Valuations

Highlights Q3/2019 Performance Breakdown: Our recommended model bond portfolio underperformed the custom benchmark by -30bps during the third quarter of the year. Winners & Losers: The biggest underperformance came from underweight positions in U.S. Treasuries (-28bps) and Italian government bonds (-18bps) as yields plunged, dwarfing gains from overweights in corporate bonds in the U.S. (+11bps) and euro area (+4bps). Scenario Analysis For The Next Six Months: We are maintaining our current positioning, staying below-benchmark on duration while overweighting U.S. and euro area corporates vs. government debt. In our base case scenario, global growth will begin to stabilize but the Fed will deliver one more “insurance” rate cut by year-end, leading to corporate bond outperformance. Feature Global bond markets have enjoyed a powerful bull run throughout 2019, as yields have plummeted alongside weakening global growth and growing political uncertainty. Those two forces came to a head in the third quarter of the year, with U.S.-China trade tensions ratcheting up another notch after the imposition of higher U.S. tariffs in early August and global manufacturing PMI data moving into contraction territory – especially in the U.S. The result was a significant fall in government bond yields as markets discounted both lower inflation expectations and more aggressive monetary easing from global central banks, led by the Fed and ECB. The benchmark 10-year U.S. Treasury yield and 10-year German Bund yield plunged -40bps and -25bps, respectively, during the July-September period. Yet at the same time, global credit markets remained surprisingly stable, as the option-adjusted spread on the Bloomberg Barclays Global Corporates index was unchanged over the same three months. In this report, we review the performance of the BCA Global Fixed Income Strategy (GFIS) model bond portfolio during the eventful third quarter of 2019. We also present our updated scenario analysis, and total return projections, for the portfolio over the next six months. As a reminder to existing readers (and to new clients), the model portfolio is a part of our service that complements the usual macro analysis of global fixed income markets. The portfolio is how we communicate our opinion on the relative attractiveness between government bond and spread product sectors. This is done by applying actual percentage weightings to each of our recommendations within a fully invested hypothetical bond portfolio. Q3/2019 Model Portfolio Performance Breakdown: Good News On Credit Trumped By Bad News On Duration Chart of the WeekDuration Losses Dwarf Credit Gains In Q3/19 Duration Losses Dwarf Credit Gains In Q3/19 Duration Losses Dwarf Credit Gains In Q3/19 The total return for the GFIS model portfolio (hedged into U.S. dollars) in the third quarter was 2.0%, lagging the custom benchmark index by -30 bps (Chart of the Week).1 This brings the cumulative year-to-date total return of the portfolio to +7.8%, which has underperformed the benchmark by a disappointing –67bps. The Q3 drag on relative returns came entirely from the government bond side of the portfolio; specifically, the underweight allocation to U.S. Treasuries and Italian government bonds (Table 1). Those allocations reflected our views on overall portfolio duration (below benchmark) and a relative value consideration within European spread product (preferring corporates to Italy). Both those recommendations went against us as global bond yields dropped during Q3, with Italian yields collapsing (the benchmark 10-year yield was down –126bps) as investors chased any positive yield denominated in euros after the ECB signaled a new round of policy easing. The total return for the GFIS model portfolio (hedged into U.S. dollars) in the third quarter was 2.0%, lagging the custom benchmark index by -30 bps  Table 1GFIS Model Bond Portfolio Q3/2019 Overall Return Attribution Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Providing some partial offset to the U.S. and Italy allocations were gains from overweight positions in government bonds in the U.K., Australia and Japan. More importantly, our overweights in corporate debt in the U.S. and euro area made a strong positive contribution to the performance of the portfolio. The bar charts showing the total and relative returns for each individual government bond market and spread product sector are presented in Charts 2 and 3. The most significant movers were: Chart 2GFIS Model Bond Portfolio Q3/2019 Government Bond Performance Attribution Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Chart 3GFIS Model Bond Portfolio Q3/2019 Spread Product Performance Attribution By Sector Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Biggest outperformers Overweight U.S. high-yield Ba-rated (+4bps) Overweight U.S. high-yield B-rated (+3bps) Overweight U.S. investment grade industrials (+3bps) Overweight Japanese government bonds with maturity of 5-7 years (+2bps) Overweight euro area corporates, both investment grade (+2bps) and high-yield (+2bps) Biggest underperformers Underweight U.S. government bonds with maturity beyond 10+ years (-15bps) Underweight Italy government bonds with maturity beyond 10+ years (-10bps) Underweight U.S. government bonds with maturity of 7-10 years (-5bps) Underweight Japanese government bonds with maturity beyond 10+ years (-4bps) Underweight U.S. government bonds with maturity of 3-5 years (-4bps) Chart 4 presents the ranked benchmark index returns of the individual countries and spread product sectors in the GFIS model bond portfolio for Q3/2019. The returns are hedged into U.S. dollars (we do not take active currency risk in this portfolio) and are adjusted to reflect duration differences between each country/sector and the overall custom benchmark index for the model portfolio. We have also color-coded the bars in each chart to reflect our recommended investment stance for each market during Q3/2019 (red for underweight, blue for overweight, gray for neutral).2 Ideally, we would look to see more blue bars on the left side of the chart where market returns are highest, and more red bars on the right side of the chart were returns are lowest. Chart 4Ranking The Winners & Losers From The Model Bond Portfolio In Q3/2019 Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence One thing that stands out from Chart 4 is that every fixed income sector generated a positive return, except for EM USD-denominated corporates. This is a fascinating outcome given the sharp falls in risk-free government bond yields which typically would correlate to a selloff in risk assets and widening of credit spreads. The soothing balm of looser global monetary policy seems to have offset the impact of elevated uncertainty on trade and future economic growth, allowing both bond yields and credit spreads to stay low. The soothing balm of looser global monetary policy seems to have offset the impact of elevated uncertainty on trade and future economic growth, allowing both bond yields and credit spreads to stay low.  We maintained an overweight stance on global spread product throughout Q3, as we felt that the monetary policy effect would continue to overwhelm uncertainty. We did, however, make some tactical adjustments to our duration stance after the U.S. raised tariffs on Chinese imports, upgrading to neutral on August 6th.3 We had felt that higher tariffs were a sign that a potential end to the U.S.-China trade conflict was now even less likely, which raised the odds of a potential risk-off financial market event that would temporarily push bond yields lower. We shifted back to a below-benchmark duration stance on September 17th, given signs of de-escalation in the trade dispute and, more importantly, some improvement evident in global leading economic indicators.4 Bottom Line: Our recommended model bond portfolio underperformed the custom benchmark index during the third quarter of the year, with the drag on performance from an underweight stance on U.S. Treasuries and Italian BTPs overwhelming the gains from corporate credit overweights in the U.S. and euro area. Future Drivers Of Portfolio Returns Looking ahead, the performance of the model bond portfolio will continue to be driven by two main factors: our below-benchmark duration bias and our overweight stance on global corporate debt versus government bonds. Chart 5Overall Portfolio Allocation: Overweight Credit Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence In terms of the specific high-level weightings in the model portfolio, we currently have a moderate overweight, equal to eight percentage points, on spread product versus government debt (Chart 5). This reflects a more constructive view on future global growth. Early leading economic indicators are starting to bottom out and global central bankers are maintaining a dovish policy bias despite low unemployment rates – both factors that will continue to benefit growth-sensitive assets like corporate debt. Early leading economic indicators are starting to bottom out and global central bankers are maintaining a dovish policy bias despite low unemployment rates – both factors that will continue to benefit growth-sensitive assets like corporate debt. We are maintaining our below-benchmark duration tilt at 0.6 years short of the custom benchmark (Chart 6). We recognize, however, that the underperformance from duration in the model portfolio will not begin to be clawed back until there are signs of a bottoming in widely-followed cyclical economic indicators like the U.S. ISM index and the German ZEW. We think that will happen given the uptick in our global leading economic indicator (LEI), but that may take a few more months to develop based on the usual lead time from the LEI to the survey data like the ISM. The hook up in the global LEI does still gives us more confidence that the big decline in global bond yields seen this year is over, especially if a potential truce in the U.S.-China trade war is soon reached, as our political strategists believe to be increasingly likely. Chart 6Overall Portfolio Duration: Moderately Below Benchmark Overall Portfolio Duration: Moderately Below Benchmark Overall Portfolio Duration: Moderately Below Benchmark Turning to country allocation, we are sticking with overweights in countries where central banks are likely to be more dovish than the Fed over the next 6-12 months (Germany, France, the U.K., Japan, and Australia). We are staying underweight the U.S. where inflation expectations appear too low and Fed rate cut expectations look too extreme. The Italy underweight has become a trickier call. We have long viewed Italian debt as a growth-sensitive credit instrument rather than the yield-driven rates vehicle it became in Q3 as markets priced in fresh monetary easing measures from the ECB (including restarting government purchases). We will revisit our Italy views in an upcoming report but, until then, we will continue to view Italian BTPs within the context of our European spread product allocation. Thus, we are maintaining an overweight on euro area corporate debt (by 1% each in investment grade and high-yield) while having an equal-sized underweight (-2%) in Italian government bonds. Our combined positioning generates a portfolio that has “positive carry”, with a yield of 3.1% (hedged into U.S. dollars) that is +25bps over that of the custom benchmark index (Chart 7). That same portfolio, however, generates an estimated tracking error (excess volatility of the portfolio versus its benchmark) of 55bps - well below our self-imposed 100bps ceiling and still within the 40-60bps range we have targeted since the start of 2019 (Chart 8). Chart 7Portfolio Yield: Positive Carry From Credit Portfolio Yield: Positive Carry From Credit Portfolio Yield: Positive Carry From Credit Chart 8Portfolio Risk Budget Usage: Cautious Portfolio Risk Budget Usage: Cautious Portfolio Risk Budget Usage: Cautious Scenario Analysis & Return Forecasts In April 2018, we introduced a framework for estimating total returns for all government bond markets and spread product sectors, based on common risk factors.5 For credit, returns are estimated as a function of changes in the U.S. dollar, the Fed funds rate, oil prices and market volatility as proxied by the VIX index (Table 2A). For government bonds, non-U.S. yield changes are estimated using historical betas to changes in U.S. Treasury yields (Table 2B). Table 2AFactor Regressions Used To Estimate Spread Product Yield Changes Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Table 2BEstimated Government Bond Yield Betas To U.S. Treasuries Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence This framework allows us to conduct scenario analysis of projected returns for each asset class in the model bond portfolio by making assumptions on those individual risk factors. In Tables 3A & 3B, we present our three main scenarios for the next six months, defined by changes in the risk factors, and the expected performance of the model bond portfolio in each case. The scenarios, described below, all revolve around our expectation that the most important drivers of future market returns will continue to be the momentum of global growth and the path of U.S. monetary policy. The scenario inputs for the four main risk factors (the fed funds rate, the price of oil, the U.S. dollar and the VIX index) are shown visually in Chart 9. Table 3AScenario Analysis For The GFIS Model Bond Portfolio For The Next Six Months Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Table 3BU.S. Treasury Yield Assumptions For The 6-Month Forward Scenario Analysis Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Chart 9Risk Factor Assumptions For The Scenario Analysis Risk Factor Assumptions For The Scenario Analysis Risk Factor Assumptions For The Scenario Analysis Base Case (Global Growth Bottoms): The Fed delivers one more -25bp rate cut by the end of 2019, the U.S. dollar weakens by -3%, oil prices rise by +10%, the VIX hovers around 15, and there is a bear-steepening of the UST curve. This is a scenario where the U.S. economy ends up avoiding recession and grows at roughly a trend-like pace. The Fed, however, still delivers one more “insurance” rate cut to mitigate the risk of low inflation expectations becoming more entrenched. Global growth is expected to bottom out as heralded by the global leading indicators. A truce (but not a full deal) is expected on the U.S.-China trade front, helping to moderately soften the U.S. dollar through reduced risk aversion. The model bond portfolio is expected to beat the benchmark index by +91bps in this case. Global Growth Strongly Rebounds: The Fed stays on hold, the U.S. dollar weakens by -5%, oil prices rise by +20%, the VIX declines to 12, there is a modest bear-steepening of the UST curve. In this tail-risk scenario, global growth starts to reaccelerate in lagged response to the global monetary easing seen this year, combined with some fiscal stimulus in major countries (China, the U.S., perhaps even Germany). The U.S. dollar weakens as global capital flows shift to markets which are more sensitive to global growth. The model bond portfolio is expected to beat the benchmark index by +106bps in this case. U.S. Downturn Intensifies: The Fed cuts rates by -75bps, the U.S. dollar is flat, oil prices fall by -15%, the VIX rises to 30; there is a bull-steepening of the UST curve. Under this tail-risk scenario, the current slowing of U.S. growth momentum gains speed, pushing the economy towards recession. The Fed cuts rates aggressively in response, helping weaken the U.S. dollar, but not before global risk assets sell off sharply to discount a worldwide recession. The model portfolio will underperform the benchmark by -38bps in this scenario. In terms of our conviction level among the main drivers of the model portfolio returns – duration allocation (across yield curves and countries) and asset allocation (credit versus government bonds) – we are most confident that credit returns will exceed those of sovereign debt over the next six months. In terms of our conviction level among the main drivers of the model portfolio returns – duration allocation (across yield curves and countries) and asset allocation (credit versus government bonds) – we are most confident that credit returns will exceed those of sovereign debt over the next six months. The underweight duration position, however, will also eventually begin to pay off if the message from the budding improvement in global leading economic indicators turns out to be correct. A collapse of the U.S.-China trade negotiations is the biggest threat to our base case, which would make the “U.S. Downturn Intensifies” scenario a more likely outcome. Bottom Line: We are maintaining our current positioning, staying below-benchmark on duration while overweighting U.S. and euro area corporates governments. In our base case scenario, global growth will begin to stabilize but the Fed will deliver one more “insurance” rate cut by year-end, leading to spread product outperformance.   Robert Robis, CFA, Chief Fixed Income Strategist rrobis@bcaresearch.com Ray Park, CFA, Research Analyst ray@bcaresearch.com Footnotes 1 The GFIS model bond portfolio custom benchmark index is the Bloomberg Barclays Global Aggregate Index, but with allocations to global high-yield corporate debt replacing very high quality spread product (i.e. AA-rated). We believe this to be more indicative of the typical internal benchmark used by global multi-sector fixed income managers. 2 Note that sectors where we made changes to our recommended weightings during Q3/2019 will have multiple colors in the respective bars in Chart 4. 3 Please see BCA Global Fixed Income Strategy Weekly Report, “Trade War Worries: Once More, With Feeling”, dated August 6, 2019, available at gfis.bcaresearch.com. 4 Please see BCA Global Fixed Income Strategy Weekly Report, “The World Is Not Ending: Return To Below-Benchmark Portfolio Duration”, dated September 17, 2019, available at gfis.bcaresearch.com. 5 Please see BCA Global Fixed Income Strategy Weekly Report, “GFIS Model Bond Portfolio Q1/2018 Performance Review: A Rough Start”, dated April 10th 2018, available at gfis.bcareseach.com. Recommendations The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Q3/2019 GFIS Model Bond Portfolio Performance Review: More Duration/Credit Divergence Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Highlights Chart 1Contagion? Contagion? Contagion? Until last week, global growth weakness had been wholly confined to the manufacturing sector. But the drop to 52.6 in September’s Non-Manufacturing PMI (from 56.4 in August) raises the specter of contagion from manufacturing into the broader U.S. economy. A further drop would be consistent with an economy headed toward recession, and run contrary to the 2015/16 roadmap that has been our base case (Chart 1). We think it is still premature to abandon the 2015/16 episode as an appropriate comparable for the current period. For one thing, the hard economic data paint a rosier picture than the PMI surveys. Industrial production and core durable goods new orders are up 2.5% and 2.3% (annualized), respectively, during the past 3 months. These data have helped drive the economic surprise index above zero, an event that usually coincides with rising yields (bottom panel). The divergence between soft and hard data makes it clear that trade uncertainties are so far having a greater impact on business sentiment than on actual production, but history tells us that these divergences don’t last long. Some positive news on the trade front will be required during the next few months to raise business sentiment and push bond yields higher. Stay tuned. Feature Investment Grade: Overweight Chart 2Investment Grade Market Overview Investment Grade Market Overview Investment Grade Market Overview Investment grade corporate bonds outperformed the duration-equivalent Treasury index by 42 basis points in September, before giving back 37 bps in the first week of October. We consider three main factors in our credit cycle analysis: (i) corporate balance sheet health, (ii) monetary conditions, and (iii) valuation. At present, the chief conundrum for investors is that while corporate balance sheet health is weak, the monetary environment is extraordinarily accommodative.1 On balance sheets, our top-down measure of gross leverage is elevated and rising (Chart 2). In contrast, interest coverage ratios remain solid, propped up by the Fed’s accommodative stance. With inflation expectations still very low, the Fed can maintain its “easy money” policy for some time yet. This will ensure that interest coverage stays solid and that bank lending standards continue to ease (bottom panel). This is an environment where corporate bond spreads should tighten. How low can spreads go? Our assessment of reasonable spread targets for the current environment suggests that Aaa, Aa and A-rated spreads are already fully valued, while Baa-rated spreads are 13 bps cheap (panels 2 & 3).2 We recommend focusing investment grade corporate bond exposure on the Baa credit tier, and subbing some Agency MBS into your portfolio in place of corporate bonds rated A or higher. Table 3ACorporate Sector Relative Valuation And Recommended Allocation* Crunch Time Crunch Time Table 3BCorporate Sector Risk Vs. Reward* Crunch Time Crunch Time High-Yield: Overweight Chart 3High-Yield Market Overview High-Yield Market Overview High-Yield Market Overview High-Yield outperformed the duration-equivalent Treasury index by 66 basis points in September, before giving back 117 bps in the first week of October. The junk index’s option-adjusted spread (OAS) has been fairly stable for most of the year, but the sector has become increasingly attractive from a risk/reward perspective.3 This is because the index’s negatively convex nature has caused its average duration to fall alongside declining Treasury yields. Chart 3 shows that while the index OAS has been rangebound, the 12-month breakeven spread has widened considerably.4 In other words, while junk expected returns have been stable, those expected returns now come with considerably less risk. As a result, the junk index OAS looks increasingly attractive relative to our spread target.5 Specifically, we now view the junk index OAS as 171 bps cheap (panel 3). Falling index duration also explains the divergence between quality spreads and the index OAS. Many have observed that the spread differential between Caa and Ba-rated junk bonds has widened in recent months, while the overall index OAS has been stable (panel 4). However, the divergence evaporates when we look at 12-month breakeven spreads instead of OAS (bottom panel). MBS: Neutral Chart 4MBS Market Overview MBS Market Overview MBS Market Overview Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 24 basis points in September, before giving back 25 bps in the first week of October. MBS have underperformed Treasuries by 31 bps, year-to-date. The conventional 30-year zero volatility spread held flat at 82 bps in September, as a 3 bps increase in expected prepayment losses (option cost) was offset by a 3 bps tightening in the option-adjusted spread (OAS). In last week’s report, we recommended favoring Agency MBS over Aaa, Aa and A-rated corporate bonds.6 We have three main reasons for this recommendation. First, expected compensation is competitive. The conventional 30-year MBS OAS is now 57 bps. This is above the pre-crisis average (Chart 4), and only 4 bps below the spread offered by a Aa-rated corporate bond. Aaa, Aa and A-rated corporate bond spreads also all look expensive relative to our targets. Second, risk-adjusted compensation heavily favors MBS. The 12-month breakeven spread for a conventional 30-year MBS is 21 bps. This compares to 6 bps, 8 bps and 12 bps for Aaa, Aa and A-rated corporates, respectively. Finally, the macro environment for MBS remains supportive. Mortgage lending standards have barely eased since the financial crisis (bottom panel), and most people have already had at least one opportunity to refinance their mortgage. This burnout will keep refi activity low, and MBS spreads tight (panel 2), going forward. Government-Related: Underweight Chart 5Government-Related Market Overview Government-Related Market Overview Government-Related Market Overview The Government-Related index outperformed the duration-equivalent Treasury index by 10 basis points in September, bringing year-to-date excess returns up to +163 bps. September returns were concentrated in the Foreign Agency sub-sector. These securities outperformed the Treasury benchmark by 55 bps on the month, bringing year-to-date excess returns up to +197 bps. Sovereign bonds underperformed duration-equivalent Treasuries by 6 bps in September, dragging year-to-date excess returns down to +436 bps. Local Authority and Domestic Agency debt underperformed by 1 bp and 2 bps on the month, respectively. Meanwhile, Supranationals bested the Treasury benchmark by a single basis point. Sovereign debt remains very expensive relative to equivalently-rated U.S. corporate credit (Chart 5). While the sector would benefit if the Fed’s dovish pivot eventually results in a weaker dollar, U.S. corporate bonds would also perform well in such an environment. Given the much more attractive starting point for U.S. corporate bond spreads, we find it difficult to recommend sovereign debt as an alternative. While sovereign debt in general looks expensive. USD-denominated Mexican sovereign bonds continue to look attractive relative to U.S. corporates (bottom panel). Investors should favor Mexican sovereigns within an otherwise underweight allocation to the sector as a whole. Municipal Bonds: Overweight Chart 6Municipal Market Overview Municipal Market Overview Municipal Market Overview Municipal bonds underperformed the duration-equivalent Treasury index by 10 basis points in September, dragging year-to-date excess returns down to -57 bps (before adjusting for the tax advantage). We recommended upgrading municipal bonds from neutral to overweight in last week’s report.7  We based the decision on the increasing attractiveness of yield ratios, despite an underlying credit environment that remains supportive for munis. Municipal bond yields failed to keep pace with falling Treasury yields in recent months, and now look quite attractive as a result (Chart 6). The average Aaa-rated Municipal / Treasury (M/T) yield ratio rose 4% in September and is now back above 90%. This is well above the 81% average that prevailed in the late stages of the previous cycle, between mid-2006 and mid-2007. In fact, Aaa M/T yield ratios for every maturity are now above average pre-crisis levels. Though yield ratios still look best at the long-end of the Aaa curve (panel 2), we now recommend owning munis in place of Treasuries across the entire maturity spectrum. Fundamentally, state & local government balance sheets remain solid. We showed in last week’s report that our Municipal Health Monitor is in “improving health” territory, and noted that state & local government interest coverage is positive (bottom panel). Both of those trends are consistent with muni ratings upgrades continuing to outnumber downgrades going forward. Treasury Curve: Maintain A Barbell Curve Positioning Chart 7Treasury Yield Curve Overview Treasury Yield Curve Overview Treasury Yield Curve Overview The Treasury curve bear-steepened in September, and then bull-steepened sharply last week. All in all, the 2/10 Treasury slope is +12 bps, 12 bps steeper than it was at the end of August. The 5/30 slope is +67 bps, 10 bps steeper than at the end of August. Our fair value models (see Appendix B) continue to show that bullets are expensive relative to barbells across the entire Treasury curve. In particular, 5-year and 7-year maturities look very expensive compared to the short and long ends of the curve. Notice that the 2/5/10 butterfly spread, the spread between the 5-year bullet and a duration-matched 2/10 barbell, remains negative despite the recent 2/10 steepening (Chart 7). We have shown in prior research that the 5-year and 7-year maturities are the most highly correlated with our 12-month Fed Funds Discounter. Our discounter is currently at -74 bps, meaning that the market is priced for nearly three more Fed rate cuts during the next 12 months (top panel). We expect fewer cuts than that, and as such, think the Discounter is more likely to rise. 5-year and 7-year maturities would underperform the rest of the curve in that scenario. We also continue to hold our short position in the February 2020 fed funds futures contract. That contract is currently priced for 2 more rate cuts during the next 3 FOMC meetings. That outcome is possible, but our base case economic outlook is more consistent with 1 further cut, likely occurring this month. TIPS: Overweight Chart 8Inflation Compensation Inflation Compensation Inflation Compensation TIPS underperformed the duration-equivalent nominal Treasury index by 38 basis points in September, dragging year-to-date excess returns down to -142 bps. The 10-year TIPS breakeven inflation rate fell 3 bps in September, and then another 2 bps last week. It currently sits at 1.51%, well below levels consistent with the Fed’s target. The divergence between the actual inflation data and inflation expectations is becoming increasingly stark. Trimmed mean PCE inflation has been fluctuating around the Fed’s target for most of the year (Chart 8). However, long-maturity TIPS breakeven inflation rates remain stubbornly low, nowhere near the 2.3% - 2.5% range that is consistent with the Fed’s target. As we have pointed out in prior research, it can take time for expectations to adapt to a changing macro environment.8 That being said, the 10-year TIPS breakeven inflation rate is currently 43 bps too low according to our Adaptive Expectations Model, a model whose primary input is 10-year trailing core inflation (panel 4). It is highly likely that the Fed will have to tolerate some overshoot of its 2% inflation target in order to re-anchor inflation expectations near desired levels. We anticipate that the committee will do so, and we maintain our view that long-dated TIPS breakevens will move above 2.3% before the end of the cycle. ABS: Underweight Chart 9ABS Market Overview ABS Market Overview ABS Market Overview Asset-Backed Securities underperformed the duration-equivalent Treasury index by 2 basis points in September, dragging year-to-date excess returns down to +72 bps. The index option-adjusted spread for Aaa-rated ABS widened 2 bps on the month. It currently sits at 36 bps, very close to its minimum pre-crisis level (Chart 9). ABS also appear unattractive on a risk/reward basis, as both Aaa-rated auto loans and credit cards have moved into the “Avoid” quadrant of our Excess Return Bond Map (Appendix C). The Map uses each bond sector’s spread, duration and volatility to calculate the likelihood of earning or losing 100 bps of excess return versus Treasuries on a 12-month horizon. At present, the Map shows that ABS offer poor expected return for their level of risk. In addition to poor valuation, the ABS sector’s credit fundamentals are shifting in a negative direction. Household interest payments continue to trend up, suggesting a higher delinquency rate in the future (panel 3). Meanwhile, senior loan officers continue to tighten lending standards for both credit cards and auto loans. Tighter lending standards usually coincide with rising delinquencies (bottom panel). All in all, the combination of poor value and deteriorating credit quality leads us to recommend an underweight allocation to consumer ABS. Non-Agency CMBS: Neutral Chart 10CMBS Market Overview CMBS Market Overview CMBS Market Overview Non-Agency Commercial Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 9 basis points in September, bringing year-to-date excess returns up to +227 bps. The index option-adjusted spread for non-agency Aaa-rated CMBS held flat on the month, before widening 4 bps last week. It currently sits at 75 bps, below average pre-crisis levels but above levels seen in 2018 (Chart 10). The macro outlook for commercial real estate is somewhat unfavorable, with lenders tightening loan standards (panel 4) amidst falling demand (bottom panel). Commercial real estate prices have accelerated of late, but are still not keeping pace with CMBS spreads (panel 3). Despite the poor fundamental picture, our Excess Return Bond Map shows that CMBS offer a reasonably attractive risk/reward trade-off compared to other bond sectors (see Appendix C). Agency CMBS: Overweight Agency CMBS outperformed the duration-equivalent Treasury index by 2 basis points in September, bringing year-to-date excess returns up to +90 bps. The index option-adjusted spread held flat on the month, before widening by 5 bps last week. It currently sits at 61 bps. The Excess Return Bond Map in Appendix C shows that Agency CMBS offer high potential return compared to other low-risk spread products. Appendix A - The Golden Rule Of Bond Investing We follow a two-step process to formulate recommendations for bond portfolio duration. First, we determine the change in the federal funds rate that is priced into the yield curve for the next 12 months. Second, we decide – based on our assessments of the economy and Fed policy – whether the change in the fed funds rate will exceed or fall short of what is priced into the curve. Most of the time, a correct answer to this question leads to the appropriate duration call. We call this framework the Golden Rule Of Bond Investing, and we demonstrated its effectiveness in the U.S. Bond Strategy Special Report, “The Golden Rule Of Bond Investing”, dated July 24, 2018, available at usbs.bcaresearch.com. Chart 11 illustrates the Golden Rule’s track record by showing that the Bloomberg Barclays Treasury Master Index tends to outperform cash when rate hikes fall short of 12-month expectations, and vice-versa. Chart 11The Golden Rule's Track Record The Golden Rule's Track Record The Golden Rule's Track Record At present, the market is priced for 74 basis points of cuts during the next 12 months. We anticipate fewer rate cuts over that time horizon, and therefore anticipate that below-benchmark portfolio duration positions will profit. We can also use our Golden Rule framework to make 12-month total return and excess return forecasts for the Bloomberg Barclays Treasury index under different scenarios for the fed funds rate. Excess returns are relative to the Bloomberg Barclays Cash index. To forecast total returns we first calculate the 12-month fed funds rate surprise in each scenario by comparing the assumed change in the fed funds rate to the current value of our 12-month discounter. This rate hike surprise is then mapped to an expected change in the Treasury index yield using a regression based on the historical relationship between those two variables. Finally, we apply the expected change in index yield to the current characteristics (yield, duration and convexity) of the Treasury index to estimate total returns on a 12-month horizon. The below tables present those results, along with 95% confidence intervals. Excess returns are calculated by subtracting assumed cash returns in each scenario from our total return projections. Crunch Time Crunch Time Crunch Time Crunch Time Appendix B - Butterfly Strategy Valuation The following tables present the current read-outs from our butterfly spread models. We use these models to identify opportunities to take duration-neutral positions across the Treasury curve. The following two Special Reports explain the models in more detail: U.S. Bond Strategy Special Report, “Bullets, Barbells And Butterflies”, dated July 25, 2017, available at usbs.bcaresearch.com U.S. Bond Strategy Special Report, “More Bullets, Barbells And Butterflies”, dated May 15, 2018, available at usbs.bcaresearch.com Table 4 shows the raw residuals from each model. A positive value indicates that the bullet is cheap relative to the duration-matched barbell. A negative value indicates that the barbell is cheap relative to the bullet. Table 5 scales the raw residuals in Table 4 by their historical means and standard deviations. This facilitates comparison between the different butterfly spreads. Table 6 flips the models on their heads. It shows the change in the slope between the two barbell maturities that must be realized during the next six months to make returns between the bullet and barbell equal. For example, a reading of +48 bps in the 5 over 2/10 cell means that we would only expect the 5-year to outperform the 2/10 if the 2/10 slope steepens by more than 48 bps during the next six months. Otherwise, we would expect the 2/10 barbell to outperform the 5-year bullet. Table 4Butterfly Strategy Valuation: Raw Residuals In Basis Points (As of October 4, 2019) Crunch Time Crunch Time Table 5Butterfly Strategy Valuation: Standardized Residuals (As of October 4, 2019) Crunch Time Crunch Time Table 6 Crunch Time Crunch Time Appendix C - Excess Return Bond Map The Excess Return Bond Map is used to assess the relative risk/reward trade-off between different sectors of the U.S. fixed income market. The Map employs volatility-adjusted breakeven spread analysis to show how likely it is that a given sector will earn/lose money during the subsequent 12 months. The Map does not incorporate any macroeconomic view. The horizontal axis of the Map shows the number of days of average spread widening required for each sector to lose 100 bps versus a position in duration-matched Treasuries. Sectors plotting further to the left require more days of average spread widening and are therefore less likely to see losses. The vertical axis shows the number of days of average spread tightening required for each sector to earn 100 bps in excess of duration-matched Treasuries. Sectors plotting further toward the top require fewer days of spread tightening and are therefore more likely to earn 100 bps of excess return. Chart 12Excess Return Bond Map (As Of October 4, 2019) Crunch Time Crunch Time Ryan Swift, U.S. Bond Strategist rswift@bcaresearch.com Footnotes 1 Please see U.S. Bond Strategy Weekly Report, “Corporate Bond Investors Should Not Fight The Fed”, dated September 17, 2019, available at usbs.bcaresearch.com 2 For more details on how we arrive at our spread targets please see U.S. Bond Strategy Weekly Report, “The Value In Corporate Bonds”, dated February 19, 2019, available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, “Corporate Bond Investors Should Not Fight The Fed”, dated September 17, 2019, available at usbs.bcaresearch.com 4 The 12-month breakeven spread is the spread widening required to break even with a duration-matched position in Treasuries on a 12-month horizon. It can be approximated by OAS divided by duration. 5 For more details on how we arrive at our spread targets please see U.S. Bond Strategy Weekly Report, “The Value In Corporate Bonds”, dated February 19, 2019, available at usbs.bcaresearch.com 6 Please see U.S. Bond Strategy Weekly Report, “Two Themes And Two Trades”, dated October 1, 2019, available at usbs.bcaresearch.com 7 Please see U.S. Bond Strategy Weekly Report, “Two Themes And Two Trades”, dated October 1, 2019, available at usbs.bcaresearch.com 8 Please see U.S. Bond Strategy Weekly Report, “Adaptive Expectations In The TIPS Market”, dated November 20, 2018, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification Corporate Sector Relative Valuation And Recommended Allocation

Highlights Corporate Bonds: High corporate debt levels will be a problem for corporate bond investors during the next downturn, but spreads will not respond to them until inflationary pressures mount and monetary policy turns restrictive. Maintain an overweight allocation to corporate bonds versus Treasuries, with a preference for the Baa and high-yield credit tiers. MBS: Agency MBS spreads are competitive with high-rated (Aaa, Aa, A) corporate bonds, and look even more attractive on a risk-adjusted basis. We recommend that investors swap the Aaa, Aa and A-rated corporate bonds in their portfolios for agency MBS. Municipal Bonds: Investors should upgrade municipal bonds from neutral to overweight, given the recent back-up in Municipal / Treasury yield ratios. Within munis, investors should retain a preference for long-maturity Aaa-rated bonds, where yields are most compelling. Feature We attended BCA’s annual Investment Conference last week. The event always provides a good opportunity to hear from some expert panelists and find out what issues are front and center in our clients’ minds. More than anything else, two themes kept popping up in the different presentations and in conversations with attendees: Large corporate debt balances Under-priced inflation risk We can’t help but see a strong connection between the two. On Corporate Debt The consensus among panelists and attendees was very much in line with our own view: Highly levered balance sheets will be a problem for corporate bond investors during the next default cycle, but don’t help us determine when that default cycle will occur. Chart 1 shows that, despite the persistent increase in the debt-to-profits ratio, corporate bankruptcies are well contained. We examined the reasons for this divergence in a recent report, concluding that accommodative monetary policy is holding down the default rate by keeping interest costs low and giving banks the confidence to roll over maturing debt.1 Essentially, banks will look through signs of deteriorating corporate balance sheet health until the Fed shifts to a more restrictive policy stance. Chart 1Corporate Balance Sheets Are In Bad Shape, But Defaults Are Low Corporate Balance Sheets Are In Bad Shape, But Defaults Are Low Corporate Balance Sheets Are In Bad Shape, But Defaults Are Low On Inflation This is where inflation becomes important. The Fed is currently running an accommodative monetary policy because many years of low prices have convinced investors that inflation might never return. As a result, the 10-year TIPS breakeven inflation rate is only 1.53%, well below the 2.3% - 2.5% range consistent with the Fed’s target. The Fed must maintain an accommodative policy stance until it achieves its goal of re-anchoring inflation expectations. Only then will monetary policy turn restrictive, raising the risk of a corporate default cycle. We have long held the view that a 10-year TIPS breakeven inflation rate above 2.3% would cause us to turn much more cautious on corporate credit. It might take many months of core inflation printing near the Fed’s target before investors start to believe that it will stay there indefinitely. Many conference panelists thought that inflation risks are currently under-priced, and while we tend to agree that it is premature to declare the death of the Phillips curve, we expect it will still take some time before inflation expectations hit our 2.3% - 2.5% target range. We have shown in prior research that inflation expectations adapt only slowly to changes in the actual inflation data.2  At present, the fair value reading from our Adaptive Expectations Model of the 10-year TIPS breakeven inflation rate is only 1.94% (Chart 2). This fair value will move higher if inflation continues to print near current levels, but that process will take some time. In other words, it might take many months of core inflation printing near the Fed’s target before investors start to believe that it will stay there indefinitely. Chart 2Adaptive Expectations Model Adaptive Expectations Model Adaptive Expectations Model Chart 3Inflation Not Far From Target Inflation Not Far From Target Inflation Not Far From Target While the adaptive process might take a long time, it’s important to note that inflation is already quite close to the Fed’s target. Trailing 12-month trimmed mean PCE inflation came in at 1.96% in August, while year-over-year core PCE hit 1.77% (Chart 3). Trimmed mean inflation has been more stable than other inflation measures since the financial crisis, and core PCE has tended to drift toward the trimmed mean over time.      On Corporate Debt & Inflation In our view, the two themes of high corporate debt and under-priced inflation risk are tightly linked. It has taken a very long time for the economy to recover from the financial crisis. As a result, inflation has been low for a prolonged period and the Fed has been forced to maintain an accommodative policy stance. That accommodative policy stance encourages banks to extend credit, and encourages firms to issue debt. Eventually, inflation pressures will mount, the Fed’s policy will turn restrictive and weak corporate balance sheets will be exposed. Only then, will corporate spreads widen significantly. Until that time, the pertinent question is whether corporate spreads offer adequate compensation for the risk that inflationary pressures emerge earlier than anticipated. For now, our answer is yes, with the caveat that the risk/reward trade-off is more attractive in the lower credit tiers. The 12-month high-yield breakeven spread is very attractive, well above its historical median (Chart 4). But within investment grade, we view only the Baa-rated credit tier as offering adequate compensation (Chart 4, bottom panel). There are better alternatives to owning Aaa, Aa and A-rated corporate bonds, as discussed in the next section. Chart 4Corporate Bond Valuation Corporate Bond Valuation Corporate Bond Valuation Favor Agency MBS Over High-Rated Corporate Credit Chart 5MBS More Attractive Than High-Rated Corporate Bonds MBS More Attractive Than High-Rated Corporate Bonds MBS More Attractive Than High-Rated Corporate Bonds As noted above, investment grade corporate bonds rated A or higher don’t offer much expected compensation at current spread levels. In fact, our prior research notes that their spreads are already below our cyclical targets.3 But on the plus side, the average option-adjusted spread (OAS) for conventional 30-year agency MBS has widened in recent months and now looks like an attractive alternative to high-rated corporate credit. We recommend that investors shift out of Aaa, Aa and A-rated corporate credit and into agency MBS for three reasons. 1) Expected Compensation Is Competitive The average OAS for conventional 30-year agency MBS now stands at 52 bps. This is only 6 bps below the average OAS offered by a Aa-rated corporate bond, and 37 bps less than that offered by an A-rated credit (Chart 5). That’s not bad for a Aaa-rated bond with agency backing. 2) Risk-Adjusted Compensation Is Stellar MBS spreads look much more attractive when we consider the risk profile. Specifically, when we consider that the average duration of the MBS index has fallen sharply this year, while the average duration of the investment grade corporate bond index has risen (Chart 5, panel 2). In fact, the average duration of the MBS index is only 2.9, compared to 7.8 for an A-rated corporate bond. This means that the MBS spread needs to widen by 18 bps over the next 12 months for an investor to see losses, while the A-rated spread needs to widen by only 11 bps (Chart 5, bottom panel). We recommend that investors shift out of Aaa, Aa and A-rated corporate credit and into agency MBS. Because MBS exhibit negative convexity, their duration declines when yields fall. By contrast, non-callable investment grade corporate bonds have positive convexity and have seen their durations rise. This means that, all else equal, negatively convex securities start to look more attractive on a risk-adjusted basis after a large decline in bond yields. This is also the main reason why negatively convex high-yield corporate bonds currently look much more attractive than investment grade corporate bonds.4 Interestingly, MBS did not look so attractive relative to corporate bonds in 2015/16, the last time that MBS index duration fell sharply. That’s because corporate bond spreads also widened during that period. This time around, corporate bond spreads have been stable as MBS index duration has plunged. Unless you think that Treasury yields have further downside, which we do not,5 agency MBS look like a good buy. 3) Macro Risks Are Lower While, as discussed above, we are not yet sounding the alarm about the macro risks to corporate bonds, we are even less concerned about the macro risks surrounding agency MBS. Mortgage refinancing activity is the most important macro driver of MBS spreads, and it should stay relatively low for a very long time. At such low mortgage rates, most homeowners have already had an opportunity to refinance, so refi burnout is currently very high. This is obvious when we observe that there was only a small spike in refi activity this year, despite a very large decline in mortgage rates (Chart 6). Chart 6Muted Refi Activity Will Keep Nominal Spreads Low Muted Refi Activity Will Keep Nominal Spreads Low Muted Refi Activity Will Keep Nominal Spreads Low Chart 6 also shows that the nominal MBS spread is highly correlated with refi activity, and that it remains near its historical tights. This spread contains both the OAS – which is a proxy for an MBS investor’s expected return – and the portion of the spread that is expected to be lost as a result of prepayment activity. The fact that the OAS is reasonably elevated compared to history while the overall nominal spread remains low means that MBS are pricing-in very little buffer for prepayment losses. Given the macro back-drop, this seems appropriate. Beyond refi risk, we also note that the credit quality of outstanding mortgages remains very high. The median FICO score on new mortgages has barely come down since the financial crisis (Chart 7). Further, while mortgage lending standards have been easing for the bulk of the post-crisis period, the Fed’s July Senior Loan Officer survey reported that the banks that view lending standards as tighter than the post-2005 average outnumber those that view standards as easier. Stronger housing activity data generally lead to higher mortgage rates, which in turn limit refi activity. Finally, there is very little reason to be concerned about significant weakness in housing activity. Of the six major housing activity data series that we track, all have rebounded sharply since this year’s drop in mortgage rates (Chart 8). Stronger housing activity data generally lead to higher mortgage rates, which in turn limit refi activity. Chart 7Mortgage Lending Standards Are Tight Mortgage Lending Standards Are Tight Mortgage Lending Standards Are Tight Chart 8Housing Activity Hooking Up Housing Activity Hooking Up Housing Activity Hooking Up   Bottom Line: Agency MBS spreads are competitive with high-rated (Aaa, Aa, A) corporate bonds, and look even more attractive on a risk-adjusted basis. We recommend that investors swap the Aaa, Aa and A-rated corporate bonds in their portfolios for agency MBS. Upgrade Municipal Bonds On July 23, we advised investors to reduce municipal bond exposure from overweight to neutral.6 The rationale was purely valuation driven. We saw no immediate signs of municipal credit distress, but noted that yields were simply too low relative to the alternatives. Today, we similarly see no signs of immediate credit distress. In fact, municipal bond ratings upgrades continue to outpace downgrades, our Municipal Health Monitor remains in “improving health” territory and state & local government interest coverage is strong (Chart 9).7 Chart 9Muni Credit Quality Is Not A Concern Muni Credit Quality Is Not A Concern Muni Credit Quality Is Not A Concern The difference, however, is that yield ratios have rebounded dramatically since early August, and municipal bonds have once again become attractive (Chart 10). Chart 10Munis Attractive Once Again Munis Attractive Once Again Munis Attractive Once Again Bottom Line: Investors should upgrade municipal bonds from neutral to overweight, given the recent back-up in Municipal / Treasury yield ratios. Within munis, investors should retain a preference for long-maturity Aaa-rated bonds, where yields are most compelling.   Ryan Swift, U.S. Bond Strategist rswift@bcaresearch.com Footnotes 1 Please see U.S. Bond Strategy Weekly Report, “Corporate Bond Investors Should Not Fight The Fed”, dated September 17, 2019, available at usbs.bcaresearch.com 2 Please see U.S. Bond Strategy Weekly Report, “Adaptive Expectations In The TIPS Market”, dated November 20, 2018, available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, “Corporate Bond Investors Should Not Fight The Fed”, dated September 17, 2019, available at usbs.bcaresearch.com 4 The high-yield bond index is negatively convex because most high-yield credits carry embedded call options. Investment grade corporate bonds tend to be non-callable. 5 Please see U.S. Bond Strategy Weekly Report, “What’s Up In U.S. Money Markets?”, dated September 24, 2019, available at usbs.bcaresearch.com 6 Please see U.S. Bond Strategy Weekly Report, “A Message To The TIPS Market”, dated July 23, 2019, available at usbs.bcaresearch.com 7 For further details on our Municipal Health Monitor please see U.S. Bond Strategy Special Report, “Trading The Municipal Credit Cycle”, dated October 18, 2016, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Investors should pay particular attention to definition and methodology when evaluating value versus growth strategies, both academically and in practice. Value investors should focus on non-U.S. markets, especially the emerging market small-cap universe. Growth investors should focus on large caps, especially the U.S. large-cap universe. Small-cap investors should focus on value. Large- and mid-cap investors should not be making bets between value and growth strategically. Tactical style rotation should be done only when valuation spreads reach extreme levels.  GAA remains neutral on value versus growth, but prefers to use sector positioning (cyclicals versus defensives, financials versus tech and health care) and country positioning (euro area versus U.S.) to implement style tilts. Investing by way of style is as old as investing itself. Value versus growth has been one of the most frequently asked questions among our clients of late, particularly given the sharp style reversal in recent weeks. In this report, we attempt to answer some of the most often-asked questions on value versus growth. We have arranged these questions into five separate sections: First, we look at 93 years of history of the Fama-French value and growth portfolios to see how value, growth, and size have interacted over time, because academics have mostly used the Fama-French framework. Second, we look at how comparable U.S. style indices are, including the S&P, the Russell and the MSCI, since practitioners mostly use these commercial indices as their benchmarks. Third, we investigate if international markets share the same value-growth performance cycles as the U.S., using the MSCI suite of value-growth indices (since MSCI is the only index provider that produces value-growth indices for each market under its global coverage). Fourth, we investigate if pure exposure to value and growth can actually improve the value-growth performance spread by comparing the pure style indices from the S&P and the Russell to their standard counterparts. Finally, we present the GAA approach to style tilts in a section on our investment conclusions. 1. Is It True That Value Outperforms Growth In The Long Run? There has been overwhelming academic evidence supporting the existence of the value premium.1 Academically, the “value premium”, also known as the HML (high minus low) factor premium, or the value outperformance, is defined as the return differential between the cheapest stocks and the most expensive. Even though Fama and French used book-to-price as the sole valuation criterion,2 many researchers have combined book-to-price with other valuation measures such as earnings-to-price, sales-to-price, dividend yield,3 and so on.  There is also academic evidence suggesting that “value outperformance is almost non-existent among large-cap stocks.”4 What is more, in 2014 Fama and French caused a huge stir by publishing “A Five-Factor Asset Pricing Model” working paper demonstrating that “HML is a redundant factor” because “the average HML return is captured by the exposure of the HML to other factors” (such as size, profitability, and investment pattern) based on U.S. data from 1963 to 2013.5 Asset owners and allocators should pay special attention when selecting benchmarks for value and growth. For non-quant practitioners, especially the long-only investors, value and growth are two separate investment styles, even though the style classification shares the same principle as the academic “value factor.” Their definitions vary, as evidenced by how S&P Dow Jones, FTSE Russell, and MSCI define their value and growth indexes (see next section on page 7). In general, value stocks are cheap, with lower-than-average earnings growth potential, while growth stocks have higher-than-average earnings growth potential but are very expensive. The indices published by commercial index providers do not have very long histories, however. Fortunately, Fama and French also provide value-growth-size portfolios on their publicly available website.6 Table II-1 shows that for 93 years, from July 1926 to June 2019, U.S. value portfolios in both large-cap and small-cap buckets based on the well-known Fama-French approach have returned more than their growth counterparts, no matter whether the portfolios are equal-weighted or market-cap-weighted. Most strikingly, equal-weighted small-cap value outperformed its growth counterpart by over 10% a year in absolute terms, and has more than doubled the risk-adjusted return compared to its growth counterpart. Table II-1Fama-French Value-Growth-Size Portfolio Performance* October 2019 October 2019 Some media reports have claimed that value stocks are “less volatile” because they are on average “larger and better-established companies.”7 This may be true for some specific time periods. For the 93 years covered by Fama and French, however, this common belief is not supported. In fact, value portfolios in both the large- and small-cap universes have consistently had higher volatility than growth portfolios, no matter how the components are weighted. The excess returns, however, have more than offset the higher volatilities in three out of four pairs, with the exception being market cap-weighted large-cap growth, which has a slightly higher risk-adjusted return due to much lower volatility than its value counterpart. From a very long-term perspective, the value outperformance does come from taking higher risk. Further investigation shows that the superior long-run outperformance of value relative to growth came mostly in the first 80 years of Fama and French’s 93-year sample. In more recent years since 2007, however, value has underperformed growth significantly in three out of the four Fama-French value-growth pairs, with the equal-weighted small-cap value-growth pair being the sole exception, as shown in Table II-2. Even though the equal-weighted small-cap value has still outperformed its growth counterpart in the most recent period, the hit ratio drops to 54% compared to 76% in the first 80 years, while the magnitude of average calendar-year outperformance drops to a meager 1.3%, compared to 12.5% in the first 80 years. Table II-2The Fight Between Value And Growth* October 2019 October 2019 Statistical analysis is sensitive to the time period chosen. How have value and growth been performing over time? Chart II-1 shows the long-term dynamics among value, growth, and size. The following conclusions are clear: Chart II-1Fama-French Value-Growth-Size Peformance Dynamics* Fama-French Value-Growth-Size Peformance Dynamics* Fama-French Value-Growth-Size Peformance Dynamics* Value investors should favor small caps over large caps, while growth investors should do the opposite, favoring large caps over small caps, albeit with much less potential success (Chart II-1, panel 1). Small-cap investors should favor value stocks over growth stocks (panel 2). Value outperformance in the large-cap space (panel 3) is much weaker than in the small-cap space (panel 2). Fama and French define small and large caps based on the median market cap of all NYSE stocks on CRSP (Center for Research In Security Prices), then use the NYSE median size to split NYSE, AMEX and NASDAQ (after 1972) into a small-cap group and a large-cap group. The value and growth split is based on book-to-price, with stocks in the lowest 30% classified as growth, and the highest 30% as value. Interestingly, small-cap value and small-cap growth account for only a very small portion of the entire universe, as shown in Charts II-2A and II-2B. Value stocks’ average market cap is about half of that of growth stocks, in both the large- and small-cap universes (panel 3 in Charts II-2A and II-2B). Again, this does not support some media claims that value stocks are larger and better-established companies. However, it does add further support to the claim that all investors should favor small-cap value stocks. Unfortunately, “small-cap value” is a very small universe. As of June 2019, the CRSP total U.S. equity market cap was $26.2 trillion, with small-cap value accounting for only 1.5% (about $383 billion); even large-cap value comprises only a relatively small weight, 13% (US$3.5 trillion). Chart II-2ASmall-Cap Value-Growth Portfolios* Small-Cap Value Growth Portfolios Small-Cap Value Growth Portfolios Chart II-2BLarge-Cap Value-Growth Portfolios* Large-Cap Value Growth Portfolios Large-Cap Value Growth Portfolios   The U.S. market is dominated by large-cap growth stocks with a heavy weight of 56% (US$14.7 trillion, as of June 2019). This is encouraging because academic research does show that the value premium among large caps is weak. But the large-cap value weakness mostly started from 2007, after 80 years of strength relative to large-cap growth (Chart II-1, panel 3). The Fama-French approach is widely used in academic research, partly due to its long history from 1926. For non-quant practitioners, especially long-only investors, however, commercial indexes from FTSE Russell, S&P Dow Jones, and MSCI are more often used as performance benchmarks. In this report, we study a series of commercial value-growth indexes in the U.S. and globally to shed light on value-growth dynamics, and how asset allocators can incorporate them into their decision-making processes. 2. Not All U.S. Style Indexes Are Created Equal Three major index providers have style indices. They are FTSE Russell (which launched the industry’s first set of value-growth indexes in 1987), S&P Dow Jones, and MSCI. MSCI is the only provider that has a full suite of value-growth indices for all individual markets under coverage. While all three provide “standard” style indices that include the full component of the parent index, the FTSE Russell and the S&P Dow Jones also provide “pure” style indices. There are two major differences between “standard” and “pure” style indices: 1) the standard indices are market-cap weighted, while the “pure” indices are weighted based on style score. 2) Standard value and standard growth have overlapping components, while pure value and pure growth do not share any common components. We prefer to use sector and country positioning to implement style tilts tactically. Other than book-to-price, the value variable used by the Fama-French approach, the three providers have added different variables in the determination of value and growth, as shown in Table II-3. This also reflects the evolution of the industry’s understanding on value and growth. For example, when MSCI first launched its style index in 1997, it used only book-to-price, but changed its approach in May 2003 to the current “multi-factor two-dimension” framework. Table II-3Value-Growth Index Criteria October 2019 October 2019 Because of the differences in index construction methodology, value-growth indices for the U.S. have behaved differently. The S&P 500, the Russell 1000, and the MSCI standard (large and mid-cap) indices are widely followed institutional benchmarks, with back-tested history dating to the 1970s. Chart II-3 shows the relative value/growth performance dynamics from the three index providers, together with that from Fama and French (market value-weighted, to be consistent with the approach from the index providers). One can observe the following: Chart II-3Which Value/Growth? Which Value/Growth? Which Value/Growth? None of the three pairs looks exactly like Fama-French’s market-cap value-weighted value/growth. This raises the question of how historical analysis based on the long history of Fama-French value/growth portfolios can be applied to the commercial indices. In the first cycle from 1975 to February 2000, all three index pairs made a round trip, with flat performance between value and growth. Also, even though the S&P 500 and Russell 1000 were more closely correlated with one another than with the MSCI, the three were quite similar. In the current cycle that began in February 2000, however, Russell value/growth has rebounded much more strongly than the other two. But in the down period that started in 2007, the three indices performed in line with each other, as shown in Table II-4. Table II-4U.S. Style Index Performance* October 2019 October 2019 In addition, the difference between S&P and Russell does not just lie between the S&P 500 and the Russell 1000. It actually exists in every market-cap segment, as shown in Chart II-4. Unfortunately, MSCI does not provide history from 1975 for the detailed cap segments. In the current cycle since February 2000, S&P value rebounded the least between 2000 and 2006. Why? Chart II-4Know Your Benchmark Know Your Benchmark Know Your Benchmark Chart II-5Value/Growth: Russell Vs. S&P Value/Growth: Russell Vs. S&P Value/Growth: Russell Vs. S&P Further investigation reveals some interesting observations, as shown in Chart II-5. At the aggregate level, the S&P 1500, the Russell 3000 and their respective style indices have performed largely in line with one another in the most recent cycle starting from February 2000 (Chart II-5, panel 4), reflecting the industry trend of index convergence. In different market cap segments, however, the divergence is still prominent, especially in the small-cap space (panel 1). The S&P 600 has consistently outperformed the Russell 2000 in both the value and growth categories. In addition to different style factors, this consistency also reflects different universes, size distribution, and sector exposure, as explained in an earlier GAA Special Report on small caps.8 Managers with Russell 2000 as their performance benchmark could simply beat it by doing a total-return-performance swap between the Russell 2000 and the S&P 600. Bottom Line: Asset owners and allocators should pay special attention when selecting benchmarks for value and growth.  3. How Have Value And Growth Performed Globally? MSCI is the only index provider that also produces value-growth indices for each equity market under its global coverage, using the same methodology. Unfortunately, only the “standard” (i.e., large- and mid-cap) universe has a long history, dating from December 1974. Charts II-6A and II-6B show the value/growth dynamics in major DM and EM markets. The relative performance of MSCI DM value versus growth shares a similar pattern to that of the U.S. in the latest cycle since 2000, but looks very different in the period before 2000 (Chart II-6A). The ratio of EM large- and mid-cap value versus growth did not peak until February 2012, about five years after the peak of its DM peer (Chart II-6B, panel 1). On the other hand, EM small-cap value has resumed its outperformance versus growth since early 2016 after having peaked around the same time as its large-cap counterpart. Chart II-6AIs Value Dead In DM? Is Value Dead In DM? Is Value Dead In DM? Chart II-6BIs Value Dead In EM? Is Value Dead In EM? Is Value Dead In EM?   The global value/growth dynamics also show that the “value outperforming growth” effect is more prominent in the small-cap space. But why has small value also underperformed small growth in most DM markets? Our explanation is that the EM universe is much less efficient than the DM universe because there are not many quant funds dedicated to the EM small-cap space – in addition to the fact that, in general, EM small caps are much smaller than those in DM markets. This is also in line with our finding that, in general, factor premia are more prominent in the EM universe.9 Bottom Line: Value premium is more prominent in non-U.S. markets, especially the EM small-cap universe. 4. Do Pure Style Indices Improve Performance? Both S&P Dow Jones and FTSE Russell provide pure-value and pure-growth indices. Unlike the standard value-growth indices, which target about 50% of the parent market cap, the pure-style indices include only stocks with the strongest value and growth characteristics. There is no overlap between the two. In theory, the pure-style indices should outperform the standard-style indices because of their concentrated exposure to style factors. How do they do in reality? Table II-5 shows that in terms of absolute return, this is indeed the case for 14 out of the 18 pairs of indices from S&P and Russell for the period between 1998 and 2019. However, the higher returns from greater exposure to style factors have largely come from much higher volatility in 17 out of the 18 pairs. Pure style has higher volatility than standard style in general, the only exception being the Russell mid-cap value space. As such, on a risk-adjusted basis, pure style is not necessarily better. Table II-5Purer Is Not Necessarily Better October 2019 October 2019 Charts II-7A and II-7B show the different performance dynamics for the S&P and Russell families of style indices. For the S&P indices, pure growth has outperformed standard growth for the entire period in all three market-cap segments, but only the S&P 500 pure value outperformed its standard counterpart. Therefore, more concentrated exposure to style characteristics has improved the value-growth spread only in the large-cap space, but it has actually worsened the value-growth spread in the mid- and small-cap universes (Chart II-7A). Chart II-7AS&P Pure Styles* S&P Pure Styles* S&P Pure Styles* Chart II-7BRussell Pure Styles* Russell Pure Styles* Russell Pure Styles*   For the Russell indices, it’s clear that there were a lot more tech stocks in its pure-growth indices leading up to the 2000 tech bubble, because pure growth shot up significantly more than the standard growth before the bubble burst, and also crashed more severely following it. Overall, only in the small-cap space did the value-growth spread improve by the more concentrated exposure to style factors. However, this improvement was not because of the outperformance of the pure-style relative to the standard indices. In fact, both pure value and pure growth in the small-cap universe underperformed their standard counterparts, but pure growth performed even worse (Chart II-7B and Table II-5). 5. Investment Conclusions Value and growth can mean very different things and behave very differently. Investors should pay special attention to the definitions and methodologies when evaluating style indices or strategies, both academically and in practice. Depending on an investor’s mandate, the following is recommended: Value investors should focus on non-U.S. markets, especially the emerging market small-cap universe. Growth investors should focus on large caps, especially the U.S. large-cap space. Small-cap investors should focus on value. Large-and mid-cap investors should not make bets between value and growth strategically. Tactical style rotation should be done only when valuation spreads reach extreme levels. Price-to-book is the only common variable used in the determination of value and growth by academics and practitioners. Its track record as a systematic return predictor has been poor, as shown in panel 2 of Charts II-8A and II-8B. Another factor we have a long history for is dividend yield. Its predictive power is even worse than that of price-to-book (panel 3). Chart II-8AValuation Is A Poor Timing Tool In The U.S. Valuation Is A Poor Timing Tool In The U.S. Valuation Is A Poor Timing Tool In The U.S. Chart II-8BValuation Is A Poor Timing Tool Globally Valuation Is A Poor Timing Tool Valuation Is A Poor Timing Tool   Many factors have been used in conjunction with price-to-book by both academics and practitioners to time the rotation between value and growth. However, the results have been mixed. Regression models that correctly predicted in the past may not work in the future. For example, a regression model based on valuation spread and earnings-growth spread using data from January 1982 to October 1999 successfully predicted the rebound of value outperformance starting in early 2000,10 but the universal suffering of value funds over the past several years implies that this model may have given many false signals. Chart II-9 demonstrates how difficult it is to use regression models as a timing tool for value and growth rotation. A simple regression is conducted between value and growth return differentials (subsequent 60-month returns) and relative price-to-book. For data from December 1974 to July 2019, the r-squared for the MSCI world is 0.38 and for the U.S. it is 0.09. In hindsight, both models predicted the value outperformance starting in early 2000. However, the gaps between actual value and fitted value started to open, long before 2000. By late 1998, the gaps were already wider than the previous cycle lows, yet they continued to widen as value continued to underperform growth until February 2000. Chart II-9How Good Is The Fit? How Good Is The Fit? How Good Is The Fit? What should investors currently do, based on these models? The gaps are large, but not as large as in early 2000. At which point should investors start to shift into value given its more than 12 years of underperformance? We have often written that we prefer to use sector and country positioning to implement style tilts.11, 12  This preference has not changed. Value and growth indices have sector tilts that change over time. Currently, the S&P Dow Jones large- and mid-cap value indices have a clear overweight in financials but an underweight in tech and health care compared to their growth counterparts (Table II-6). Table II-6Sector Bets In Value And Growth Indices* October 2019 October 2019 Chart II-10Prefer Sector And Country Positioning To Style Prefer Sector and Country Positioning To Style Tilts Prefer Sector and Country Positioning To Style Tilts We have been neutral on value and growth, but would likely change this view if we change our country equity allocation between the U.S. and the euro area, and our equity sector allocation between cyclicals and defensives as well as between financials and information technology (Chart II-10). Xiaoli Tang Associate Vice President​​​​​​​ Global Asset Allocation   Footnotes 1     Antti Ilmanen, Ronen Israel, Tobias J. Moskowitz, Ashwin Thapar, Franklin Wang, “Factor Premia and Factor Timing: A Century of Evidence,” AQR Working Paper, July 2, 2019. 2     Eugene F. Fama and Kenneth R. French, “Common risk factors in the return on stocks and bonds,” Journal of Financial Economics, 33 (1993). 3     Clifford Asness, Andrea Frazzini, Ronen Israel and Tobias Moskowitz, “Fact, Fiction, and Value Investing,” The Journal of Portfolio Management, Vol. 42 No.1, Fall 2015. 4     Ronen Israel and Tobias J. Moskowitz, “The Role of Shorting, Firm Size and Time on Market Anomalies,” Journal of Financial Economics, Vol 108, Issue 2, May 2013 5      Eugene F. Fama and Kenneth R. French, “A Five-Factor Asset Pricing Model,” Working Paper, University of Chicago, September 2014. 6             Fama-French value-growth-size portfolios. 7     Mark P. Cussen, “Value or growth Stocks: Which are Better?” Investopedia, Jun 25, 2019. 8     Please see Global Asset Allocation Special Report titled “Small Cap Outperformance: Fact or Myth?” dated April 7, 2017, available at gaa.bcaresearch.com. 9     Please see Global Asset Allocation Special Report titled, “Is Smart Beta A Useful Tool In Global Asset Allocation?” dated July 8, 2016, available at gaa.bcaresearch.com. 10    Clifford S. Asness, Jacques A Friedman, Robert J. Krail and John M Liew, “Style Timing: Value versus Growth,” The Journal of Portfolio Management, Spring 2000. 11     Please see Global Asset Allocation Quarterly Portfolio Outlook, “Quarterly - March 2016,” dated March 31, 2016, and available at gaa. bcaresearch.com. 12     Please see Global Asset Allocation Quarterly Portfolio Outlook, “Quarterly - April 2019,” dated April 1, 2019 available at gaa.bcaresearch.com.

Highlights Investors should pay particular attention to definition and methodology when evaluating value versus growth strategies, both academically and in practice. Value investors should focus on non-U.S. markets, especially the emerging market small-cap universe. Growth investors should focus on large caps, especially the U.S. large-cap universe. Small-cap investors should focus on value. Large- and mid-cap investors should not be making bets between value and growth strategically. Tactical style rotation should be done only when valuation spreads reach extreme levels.  GAA remains neutral on value versus growth, but prefers to use sector positioning (cyclicals versus defensives, financials versus tech and health care) and country positioning (euro area versus U.S.) to implement style tilts. Feature Investing by way of style is as old as investing itself. Value versus growth has been one of the most frequently asked questions among our clients of late, particularly given the sharp style reversal in recent weeks. In this report, we attempt to answer some of the most often-asked questions on value versus growth. We have arranged these questions into five separate sections: First, we look at 93 years of history of the Fama-French value and growth portfolios to see how value, growth, and size have interacted over time, because academics have mostly used the Fama-French framework. Second, we look at how comparable U.S. style indices are, including the S&P, the Russell and the MSCI, since practitioners mostly use these commercial indices as their benchmarks. Third, we investigate if international markets share the same value-growth performance cycles as the U.S., using the MSCI suite of value-growth indices (since MSCI is the only index provider that produces value-growth indices for each market under its global coverage). Fourth, we investigate if pure exposure to value and growth can actually improve the value-growth performance spread by comparing the pure style indices from the S&P and the Russell to their standard counterparts. Finally, we present the GAA approach to style tilts in a section on our investment conclusions. 1. Is It True That Value Outperforms Growth In The Long Run? There has been overwhelming academic evidence supporting the existence of the value premium.1 Academically, the “value premium”, also known as the HML (high minus low) factor premium, or the value outperformance, is defined as the return differential between the cheapest stocks and the most expensive. Even though Fama and French used book-to-price as the sole valuation criterion,2 many researchers have combined book-to-price with other valuation measures such as earnings-to-price, sales-to-price, dividend yield,3 and so on.  There is also academic evidence suggesting that “value outperformance is almost non-existent among large-cap stocks.”4 What is more, in 2014 Fama and French caused a huge stir by publishing “A Five-Factor Asset Pricing Model” working paper demonstrating that “HML is a redundant factor” because “the average HML return is captured by the exposure of the HML to other factors” (such as size, profitability, and investment pattern) based on U.S. data from 1963 to 2013.5 For non-quant practitioners, especially the long-only investors, value and growth are two separate investment styles, even though the style classification shares the same principle as the academic “value factor.” Their definitions vary, as evidenced by how S&P Dow Jones, FTSE Russell, and MSCI define their value and growth indexes (see next section on page 7). In general, value stocks are cheap, with lower-than-average earnings growth potential, while growth stocks have higher-than-average earnings growth potential but are very expensive. The indices published by commercial index providers do not have very long histories, however. Fortunately, Fama and French also provide value-growth-size portfolios on their publicly available website.6  Table 1 shows that for 93 years, from July 1926 to June 2019, U.S. value portfolios in both large-cap and small-cap buckets based on the well-known Fama-French approach have returned more than their growth counterparts, no matter whether the portfolios are equal-weighted or market-cap-weighted. Most strikingly, equal-weighted small-cap value outperformed its growth counterpart by over 10% a year in absolute terms, and has more than doubled the risk-adjusted return compared to its growth counterpart. Table 1Fama-French Value-Growth-Size Portfolio Performance* Value? Growth? It Really Depends! Value? Growth? It Really Depends! Some media reports have claimed that value stocks are “less volatile” because they are on average “larger and better-established companies.”7 This may be true for some specific time periods. For the 93 years covered by Fama and French, however, this common belief is not supported. In fact, value portfolios in both the large- and small-cap universes have consistently had higher volatility than growth portfolios, no matter how the components are weighted. The excess returns, however, have more than offset the higher volatilities in three out of four pairs, with the exception being market cap-weighted large-cap growth, which has a slightly higher risk-adjusted return due to much lower volatility than its value counterpart. From a very long-term perspective, the value outperformance does come from taking higher risk. Further investigation shows that the superior long-run outperformance of value relative to growth came mostly in the first 80 years of Fama and French’s 93-year sample. In more recent years since 2007, however, value has underperformed growth significantly in three out of the four Fama-French value-growth pairs, with the equal-weighted small-cap value-growth pair being the sole exception, as shown in Table 2. Even though the equal-weighted small-cap value has still outperformed its growth counterpart in the most recent period, the hit ratio drops to 54% compared to 76% in the first 80 years, while the magnitude of average calendar-year outperformance drops to a meager 1.3%, compared to 12.5% in the first 80 years. Table 2The Fight Between Value And Growth* Value? Growth? It Really Depends! Value? Growth? It Really Depends! Statistical analysis is sensitive to the time period chosen. How have value and growth been performing over time? Chart 1 shows the long-term dynamics among value, growth, and size. The following conclusions are clear: Value investors should favor small caps over large caps, while growth investors should do the opposite, favoring large caps over small caps, albeit with much less potential success (Chart 1, panel 1). Small-cap investors should favor value stocks over growth stocks (panel 2). Value outperformance in the large-cap space (panel 3) is much weaker than in the small-cap space (panel 2). Chart 1Fama-French Value-Growth-Size Peformance Dynamics* Fama-French Value-Growth-Size Peformance Dynamics* Fama-French Value-Growth-Size Peformance Dynamics* Asset owners and allocators should pay special attention when selecting benchmarks for value and growth. Fama and French define small and large caps based on the median market cap of all NYSE stocks on CRSP (Center for Research In Security Prices), then use the NYSE median size to split NYSE, AMEX and NASDAQ (after 1972) into a small-cap group and a large-cap group. The value and growth split is based on book-to-price, with stocks in the lowest 30% classified as growth, and the highest 30% as value. Interestingly, small-cap value and small-cap growth account for only a very small portion of the entire universe, as shown in Charts 2A and 2B. Chart 2ASmall-Cap Value-Growth Portfolios* Small-Cap Value Growth Portfolios* Small-Cap Value Growth Portfolios* Chart 2BLarge-Cap Value-Growth Portfolios* Large-Cap Value Growth Portfolios* Large-Cap Value Growth Portfolios* Value stocks’ average market cap is about half of that of growth stocks, in both the large- and small-cap universes (panel 3 in Charts 2A and 2B). Again, this does not support some media claims that value stocks are larger and better-established companies. However, it does add further support to the claim that all investors should favor small-cap value stocks. Unfortunately, “small-cap value” is a very small universe. As of June 2019, the CRSP total U.S. equity market cap was $26.2 trillion, with small-cap value accounting for only 1.5% (about $383 billion); even large-cap value comprises only a relatively small weight, 13% (US$3.5 trillion). The U.S. market is dominated by large-cap growth stocks with a heavy weight of 56% (US$14.7 trillion, as of June 2019). This is encouraging because academic research does show that the value premium among large caps is weak. But the large-cap value weakness mostly started from 2007, after 80 years of strength relative to large-cap growth (Chart 1, panel 3). The Fama-French approach is widely used in academic research, partly due to its long history from 1926. For non-quant practitioners, especially long-only investors, however, commercial indexes from FTSE Russell, S&P Dow Jones, and MSCI are more often used as performance benchmarks. In this report, we study a series of commercial value-growth indexes in the U.S. and globally to shed light on value-growth dynamics, and how asset allocators can incorporate them into their decision-making processes. 2. Not All U.S. Style Indexes Are Created Equal Three major index providers have style indices. They are FTSE Russell (which launched the industry’s first set of value-growth indexes in 1987), S&P Dow Jones, and MSCI. MSCI is the only provider that has a full suite of value-growth indices for all individual markets under coverage. While all three provide “standard” style indices that include the full component of the parent index, the FTSE Russell and the S&P Dow Jones also provide “pure” style indices. There are two major differences between “standard” and “pure” style indices: 1) the standard indices are market-cap weighted, while the “pure” indices are weighted based on style score. 2) Standard value and standard growth have overlapping components, while pure value and pure growth do not share any common components. Other than book-to-price, the value variable used by the Fama-French approach, the three providers have added different variables in the determination of value and growth, as shown in Table 3. This also reflects the evolution of the industry’s understanding on value and growth. For example, when MSCI first launched its style index in 1997, it used only book-to-price, but changed its approach in May 2003 to the current “multi-factor two-dimension” framework. Table 3Value-Growth Index Criteria Value? Growth? It Really Depends! Value? Growth? It Really Depends! Because of the differences in index construction methodology, value-growth indices for the U.S. have behaved differently. The S&P 500, the Russell 1000, and the MSCI standard (large and mid-cap) indices are widely followed institutional benchmarks, with back-tested history dating to the 1970s. Chart 3 shows the relative value/growth performance dynamics from the three index providers, together with that from Fama and French (market value-weighted, to be consistent with the approach from the index providers). One can observe the following: Chart 3Which Value/Growth? Which Value/Growth? Which Value/Growth? None of the three pairs looks exactly like Fama-French’s market-cap value-weighted value/growth. This raises the question of how historical analysis based on the long history of Fama-French value/growth portfolios can be applied to the commercial indices. In the first cycle from 1975 to February 2000, all three index pairs made a round trip, with flat performance between value and growth. Also, even though the S&P 500 and Russell 1000 were more closely correlated with one another than with the MSCI, the three were quite similar. In the current cycle that began in February 2000, however, Russell value/growth has rebounded much more strongly than the other two. But in the down period that started in 2007, the three indices performed in line with each other, as shown in Table 4. Table 4U.S. Style Index Performance* Value? Growth? It Really Depends! Value? Growth? It Really Depends! In addition, the difference between S&P and Russell does not just lie between the S&P 500 and the Russell 1000. It actually exists in every market-cap segment, as shown in Chart 4. Unfortunately, MSCI does not provide history from 1975 for the detailed cap segments. In the current cycle since February 2000, S&P value rebounded the least between 2000 and 2006. Why? Chart 4Know Your Benchmark Know Your Benchmark Know Your Benchmark Further investigation reveals some interesting observations, as shown in Chart 5. Chart 5Value/Growth: Russell Vs. S&P Value/Growth: Russell Vs. S&P Value/Growth: Russell Vs. S&P At the aggregate level, the S&P 1500, the Russell 3000 and their respective style indices have performed largely in line with one another in the most recent cycle starting from February 2000 (Chart 5, panel 4), reflecting the industry trend of index convergence. In different market cap segments, however, the divergence is still prominent, especially in the small-cap space (panel 1). The S&P 600 has consistently outperformed the Russell 2000 in both the value and growth categories. In addition to different style factors, this consistency also reflects different universes, size distribution, and sector exposure, as explained in an earlier GAA Special Report on small caps.8 Managers with Russell 2000 as their performance benchmark could simply beat it by doing a total-return-performance swap between the Russell 2000 and the S&P 600. Bottom Line:  Asset owners and allocators should pay special attention when selecting benchmarks for value and growth.  3. How Have Value And Growth Performed Globally? MSCI is the only index provider that also produces value-growth indices for each equity market under its global coverage, using the same methodology. Unfortunately, only the “standard” (i.e., large- and mid-cap) universe has a long history, dating from December 1974. Charts 6A and 6B show the value/growth dynamics in major DM and EM markets. The relative performance of MSCI DM value versus growth shares a similar pattern to that of the U.S. in the latest cycle since 2000, but looks very different in the period before 2000 (Chart 6A). The ratio of EM large- and mid-cap value versus growth did not peak until February 2012, about five years after the peak of its DM peer (Chart 6B, panel 1). On the other hand, EM small-cap value has resumed its outperformance versus growth since early 2016 after having peaked around the same time as its large-cap counterpart. Chart 6AIs Value Dead In DM? Is Value Dead In DM? Is Value Dead In DM? Chart 6BIs Value Dead In EM? Is Value Dead In EM? Is Value Dead In EM? The global value/growth dynamics also show that the “value outperforming growth” effect is more prominent in the small-cap space. But why has small value also underperformed small growth in most DM markets? Our explanation is that the EM universe is much less efficient than the DM universe because there are not many quant funds dedicated to the EM small-cap space –  in addition to the fact that, in general, EM small caps are much smaller than those in DM markets. This is also in line with our finding that, in general, factor premia are more prominent in the EM universe.9 Bottom Line: Value premium is more prominent in non-U.S. markets, especially the EM small-cap universe. 4. Do Pure Style Indices Improve Performance? Both S&P Dow Jones and FTSE Russell provide pure-value and pure-growth indices. Unlike the standard value-growth indices, which target about 50% of the parent market cap, the pure-style indices include only stocks with the strongest value and growth characteristics. There is no overlap between the two. We prefer to use sector and country positioning to implement style tilts tactically. In theory, the pure-style indices should outperform the standard-style indices because of their concentrated exposure to style factors. How do they do in reality? Table 5 shows that in terms of absolute return, this is indeed the case for 14 out of the 18 pairs of indices from S&P and Russell for the period between 1998 and 2019. However, the higher returns from greater exposure to style factors have largely come from much higher volatility in 17 out of the 18 pairs. Pure style has higher volatility than standard style in general, the only exception being the Russell mid-cap value space. As such, on a risk-adjusted basis, pure style is not necessarily better. Table 5Purer Is Not Necessarily Better Value? Growth? It Really Depends! Value? Growth? It Really Depends! Charts 7A and 7B show the different performance dynamics for the S&P and Russell families of style indices. For the S&P indices, pure growth has outperformed standard growth for the entire period in all three market-cap segments, but only the S&P 500 pure value outperformed its standard counterpart. Therefore, more concentrated exposure to style characteristics has improved the value-growth spread only in the large-cap space, but it has actually worsened the value-growth spread in the mid- and small-cap universes (Chart 7A). Chart 7AS&P Pure Styles* S&P Pure Styles* S&P Pure Styles* Chart 7BRussell Pure Styles* Russell Pure Styles* Russell Pure Styles* For the Russell indices, it’s clear that there were a lot more tech stocks in its pure-growth indices leading up to the 2000 tech bubble, because pure growth shot up significantly more than the standard growth before the bubble burst, and also crashed more severely following it. Overall, only in the small-cap space did the value-growth spread improve by the more concentrated exposure to style factors. However, this improvement was not because of the outperformance of the pure-style relative to the standard indices. In fact, both pure value and pure growth in the small-cap universe underperformed their standard counterparts, but pure growth performed even worse (Chart 7B and Table 5). 5. Investment Conclusions Value and growth can mean very different things and behave very differently. Investors should pay special attention to the definitions and methodologies when evaluating style indices or strategies, both academically and in practice.  Depending on an investor’s mandate, the following is recommended: Value investors should focus on non-U.S. markets, especially the emerging market small-cap universe. Growth investors should focus on large caps, especially the U.S. large-cap space. Small-cap investors should focus on value. Large-and mid-cap investors should not make bets between value and growth strategically. Tactical style rotation should be done only when valuation spreads reach extreme levels. Price-to-book is the only common variable used in the determination of value and growth by academics and practitioners. Its track record as a systematic return predictor has been poor, as shown in panel 2 of Charts 8A and 8B. Another factor we have a long history for is dividend yield. Its predictive power is even worse than that of price-to-book (panel 3). Chart 8AValuation Is A Poor Timing Tool In The U.S. Value? Growth? It Really Depends! Value? Growth? It Really Depends! Chart 8BValuation Is A Poor Timing Tool Globally Value? Growth? It Really Depends! Value? Growth? It Really Depends! Many factors have been used in conjunction with price-to-book by both academics and practitioners to time the rotation between value and growth. However, the results have been mixed. Regression models that correctly predicted in the past may not work in the future. For example, a regression model based on valuation spread and earnings-growth spread using data from January 1982 to October 1999 successfully predicted the rebound of value outperformance starting in early 2000,10 but the universal suffering of value funds over the past several years implies that this model may have given many false signals. Chart 9 demonstrates how difficult it is to use regression models as a timing tool for value and growth rotation. A simple regression is conducted between value and growth return differentials (subsequent 60-month returns) and relative price-to-book. For data from December 1974 to July 2019, the r-squared for the MSCI world is 0.38 and for the U.S. it is 0.09. In hindsight, both models predicted the value outperformance starting in early 2000. However, the gaps between actual value and fitted value started to open, long before 2000. By late 1998, the gaps were already wider than the previous cycle lows, yet they continued to widen as value continued to underperform growth until February 2000.  Chart 9How Good Is The Fit? How Good Is The Fit? How Good Is The Fit? What should investors currently do, based on these models? The gaps are large, but not as large as in early 2000. At which point should investors start to shift into value given its more than 12 years of underperformance? We have often written that we prefer to use sector and country positioning to implement style tilts.11, 12 This preference has not changed. Value and growth indices have sector tilts that change over time. Currently, the S&P Dow Jones large- and mid-cap value indices have a clear overweight in financials but an underweight in tech and health care compared to their growth counterparts (Table 6). Table 6Sector Bets In Value And Growth Indices* Value? Growth? It Really Depends! Value? Growth? It Really Depends! Chart 10Prefer Sector And Country Positioning To Style Tilts Prefer Sector and Country Positioning To Style Tilts Prefer Sector and Country Positioning To Style Tilts We have been neutral on value and growth, but would likely change this view if we change our country equity allocation between the U.S. and the euro area, and our equity sector allocation between cyclicals and defensives as well as between financials and information technology (Chart 10).     Xiaoli Tang, Associate Vice President xiaoliT@bcaresearch.com Footnotes 1Antti Ilmanen, Ronen Israel, Tobias J. Moskowitz, Ashwin Thapar, Franklin Wang, “Factor Premia and Factor Timing: A Century of Evidence,” AQR Working Paper, July 2, 2019. 2Eugene F. Fama and Kenneth R. French, “Common risk factors in the return on stocks and bonds,” Journal of Financial Economics, 33 (1993). 3Clifford Asness, Andrea Frazzini, Ronen Israel and Tobias Moskowitz, “Fact, Fiction, and Value Investing,” The Journal of Portfolio Management, Vol. 42 No.1, Fall 2015.  4Ronen Israel and Tobias J. Moskowitz, “The Role of Shorting, Firm Size and Time on Market Anomalies,”Journal of Financial Economics, Vol 108, Issue 2, May 2013 5Eugene F. Fama and Kenneth R. French, “A Five-Factor Asset Pricing Model,” Working Paper, University of Chicago, September 2014. 6Fama-French value-growth-size portfolios. 7Mark P. Cussen, “Value or growth Stocks: Which are Better?” Investopedia, Jun 25, 2019. 8Please see Global Asset Allocation Special Report titled “Small Cap Outperformance: Fact or Myth?” dated April 7, 2017, available at gaa.bcaresearch.com. 9Please see Global Asset Allocation Special Report titled, “Is Smart Beta A Useful Tool In Global Asset Allocation?” dated July 8, 2016, available at gaa.bcaresearch.com 10Clifford S. Asness, Jacques A Friedman, Robert J. Krail and John M Liew, “Style Timing: Value versus Growth,” The Journal of Portfolio Management, Spring 2000. 11Please see Global Asset Allocation Quarterly Portfolio Outlook, “Quarterly - March 2016,” dated March 31, 2016, and available at gaa. bcaresearch.com. 12Please see Global Asset Allocation Quarterly Portfolio Outlook, “Quarterly - April 2019,” dated April 1, 2019 available at gaa.bcaresearch.com.  

Highlights Portfolio Strategy The contracting manufacturing sector that rekindled recession fears, the harsh reality of the Sino-American trade war weighing on profits, downbeat business confidence and mushrooming capex slowdown signals all warn that investors should tread carefully in the historically difficult equity market months of September and October. It no longer pays to be overweight gold mining equities as sentiment is stretched, the restarting of global QE will likely reverse or at least halt the drubbing in global yields and the U.S. dollar inverse correlation should reassert itself and weigh on global gold miners. EM and China ills, deflating global producer pricing power, export blues and souring financial statement metrics underscore that materials stocks have ample downside. Recent Changes Trim the Global Gold Mining index to neutral, today. Downgrade the S&P Materials sector to underweight, today. Table 1 Extend And Pretend? Extend And Pretend? Feature Equities broke out of their trading range last week, but in order for this short-covering rally to become durable, and for volatility to subside, either global growth needs to turn the corner and alleviate recession fears or the trade war needs to de-escalate materially. On the recession front Central Banks (CBs) are doing their utmost to reflate their respective economies, but the early stages of looser monetary policy have been insufficient to change the global growth trajectory. With regard to the trade war, markets cheered the news that talks between the U.S. and China will resume in September and October. The dates for talks are conveniently chosen to follow the September FOMC meeting and the October 1 70th anniversary of the People's Republic of China. The latter date implies that Washington is considering delaying the October 1 tariff hike – and it could imply that Washington does not anticipate any violent suppression of Hong Kong protesters by that time. However, the harsh reality is that the two sides are just “kicking the can down the road”. The longer the Sino-American trade war takes to conclude, the more likely it will serve as a catalyst for a repricing of risk significantly lower (top panel, Chart 1). A technical correction may be necessary to force Trump to reduce the trade pressure significantly. Even if the October 1 tariff hike is postponed it will remain a source of uncertainty ahead of the final tariff tranche slated for December 15. The bond market may offer some clues as to the extent that the escalating trade war will eventually get reflected into stocks (bottom panel, Chart 1). The equity transmission mechanism is through the earnings avenue. Simply put, rising trade uncertainty deals a blow to global trade that boosts the U.S. dollar which in turn makes U.S. exports uncompetitive in global markets, deflates the commodity complex and with a lag weighs on SPX earnings. Chart 1Tracking Trade Uncertainty Tracking Trade Uncertainty Tracking Trade Uncertainty Speaking of the economically hypersensitive manufacturing sector, last week’s ISM release made for grim reading, further fueling recession fears (the New York Fed now pegs the recession probability just shy of 38% by next August). Not only did the overall survey fall below the boom/bust line (middle panel, Chart 2), but also new orders collapsed. In fact, the drubbing in new orders is worrying and it signals that the economy is going to get worse before it gets better (top panel, Chart 2). Tack on the simultaneous rise in inventories, and the sinking new orders-to-inventories ratio (not shown) warns of additional manufacturing ills in the coming months. Importantly, export orders suffered the steepest losses plunging to 43.3. The last three times that this trade-sensitive survey subcomponent was in such a steep freefall were in 1998, 2001 and 2008, when the SPX suffered peak-to-trough losses of 20%, 49% and 57%, respectively. In fact, since the history of the data, ISM manufacturing export orders have never been lower with the exception of the GFC (Chart 3). Such a retrenchment will either mark the bottom for equities or is a harbinger of a steep equity market correction. We side with the latter as the odds of President Trump striking a real trade deal (including tech) with China any time soon are low. Chart 2Like Night Follows Day Like Night Follows Day Like Night Follows Day Similar to the ISM manufacturing/non-manufacturing divergence (bottom panel, Chart 2), business confidence is trailing consumer conference by a wide mark. Historically this flaring chasm has been synonymous with a sizable loss of momentum in the broad equity market (Chart 4). One plausible explanation is that as business animal spirits suffer a setback, CEOs are quick to prune/postpone capex plans and, at the margin, corporations retrench and short-circuit the capex upcycle. Chart 3Export Carnage Export Carnage Export Carnage Chart 4Mind The Gap Mind The Gap Mind The Gap Circling back to last week’s capex update, national accounts corroborate the financial statement data deceleration, and in some cases contraction, in capital outlays (Chart 5). As a reminder our thesis is that the EPS-to-capex virtuous upcycle is morphing into a vicious down cycle.1 This week, we downgrade a deep cyclical sector by taking profits in a niche subgroup that has served as a reliable portfolio hedge. Crucially, tech investment, that comprises almost 30% of total investment according to national accounts, is decelerating, R&D and other intellectual property investment have also hooked down, non-residential structures are on the verge of contraction, and industrial, transportation and other equipment –that have the largest weight in U.S. capex – are also quickly losing steam (Chart 6). Chart 5Capex Blues Capex Blues Capex Blues Chart 6All Capex Segments… All Capex Segments… All Capex Segments… In more detail, Charts 7 & 8 further break down capital outlays in the respective categories and reveal that worrisomely the investment spending slowdown is broad based. Chart 7…Have Rolled Over… …Have Rolled Over… …Have Rolled Over… Chart 8…Except For One …Except For One …Except For One Adding it all up, the contracting manufacturing sector that rekindled recession fears, the harsh reality of the Sino-American trade war weighing on profits, downbeat business confidence and mushrooming capex slowdown signals all warn that investors should tread carefully in the historically difficult equity market months of September and October. As a reminder, this is U.S. Equity Strategy service’s view and it contrasts with BCA’s sanguine equity market house view. This week, we downgrade a deep cyclical sector by taking profits in a niche subgroup that has served as a reliable portfolio hedge. Downgrade Materials To Underweight… Heightened economic and trade policy uncertainty has claimed the S&P materials sector as one of its victims (Chart 9). Given that our Geopolitical Strategy service’s base case remains that there will be no Sino-American trade deal by the U.S. November 2020 election, there is more downside for materials stocks and we are downgrading this niche deep cyclical sector to a below benchmark allocation.2 Beyond the U.S./China trade war inflicted wounds that materials stocks have to nurse, there are four major headwinds that they will also have to contend with in the coming months. Chart 9Trade Uncertainty Sinking Materials Trade Uncertainty Sinking Materials Trade Uncertainty Sinking Materials First, the emerging markets (EM) in general and China in particular are in a prolonged soft patch that predates the Sino-American trade war. EM stocks and EM currencies are both deflating at an accelerating pace warning that relative share prices will suffer the same fate (Chart 10). Nothing epitomizes the infrastructure spending/capex cycle more than China’s insatiable appetite for commodities and the news on that front remains dire. The Li Keqiang index continues to emit a distress signal and that is negative for materials top line growth (bottom panel, Chart 10). Second, global inflation is in hibernation and select EM producer price inflation growth series are on the verge of contraction or already outright contracting. Chinese raw materials wholesale prices are in the deflation zone and warn that U.S. materials sector profits will underwhelm (Chart 11). Chart 10Bearish EM… Bearish EM… Bearish EM… Chart 11…And China Backdrops …And China Backdrops …And China Backdrops Base metal prices are a real time indicator of the wellness of the S&P materials sector. Currently, base metals are deflating both on the back of a firming U.S. dollar and contracting global manufacturing. Such a commodity price backdrop is dampening prospects for a profit-led materials sector relative share price recovery (top & middle panels, Chart 12). Third, the materials exports outlook is darkening. Apart from the deflating effect the appreciating U.S. dollar has on commodities it also clips basic materials companies’ exports prospects. How? It renders materials related exports uncompetitive in international markets leading to market share losses. Netting it all out, EM and China ills, deflating global producer pricing power, export blues and souring financial statement metrics underscore that materials stocks have ample downside. Chart 12Weak Pricing Power And Declining Exports Weak Pricing Power And Declining Exports Weak Pricing Power And Declining Exports In addition, the latest ISM export order subcomponent plunged to multi-year lows reflecting trade war pessimism and falling global end-demand. The implication is that the export relief valve is closed for materials equities (bottom panel, Chart 12). Finally, materials sector financial statement metrics are moving in the wrong direction. Net debt-to-EBITDA is rising anew and interest coverage has likely peaked for the cycle at a time when free cash flow generation has ground to a halt (Chart 13). U.S. Equity Strategy’s S&P materials sector profit growth model encapsulates all these moving parts and warns that a severe profit contraction phase looms (Chart 14). Chart 13Financial Statement Red Flags Financial Statement Red Flags Financial Statement Red Flags Chart 14Model Says Sell Model Says Sell Model Says Sell Netting it all out, EM and China ills, deflating global producer pricing power, export blues and souring financial statement metrics underscore that materials stocks have ample downside. Bottom Line: The time is ripe to downgrade the S&P materials sector to underweight. …Via Trimming Gold Miners To Neutral The way we are executing this downgrade in the materials sector to an underweight stance is by trimming the global gold mining index to a benchmark allocation. Our thesis that gold stocks serve as a sound portfolio hedge remains intact and underpinned when: economic and trade policy uncertainty are on the rise (top panel, Chart 15) global CBs start cutting interest rates and in some cases doubling down on negative interest rates currency wars are overheating Nevertheless, what has changed is the price, and we deem that global gold miners that have gone parabolic are in desperate need of a breather. The top panel of Chart 16 shows that gold stocks have rallied 58% since the May 5, 2019 Trump tweet. This outsized four-month relative return is remarkable and likely almost fully reflects a very dovish Fed and melting real U.S. Treasury yields (TIPS yield shown inverted, bottom panel, Chart 15). A much needed pause for breath is required before the next leg of the relative rally resumes, and we opt to move to the sidelines. Chart 15Positive Backdrop… Positive Backdrop… Positive Backdrop… Chart 16…But Reflected In Prices …But Reflected In Prices …But Reflected In Prices Moreover, on the eve of the ECB’s September meeting, were President Mario Draghi to re-commence QE in the form of sovereign and corporate bond purchases as markets participants expect, counterintuitively a selloff in the bond markets would confirm that QE and its signaling is working (bottom panel, Chart 16). Ergo, this would likely exert upward pressure on global interest rates including the U.S., especially given the one-sided positioning in the respective global risk free assets. The implication is that the shiny metal and global gold miners would suffer a setback as real yields would rise further. As a reminder, gold bullion yields nothing and gold mining equities next to nothing, thus when competing safe haven assets at the margin start yielding higher, investors flee gold and gold miners and flock to risk free assets. Sentiment toward gold and global gold miners is stretched. Gold ETF holdings are at multi-year highs (second panel, Chart 17) and gold net speculative positions are at a level that has marked previous reversals. In addition, bullish consensus on gold is near 72%, a percentage last reached in 2012 (third & bottom panels, Chart 17). Similarly, relative share price momentum is also warning that global gold mining equities are currently extended (bottom panel, Chart 18). Chart 17Extreme… Extreme… Extreme… Chart 18…Sentiment …Sentiment …Sentiment Finally, while the bond market’s view of 100bps in Fed cuts in the next 12 months should have undermined the trade-weighted U.S. dollar, it has actually defied gravity and slingshot to fresh cycle highs. This is a net negative both for gold and gold mining equities as the underlying commodity is priced in U.S. dollars and enjoys an inverse correlation with the greenback. The implication is that the multi-decade inverse correlation will hold and will likely pull down gold and gold mining equities at least in the short-run (U.S. dollar shown inverted, Chart 19). In sum, the exponential rise in global gold miners is in need of a breather. Sentiment is stretched, the restating of global QE will likely reverse or at least halt the drubbing in global yields and the U.S. dollar inverse correlation should reassert itself and weigh on relative share prices Chart 19Gold Miners/Dollar Correlation Re-establishment Risk Gold Miners/Dollar Correlation Re-establishment Risk Gold Miners/Dollar Correlation Re-establishment Risk Bottom Line: Downgrade the global gold mining index to neutral, but stay tuned.   Anastasios Avgeriou, U.S. Equity Strategist anastasios@bcaresearch.com Footnotes 1      Please see U.S. Equity Strategy Weekly Report, “Capex Blues” dated September 3, 2019, available at uses.bcaresearch.com 2      Please see The Bank Credit Analyst Special Report, “Big Trouble In Greater China” dated August 29 , 2019, available at bca.bcaresearch.com Current Recommendations Current Trades Size And Style Views Stay neutral cyclicals over defensives   (downgrade alert) Favor value over growth Favor large over small caps

Highlights The lingering global manufacturing recession and the substantial drop in U.S. bond yields have been behind the decoupling between both EM stocks and the S&P 500, and cyclical and defensive equities. Neither the most recent economic data, nor the relative performance of global cyclicals, China-related plays and high-beta markets herald a broad-based and lasting risk-on phase in global markets. On the contrary, economic and market signposts continue to indicate either further bifurcation in global markets or a risk-off period. We review some of our long-standing themes and associated recommendations. Feature Global financial markets have become bifurcated. On one hand, numerous segments of global financial markets leveraged to global growth, including EM stocks, have already sold off (Chart I-1). On the other hand, share prices of growth companies, defensive stocks and global credit markets have remained resilient. Chart I-2 shows that a similar divergence has taken place within EM asset classes: EM share prices have plummeted while EM corporate credit excess returns have not dropped much. Chart I-1Bifurcated Equity Markets Bifurcated Equity Markets Bifurcated Equity Markets Chart I-2Bifurcated Markets In EM Bifurcated Markets In EM Bifurcated Markets In EM   How to explain this market bifurcation? Financial markets sensitive to global trade and manufacturing cycles have been mirroring worsening conditions in global trade and manufacturing. Some of the affected segments include: Global cyclical equity sectors. Emerging Asia manufacturing-related currencies (KRW, TWD and SGD) versus the U.S. dollar (Chart I-3). EM and DM commodity currencies (Chart I-4). Chart I-3Total Return (Including Carry): KRW, TWD And SGD Vs. USD bca.ems_wr_2019_09_05_s1_c3 bca.ems_wr_2019_09_05_s1_c3 Chart I-4EM And DM Commodity Currencies EM And DM Commodity Currencies EM And DM Commodity Currencies   Industrial and energy commodities prices. U.S. high-beta stocks as well as U.S. small caps (Chart I-5). Chart I-5U.S. High-Beta Stocks U.S. High-Beta Stocks U.S. High-Beta Stocks DM bond yields.  Crucially, the current global trade and manufacturing downturns have taken place despite robust U.S. consumer spending. In fact, our theme for the past several years has been that a global business cycle downturn would occur despite ongoing strength in American household spending. The rationale has been that China and the rest of EM combined are large enough on their own to bring down global trade and manufacturing, irrespective of strength in U.S. consumer spending. At the current juncture, one wonders whether such a market bifurcation is justified. It is not irrational. The basis for decoupling between cyclical and defensive equities has been U.S. bond yields. The substantial downshift in U.S. interest rate expectations has led to a re-rating of non-cyclicals and growth company stocks. Corporate bonds have also done well, given the background of a falling risk-free rate. Will the current market bifurcation continue? Or will these segments in global financial markets recouple and in which direction? What To Watch China rather than the U.S. has been the epicenter of this slowdown, as we have argued repeatedly in the past. Hence, a major rally in global cyclical equities and EM risk assets all hinge on a recovery in the Chinese business cycle. The basis for decoupling between cyclical and defensive equities has been U.S. bond yields. The substantial downshift in U.S. interest rate expectations has led to a re-rating of non-cyclicals and growth company stocks. Even though Caixin’s PMI for China was slightly up in August, many other economic indicators remain downbeat: The latest hard economic data out of Asia suggest that global trade/manufacturing continues to contract. Korea’s total exports in August contracted by 12.5% from a year ago, and its shipments to China plunged by 20% (Chart I-6). The import sub-component of China’s manufacturing PMI is not showing signs of amelioration (Chart I-7). The mainland’s import recovery is very critical to a revival in global trade and manufacturing. Chart I-6Korean Exports: No Recovery Korean Exports: No Recovery Korean Exports: No Recovery Chart I-7Chinese Imports To Remain Weak Chinese Imports To Remain Weak Chinese Imports To Remain Weak Chart I-8German Manufacturing Confidence German Manufacturing Confidence German Manufacturing Confidence German manufacturing IFO business expectations and current conditions both suggest that it is still early to bet on a global trade recovery (Chart I-8). Newly released August data points reveal that U.S., Taiwanese, and Swedish manufacturing new export orders continue to tumble. To gauge whether bifurcated markets will recouple and whether it will occur to the downside or the upside, investors should watch the relative performance of China-exposed markets, global cyclicals and high-beta plays – the ones that have already sold off substantially. The notion is as follows: These markets’ relative performance will likely bottom before their absolute performance recovers. If so, their relative performance will likely foretell the outlook for their absolute performance. Concerning share prices of growth companies, defensive equity sectors and credit markets, these segments are at risk because of expensive valuations and crowded investor positioning. In other words, they could sell off even if a global recession is avoided. Concerning share prices of growth companies, defensive equity sectors and credit markets, these segments are at risk because of expensive valuations and crowded investor positioning. To assess the outlook for global cyclicals and China-related plays, we are monitoring the following financial market indicators: The Risk-On/Safe-Haven currency ratio is the average of high-beta commodity currencies such as the CAD, AUD, NZD, BRL, CLP and ZAR total return (including carry) indices relative to the average of JPY and CHF total returns (including carry). This ratio is dollar-agnostic. This ratio is making a new cyclical low (Chart I-9). Hence, it presently warrants a negative view on global growth, China’s industrial sector and commodities. Global cyclical equity sectors seem to be on the edge of breaking down versus defensives (Chart I-10). This ratio does not signal ameliorating global growth conditions. Chart I-9The Risk-On/Safe-Haven Currency Ratio bca.ems_wr_2019_09_05_s1_c9 bca.ems_wr_2019_09_05_s1_c9 Chart I-10Global Cyclicals Versus Defensives Global Cyclicals Versus Defensives Global Cyclicals Versus Defensives Chart I-11U.S. High-Beta Stocks Versus S&P 500 U.S. High-Beta Stocks Versus S&P 500 U.S. High-Beta Stocks Versus S&P 500 Finally, U.S. high-beta stocks continue to underperform the S&P 500 (Chart I-11). This is consistent with overall U.S. growth deceleration. Bottom Line: Neither the most recent economic data, nor the relative performance of global cyclicals, China-related plays and high-beta markets herald a broad-based and lasting risk-on phase in global markets. On the contrary, economic and market signposts continue to foreshadow either further bifurcation in global markets or a risk-off period. Continue trading EM stocks and currencies on the short side, and underweighting EM risk assets versus DM. Our Investment Themes And Positions Some of our open positions often run for years because they reflect our long-standing themes. Our core theme has for some time been that a global trade/manufacturing recession will be generated by a growth relapse in China. To capitalize on this theme, we have been recommending a short EM stocks / long 30-year U.S. Treasurys strategy since April 2017. This recommendation has produced a 25% gain since its initiation (Chart I-12). Continue betting on lower local interest rates in emerging economies where the central bank can cut rates despite currency depreciation. To implement this theme, we have been recommending receiving swap rates in Korea and Chile for the past several years. Our reluctance to recommend an outright buy on local bonds stems from our bearish view on both currencies – the Korean won and Chilean peso. In fact, we have been shorting both the KRW and the CLP against the U.S. dollar. Chart I-13 shows that swap rates in Korea and Chile have dropped substantially since our recommendations to receive rates in these countries. More rate cuts are forthcoming in these economies, and we are maintaining these positions. Chart I-12EM Stocks Have Massively Underperformed U.S. Bonds EM Stocks Have Massively Underperformed U.S. Bonds EM Stocks Have Massively Underperformed U.S. Bonds Chart I-13Continue Receiving Rates In Korea And Chile Continue Receiving Rates In Korea And Chile Continue Receiving Rates In Korea And Chile   We have been bearish on EM banks in general and Chinese banks in particular. We have expressed these themes in a number of ways: Short EM and Chinese / long U.S. bank stocks. Short EM banks / long EM consumer staples (Chart I-14). Within Chinese banks, we have been short Chinese medium and small banks / long large ones. All these strategies remain valid. In credit markets, we have been favoring U.S. corporate credit versus EM sovereign and corporate credit. Ability to service debt is better among U.S. debtors than EM/Chinese borrowers. We have been playing this theme in the following ways: Underweight EM sovereign and corporate credit / overweight U.S. investment-grade corporates (Chart I-15). Chart I-14Short EM Banks / Long EM Consumer Staples Short EM Banks / Long EM Consumer Staples Short EM Banks / Long EM Consumer Staples Chart I-15Underweight EM Credit / Overweight U.S. Investment-Grade Corporates Underweight EM Credit / Overweight U.S. Investment-Grade Corporates Underweight EM Credit / Overweight U.S. Investment-Grade Corporates   Underweight Asian high-yield corporate credit / overweight emerging Asian investment-grade corporates. As a bet on a deteriorating political and business climate in Hong Kong, in our Special Report on Hong Kong SAR from June 27, we reiterated the following positions: Short Hong Kong property stocks / long Singapore equities. Arthur Budaghyan Chief Emerging Markets Strategist arthurb@bcaresearch.com   Mexico: Crying Out For Policy Easing The Mexican economy is heading into a full-blown recession. Most segments of the economy are in contraction, and leading indicators point to further downside. Both manufacturing and non-manufacturing PMIs are well below 50 (Chart II-1). Monetary policy remains too restrictive: Nominal and real interest rates are both very high and plunging narrow money (M1) growth is signaling  further downside in economic activity (Chart II-2). Chart II-1The Economy Is Deteriorating The Economy Is Deteriorating The Economy Is Deteriorating Chart II-2Narrow Money Points To Negative Growth Narrow Money Points To Negative Growth Narrow Money Points To Negative Growth   An inverted yield curve signifies that the central bank is behind the curve and foreshadows growth contraction (Chart II-3). Fiscal policy has tightened as the government has remained committed to achieving a primary fiscal surplus of 1% of GDP in 2019 (Chart II-4, top panel). Consequently, nominal government expenditures have been curbed (Chart II-4, bottom panel). The government’s fiscal stimulus has not been large and has been implemented too late. Chart II-3A Message From The Inverted Yield Curve A Message From The Inverted Yield Curve A Message From The Inverted Yield Curve Chart II-4Fiscal Policy Has Tightened A Lot Fiscal Policy Has Tightened A Lot Fiscal Policy Has Tightened A Lot   Finally, business confidence is extremely low due to uncertainty over President Andrés Manuel López Obrador’s (AMLO) policies towards the private sector. The president is attempting to revive business confidence, but it will take time. Chart II-5Mexico Versus EM: Domestic Bonds And Sovereign Credit Mexico Versus EM: Domestic Bonds And Sovereign Credit Mexico Versus EM: Domestic Bonds And Sovereign Credit Our major theme for Mexico has been that both monetary and fiscal policies are very tight. Consequently, we have been recommending overweight positions in Mexican domestic bonds and sovereign credit relative to their respective EM benchmarks. (Chart II-5). Recessions are bad for share prices, but in tandem with prudent macro policies, they can be positive for fixed-income markets. Meanwhile, we have been favoring the Mexican peso relative to other EM currencies due to the fact that AMLO is not as negative for the country as was initially perceived by markets. With inflation falling and the Federal Reserve cutting rates, Banxico will ease further. Yet, it will likely cut rates slower than warranted by the economy. The longer the central bank takes to ease, the lower domestic bond yields will drop. Concerning sovereign credit, investors should remain overweight Mexico within an EM credit portfolio. Mexico’s fiscal position is healthier, and macroeconomic policies will be more prudent relative to what the market is currently pricing. We continue to believe concerns about Pemex’s financing and its impact on government debt are overblown, as we discussed in detail in our previous Special Report. In July, the government released an action plan for Pemex financing. We view this plan as marginally positive. To supplement this plan, the government can use the $14.5 billion federal budget stabilization fund to fill in financing shortfalls in the coming years. Importantly, the starting point of Mexican public debt is quite low, which will allow the government to finance Pemex in the years to come by borrowing more from markets. Recessions are bad for share prices, but in tandem with prudent macro policies, they can be positive for fixed-income markets. Lastly, our overweight recommendation in Mexican stocks has not played out. However, we are maintaining it for the following reasons: Chart II-6 illustrates that when Mexican domestic bond yields decline relative to EM ones (shown inverted on Chart II-6), Mexican share prices usually outperform their EM counterparts in common currency terms. Consistent with our view that Mexican local currency bonds will outperform their EM peers, we expect Mexican stocks to outpace the EM equity benchmark. The Mexican bourse’s relative performance against EM often swings with the relative performance of EM consumer staples versus the EM equity benchmark. This is due to the large share of consumer staples stocks in Mexico (34.5%) compared to that in the EM benchmark (7%). Consumer staples stocks are beginning to outpace the EM equity index, raising the odds of Mexican equity outperformance versus its EM peers (Chart II-7). Chart II-6Local Bond Yields And Relative Stocks: Mexico Versus EM Local Bond Yields And Relative Stocks: Mexico Versus EM Local Bond Yields And Relative Stocks: Mexico Versus EM Chart II-7Consumer Staples Have A Large Weight In Mexican Bourse Consumer Staples Have A Large Weight In Mexican Bourse Consumer Staples Have A Large Weight In Mexican Bourse   We do not expect a major rally in this nation’s stock market given the negative growth outlook. Our bet is that Mexican share prices - having already deflated considerably - will drop less in dollar terms than the overall EM equity index. Bottom Line: We continue to recommend an overweight stance on Mexican sovereign credit, domestic bonds and equities relative to their respective EM benchmarks. The main risk to the Mexican peso stems from persisting selloff in EM currencies. Traders’ net long positions in the MXN are elevated posing non-trivial risk (Chart II-8). We have been long MXN versus ZAR but are taking profit today. This trade has generated a 9.7% gain since March 29, 2018. A plunging oil-gold ratio warrants a caution on this cross rate in the near term (Chart II-9). Chart II-8Investors Are Long MXN Investors Are Long MXN Investors Are Long MXN Chart II-9Take Profits On Long MXN / Short ZAR Trade Take Profits On Long MXN / Short ZAR Trade Take Profits On Long MXN / Short ZAR Trade   Juan Egaña, Research Associate juane@bcaresearch.com Arthur Budaghyan Chief Emerging Markets Strategist arthurb@bcaresearch.com Footnotes   Equities Recommendations Currencies, Credit And Fixed-Income Recommendations

Dear Client, We will not be publishing a report next week as we take an end-of-summer break. Our next report will be published on Tuesday, September 10th. Best regards, Robert Robis  Highlights Canadian Corporates: The small but growing Canadian corporate bond market has delivered performance comparable to other developed market credit over the past decade, with less duration risk and higher average credit quality compared to the larger U.S. corporate debt market. Returns: Our new regression model for Canadian corporate bond excess returns is calling for modest positive gains for Canadian corporate debt over the next year.  Corporate Health: Canadian companies’ financial health remains a positive for corporate bond returns on a cyclical basis, but high leverage and mediocre profitability are longer-term concerns. Allocation: We recommend overweight allocations into Canadian investment grade corporates, versus both Canadian government bonds and U.S. investment grade corporates. Amid elevated global policy uncertainty, favor the moderate spread volatility and attractive valuation in Canadian corporates. Feature Canadian corporate bonds do not get much attention from global fixed income investors due to the relatively small size of the market. Yet Canadian corporates have delivered returns in line with their global peers over the past decade, delivering an average excess return over Canadian government bonds (hedged into U.S. dollars) of 2.8% (Chart  of the Week). Chart 1 Looking ahead, Canadian corporates may present an opportunity for diversification in what is becoming an increasingly challenging environment for corporate bond investors, offering relatively higher yields and better credit quality with an economy that has held up well relative to the current weakening trend in global growth. In this Special Report, we outline the contours of the Canadian corporate bond market, assess the macroeconomic factors driving Canadian corporate bond returns, and survey the current overall financial health of Canadian companies. We also take a high-level look at the state of Canadian corporate debt at the sector level, while offering our recommendations on which ones to favor over the next 6-12 months. A Brief Overview Chart 2 The bulk of outstanding Canadian corporate debt is rated investment grade (IG), but this represents only 5% of the global IG market (Chart 2), using the Bloomberg Barclays Global Corporates Index as a proxy.1 However, the total market capitalization of Canadian corporate bonds is 30% of Canadian GDP – a ratio as large as seen in other major developed countries like the U.S., U.K. and Switzerland (Chart 3).  Like those other markets, Canadian companies have taken advantage of historically low borrowing rates and increased demand for income-generating assets to add leverage to their balance sheets.   Chart 3 On the demand side, Canadian corporates have traditionally been more of an institutional investment product, although domestic retail investor interest has picked up in recent years (mostly through mutual funds and exchange traded funds). The buy-and-hold nature of those local institutional investors reduces liquidity, particularly in comparison to the more widely-traded debt of Canadian federal and provincial governments. Yet according to a September 2018 Bank of Canada (BoC) report, domestic investor concerns over a perceived deterioration of Canadian corporate bond market liquidity appeared overstated.2 The report concluded that corporate bond market liquidity had generally been improving since 2010, with only short-lived bouts of illiquidity around events such as the 2011 European Debt Crisis and the 2014/15 collapse in oil prices. That medium-term improvement in liquidity was especially concentrated in high-grade corporate debt and bonds issued by banks, although the BoC concluded that liquidity and trading activity in low-grade and non-bank bonds have generally been stable. Issuance is dominated by financials, utilities, and energy companies. Unsurprisingly, the defensive utilities sector, which has high borrowing requirements, has been the top-performing industry group in 2019 (total return of +14% year-to-date) against a backdrop of falling bond yields and increased investor nervousness about future global growth (Chart 4). Yet all Canadian corporate bonds have generally performed well, with the overall Bloomberg Barclays Canadian Corporate Index delivering a total return of +8.2% so far in 2019, compared to 11.4% for Canadian equities and 5.6% for Canadian government bonds. Chart 4 Canadian corporate credit spreads have been remarkably stable since the 2008 Global Financial Crisis.  The overall index option-adjusted spread (OAS) has stayed in a range between 100-200bps, while both total and excess (duration-matched versus government debt) returns exhibiting fairly low volatility since 2008 (Chart 5). Canadian corporate credit spreads have been remarkably stable since the 2008 Global Financial Crisis.  The overall index option-adjusted spread (OAS) has stayed in a range between 100-200bps, while both total and excess (duration-matched versus government debt) returns exhibiting fairly low volatility since 2008 The duration of the benchmark Canadian IG corporate index is now 6.4 years, well below the equivalent level for U.S IG (7.9 years) even though it has steadily increased over the past decade.  Over that same period, the average credit quality has deteriorated, with 40% of the Canadian corporate index now rated BBB (Chart 6). This is below the BBB share seen in the U.S. (50%) and euro area (52%), though, making Canadian IG relatively less exposed to potential downgrades to junk bond status. Chart 5Low Volatility Of Spreads & Returns Since 2008 Low Volatility Of Spreads & Returns Since 2008 Low Volatility Of Spreads & Returns Since 2008 Chart 6Lower Share Of BBBs Compared To The U.S. & Europe Lower Share Of BBBs Compared To The U.S. & Europe Lower Share Of BBBs Compared To The U.S. & Europe Bottom Line: The small but growing Canadian corporate bond market has delivered performance comparable to other developed market credit over the past decade, with less duration risk and higher average credit quality compared to the large U.S. corporate debt market. A Fundamental Model To Forecast Canadian Corporate Bond Returns In order to help forecast Canadian corporate bond performance, we have developed a factor-based regression model of Canadian IG excess returns (in local currency terms). We first determined the independent variables in the regression by compiling a list of potential drivers of bond returns which map to four factor groups: growth, inflation, financial variables, and other miscellaneous factors. After statistically testing those factors, the insignificant and unrelated ones were dropped. The final result of this analysis is shown in Table 1. Table 1Regression Details Of The Fundamental Canadian Corporate Bond Return Model The Great White North: A Framework For Analyzing Canadian Corporate Bonds The Great White North: A Framework For Analyzing Canadian Corporate Bonds We concluded that five variables explain the bulk of Canadian corporate bond returns:  the annual percentage change in oil prices (using the Canadian benchmark, Western Canadian Select), non-residential fixed investment growth, the M3 measure of money supply growth, the Canadian dollar trade-weighted index (CAD TWI), and the level of Canadian industrial capacity utilization. We concluded that five variables explain the bulk of Canadian corporate bond returns:  the annual percentage change in oil prices (using the Canadian benchmark, Western Canadian Select), non-residential fixed investment growth, the M3 measure of money supply growth, the Canadian dollar trade-weighted index (CAD TWI), and the level of Canadian industrial capacity utilization. Chart 7A Fundamental Model Of Canadian Corporate Bond Returns A Fundamental Model Of Canadian Corporate Bond Returns A Fundamental Model Of Canadian Corporate Bond Returns Looking at recent excess return history (Chart 7), it is not surprising that oil prices significantly affect returns given the importance of Canada’s energy sector to the overall Canadian economy.  Moreover, growth in non-residential fixed asset investment also positively influences excess returns as faster capital spending can potentially increase the profitability of Canadian firms. In contrast, the inflation factors - money supply and capacity utilization – are detrimental to returns. Increases in both of those factors can result in higher inflation and rising bond yields as the BoC is forced to tighten monetary policy, which often results in rising risk premiums and wider corporate credit spreads (falling excess returns). Finally, the CAD TWI is (weakly) positively correlated to corporate bond excess returns. A stronger currency is a reflection of a strong domestic economy, but it also helps lower imported input costs for Canadian companies – both of which boost corporate profits and corporate bond returns. We now turn to the outlook for these factors over the next 6-12 months, which remain generally supportive for moderate positive excess returns for Canadian corporates: Oil prices: BCA’s commodity strategists expect global oil prices to increase moderately over the next year as global inventory drawdowns outpace expectations (Iran sanctions, Venezuela production collapsing and OPEC 2.0 production discipline are likely sources of supply restraint). In addition, if global growth starts to rebound from the end of this year, as we expect, oil demand will also rise. Non-residential fixed investment: According to the BoC’s most recent Business Outlook Survey of Canadian companies, investment spending plans of firms remain healthy – although that survey was taken at the end of June before the latest increase in uncertainty over global trade and economic growth.3  Moreover, relatively easy credit conditions have made it easier for firms to finance capex. Therefore, our baseline scenario is still to expect moderate growth in non-residential fixed capital investment, although risks are to the downside given the global macro uncertainties. Money supply: The most recent reading of the annual growth of Canadian M3 from June was a solid +7.5%. The BoC is expected to maintain an accommodative monetary policy stance, keeping the current policy rate on hold until the end of 2020. Therefore, money supply growth is likely to remain firm – a negative for Canadian corporate bond returns in our model, although perhaps less so than in the past since rapid money growth will not generate the same type of monetary tightening response from the BoC. Capacity utilization: The Canadian capacity utilization rate is currently at 81%, a meaningful pullback from the 84% level seen in early 2018. According to the latest BoC Monetary Policy Report, the Canadian economy is operating below potential (the output gap in Q1 was estimated to be between -1.25% to -0.25% of potential GDP) and that gap is only expected to close over the next two years.  Thus, capacity utilization is not expected to have a major impact on corporate excess returns over the next 6-12 months. Canadian Dollar: The CAD TWI has shown no change over the past year, and will likely remain near current levels in the short term. Although we do not expect the BoC to cut interest rates as much as currently discounted by markets (-40bps over the next twelve months), Canadian monetary policy will still remain accommodative and will likely keep the CAD relatively soft until global manufacturing growth and trade activity stabilize and begin to revive. The CAD is likely to be a neutral factor for Canadian corporate returns over the next year. Bottom Line: Our new regression model for Canadian corporate bond excess returns is calling for modest positive gains for Canadian corporate debt over the next year.  Canadian Corporate Balance Sheet Health: OK For Now, But At Risk If The Economy Weakens Chart 8The BCA Canadian Corporate Health Monitors The BCA Canadian Corporate Health Monitors The BCA Canadian Corporate Health Monitors Regular readers of our work will be familiar with our Corporate Health Monitor (CHM) framework. In this approach, we combine financial ratios that are most important for corporate creditworthiness of the entire non-financial corporate sector of a given country into a summary indicator that is designed to track corporate credit spreads. We introduced a Canadian CHM in April 2018, both using top-down national accounts data and aggregated bottom-up ratios from actual company financial statements.4  The latest reading from our top-down and bottom-up Canadian CHMs suggest that the overall health of Canadian corporates is decent, with the CHMs both below the zero line (Chart 8).5  Digging into the individual ratios, however, does reveal some potential signs of future weakness.  Leverage is relatively high, while profitability metrics and interest coverage ratios are at the low end of the historical range.  However, in our CHM framework, how the latest data compares to the medium-term trend – rather than the absolute level of the ratios - is most relevant for corporate bond performance. On that front, the latest data points for the CHM ratios do represent modest improvements versus the levels seen in 2014 and 2015, which is why our CHMs remain in the “improving health” zone. The more cyclically-driven ratios (profit margins, return on capital, interest coverage) declined amid the sharp plunge in Canadian economic growth at the end of 2018. However, given the recent reacceleration visible in some Canadian economic data, those cyclically-driven ratios may end up showing signs of stabilization, if not improvement, once the underlying CHM data for Q2/2019 and Q3/2019 are available. Looking ahead, Canadian corporate debt would be vulnerable to spread widening (rising risk premiums) in the event of a sustained slowing of the Canadian economy, given the poor absolute levels of the CHM component ratios. With the BoC maintaining an accommodative monetary policy stance, however, and the Canadian economy likely to continue growing at a trend-like pace supported by consumer spending, we think the backdrop will remain conducive to credit spread stability in Canada over the next 6-12 months. With the BoC maintaining an accommodative monetary policy stance, however, and the Canadian economy likely to continue growing at a trend-like pace supported by consumer spending, we think the backdrop will remain conducive to credit spread stability in Canada over the next 6-12 months. Bottom Line: The financial health of Canadian companies remains a positive for corporate bond returns on a cyclical basis, but there are longer-term concerns given high leverage and mediocre profitability. Canadian Corporate Bond Sector Valuation For IG corporate sectors in the U.S., euro area and the U.K., we utilize a relative value framework to rank credit spreads within the benchmark corporate universe.  We can apply that same approach to assess valuations of Canadian corporate bond sectors. In our sector relative value model, the “fair value” option-adjusted spread (OAS) for each sector within the Bloomberg Barclays Canadian IG Corporate index is estimated based on a panel regression. The explanatory variables in the regression are the modified duration, convexity and credit rating of each industry sub-sector within the index. The regression produces a set of common coefficients for all sectors that can be used to estimate a fair value OAS for each industry group as a function of its own interest rate duration, convexity and credit quality – all important drivers of corporate bond returns. The Risk-Adjusted Valuation is the difference between each sector’s current OAS and the model estimate of the sector’s fair value OAS.  A positive Risk-Adjusted Valuation implies undervaluation for the sector in question, and a negative reading implies overvaluation. Table 2 shows the recommended positioning of the Canadian IG industry sectors based on our relative value model. Sectors with positive Risk-Adjusted Valuations have overweight allocations versus the benchmark, with the opposite holds true for sectors with negative valuations. Sectors with spreads that are very close to fair value (within a range of +5bps to -5bps) have only a neutral recommended weighting versus the benchmark. Table 2Canada Investment Grade Corporate Bond Aggregate: Sector Relative Valuation* The Great White North: A Framework For Analyzing Canadian Corporate Bonds The Great White North: A Framework For Analyzing Canadian Corporate Bonds Chart 9 depicts the risk/reward tradeoff between the valuation metric and the riskiness of each sector as measured by its duration-times-spread (DTS).  Valuation is measured along the vertical axis of the chart, while DTS is measured along the horizontal axis. Sectors with higher DTS exhibit greater excess return volatility and are thus riskier. Chart 9 In the current environment of heightened uncertainty and slowing global growth, but with the BoC and other global central banks responding with a more dovish monetary policy stance, targeting cheap sectors that are less risky (i.e. DTS scores close to or below the average DTS of all sectors) is a prudent strategy.  Those would be sectors that appear in the upper left quadrant of Chart 9, like Metals & Mining, Finance Companies and Office REITs. Chart 10Positive Support For Canadian Consumer Cyclicals Positive Support For Canadian Consumer Cyclicals Positive Support For Canadian Consumer Cyclicals We also see a case for overweighting the cheap Consumer Cyclical Services sector, even with a DTS that is modestly higher than the overall index, given the continued strength in the Canadian labor market which supports consumer confidence through rising earning power (Chart 10). Recommended underweights are in the bottom right quadrant of Chart 9, with expensive valuations and high DTS scores, like Utilities: Natural Gas, Utilities: Electric, Supermarkets and Food & Beverage. Bottom Line: Favor Canadian corporate bond sectors with cheap valuations and spread volatility close to that of the overall benchmark index. Investment Conclusions Chart 11Canadian Corporates Outperformance Vs U.S. Will Continue Canadian Corporates Outperformance Vs U.S. Will Continue Canadian Corporates Outperformance Vs U.S. Will Continue Canadian IG corporates now offer a potential opportunity to diversify corporate bond exposure away from the larger markets in the U.S. and Europe.  The Canadian economy remains resilient despite slowing global growth, while the fundamental drivers of Canadian corporate bond returns are stabilizing or even improving. At the same time, the economic weakness abroad and heightened trade/political uncertainty will ensure that the BoC maintains an accommodative monetary stance over the next 6-12 months. That is not to say that Canadian corporates are not without risk. Canada is not a low-beta market - spreads do widen during “risk-off” periods in global financial markets. Also, underlying Canadian corporate credit fundamentals look poor on a long-term basis; Canadian private sector debt levels are high (especially for households); and the export-intensive Canadian economy is vulnerable to any incremental deceleration of global growth in particular, and the US more specifically.  Yet as a relative value trade versus the much larger corporate bond market to the south, Canadian corporates are well positioned to continue their recent bout of outperformance versus U.S. equivalents over the next 6-12 months, for the following reasons (Chart 11): While markets are priced for rate cuts from both the Fed and the BoC, the starting point for monetary conditions is easier in Canada than in the U.S. given the much weaker level of the Canadian dollar compared to the U.S. dollar. There is a wide gap between the corporate credit fundamentals in Canada and the U.S. according to our top-down Corporate Health Monitors for both countries, such that Canadian balance sheets are more robust. There is a wide gap between the corporate credit fundamentals in Canada and the U.S. according to our top-down Corporate Health Monitors for both countries, such that Canadian balance sheets are more robust. Bottom Line: We recommend that domestic Canadian investors continue to stay overweight Canadian corporates versus Canadian government bonds, while keeping an overall level of spread risk close to benchmark. Global credit investors that have access to the Canadian corporate bond market should consider allocations out of U.S. investment grade corporates into Canadian equivalents. Ray Park, CFA, Research Analyst ray@bcaresearch.com Robert Robis, CFA,  Chief Fixed Income Strategist rrobis@bcaresearch.com  Footnotes 1 Throughout this report, we solely use data on Canadian corporate debt from the Bloomberg Barclays bond indices, which is the main index data we use in all our global bond research. Comprehensive data is also available from other providers such as FTSE Russell and S&P Global. 2 Bank of Canada September 2018 Staff Analytical Note 2018-31, “Have Liquidity and Trading Activity in the Canadian Corporate Bond Market Deteriorated?” 3 https://www.bankofcanada.ca/2019/06/business-outlook-survey-summer-2019/ 4 Please see BCA Global Fixed Income Strategy Weekly Report, “BCA Corporate Health Monitor Chartbook Update: Growth Is Papering Over The Cracks”, dated April 24, 2018, available at gfis.bcaresearch.com 5 A CHM below zero implies improving financial health, while a CHM above zero indicates deteriorating financial health. Thus, the direction of the CHM is designed to be positively correlated with corporate credit spreads.

Highlights Portfolio Strategy The sustained global growth slowdown, widening junk spreads, along with the risk of a U.S. recession becoming a self-fulfilling prophecy suggest that caution is still warranted in the broad equity market on a 3-12 month time horizon. Weakening consumer sentiment, softening hotel industry operating metrics that point to a margin squeeze, anemic relative outlays on lodging and a decelerating ISM non-manufacturing index, all signal that more pain lies ahead for the S&P hotels, resorts & cruise lines index.   Waning industry operating metrics, a bearish signal from our EPS growth model along with the mighty U.S. dollar warns against bottom fishing in the S&P electrical components & equipment (EC&E) index.  Recent Changes There are no changes to the portfolio this week. Table 1 Elusive Growth Elusive Growth Feature The S&P 500 traded in an uncharacteristically tight range last week before falling apart on Friday on the back of a re-escalation in the U.S./China trade war. Worries of recession also resurfaced. Not only did the MARKIT flash manufacturing PMI break below the 50 expansion/contraction line, but it also pulled down the MARKIT flash services PMI survey that barely held above the boom/bust line. Adding insult to injury, the 10/2 yield curve slope inverted anew last week further fanning these recession fears. Worrisomely, consumer sentiment took a hit recently according to the University of Michigan survey (top panel, Chart 1). Importantly, what caught our attention was the following commentary: “The main takeaway for consumers from the first cut in interest rates in a decade was to increase apprehensions about a possible recession. Consumers concluded, following the Fed’s lead, that they may need to reduce spending in anticipation of a potential recession.” While the consumer is the last and most significant pillar standing for the U.S. economy, reflexivity may spoil the party and a recession may become a self-fulfilling prophecy. This is the message the bond market is sending and it is warning that the path of least resistance is a lot lower for stocks (bottom panel, Chart 1). Chart 1“The First Cut Is The Deepest” “The First Cut Is The Deepest” “The First Cut Is The Deepest” Economists are also downgrading their U.S. real GDP growth estimates and that forecast now stands at 2.3% for the current year according to Bloomberg. While the recession alarm bells are not sounding off, these downward revisions bode ill for stocks (Chart 2)  Chart 2Watch Out Down Below Watch Out Down Below Watch Out Down Below Moving to another part of the fixed income market, stress is slowly building in the high yield market especially given the recent tick up in bankruptcies and the blind sides that cove-lite loans now pose to bond investors. As a reminder, the U.S. high yield option adjusted spread (OAS) troughed last September and continues to emit a distress signal for the broad equity market (junk OAS shown inverted, top panel, Chart 3). Chart 3Mind The Gaps Mind The Gaps Mind The Gaps With regard to global growth, it is still missing in action, and given that Dr. Copper is on the verge of a breakdown, a global growth recovery is a Q1/2020 story at the earliest. This week we update a consumer discretion­ary subindex and also highlight an industrials sector subgroup. Chart 4SPX: The Next Shoe To Drop? SPX: The Next Shoe To Drop? SPX: The Next Shoe To Drop? Chart 5Risk To View Risk To View Risk To View Other financial market variables concur that global growth is elusive. J.P. Morgan’s EM FX index has broken down and EM equities are also hanging from a thread. The EM high yield OAS has broken out signaling that the risk off phase has yet to fully run its course (EM junk OAS shown inverted, bottom panel, Chart 4). Finally, there is a short-term risk to our cautious equity market view. Indiscriminate buying in U.S. Treasurys has now pushed the 10-year yield down almost 180bps from last November’s peak deeply in overvalued territory. While such a move is not unprecedented, buying may be exhausted and in need of at least a short-term breather (Chart 5).     Netting it all out, the sustained global growth slowdown, widening junk spreads, along with the risk of a U.S. recession becoming a self-fulfilling prophecy suggest that caution is still warranted in the broad equity market on a 3-12 month time horizon. As a reminder, this is U.S. Equity Strategy’s view, which contrasts BCA’s sanguine equity market house view. This week we update a consumer discretionary subindex and also highlight an industrials sector subgroup. Empty Spaces When the consumer is worried about a possible recession as the latest survey revealed, the knee jerk reaction is to tighten the purse strings and marginally retrench. The latest University of Michigan consumer sentiment survey made for grim reading and such souring in confidence will continue to weigh on lodging equities (Chart 6). As a result, we remain underweight the niche S&P hotels, resorts & cruise lines consumer discretionary subgroup. When the consumer is worried about a possible recession as the latest survey revealed, the knee jerk reaction is to tighten the purse strings and marginally retrench. Chart 6Stay Checked Out Of Hotels Stay Checked Out Of Hotels Stay Checked Out Of Hotels   Already discretionary retail sales have taken the back seat and non-discretionary retail sales are in the driver’s seat. In fact, the top panel of Chart 7 shows that the relative retail sales backdrop has plunged to levels last seen during the GFC, warning that relative share prices have ample room to fall. Drilling deeper in the consumption data is instructive. Lodging outlays are decelerating and are also trailing overall PCE. The implication is that relative profits will likely underwhelm sustaining the 18-month long de-rating phase (middle & bottom panels, Chart 7). On the operating front the news is equally dour. While selling prices are expanding, the relentless construction binge will lead to a mean reversion sooner rather than later (bottom panel, Chart 8).   Chart 7De-rating Phase To Gain Steam De-rating Phase To Gain Steam De-rating Phase To Gain Steam Chart 8Margin Squeeze Looming Margin Squeeze Looming Margin Squeeze Looming   Tack on the ongoing assault from the new sharing economy unicorns like Airbnb, and industry pricing power will remain in check in coming quarters. Similarly, the ISM non-manufacturing price subcomponent is warning that a deflation scare is looming in the lodging industry (second panel, Chart 8). Not only are selling prices under attack, but also labor-related input costs are on fire. The sector’s wage inflation is climbing at a 3.9%/annum pace or roughly 120bps higher that the overall employment cost index (third panel, Chart 8). Taken together, there are high odds that a profit margin squeeze will weigh on profits and on relative share prices (top panel, Chart 8). Importantly, the overall ISM services survey best encapsulates the bearish backdrop of the S&P hotels, resorts & cruise lines index. Historically, relative share prices have been moving in tandem with the ISM non-manufacturing survey and the current message is that selling pressures on relative share prices will persist in the coming months (Chart 9). Chart 9Heed The Message From The ISM Services Survey Heed The Message From The ISM Services Survey Heed The Message From The ISM Services Survey In sum, weakening consumer sentiment, softening hotel industry operating metrics that point to a margin squeeze, anemic relative outlays on lodging and a decelerating ISM non-manufacturing index signal that more pain lies ahead for the S&P hotels, resorts & cruise lines index. Bottom Line: Continue to avoid the S&P hotels, resorts & cruise lines index. The ticker symbols for the stocks in this index are: BLBG: S5HOTL – MAR, HLT, RCL, CCL, NCLH. Short Circuited The S&P EC&E index broke down recently (top panel, Chart 10) and we reiterate our underweight recommendation in this industrials sector subgroup. While it is tempting to bottom fish here especially given oversold technical and bombed out valuations (bottom panel, Chart 11), a number of the indicators we track suggest that more losses are around the corner. Chart 10Sell The Weakness Sell The Weakness Sell The Weakness Chart 11Good Reasons For Valuation Discount Good Reasons For Valuation Discount Good Reasons For Valuation Discount   First the trade-weighted dollar has broken out to fresh cyclical highs despite the collapse in the 10-year yield. Historically, relative share prices and the greenback are tightly inversely correlated and the current weak global growth message the U.S. dollar is emitting is bearish for the S&P EC&E index (U.S. dollar shown inverted, middle panel, Chart 10). This global growth soft patch is not only negative for new orders owing to deficient foreign demand, but the appreciating currency also makes EC&E exports less competitive in the global market place (U.S. dollar shown inverted, bottom panel, Chart 10). Second, while industry new orders have been resilient, the massive inventory buildup dwarfs new order growth and warns that a deflationary liquidation phase is looming (middle panel, Chart 11). In fact, the recent drubbing in the ISM manufacturing prices paid subcomponent portends a deflationary industry phase (third panel, Chart 12). Adding it all up, waning industry operating metrics, a bearish signal from our EPS growth model along with the mighty U.S. dollar warns against bottom fishing in the S&P EC&E index. Other operating metrics are also warning that EC&E profits will underwhelm. Industry weekly hours worked have plunged and sell-side analysts have been aggressively cutting EPS estimates (bottom panel, Chart 13). On the productivity front, executives have not adjusted labor cost structures to lower running rates yet (second panel, Chart 13) and, thus, our EC&E productivity gauge (industrials production versus employment) is contracting which bodes ill for industry earnings (third panel, Chart 13). Chart 12Weak Profit Backdrop Weak Profit Backdrop Weak Profit Backdrop Chart 13Deteriorating Operating Metrics Deteriorating Operating Metrics Deteriorating Operating Metrics   Finally, our S&P EC&E EPS growth model does an excellent job in encapsulating all these moving parts and is signaling that the path of least resistance is lower for EPS growth in the coming months (bottom panel, Chart 12). Adding it all up, waning industry operating metrics, a bearish signal from our EPS growth model along with the mighty U.S. dollar warns against bottom fishing in the S&P EC&E index. Bottom Line: Stay underweight the S&P EC&E index. BLBG: S5ELCO – AME, EMR, ETN, ROK.     Anastasios Avgeriou, U.S. Equity Strategist anastasios@bcaresearch.com   Current Recommendations Current Trades Size And Style Views Favor value over growth Favor large over small caps

Highlights The plunge in government bond yields means that There Is No Alternative to stocks: TINA. As long as bond yields stay reasonably low, stocks will continue to climb the proverbial wall of worry. Global equities are quite cheap compared with bonds. This suggests that stock market returns could be quite strong over the next couple of years, as PE multiples rise in order to narrow the gap between bond yields and earnings yields. While cheap in relative terms, global equities are modestly expensive in absolute terms. Thus, long-term absolute stock market returns are likely to be subpar, even if they are reasonably high in the near term. U.S. stocks are noticeably more expensive than their overseas peers. Differences in sector composition can explain some of the valuation gap, but not all of it. We intend to upgrade EM and European stocks later this year once global growth begins to reaccelerate. Feature Falling Bond Yields Have Made Stocks More Attractive After peaking last year, global bond yields have plunged anew (Chart 1). To a large extent, the decline in yields has been driven by the slowdown in global growth. Chart 2 shows that there is a strong correlation between government bond yields and manufacturing purchasing manager indices. Chart 1Global Bond Yields Sinking Global Bond Yields Sinking Global Bond Yields Sinking Chart 2The Decline In Bond Yields Driven By Slower Global Growth The Decline In Bond Yields Driven By Slower Global Growth The Decline In Bond Yields Driven By Slower Global Growth   Chart 3Dividend Yields Are Higher Than Bond Yields Globally Dividend Yields Are Higher Than Bond Yields Globally Dividend Yields Are Higher Than Bond Yields Globally As we discussed last week, global growth should stabilize and recover over the remainder of the year, which will cause bond yields to move somewhat higher. Nevertheless, yields are poised to stay low by historic standards – at least until inflation picks up significantly, which is unlikely to occur during the next 12 months. The fact that government bond yields in many countries are negative in real terms – and indeed, negative in nominal terms in Japan and across much of Europe – implies that the only way investors can hope to generate decent returns is by taking on more risk. This means moving further down the quality ladder in the fixed-income space, as well as turning to historically riskier asset classes such as equities. The allure that equities have in today’s low rate environment even has a name: TINA – There Is No Alternative. The S&P 500 dividend yield is currently 1.98%, 37 bps above the yield on 10-year Treasury notes. To put things in perspective, even if S&P 500 companies did not increase cash dividends at all for the next ten years, the real value of the index would still have to fall by 26% (assuming 2% inflation) for bonds to outperform stocks. The gap between dividend yields and bond yields is even greater abroad (Chart 3).   TINA’s Critics That may seem like a very low bar for success, but there are plenty of prognosticators who think stocks will still fail to clear it. TINA’s detractors offer two arguments to justify their skepticism: Today’s low interest rates are simply a reflection of poor economic growth prospects. Even if one believed that lower interest rates warrant higher equity multiples, the stock market has already priced that in. John Hussman eloquently summarized these arguments in a recent report: Another danger for investors here is the willingness to accept offensively speculative valuations on the argument that bond yields are low. The empirical fact is that bond yields are tightly correlated with nominal GDP growth. But as one can demonstrate using any discounted cash flow model, if interest rates are low because growth rates are also low, no valuation premium is “justified” by the low interest rates at all. Long-term returns will already be lower, commensurate with the lower interest rates, by virtue of the lower growth rate itself. A valuation premium then just adds insult to injury.           – John Hussman, “Questions we hear a lot,” Hussman Market Comment, (January 2019). John Hussman is a perspicuous market observer, but there are two flaws in his logic here. The first one is factual. Despite his claim to the contrary, U.S. bond yields have actually fallen more than trend nominal GDP growth over the past decade (Chart 4). The current gap between U.S. potential nominal GDP growth, as estimated by the Congressional Budget Office, and the 10-year Treasury yield is over two percentage points, the highest since 1979. Chart 4Bond Yields Have Fallen More Than Trend Nominal GDP Growth Bond Yields Have Fallen More Than Trend Nominal GDP Growth Bond Yields Have Fallen More Than Trend Nominal GDP Growth Chart 5The Trend In Global Growth Has Remained Steady Thanks To Faster-Growing EM The Trend In Global Growth Has Remained Steady Thanks To Faster-Growing EM The Trend In Global Growth Has Remained Steady Thanks To Faster-Growing EM Chart 6 At the global level, trend GDP growth has barely changed, largely because faster-growing emerging markets now make up a larger share of the global economy (Chart 5). For large multinational companies, global growth, rather than domestic growth, is the more relevant measure. If trend global GDP growth has not fallen, why are real interest rates so low? The answer is that the world is still awash with savings. As Chart 6 illustrates, an increase in desired savings will lead to a decline in real rates, even if underlying growth does not change.   No Free Lunch The second error is more subtle. Hussman discusses earnings growth and GDP growth as though they are one in the same. However, there is no long-term mechanism that magically bestows faster earnings growth on firms just because trend economic growth accelerates. Earnings growth and GDP growth are tightly correlated over the course of a business cycle because rising demand allows firms to spread fixed costs over a larger amount of output, thus increasing so-called operational leverage. But if a firm wishes to grow earnings over the long run, it has to invest in new technology and new capacity. That takes real resources — resources that could otherwise be distributed to shareholders. An example might clarify this point. Consider two firms, each with a market value of $1 million and $100,000 in earnings. Firm A distributes all its earnings to shareholders in the form of dividends. Firm B pays no dividend. Rather, it invests all its earnings in trying to grow the business. Which firm would you rather own? There is actually no simple answer to this question. If you invest in Firm A and the share price remains unchanged because the company has done nothing to grow its business, you will still receive a 10% dividend yield. You will not receive a dividend by investing in Firm B. However, if the company is able to grow earnings by 10% and the price-earnings ratio remains unchanged, the share price will rise by 10%. In both cases, you make a 10% return. The point is that there is no free lunch. Firms in fast-growing economies will be able to avail themselves of expanding domestic markets, but they will need to spend money to grow their businesses. They are also likely to face stiffer competition from new entrants. This is a key reason why Jay Ritter and other economists have shown that there is no clear correlation between long-term economic growth and shareholder returns (Chart 7). Chart 7 Priced For Perfection? Chart 8 One thing that Hussman gets completely right is that absolute long-term equity returns depend on the absolute level of valuations rather than how expensive stocks are in relation to bonds. A decline in the discount rate will increase the present value of future earnings, thus leading to a temporary period of abnormally high returns. However, once equity valuations have reset to a higher level, returns will be permanently lower. In other words, falling interest rates simply shift returns from the future into the present (Chart 8). The key question for investors is where we are in this process. If stock valuations have yet to fully adjust to the decline in interest rates, near-term returns could still be quite strong, even if they do come at the expense of subsequent returns. There is good reason to think this adjustment has yet to play out fully. The forward PE ratio for U.S. stocks is currently 16.5. If one were to use the earnings yield as a proxy for expected returns (see Appendix A for details), one would expect U.S. equities to generate long-term annualized real total returns of 6.1%. Outside the U.S., the forward PE ratio stands at 12.7, implying an expected return of 7.8% (Chart 9). Chart 9 Chart 10 shows that the equity risk premium in the U.S. based on the forward PE ratio remains quite high, indicating that the earnings yield has not fallen as much as one would have expected based on the decline in real bond yields. The equity risk premium is even higher outside the U.S., reflecting both the fact that valuations are cheaper abroad and interest rates are generally lower. Chart 10AEquity Risk Premia Remain Quite High (I) Equity Risk Premia Remain Quite High (I) Equity Risk Premia Remain Quite High (I) Chart 10BEquity Risk Premia Remain Quite High (II) Equity Risk Premia Remain Quite High (II) Equity Risk Premia Remain Quite High (II)   Are PE Ratios Biased Down? One legitimate criticism of the forward PE ratio is that it relies on analyst earnings estimates, which tend to be too optimistic. That analysts tend to be too bullish is undeniable (Chart 11). However, even if one were to use the trailing PE ratio, the implied long-term expected real total return would still be 4.8% in the U.S. and 7.1% abroad. Chart 11Analysts Are Usually Too Optimistic Analysts Are Usually Too Optimistic Analysts Are Usually Too Optimistic Moreover, as Chart 12 illustrates, projected 12-month earnings growth is currently below its historic average both in the U.S. and abroad. Thus, to the extent that forward PE ratios are biased downwards, that bias is arguably smaller than in the past. Chart 12Earnings Growth Estimates Are Not Excessive Today Earnings Growth Estimates Are Not Excessive Today Earnings Growth Estimates Are Not Excessive Today Chart 13Cyclically-Adjusted PEs Point To Subpar Absolute Future Returns Cyclically-Adjusted PEs Point To Subpar Absolute Future Returns Cyclically-Adjusted PEs Point To Subpar Absolute Future Returns   A further criticism of both trailing and forward PE ratios is that they do not take into account cyclical factors that could either flatter or depress earnings. For example, if profit margins are temporarily inflated, standard PE ratios will tend to show that stocks are cheaper than they really are. One way to address this issue is to use a moving average for earnings. The most popular tool for doing so is the Shiller PE ratio (also known as the cyclically-adjusted PE ratio, or CAPE). It divides the value of today’s stock market index by the average of real earnings over the past ten years. The Shiller PE ratio currently points to below-average, but still positive, real returns for stocks over the coming years (Chart 13). S&P 500 Margins Versus Economy-Wide Margins Some stock market bears charge that the Shiller PE ratio does not go far enough in eliminating the upward bias to earnings. They point out that even though S&P 500 profit margins were very depressed following the Global Financial Crisis, the 10-year average of margins is now well above its historic norm (Chart 14). Chart 14U.S.: 10-Year Average Of Margins Is Now Well Above Its Historic Norm U.S.: 10-Year Average Of Margins Is Now Well Above Its Historic Norm U.S.: 10-Year Average Of Margins Is Now Well Above Its Historic Norm John Hussman’s preferred measure, the Margin-Adjusted PE ratio, tries to control for this alleged problem by adjusting earnings using the economy-wide profit-to-GDP ratio. It suggests that future returns will be lower than those implied by the Shiller PE ratio. The problem with Hussman’s approach is that S&P 500 profits have increasingly become disconnected from economy-wide profits. Chart 15 shows that pre-tax profits have trended lower as a share of GDP in recent years, a move that has been mirrored in the rise in employee compensation. However, no such decline has occurred for S&P 500 profits. Chart 15The Recent Decline In U.S. Pre-Tax Profits Has Been Mirrored In The Rise In Employee Compensation The Recent Decline In U.S. Pre-Tax Profits Has Been Mirrored In The Rise In Employee Compensation The Recent Decline In U.S. Pre-Tax Profits Has Been Mirrored In The Rise In Employee Compensation Chart 16S&P 500: All Of The Increase In Margins Has Occurred In The IT Sector S&P 500: All Of The Increase In Margins Has Occurred In The IT Sector S&P 500: All Of The Increase In Margins Has Occurred In The IT Sector S&P 500 margins have stayed elevated partly because U.S. multinationals are less exposed to domestic wage pressures. Margins have also been propped up by the fact that the largest companies in the index increasingly operate like natural monopolies. It is perhaps no surprise that all of the increase in S&P 500 margins over the past few decades has been due to soaring profits in the IT sector (Chart 16). If this represents a true structural change, S&P margins could remain high by historic standards. Investment Implications Different valuation measures can generate different results. As such, we would not get too hung up on the precise expected return forecast that any one particular valuation indicator produces. Appendix B shows long-term return projections for various stock markets using a composite valuation measure based on price-to-trailing earnings, price-to-forward earnings, price-to-cash flow, price-to-sales, price-to-book, Tobin’s Q, stock market capitalization-to-GDP, and dividend yield. Chart 17 Three broad conclusions stand out: First, global equities are quite cheap compared with bonds. This suggests that stock market returns could be quite strong over the next couple of years, as PE multiples rise in order to narrow the gap between bond yields and earnings yields. We recommend that asset allocators overweight stocks relative to government bonds on a 12-to-18 month horizon. Second, global equities are modestly expensive in absolute terms. Thus, long-term stock market returns are likely to be subpar, even if they are reasonably high in the near term. Put differently, while equities will trump bonds over the long haul, both asset classes will deliver uninspiring absolute returns compared with their long-term history. Third, U.S. stocks are noticeably more expensive than their overseas peers. Differences in sector composition can explain some of the valuation gap, but not all of it. As Chart 17 illustrates, if one were to calculate the forward PE of say, European stocks, using U.S. sector weights, the former would still be significantly cheaper than the latter. We intend to upgrade EM and European stocks later this year once global growth begins to reaccelerate.   APPENDIX A The Earnings Yield As A Proxy For Expected Shareholder Returns  Chart 18 APPENDIX B CHART 1 Equities: Composite Valuation Indicators Versus Subsequent 10-Year Real Annualized Returns Chart 19 Peter Berezin,  Chief Global Strategist Global Investment Strategy  peterb@bcaresearch.com MacroQuant Model And Current Subjective Scores Chart 20