch Appears Over... AHE SoftPatch Appears Over... AHE SoftPatch Appears Over... Chart I-8...And The ECI Marches Higher ...And The ECI Marches Higher ...And The ECI Marches Higher We conclude from these and other wage measures that the Phillips curve is still operating in the U.S. Admittedly, the curve appears to be quite flat, which means it is difficult to generate inflation even with a tight labor market. Nonetheless, the relationship between the ECI and various measures of labor market tightness shown in Chart I-8 does not appear to have broken down. The percentage of U.S. states with unemployment below the Fed's estimate of full employment jumped to 70% in May. Anything over 60% in the past has been associated with wage pressure (Chart I-9). The bottom line is that, while we are concerned about the breadth of the soft patch in the consumer price data, we are in agreement with the Fed hawks that the labor market is tight enough to gradually push up inflation. We are willing at this point to chalk up the recent drop in core inflation partly to randomness in the data, and partly to lagged effects of the slowdown in real GDP growth in the first half of 2016 (Chart I-10). Admittedly, however, the U.S. inflation reports in the coming months are a key risk to our reflation-related asset allocation. Chart I-9More Than 70% Of U.S. States Have Excess Labor Demand More Than 70% Of U.S. States Have Excess Labor Demand More Than 70% Of U.S. States Have Excess Labor Demand Chart I-10Financial Conditions Point To Faster Growth And Inflation Financial Conditions Point To Faster Growth And Inflation Financial Conditions Point To Faster Growth And Inflation What Will The Fed Do? The CPI data have certainly rattled some members of the FOMC. Federal Reserve Bank Presidents Kaplan and Kashkari, for example, believe that the Fed needs to be patient to ensure that the inflation pullback is temporary. However, the June FOMC Statement and Yellen's press conference suggested that the consensus is determined to stick with the current tightening timetable in terms of rate hikes and balance sheet adjustment. She stressed that the FOMC makes policy for the "medium term," and should not over-react to short-term wiggles in the data. Vice President Dudley echoed this view in recent comments he made to the press. The Fed has been quick to back away from planned rate hikes at the first hint of trouble in recent years. However, it appears that the reaction function has changed, now that the labor market is at full employment. This is especially the case because financial conditions have eased further, despite the June rate hike. Unemployment will edge further below the full-employment level if the FOMC does not slow the pace of job creation. Policymakers know that the Fed has had little success in the past when it tried to nudge unemployment higher in order to relieve budding inflation pressure; these attempts almost always ended in recession. Dudley added that "...pausing policy now could raise the risk of inflation surging and hurting the economy." Other FOMC members are worried that financial stability risk will build if the low-rate environment extends much further. The bottom line is that we expect the Fed to stick with the game-plan for now. The FOMC will begin shrinking the balance sheet in September, but will wait until December for the next rate hike. That said, a stubbornly low inflation rate in the coming months would likely see the FOMC postpone the next rate increase into next year. Where Next For Bonds? We see three possible scenarios for the bond market: Reflation Returns: Weak recent inflation readings are nothing more than a lagged response to last year's deceleration in economic growth. U.S. growth accelerates in the second half, unemployment falls further and both wage growth and inflation pick up. Oil inventories begin to contract and prices head higher. The FOMC is vindicated in its inflation view and proceeds with the current rate hike and balance sheet adjustment agenda. Investors receive a "wake up call" from the Fed, bond prices get hit and recent curve-flattening trend reverses. Fed Capitulates: The U.S. labor market continues to tighten, but core PCE inflation is still close to 1½% by the September FOMC meeting. We would expect the Fed to lower its forecasted rate hike path, signaling that no further rate hikes are likely in 2017. Long-maturity real yields would fall in this scenario, although long-term inflation expectations could rise to the extent that the Fed's more dovish tilt will weaken the dollar and generate more inflation in the medium term. Nominal yields may not end up moving much in this scenario. A Policy Mistake: If core inflation remains low between now and the September FOMC meeting and the Fed continues to write-off low inflation as transitory, signaling its intention to stick to its current projected rate hike path, then the market would begin to discount a "policy mistake" scenario. The yield curve would flatten and long-maturity nominal yields would fall, led by tighter TIPS breakevens. In terms of probabilities, we would characterize Scenario 1 as our base case, Scenario 2 as unlikely and Scenario 3 as a tail risk. We remain short-duration in anticipation of a rebound in long-term inflation expectations and higher yields. A bond selloff, however, should not present a major headwind for stocks as long as the earnings backdrop remains constructive. Will The Real Profit Margin Please Stand Up For some time we have been highlighting the importance of the mini-cycle in U.S. earnings growth; the corporate sector is in a catch-up phase following last year's profit recession, a trend that extends beyond the energy patch. EPS growth has surged this year on the back of somewhat stronger sales and rising S&P 500 margins. The National Accounts (NIPA) data, however, paint a different picture. Earnings growth for the entire corporate sector fell sharply in the first quarter and margins continued to slide. If the NIPA data are telling the true story, then the equity market is in big trouble because it suggests that the earnings outlook is much weaker than what is discounted in stock prices. There are many definitional differences that make it difficult to reconcile the NIPA and S&P data.3 Nonetheless, we can make some general observations. Chart I-11 presents the 4-quarter growth rate of NIPA profits4 and a proxy for aggregate S&P earnings. For the latter, we multiplied earnings-per-share by the divisor to obtain an estimate of the level of aggregate earnings in dollar terms (i.e. not on a per-share basis). The bottom panel of Chart I-11 compares the level of profits, each indexed to be 100 in 2011 Q1. The charts highlight that, while there have been marked differences in annual growth rates between the two measures in some years, the levels ended up being close to the same point in the first quarter of 2017. The dip in NIPA profit growth in the first quarter was not reflected in the S&P measure. It appears that this is partly due to different profiles for profit growth in the energy and financials sectors. That said, broadly speaking, it does not appear that the difference in margins is due to a significant divergence in aggregate profits. It turns out that most of the margin divergence is related to the denominator of the calculation (Chart I-12). The NIPA denominator is corporate sector Gross Domestic Product (GDP). This is a value-added concept that is quite different from sales. It is not clear why, but GDP has grown much faster than sales since the end of 2014. It appears to us that the S&P data are telling the correct story at the moment. After all, sales are straight forward to measure, while value added is complicated to construct. The fact that sales are growing slowly is not a bullish point for stocks. Nonetheless, it does not appear that financial engineering has distorted bottom-up company data to such an extent that the S&P data are signaling strong profit growth when the reality is the opposite. We expect the secular mean-reversion of margins to re-assert itself in the S&P data, perhaps beginning early in 2018. Nonetheless, the profit backdrop remains positive for stocks for now. The same is true in the Eurozone and Japan, where margins are also rising. It is worrying that a large part of this year's U.S. equity market advance has been concentrated in a small number of stocks, but that belies the breadth of the profit recovery (Chart I-13). The proportion of S&P industry groups with rising earnings estimates is above 75%. Such widespread participation is consistent with ongoing upward revisions to 12-month forward earnings estimates. Chart I-11S&P And NIPA Profit Comparison S&P And NIPA Profit Comparison S&P And NIPA Profit Comparison Chart I-12Denominator Explains S&P/NIPA Margin Divergence Denominator Explains S&P/NIPA Margin Divergence Denominator Explains S&P/NIPA Margin Divergence Chart I-13Positive Earnings Revisions Are Broadly Based Positive Earnings Revisions Are Broadly Based Positive Earnings Revisions Are Broadly Based The solid earnings backdrop is the main reason we remain overweight stocks versus bonds and cash. Of course, given poor valuation, we must be extra vigilant in watching for warning signs of a bear market. Valuation has never been good leading indicator for bear markets, but it does provide information on the risks. Monitoring The Bear Market Barometer BCA's Chief Economist, Martin Barnes, highlighted the best "equity bear market" indicators to watch in a 2014 Special Report.5 He noted that no two bear markets are the same, and that there are no indicators that have reliably heralded bear phases. Nonetheless, there are some common elements. The safest time to invest in the market is when monetary conditions are favorable, there are no signs of a looming economic downturn, there is not extreme overvaluation, and technical indicators are not flashing red. Some indicators related to each of these fundamental factors are shown in Chart I-14: Chart I-14Equity Bear Market Indicators Equity Bear Market Indicators Equity Bear Market Indicators Monetary Conditions: The yield curve is quite flat by historical standards, but it is far from inverting. Moreover, real short-term interest rates are normally substantially higher than today, and above 2%, when bear markets commence. Excess liquidity, which we define as M2 growth less nominal GDP growth, is also still well above the zero line, a threshold that has warned of a downturn in stock prices in the past. Valuation: Our composite valuation indicator is still shy of the +1 standard deviation level that defines over-valued. However, this is because of the components that compare equity prices to bond yields. The other three components of the equity indicator, which are unrelated to bond yields, suggest that stock valuation is quite stretched. Economic Outlook: Economic data such as the leading economic indicator and ISM have been unreliable bear market signals. That said, we do not see anything that suggests that a recession is on the horizon. Indeed, U.S. growth is likely to remain above-trend in the second half of the year based on its relationship with financial conditions. Technical conditions: Sentiment is elevated, which is bearish from a contrary perspective. However, breadth, the deviation from the 40-week moving average, and our composite technical indicator are not flashing red. Earnings: Trends in earnings and margins did not provide any additional reliable signals for timing equity market downturns in the past. Still, it has been a bad sign when EPS growth topped out. And this has often been preceded by a peak in industrial production growth. We expect U.S. EPS growth to continue to accelerate for at least a few more months, but are watching industrial production closely. EPS growth in Japan and the Eurozone will likely peak after the U.S., since these markets are not as advanced in the profit rebound. The bottom line is that the equity market is vulnerable to unforeseen shocks given stretched valuation. Nonetheless, none of the main indicators that have provided some leading information in the past are warning of an equity bear market. Investment Conclusions The major world bourses remain in a sweet spot because of the mini cyclical rebound in profits. One can imagine many scenarios in which equities suffer a major correction or bear phase. However, stocks would likely perform well under the two most likely scenarios for the remainder of the year. If U.S. and global growth disappoint, the combination of low bond yields and still-robust earnings growth will continue to support prices. Conversely, if world growth remains solid and the U.S. picks up, as we expect, then bond yields will rise but investors will pencil-in an even stronger profit advance over the next year. Of course, this win-win situation for stocks will not last forever. Perhaps paradoxically, the economic cycle could be shortened if the U.S. Congress gets around to passing a bill that imparts fiscal stimulus in 2018. The Fed would have to respond with a more aggressive tightening timetable, setting the stage for the next recession. In contrast, the economic cycle would be further stretched out in the absence of fiscal stimulus, keeping alive for a while longer the lackluster growth/low inflation/low bond yield backdrop that has been favorable for the equity market. We are watching the indicators discussed above to time the exit from our pro-risk asset allocation that favors stocks and corporate bonds to government bonds and cash. As for the duration call, the whiff of deflation that has depressed bond yields over the past month is overdone. Investors have also become too complacent on the Fed. We expect that the recent drop in commodity prices, especially oil, will reverse. If this view is correct, it means that the cyclical bull phase in the dollar is not over because market expectations for the pace of Fed rate hikes will rise relative to expectations in the other major economies (with the exception of Canada). We are still looking for a 10% dollar appreciation. It also means that Treasurys will underperform JGBs and Bunds within currency-hedged fixed-income portfolios. We expect the Eurostoxx 600 and the Nikkei indexes to outperform the S&P 500 this year in local currencies, despite our constructive view on U.S. growth. Stocks are cheaper in the former two markets. Moreover, both Japan and the Eurozone are earlier in the profit mini-cycle, which means that there is room for catch-up versus the U.S. over the next 6-12 months when growth in the latter tops-out. The prospect of structural reform in France is also constructive for European stocks, following the election of a reformist legislature in June. However, the upcoming Italian election warrants close scrutiny. The key risk to this base case is our view that oil prices will rebound. This is clearly a non-consensus call. If OPEC production cuts are unable to overwhelm the rise in U.S. shale output, then inventories will remain elevated and oil prices could move even lower in the near term. Our bullish equity view would be fine in this case, but the bond bear market and dollar appreciation we expect would at least be delayed. Finally, a few words on the U.K. Our geopolitical experts highlight two key points related to June's election outcome: fiscal austerity is dead and the U.K. will pursue a "softer" variety of Brexit. This combination should provide a relatively benign backdrop for U.K. stocks and the economy over the next year. Nonetheless, the cloud of uncertainty hanging over the U.K. is large enough to keep the Bank of England (BoE) on hold. Some BoE hawks are agitating for tighter policy due to the worsening inflation overshoot, but it will probably be some time before the consensus on the Monetary Policy Committee shifts in favor of rate hikes. This means that it is too early to position for gilt underperformance within fixed-income portfolios. Sterling weakness looks overdone, although we do not see much upside either. As long as Brexit talks do not become acrimonious (which is our view), the U.K. stock market should be one of the outperformers in local currency terms among the major developed markets. Mark McClellan Senior Vice President The Bank Credit Analyst June 29, 2017 Next Report: July 27, 2017 1 For more discussion, see Alternative Facts in the Bond Market at BCA Global Fixed Income Strategy Weekly Report, dated June 13, 2017 available at gfis.bcaresearch.com 2 Please see Energy Sector Strategy Weekly Report, "Views from the Road," dated June 21, 2017, available at nrg.bcaresearch.com 3 The first problem is that the S&P data are expressed on a per-share basis. Moreover, the NIPA data adjusts for inventory and depreciation allowance. S&P margins are calculated using sales in the denominator, while we generally use GDP as the denominator for calculating NIPA profits. 4 The NIPA data shown include financials and profits earned overseas, as is the case for the S&P. 5 Please see BCA Special Report "Timing The Next Equity Bear Market," dated January 24, 2014, available at bcaresearch.com II. Preferences As Trading Constraints: A New Asset Allocation Indicator Our new Revealed Preference Indicator (RPI) is the latest installment in our ongoing research into trading rules that can augment our top-down macro approach to asset allocation. The RPI borrows from Paul Samuelson's "revealed preference" theory of consumer behavior to market behavior. It combines the idea of market momentum with valuation and the monetary policy backdrop. A trading rule for the stock/bond allocation based on the RPI outperforms traditional technical, monetary, and valuation indicators. It provides a powerful bullish signal if positive equity market momentum lines up with positive signals from policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. This model adds value on its own, but we feel that it is best used in conjunction with other indicators designed to improve performance around major market turning points. Future research will experiment with combining the RPI with other indicators to further enhance performance. In the meantime, we will present the RPI's signals each month in Section III of the monthly publication. As with all indicators and models, however, the RPI is only one input to our decision process. We base our asset allocation decision on a combination of indicators, macro themes, detailed data analysis and judgment. In 1938, economist Paul Samuelson published a paper entitled, "A Note on the Pure Theory of Consumer's Behavior," in which he outlined an alternative to the well-known economic principle, utility theory. He dubbed his work "revealed preference theory."1 His goal was to redefine utility - a measure of consumer satisfaction with a good or service - by observing behavior. He posited that when consumers reveal their preferences by buying one item rather than another, they reveal the way in which they maximize happiness or satisfaction. For instance, one can measure preferences via experiments in which a subject is given $200 and the choice between two brands of shoes at different prices. Repeating the exercise at different levels of income and relative prices generates "preference axioms." Samuelson's theory has many more layers of complexity, but this Special Report focuses on modelling investors' preferences through observed behavior. Borrowing from Samuelson's reasoning, we developed a methodology to identify investors' axioms of preference for equities and bonds at different levels of incomes and prices. Then we compared investors' real actions with those anticipated by our methodology. This allowed us to generalize our findings and analyze the effects on a portfolio of equities and bonds. The main finding of our statistical exercise is that asset allocators can profit from understanding how short-term moves are linked to the market's revealed preferences at different times during the economic cycle. We then use the results to construct an indicator and a trading rule that not only outperforms a buy-and-hold strategy by a wide margin, it outperforms other traditional trading rules as well. Building A Revealed Preference Model Our primary objective in constructing a revealed preference indicator (RPI) is to understand: (1) how market preferences shape the behavior of investors; and (2) how they ultimately affect future returns. To do so, we broke down our analysis into three key areas: Part I identifies market preferences for different levels of income and price in the economy. Part II defines a general investment strategy that utilizes historic preferences and short-term market movements as a market timing tool. Part III optimizes the RPI and compares its historical track record with a buy-and-hold asset allocation and trading rules based on other indicators. Part I - Developing The Framework The first step in building the RPI is to establish the proper control variables. We limited our basket of investable "goods" to U.S. equities and 10-year Treasurys. We also need a variable that is analogous to the income measure that Samuelson used in his study. For the choice facing investors who are deciding between buying two financial assets, we believe that a measure of market "liquidity" is more appropriate than income. By this we do not mean the ease by which financial assets can be bought and sold. Rather, it is "funding liquidity", or how easily it is to borrow to invest. BCA often uses the four phases of the Fed cycle, as interest rates fluctuate around the equilibrium level, as a measure of funding liquidity (Chart II-1):2 Chart II-1Fed Funds Rate As A Proxy For Income Fed Funds Rate As A Proxy For Income Fed Funds Rate As A Proxy For Income Phase I = Policy is accommodative but the fed funds rate is rising. Phase II = Policy is tight and the Fed is still tightening. Phase III = Policy is tight but the Fed is cutting rates. Phase IV = Policy is easy and the Fed is cutting rates. The rationale for using the fed funds-rate cycle as a proxy for income is that, when interest rates are below equilibrium, monetary conditions are accommodative. Leverage is easy to obtain because there is plenty of liquidity (income) to fund investments. When conditions are tight, funding liquidity is relatively scarce. To measure relative prices, we first divided the S&P 500 price index by its 12-month moving average and second, we took the inverse of the 10-year Treasury yield divided by 12-month moving average.3 We then used the ratio of these two deviations-from-trend to construct a relative price measure (Chart II-2). This ratio provides a single measure of how expensive stocks and bonds are, not only to each other, but to their own history as well. We then grouped the relative price data into four sets of percentiles, or buckets, shown in Table II-1. Stocks are expensive and bonds cheap at the top of the table, while the reverse is true at the bottom. Chart II-2Constructing A Single Price Measure For Equities And Bonds Constructing A Single Price Measure For Equities And Bonds Constructing A Single Price Measure For Equities And Bonds Table II-1Distribution Of Relative Price July 2017 July 2017 Table II-2A presents the average historical monthly percent change in stock prices for each combination of the four relative price and liquidity buckets over the entire dataset. In the fourth phase of the Fed cycle (when monetary conditions are easy and the Fed is still cutting interest rates), and when relative prices are in the first bucket (i.e. stocks are expensive), the average stock price increase during the month was slightly above 1% percent. Table II-2B provides the same breakdown for the average change in bond yields (shown in basis points, not returns). Tables II-2A and II-2B are recalculated at each point in time - meaning that we used an expanding sample to calculate the price buckets, and updated the results for the ensuing price or yield movements as new data are added. That way, we completely avoid the advantage of hindsight. To simplify our methodology, we coded the results to end up with the stock and bond returns for the 16 different combinations of Fed and relative price buckets. Table II-3 uses the results from Tables II-2A and II-2B in the last period of history as an example. The "Liquidity" and "Price" columns indicate the bucket (e.g. price in bucket 1 and liquidity in bucket 1). The "Stocks" and "Bonds" columns are coded as "1" if the asset appreciated during the month given the indicated liquidity/price bucket, and a "0" if it depreciated that month. Table II-2AEquity Market Reactions At Given Levels Of Price And Liquidity July 2017 July 2017 Table II-2BTreasury Market Reactions At Given Levels Of Price And Liquidity July 2017 July 2017 Chart II-3Revealing What Investors Prefer July 2017 July 2017 Part II - Habits Create Expectations It is important to keep in mind that the objective of our revealed preference model is not to use the revealed market preferences as forecasts but rather to examine what happens when investors decide to follow or ignore them. Our hypothesis in building this model is that, when investors go against their historical preferences, the result should be interpreted as short-term noise. It is only when preferences and (subsequent) short-term market moves are aligned that we should heed the signal and invest accordingly. Table II-3 can be thought of as the market's revealed preference. Again, keep in mind that we allowed revealed preferences to change over time by recalculating it under our stretching-sample approach. The following steps detail how we used investor preferences to create a trading rule that verifies our hypothesis empirically: Step 1 - Expected vs. Actual: The first step is to examine how actual equity prices and bond yields behaved relative to their expected trajectory. We created two variables - one for equities and one for bonds. If revealed preference last month (t-1) suggested that the asset's return should be positive in the subsequent month (t), and it indeed turned out to be positive in period t, then we coded month t as "1." If both the revealed preference and the actual outcome were negative, we coded it as "-1." If they did not match, the code is "0" (in other words, the market did not follow the typical historical revealed preference). Thus we have two time series, one for bonds and one for stocks, which are made up of 1s, -1s and zeros. Step 2 - Bullish, Bearish, and Neutral: We combined the coded series for stocks and bonds to encompass the nine possible outcomes in our model (i.e. both bonds and stocks can have a value in any month of 1, 0 or -1, providing 9 different combinations). Table II-4 presents the nine outcomes along with the asset allocation that would have maximized investor returns based on our historical analysis. For example, investors were paid to be overweight equities when equities and bonds have a code of "1" and "-1," respectively (top row in Table II-4). In other words, stocks tended to outperform bonds when revealed preferences from the month before predicted rising stock prices and rising bond yields, and these predictions were confirmed. Table II-4Understanding The Signals From Preferences July 2017 July 2017 If revealed preference is not confirmed for both bonds and stocks, then it is best for investors to stand aside with a benchmark allocation. Step 3 - "If It Don't Make Dollars, It Don't Make Sense": To test whether our theory would add strategic value, we computed a trading rule to see how well it performed against a benchmark portfolio of 50% equities and 50% Treasurys. The trading rule was computed as follows: when the revealed preference for equities is positive (at time t-1) and this signal is confirmed in t, then in t+1 we allocate 100% to the S&P 500 and 0% to Treasurys. When the revealed equity preference signal is correctly bearish, we removed all exposure to equities and allocated 100% to Treasurys. When the signal was neutral, we kept a benchmark allocation of 50% equities and 50% Treasurys. Chart II-3 shows that this trading rule outperforms the benchmark, confirming our initial hypothesis - one should fade the short-term movements when investors go against their preferences, and only follow the signals when those movements align with historical preferences. History shows that investors tend to underperform in terms of the stock/bond allocation when they deviate from their revealed preference. Chart II-3Correctly Gauging How Investors Behave Pays Off Correctly Gauging How Investors Behave Pays Off Correctly Gauging How Investors Behave Pays Off Part III - Validating The Results One drawback is that this trading rule would require frequent portfolio allocation changes every month, as shown in Chart II-4. As such, we constructed a smoothed version by imposing the rule that asset allocation is unchanged unless the model provides a new signal for two months in a row (Chart II-5).4 These restrictions not only dramatically reduced the frequency of the asset allocation adjustments, but it also augmented historical cumulative excess returns (Chart II-6). Chart II-4Revealed Preference Indicator Is Inherently Volatile Revealed Preference Indicator Is Inherently Volatile Revealed Preference Indicator Is Inherently Volatile Chart II-5Removing Some Of The Noise Removing Some Of The Noise Removing Some Of The Noise Any new indicator of course must be able to outperform a buy-and-hold strategy to be useful but it is also interesting to see how its performance ranks compared to a set of random portfolios. This way, we can identify if the indicator truly provides additional information. Random portfolios are generated using a monthly allocations of 100% or 0% to equities, with the remainder in Treasurys. Chart II-7 shows the performance of the smoothed indicator versus a set of 1,000 randomly generated portfolios. Chart II-6Once Smoothed, The RPI Truly Shines Once Smoothed, The RPI Truly Shines Once Smoothed, The RPI Truly Shines Chart II-7The RPI Adds A Significant Amount Of Information The RPI Adds A Significant Amount Of Information The RPI Adds A Significant Amount Of Information We compared the indicator's trading rule to simple moving averages or BCA's other indicators. We also wanted to ensure that the RPI adds value beyond investing based strictly on the four phases of the liquidity cycle or based on relative value alone. We therefore compared the track record of the RPI trading rule to rules that are based on: (1) the deviation of the S&P 500 from its 12-month moving trend; (2) BCA's monetary conditions indicator; (3) BCA's valuation indicator; (4) BCA's technical indicator; (5) the four phases of the Fed cycle; and (6) the relative price index. Charts II-8A and II-8B highlights that RPI indeed impressively dominates the other trading rules. The one exception is that, during the Great Recession, the model's performance fell to roughly match the performance of a S&P 500 technical trading rule. Chart II-8AThe RPI Outperforms The Sum Of Its Parts... July 2017 July 2017 Chart II-8B...As Well As Other Indicators July 2017 July 2017 Part IV - Conclusions The RPI is the latest installment in our ongoing research into trading rules that can augment our top-down macro approach to asset allocation. Quite simply, it combines the idea of market momentum with valuation and policy measures. It provides a powerful bullish signal if positive market momentum lines up with constructive signals from the policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. This model adds value on its own, but we feel that it will best be used in conjunction with other indicators designed to improve performance around major market turning points. Future research will experiment with combining the RPI with other indicators to further enhance performance. In the meantime, we will present the RPI's signals each month in Section III of the monthly publication. As with all indicators and models, however, the RPI is only one input to our decision process. We base our asset allocation decision on a combination of indicators, macro themes, detailed data analysis and judgment. The indicator's current reading for stocks versus bonds, at benchmark, is more conservative than our official recommendation. The benchmark reading reflects the fact that equities are overvalued and that investors have deviated from their preferences in their past two quarters. David Boucher Associate Vice President Quantitative Strategist 1 For more information, please see P. A. Samuelson, "A Note on the Pure Theory of Consumer's Behavior," Economica 5:17 (1938), pp. 61-71. 2 Please see U.S. Investment Strategy Special Report "Stocks And The Fed Funds Rate Cycle," dated December 23, 2013, available at usis.bcaresearch.com 3 We tested a few other measures, most notably the stock-to-bond total return ratio (measured by comparing each asset's total returns), but the chosen measures provided the best and most robust results. 4 We conducted a statistical exercise to validate and optimize the allocations in Table 4 to provide a smoother performance. III. Indicators And Reference Charts Thanks to the recent dollar and bond yield “relief valves”, the S&P 500 is stubbornly holding above the 2,400 level. The breakout above this level further stretched valuation metrics. Measures such as the Shiller P/E and price/book are at post tech-bubble highs. Stocks remain expensive based on our composite Valuation Index, although it is still shy of the +1 standard deviation level that demarcates over-valuation. This is because our composite indicator includes valuation measures that take into account the low level of interest rates. Of course, once interest rate normalization is well underway, these indicator will not look as favorable. It is good news for the equity market that our Monetary Indicator did not move further into negative territory over the past month. Indeed, the indicator has hooked up slightly and is sitting close to a neutral level. Our equity Technical Indicator remains constructive. Other measures, such as our Speculation Index, composite sentiment and the VIX suggest that equity investors are overly bullish from a contrary perspective. On the other hand, the U.S. earnings surprises diffusion index highlights that upside earnings surprises are broadly based. Our elevated U.S. Willingness-to-Pay (WTP) indicator ticked down from a high level this month, suggesting that ‘dry powder’ available to buy this market is depleted. This indicator tracks flows, and thus provides information on what investors are actually doing, as opposed to sentiment indexes that track how investors are feeling. Investors often say they are bullish but remain conservative in their asset allocation. In contrast to the U.S., the WTP indicators for both the Eurozone and Japan are rising from a low level. This suggests that a rotation into these equity markets is underway and has some ways to go. We remain overweight both the Eurozone and Japanese markets relative to the U.S. on a currency-hedged basis. The pull back in long-term bond yields since March was enough to “move the dial” in terms of the bond valuation or technical indicators. U.S. bond valuation has inched lower to fair value. However, we believe that fair value itself is moving higher as some of the economic headwinds fade. We also think that the FOMC is determined to stick with the current tightening timetable in terms of rate hikes and balance sheet adjustment, which support our negative view on bond prices. Now that oversold technical conditions have been unwound, it suggests that the consolidation phase for bond yields is largely complete. The trade-weighted dollar remains quite overvalued on a PPP basis, although less so by other measures. Technically, it is a bearish sign that the dollar moved lower and crossed its 200-day moving average. However, our Composite Technical Indicator highlights that overbought conditions have been worked off. We still believe the U.S. dollar’s bull phase has one more upleg left. Technical conditions are also benign in the commodity complex. Most commodities have shifted down over the last month to meet support at their 200-day moving averages. Base metals are due for a bounce, but we are most bullish on oil. EQUITIES: Chart III-1U.S. Equity Indicators U.S. Equity Indicators U.S. Equity Indicators Chart III-2Willingness To Pay For Risk Willingness To Pay For Risk Willingness To Pay For Risk Chart III-3U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators Chart III-4U.S. Stock Market Valuation U.S. Stock Market Valuation U.S. Stock Market Valuation Chart III-5U.S. Earnings U.S. Earnings U.S. Earnings Chart III-6Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance Chart III-7Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance FIXED INCOME: Chart III-8U.S. Treasurys and Valuations U.S. Treasurys and Valuations U.S. Treasurys and Valuations Chart III-9U.S. Treasury Indicators U.S. Treasury Indicators U.S. Treasury Indicators Chart III-10Selected U.S. Bond Yields Selected U.S. Bond Yields Selected U.S. Bond Yields Chart III-1110-Year Treasury Yield Components 10-Year Treasury Yield Components 10-Year Treasury Yield Components Chart III-12U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor Chart III-13Global Bonds: Developed Markets Global Bonds: Developed Markets Global Bonds: Developed Markets Chart III-14Global Bonds: Emerging Markets Global Bonds: Emerging Markets Global Bonds: Emerging Markets CURRENCIES: Chart III-15U.S. Dollar And PPP U.S. Dollar And PPP U.S. Dollar And PPP Chart III-16U.S. Dollar And Indicator U.S. Dollar And Indicator U.S. Dollar And Indicator Chart III-17U.S. Dollar Fundamentals U.S. Dollar Fundamentals U.S. Dollar Fundamentals Chart III-18Japanese Yen Technicals Japanese Yen Technicals Japanese Yen Technicals Chart III-19Euro Technicals Euro Technicals Euro Technicals Chart III-20Euro/Yen Technicals Euro/Yen Technicals Euro/Yen Technicals Chart III-21Euro/Pound Technicals Euro/Pound Technicals Euro/Pound Technicals COMMODITIES: Chart III-22Broad Commodity Indicators Broad Commodity Indicators Broad Commodity Indicators Chart III-23Commodity Prices Commodity Prices Commodity Prices Chart III-24Commodity Prices Commodity Prices Commodity Prices Chart III-25Commodity Sentiment Commodity Sentiment Commodity Sentiment Chart III-26Speculative Positioning Speculative Positioning Speculative Positioning ECONOMY: Chart III-27U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop Chart III-28U.S. Macro Snapshot U.S. Macro Snapshot U.S. Macro Snapshot Chart III-29U.S. Growth Outlook U.S. Growth Outlook U.S. Growth Outlook Chart III-30U.S. Cyclical Spending U.S. Cyclical Spending U.S. Cyclical Spending Chart III-31U.S. Labor Market U.S. Labor Market U.S. Labor Market Chart III-32U.S. Consumption U.S. Consumption U.S. Consumption Chart III-33U.S. Housing U.S. Housing U.S. Housing Chart III-34U.S. Debt And Deleveraging U.S. Debt And Deleveraging U.S. Debt And Deleveraging Chart III-35U.S. Financial Conditions U.S. Financial Conditions U.S. Financial Conditions Chart III-36Global Economic Snapshot: Europe Global Economic Snapshot: Europe Global Economic Snapshot: Europe Chart III-37Global Economic Snapshot: China Global Economic Snapshot: China Global Economic Snapshot: China

Feature The BCA Corporate Health Monitor (CHM) - designed to assess the financial well-being of companies - is one of our most reliable indicators that is also extremely popular with our clients. That is no surprise, as the CHMs have a solid track record in signaling broad turning points in company credit quality. This makes them useful in determining asset allocation recommendations on Investment Grade (IG) and High-Yield (HY) corporate bonds. Chart 1U.S. Corporates Outperforming, ##br##Despite Worsening Credit Quality U.S. Corporates Outperforming, Despite Worsening Credit Quality U.S. Corporates Outperforming, Despite Worsening Credit Quality In this Weekly Report, we present the "top-down" CHMs based on corporate data from national income (i.e. "GDP") accounts for the U.S., Euro Area and the U.K. We also show the "bottom-up" CHMs constructed using the actual reported financial data of individual companies in the U.S., Euro Area and Emerging Markets (EM). The CHMs are shown in a chartbook format that allows for quick visual analysis and comparisons. Going forward, we will publish this CHM Chartbook on a quarterly basis as a regular part of Global Fixed Income Strategy. The broad conclusion from looking at the CHMs is that corporate credit quality has been steadily improving in Europe, the U.K. and in the EM universe over the past couple of years, in sharp contrast to the worsening financial health of highly-levered U.S. companies. Bond investors seem to be ignoring the relative message sent by our CHMs, however, as U.S. corporate debt has outperformed other developed credit markets since the beginning of 2016 (Chart 1). An Overview Of The BCA Corporate Health Monitors The BCA Corporate Health Monitor (CHM) is an indicator designed to assess the underlying financial strength of the corporate sector for a country. The Monitor is an average of six financial ratios similar to those used by credit rating agencies to evaluate individual companies. However, we calculate our ratios using top-down (national accounts) data for profits, interest expense, debt levels, etc. The idea is to treat the entire corporate sector as if it were one big company, and then look at the credit metrics that would be used to assign a credit rating to it. Importantly, only data for the non-financial corporate sector is used in the CHM, as the metrics used to measure the underlying health of banks and other financial firms are different than those for the typical company. The six ratios used in the CHM are shown in Table 1 below. To construct the CHM, the individual ratios are standardized, added together, and then shown as a deviation from the medium-term trend. That last part is important, as it introduces more cyclicality into the CHM and allows it to better capture major turning points in corporate well-being. Largely because of this construction, the CHM has a very good track record at heralding trend changes in corporate credit spreads (both for Investment Grade and High-Yield) over many cycles. Top-down CHMs are now available for the U.S., Euro Area and U.K. The CHM methodology was extended in 2016 to look at corporate health by industry and by credit quality.1 The financial data of a broad set of individual U.S. and Euro Area companies was used to construct individual "bottom-up" CHMs using the same procedure as the more familiar top-down CHM. Some of the ratios differ from those used in the top-down CHM (see Table 1), largely due to definitional differences in data presented in national income accounts versus those from actual individual company financial statements. The bottom-up CHMs analyze the health of individual sectors, and can be aggregated up into broad CHMs for Investment Grade and High-Yield groupings to compare with credit spreads. An EM version of the bottom-up CHM was introduced by the BCA Emerging Markets Strategy team last September, which extends the CHM analysis to EM hard-currency corporate debt.2 Table 1Definitions Of Ratios That Go Into The CHMs BCA Corporate Health Monitor Chartbook BCA Corporate Health Monitor Chartbook U.S. Corporate Health Monitors: Still Deteriorating Chart 2Top-Down U.S. CHM: Still Deteriorating Top-Down U.S. CHM: Still Deteriorating Top-Down U.S. CHM: Still Deteriorating Our top-down CHM for the U.S. has been flashing a deteriorating state of corporate health since mid-2014 (Chart 2). That trend had been showing signs of stabilization last year, but the Q1/2017 data worsened on the back of lower profit margins and returns on capital. The latter now sits just above 5% - a level last seen during the 2009 recession. Corporate leverage, as measured in our top-down CHM using the value of debt versus equity, does not look to be a problem. The story is quite different when using alternative measures like net debt/EBITDA, however, with U.S. leverage exceeding the highs from the Telecom bubble of the early 2000s. While booming equity values certainly flatter the leverage ratio in our top-down CHM, a strong stock market should, to some degree, reflect a better backdrop for growth in corporate profits and creditworthiness. Even against this positive backdrop, however, other credit indicators are flashing some warning signs that leave our top-down CHM in the "deteriorating health" zone. Interest coverage and debt coverage ratios, while still above the lows seen during past recessions, are steadily falling. This does raise concerns for U.S. corporate health if U.S. bond yields begin to climb again, as we expect. However, given the historically low interest rate backdrop for corporate debt, a bigger threat to interest coverage ratios and overall credit quality would come from an economic slump that damages company profits. That is not going to be a problem for the rest of this year, but weaker growth is a more likely outcome in 2018 as the Fed continues its monetary tightening cycle. Our bottom-up CHMs for U.S. IG (Chart 3) and U.S. HY (Chart 4) have shown a bit of improvement in recent quarters relative to the signal from our top-down CHM. This is likely related to the growing gap between corporate profits as reported in the U.S. national accounts data, which are slowing, compared to the reported earnings of publicly traded companies, which are accelerating. Also, leverage in the bottom-up CHMs uses the book value of equity, which is more readily reported by individual companies, and is thus much higher than the measure used in our top-down CHM. Return on capital is at multi-decade lows for both IG and HY corporates, although profit margins look to be in much better shape for IG names relative to HY issuers. HY margins have enjoyed a cyclical improvement, however, largely due to better earnings from HY energy companies (Chart 4, panel 4). Interest coverage and debt coverage are depressed, with HY issuers in much worse shape than IG. Chart 3Bottom-Up U.S. Investment Grade CHM: ##br##Deteriorating Bottom-Up U.S. Investment Grade CHM: Deteriorating Bottom-Up U.S. Investment Grade CHM: Deteriorating Chart 4Bottom-Up U.S. High-Yield CHM: ##br##Some Cyclical Improvement Bottom-Up U.S. High-Yield CHM: Some Cyclical Improvement Bottom-Up U.S. High-Yield CHM: Some Cyclical Improvement The cumulative message from our top-down and bottom-up U.S. CHMs is that U.S. corporate health has enjoyed some cyclical improvement over the past few quarters, but the state of balance sheets is slowly-but-steadily worsening. High corporate leverage will become a major problem during the next U.S. recession, but is not a major factor weighing on credit spreads at the moment (Chart 5). We are maintaining our overweight stance on U.S. IG and higher-rated U.S. HY, both of which should continue to outperform Treasuries over the next few months, but a repeat performance is far less likely next year. Chart 5No Signs Of Concern##br## In U.S. Corporate Credit Spreads No Signs Of Concern In U.S. Corporate Credit Spreads No Signs Of Concern In U.S. Corporate Credit Spreads Chart 6Top-Down Euro Area CHM:##br## Improving Top-Down Euro Area CHM: Improving Top-Down Euro Area CHM: Improving Euro Area Corporate Health Monitors: Solid Improvement Our top-down Euro Area CHM has been showing steady improvement since 2013, driven by strong profit margins and rising interest and debt coverage ratios (Chart 6). The ultra-stimulative monetary policies of the European Central Bank (ECB) have likely played a large role in helping lower corporate borrowing costs and boosting the interest coverage ratio. The average coupon on bonds in the Bloomberg Barclays Euro-Aggregate Investment Grade corporate index is now down to a mere 2.3% - a far cry from the 5% level that prevailed during the peak of the 2011 Euro Debt crisis or the 3.5% level just before the ECB began its asset purchase program in 2015. Return on capital has fallen over the past decade and now sits at 8%, although profit margins remain quite strong on our top-down CHM measure. Short-term liquidity is at a record high, suggesting no imminent problems for European corporate borrowers. Our bottom-up CHMs for Euro Area IG (Chart 7) and HY (Chart 8) are telling a broadly similar story to the top-down CHM. The bottom-up CHMs have steadily improved in the past couple of years, most notably for domestic issuers of Euro-denominated debt.3 Some improvement in the bottom-up aggregates for operating margins and interest coverage ratios is providing a boost to European credit quality. Chart 7Bottom-Up Euro Area Investment Grade CHMs Bottom-Up Euro Area Investment Grade CHMs Bottom-Up Euro Area Investment Grade CHMs Chart 8Bottom-Up Euro Area High-Yield CHMs Bottom-Up Euro Area High-Yield CHMs Bottom-Up Euro Area High-Yield CHMs The bottom-up measure of leverage for domestic IG issuers has been steadily declining since the 2009 recession, a sign that European companies have been much more cautious in managing their balance sheet risk than their U.S. counterparts. The same cannot be said for Euro Area domestic HY issuers, where all the individual ratios are at weak absolute levels. When splitting our bottom-up Euro Area IG company list into issuers from core Europe versus countries on the Periphery (Chart 9), the "regional" European CHMs tell broadly similar stories of improving credit quality. The fact that even Peripheral issuers are seeing rising interest coverage and liquidity ratios, despite much higher levels of leverage than in the core, is an indication of how the ECB's low interest rate policies and asset purchase programs (which include buying corporate debt) have helped support the European corporate sector. Net-net, our Euro Area CHMs are sending a signal that there are no immediate stresses on corporate balance sheets or profitability. This is already reflected in the current low level of corporate bond yields and spreads, though (Chart 10). A bigger threat to Euro Area corporates comes from monetary policy. The ECB is under increasing pressure to consider announcing a tapering of its asset purchases - likely to include slower buying of corporates - starting in 2018. There is a risk of a negative market reaction that could undermine future Euro Area corporate bond performance. Because of this, we continue to prefer U.S. corporate debt over Euro Area equivalents, despite the large gap between the U.S. and European top-down CHMs (Chart 11). Chart 9Bottom-Up Euro Area IG CHMs: ##br##Core Vs. Periphery Bottom-Up Euro Area IG CHMs: Core vs Periphery Bottom-Up Euro Area IG CHMs: Core vs Periphery Chart 10Euro Area Corporate Bonds ##br## Have Had A Great Run Euro Area Corporate Bonds Have Had A Great Run Euro Area Corporate Bonds Have Had A Great Run Chart 11Relative CHMs Starting ##br##To Turn Less Favorable For U.S. Credit Relative CHMs Starting To Turn Less Favorable For U.S. Credit Relative CHMs Starting To Turn Less Favorable For U.S. Credit U.K. Corporate Health Monitor: Solid Balance Sheet Fundamentals The top-down U.K. CHM has steadily improved over the past couple of years, led by rising profit margins, higher interest coverage and very robust liquidity (Chart 12). Return on capital is low relative to its history, which is consistent with the trends seen in the U.S. and Euro Area and likely reflects the global low productivity backdrop. Fundamental analysis of U.K. corporates may not be of much use at the moment given the extremely accommodative monetary policy environment provided by the Bank of England (BoE). Low interest rates, combined with BoE asset purchases (which include a small amount of corporate debt) and a steep fall in the Pound in the aftermath of the Brexit-driven political uncertainty, are all helping keep the U.K. economy afloat. The BoE is now having to deal with a currency-driven climb in U.K. inflation, with three members of the BoE policy committee even calling for a rate hike at the latest policy meeting. The political backdrop after last year's Brexit vote and this month's closer-than-expected U.K. election result remains too volatile for the BoE to seriously consider any imminent tightening of monetary policy. While it can be debated how much the Brexit uncertainty is truly weighing on the U.K. economy, the BoE is unlikely to take any risks on triggering a growth slowdown by becoming too hawkish, too soon - even with the relatively high level of currency-driven inflation. A combination of a strong CHM and a dovish BoE will allow U.K. corporate debt, both IG and HY, to continue to outperform Gilts. We continue to recommend an overweight allocation to U.K. corporates even though, as in the other countries shown in this report, valuations are not cheap (Chart 13). We have not yet constructed bottom-up versions of the CHM for U.K. corporates to allow us to make any additional comments on the relative merits of U.K. IG and HY debt. This is something we intend to look into for future reports. Chart 12U.K. Corporate Balance Sheets ##br##Are In Good Shape... U.K. Corporate Balance Sheets Are In Good Shape... U.K. Corporate Balance Sheets Are In Good Shape... Chart 13...Which Is Already Reflected In ##br##Tight Credit Spreads ...Which Is Already Reflected In Tight Credit Spreads ...Which Is Already Reflected In Tight Credit Spreads Emerging Market Corporate Health Monitor: Cyclically Strong, Structurally Weak The CHM for EM corporates built by our Emerging Markets Strategy team is purely a bottom-up measure. The financial data from 220 EM companies in over 30 countries is used to construct the EM CHM. Only firms that issue U.S. dollar-denominated bonds are included, with banks and other financials also omitted in a similar fashion to the CHMs for the developed economies. A shorter list of financial ratios is used in the EM CHM than the developed CHMs, including: Profit margins Free cash flow to total debt: Liquidity Leverage Unlike the developed CHMs, the ratios are not equally weighted in the construction of the EM CHM. Profit margins and cash flow/debt combined represent 75% of the EM CHM. The weightings are designed to optimize the performance of the EM CHM versus the actual spread movements in the J.P. Morgan CEMBI benchmark index for EM corporate debt. Chart 14EM Corporate Health: Cyclically Solid... EM Corporate Health: Cyclically Solid... EM Corporate Health: Cyclically Solid... The EM CHM is currently pointing to very strong fundamental underpinnings for EM corporates with the indicator at the most credit-positive level in a decade (Chart 14). That recent strength is a modest cyclical improvement after a multi-year deterioration in all the individual EM CHM components (Chart 15). The uptick in global commodity prices in 2016 played a major role in the improvement in the growth-sensitive components (top two panels). However, the biggest structural headwind for EM corporates is the unrelenting rise in balance sheet leverage (bottom panel) - a problem that could come to the forefront if the recent slump in commodity prices persists or developed market interest rates begin to rise more sharply as central banks become marginally less accommodative. For now, we continue to recommend only a neutral allocation to EM hard currency debt, as the positive message sent by the EM CHM appears fully priced into the current low level of EM yields and spreads (Chart 16). Chart 15...But Structurally More Challenged ...But Structurally More Challenged ...But Structurally More Challenged Chart 16EM Corporate Debt Is Not Cheap EM Corporate Debt Is Not Cheap EM Corporate Debt Is Not Cheap Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 Please see Section II of The Bank Credit Analyst, "U.S. Corporate Health Gets A Failing Grade", dated February 2016, available at bca.bcaresearch.com. 2 Please see BCA Emerging Markets Strategy Special Report, "EM Corporate Health Is Flashing Red" dated September 14 2016, available at ems.bcaresearch.com. 3 Given the large share of non-European issuers in the Euro-denominated corporate debt market, we have split our sample set of companies in our bottom-up Euro Area CHMs into "domestic" and "foreign" issuer groups. This allows a more precise analysis of the corporate health of European-domiciled companies. Recommendations Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns The GFIS Recommended Portfolio Vs. The Custom Benchmark Index BCA Corporate Health Monitor Chartbook BCA Corporate Health Monitor Chartbook

Highlights Risk Budgeting: We are introducing a more formal risk measurement element to our model global bond portfolio. This is to identify if our individual views are potentially creating too much volatility, in aggregate, but also as a way to express the conviction of our individual recommendations through allocation of a "risk budget". Tracking Error Of Our Portfolio: We are setting our maximum allowable tracking error, or excess volatility of our portfolio versus our benchmark index, at 100 basis points. Our current tracking error is just under ½ of that limit. We estimate that our highest conviction views at the moment - staying below-benchmark on duration risk, overweighting U.S. corporates, underweighting both U.S. Treasuries and Italian government debt - contribute nearly 4/5ths of our overall portfolio tracking error. Feature Last September, we introduced a model portfolio framework to Global Fixed Income Strategy.1 This was done to better communicate our investment research into actionable ideas more in line with the day-to-day decisions and trade-offs made by professional bond managers. We followed that up with the addition of performance measurement tools to more accurately track the returns of our model bond portfolio versus a stated benchmark.2 We are now initiating the final piece of our model bond portfolio framework in this Special Report - introducing a risk management component to identify cumulative exposures and guide the relative sizes of our suggested tilts. Our goal is to translate our individual investment recommendations into the language of a "risk budget", i.e. how much of the desired volatility of the portfolio would we suggest placing into any single trade idea. This will allow our readers to apply our proposed tilts - based on how much conviction (i.e. "risk") we allocate to each position - to their own portfolios which may have different risk limits and return expectations. For example, our current recommendation to overweight U.S. corporate debt, both Investment Grade (IG) and High-Yield (HY) represents nearly 1/3 of our estimated total portfolio risk, by far our largest source of potential volatility both in absolute terms and versus our benchmark index (Table 1). Overweighting U.S. corporates, both versus U.S. Treasuries and Euro Area equivalents, is one of our highest conviction trades at the moment. A client who may choose to run a lower risk portfolio can still follow our recommendation by placing enough into U.S. corporates so that 33% of the desired portfolio volatility will come from those positions. Table 1Risk Allocation In Our Model Bond Portfolio Adding A Risk Management Framework To Our Model Bond Portfolio Adding A Risk Management Framework To Our Model Bond Portfolio In the rest of this Special Report, we will discuss some of the various ways to measure fixed income portfolio risk, apply them to our model portfolio, and introduce some measures to monitor our aggregate portfolio volatility. Going forward, we will closely watch our established metrics and position sizes to ensure that the combination of our individual investment recommendations that we discuss on a week-to-week basis does not create a portfolio that is potentially more volatile than desired. Risk Measurement In Fixed Income Portfolios While investors are typically focused on meeting return targets for their portfolios, the other side of the equation - managing portfolio volatility - is often less stressed. This is especially true during bull markets for any asset class. Investors may become complacent if returns meet or exceed their targets when, in fact, excess returns may have actually been earned through overly risky positions that could have easily not worked in the investors' favor. In the current macro environment, where many financial asset prices are at new highs with stretched valuations and with most of the major global central banks incrementally moving towards less accommodative monetary policy stances, risk management should be even more important for investors. Overly concentrated positioning could now lead to considerable portfolio losses, especially if measuring risk with a metric that is flawed or incomplete, which can lead to a false sense of security. With that in mind, we consider some typical risk measurement metrics used by fixed income investors: Duration: Duration is usually the most popular risk metric for fixed income portfolios as it measures interest rate sensitivity. Duration is defined as the percentage change in a portfolio or asset resulting from a one percentage point change in interest rates. While it provides a solid base understanding of interest rate risk, it does make a simplifying assumption that there is a linear relationship between interest rates and bond prices. Value-At-Risk: Value-At-Risk (VaR) is a statistical technique that measures the loss of an investment, or of an entire portfolio, over a certain period with a given level of confidence. However, there are two considerable flaws with this approach. First, the VaR output suggests a portfolio can lose at least X%, it does not actually indicate how big the potential loss could be. Instead, using a measure such as Historical VaR, if a portfolio has a long enough track record, can better quantify potential losses. Second, VaR is highly susceptible to estimation errors. Certain assumptions on correlations and the normality of return distributions can have a substantial impact on VaR readings. Table 2Value At Risk Of Our Benchmark Adding A Risk Management Framework To Our Model Bond Portfolio Adding A Risk Management Framework To Our Model Bond Portfolio In Table 2, we show the Historical VaR (HVAR) of our benchmark index, calculating the potential monthly loss using data going back to 2005. On that basis, the worst expected monthly loss for our benchmark is -1.6% (using a 95% confidence interval) and -2.1% (using a 99% confidence interval). Tracking Error: Tracking error measures the volatility of excess returns relative to a certain benchmark. It is a standard risk measure used by a typical "real money" bond manager with a benchmark performance index, like a mutual fund. Tracking error does not offer information on alpha generation (i.e. how much you can expect to beat your benchmark based on your current investments), it simply indicates how much more volatile a portfolio is expected to be versus its benchmark. As our model portfolio returns are measured on a relative basis to our stated bond benchmark index, tracking error is quite appropriate as our main risk metric. A Historical Examination Of Our Portfolio When we first created our model portfolio, we also introduced a benchmark index against which we could measure our performance. Our customized benchmark differs from typical multi-sector measures like the Barclays Global Aggregate Index in that it has a broader scope, including sectors that can have credit ratings below investment grade such as High Yield corporates. The benchmark does, however, exclude smaller regions that we only occasionally discuss such as Sweden, Portugal, Norway and New Zealand. These smaller markets offer comparatively poor liquidity and we want our benchmark to be as investible as possible. Nevertheless, our customized benchmark has been highly correlated to the Barclays Global Aggregate Index over the past decade. As our portfolio has not had a full year of return data, its history is quite limited. Still, in our first performance review conducted two months ago, we indicated that our portfolio had been very closely tracking our customized benchmark. We have since increased our positions in our highest conviction views and our tracking error has risen noticeably and now sits at just over 40bps (Chart 1). Within our model portfolio, we are setting an expected excess return target of 100bps per year. That means that we are setting a goal of beating our benchmark index returns by one full percentage point per year. Given that we are measuring our performance versus currency-hedged benchmarks that are primarily rated investment grade or better, 100bps of annual excess return is a reasonable target. We are also setting a limit where the excess return/tracking error ratio should aim to be equal to 1 each year. This is under the simple assumption that we want an equal amount of return over our benchmark for our expected excess volatility versus our benchmark. On that basis, we are setting our tracking error "limit" at 100bps per year. That suggests that our current tracking error is relatively low. However, correlations between the individual components of our benchmark index have been rising over the past couple of years (Chart 2). Therefore, running a relatively low overall level of risk at a time where diversification among the positions within our portfolio is now harder to achieve, and when the valuations on most government bond and credit markets look rich, is prudent. Chart 1Higher Tracking Error, But Still Well Below Our Target Higher Tracking Error, But Still Well Below Our Target Higher Tracking Error, But Still Well Below Our Target Chart 2Correlations Across Fixed Income Sectors Have Been Rising Correlations Across Fixed Income Sectors Have Been Rising Correlations Across Fixed Income Sectors Have Been Rising This is another way that we can control the overall riskiness of our model portfolio. Not only by how much of our risk budget (tracking error) that we want to allocate to each of our recommended positions, but also how big of a risk budget do we want to run at any given point in time. If we see more assets trading at cheap valuations, then we could choose to run a higher tracking error than when most assets look expensive. Bottom Line: We are introducing a more formal risk measurement element to our model global bond portfolio. This is to identify if our individual views are potentially creating too much volatility, in aggregate, but also as a way to express the conviction of our individual recommendations through allocation of a "risk budget". We are setting our maximum allowable tracking error, or excess volatility of our portfolio versus our benchmark index, at 100 basis points. Measuring The Contribution To Risk From Our Market Tilts In our model portfolio, we include a wide range of geographies and sectors from the global fixed income universe. Understanding the risk contribution of each position to the overall portfolio provides a clearer picture as to where our potential risks lie, and by how much. To measure the risk contribution of each of our individual recommendations to our overall portfolio volatility, we used the following formula: wA * E CovAB * wB Where W = the weight of any single asset in our portfolio and COV is the covariance between the asset and other assets in the portfolio. As such, an asset's contribution to risk is a function of its weight in the portfolio and its covariance with the other assets. Importantly, since we are measuring our model portfolio performance in terms of excess returns, we examined each position's contribution to risk relative to the benchmark. All calculations begin in late 2005, when return data is available for all of the assets in our portfolio. The results are summarized in Table 1 on Page 1. Our portfolio tilts are based off of our four highest conviction themes. They include: Stronger global growth led by the U.S. The U.S. economy should expand at a faster pace in the latter half of the year on the back of a rebound in consumption and strong capital spending, all supported by solid income growth and easy financial conditions. We have expressed this theme through our overweight allocation to U.S. corporate debt. While our U.S. Corporate Health Monitor is flashing that balance sheets are becoming increasingly strained, easy monetary conditions and an expansionary economic backdrop should continue to support excess returns for U.S. corporates. More Fed rate hikes than expected. We expect U.S. economic and corporate profit growth to remain robust due to accommodative monetary conditions, diminishing slack and resilient consumption. As such, the Fed will continue tightening policy by more than what markets are currently pricing in. This theme is expressed through an underweight position in U.S. Treasuries, which accounts for 17% of our volatility versus 24% for that of the benchmark. This wide spread relative to the benchmark is a substantial source of our tracking error, but one that we are comfortable running given our view that U.S. Treasury yields are too low. Chart 3Realized Bond Volatility Has Been Declining Realized Bond Volatility Has Been Declining Realized Bond Volatility Has Been Declining Rising tapering risks in Europe. Our expectation is that the European Central Bank (ECB) will be forced to announce a slower pace (tapering) of bond buying starting next year, given the current robust economic expansion in Europe that is rapidly absorbing spare capacity. An ECB taper announcement is expected to lead to rising longer-term global bond yields, mostly via rising term premia. We are expressing that view in our portfolio through our overall underweight interest rate duration stance. Our current portfolio duration is 5.6 years versus our benchmark duration of 7.0 years. That is a large tilt that represents a significant portion of our tracking error, but given our view that U.S. Treasuries also look overvalued, running a large overall duration underweight does correlate to our conviction level. Rising geopolitical risks and banking sector issues in Italy. Geopolitical risks remain elevated leading up to parliamentary elections in 2018, and Italian banks remain undercapitalized with non-performing loans still in an uptrend. Therefore, we are underweight Italian debt, though this is a smaller deviation of portfolio risk versus our benchmark (around 2%), given the smaller size of Italy in our benchmark. Purely looking at geography and sector selection, our four highest conviction views make up almost 80% of the active portfolio risk that we are "running" in our model portfolio. That number may seem high but, as described earlier, our realized portfolio volatility has been quite low (Chart 3). That suggests that there could be some degree of underlying diversification within our recommended portfolio given lower correlations of certain assets to the rest of the portfolio. This is a topic that we will investigate more deeply in future Weekly Reports. Bottom Line: We estimate that our highest conviction views at the moment - staying below-benchmark on duration risk, overweighting U.S. corporates, underweighting both U.S. Treasuries and Italian government debt - contribute nearly 4/5ths of our overall portfolio tracking error. Patrick Trinh, Associate Editor Patrick@bcaresearch.com Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 Please see BCA Global Fixed Income Strategy Special Report, "Introducing Our Recommended Global Fixed Income Portfolio", dated September 20 2016, available at gfis.bcaresearch.com 2 Please see BCA Global Fixed Income Strategy Special Report, "An Initial Look At The Performance Of Our Model Bond Portfolio", dated April 18 2017, available at gfis.bcaresearch.com The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Table 4 Adding A Risk Management Framework To Our Model Bond Portfolio Adding A Risk Management Framework To Our Model Bond Portfolio