Fixed Income

Special Report

Dear Client, This Special Report is the full transcript and slides of a keynote presentation I recently gave to the Sovereign Investor Institute in London titled: 'The Biggest Risks To The Global Economy Are…' The short presentation pulls together several concepts and observations which identify the ‘weak links’ in the global economy. Therefore, the presentation should serve as a useful summary of the global economy’s current vulnerabilities. The report then explains how each of the risks translates into a European investment context. I hope you find it insightful. Best regards, Dhaval Joshi, Chief European Investment Strategist Feature Full Transcript And Slides Good morning Thank you for inviting me to give today’s keynote presentation under the title: ‘The Biggest Risks To The Global Economy Are…’ (Slide 1). I will not discuss all the risks out there, but the four risks that I will present are the ones that I think are the most significant. And the biggest of these four risks I will leave to the end. So let’s begin. Risk 1 is China’s Credit Cycle (Slide 2). You can see this very clearly in this slide (Slide 3) which shows the short-term accelerations and decelerations in credit within the world’s three largest economies – Europe, the United States, and China. In essence, it is showing how much new credit was created in the last six months compared with the preceding six months. Was it more credit creation or was it less, and how much more or less? Everything is in dollars to allow a fair comparison. Now look at the red line. The red line is China. Just ten years ago, China’s credit cycle was irrelevant. It simply didn’t matter. But after the GFC, China’s short-term credit expansions and contractions suddenly became as large as those in Europe and the U.S. More recently, China’s cycle is dwarfing the others, so now it is the European and the U.S. credit cycles that are irrelevant! This means that whenever China’s short-term credit cycle turns down, as it did in late 2015, early 2017, and 2018, the global economy feels a chill. The point is that this short-term cycle is a near-perfect oscillator. Down-oscillations will occur every eighteen months or so, and any of them has the potential to turn nasty. Though we are currently in an up-oscillation, the next down-oscillation is due later this year. And I predict that it will pose a big risk to the global economy. Risk 2 is Trade Imbalances (Slide 4). This slide (Slide 5) has a mischievous title ‘Where President Trump Is Right About Europe’. The red line shows where the president is absolutely right: Europe is running a massive – a record-high – trade surplus with the United States. It is an undeniable fact. But the president is wrong about the underlying cause. The underlying cause is not unfair trade practices or tariffs, the underlying cause is the other line, the blue line, which shows the divergent monetary policies of the ECB and the Fed. The trade imbalance and monetary policy divergence are moving together tick for tick, and the transmission mechanism is of course the exchange rate. The divergent monetary policies have depressed the euro, and a depressed euro obviously makes German cars cheaper for American consumers. That is the reason that the president is seeing so many BMWs driving down Fifth Avenue! My point is that these record-high imbalances are being used to justify economic nationalism – retaliatory tariffs, restricted trade, and potentially all-out trade wars. Alternatively, this chart suggests that the imbalances would correct with large-scale movements of exchange rates. But to me, either of these options poses a big risk to the global economy. Risk 3 Is Technological Disruption (Slide 6). To understand why, I want to introduce you to a concept known as Moravec’s Paradox (Slide 7). A professor of robotics, Hans Moravec, noticed something odd. He realized that things that we find very hard are actually very easy for AI. Things like complex mathematics, speaking multiple languages, or advance pattern recognition. Typically, as few people have these skills, they are well-paid skills. Whereas things that we find very easy are incredibly difficult for AI. Things like human movement and recognizing, and responding to, emotional signals. Typically, as everybody has these skills, they are low-paid skills.  Moravec’s Paradox means that the current wave of technological progress is much more disruptive than previous waves. The steam engine destroyed low-paid jobs, forcing workers up the income ladder. But the current wave of technology, led by AI, is destroying well-paid jobs forcing workers down the income ladder. You can see it in the data. While job creation in most major economies is on the face of it very strong, just look at what type of jobs are being created (Slide 8). Food delivery, bar work, care work and social work. Now you’ll agree that this is not highly paid work with career prospects!  In essence, the current wave of technology is revealing a huge misallocation of capital. You might have invested huge amounts of time and money in say, becoming a linguist. Only to find that AI can translate languages much better than you – and your employment opportunities are limited to lower-income work. Well that misallocation of capital is very disruptive.  In my opinion, it’s one of the main reasons why even though economies are growing and unemployment is very low, people don’t feel good. Making them susceptible to simplistic fixes such as ‘take back control’ and economic nationalism. My point is that the current wave of AI-led job disruption has much further to run, and the populist backlash will remain a big risk to the global economy. But now I want to turn to what I believe is the biggest risk of all. Risk 4 Is Higher Bond Yields (Slide 9). Most people believe that economic downturns cause financial market downturns. But the truth is the complete opposite: the causality almost always runs the other way! In the vast majority of cases, it is financial market imbalances and mispricing that cause economic downturns and crises. Take the last three economic downturns – in 2001, in 2008 and in 2011. They all had their roots in financial mispricing – the dot com bubble, the U.S. mortgage market, and euro area sovereign debt. Likewise for the Great Depression in the 30s, Japan’s recession in the early 90s. I could go on. You get the point… What is the financial vulnerability today that could cause an economic downturn? (Slide 10) The answer is that the very rich valuation of equities and other risk-assets is highly sensitive to bond yields. Which means that substantially higher bond yields pose a very big risk to the global economy. You see, at very low bond yields, the bond price can no longer go up much but it can go down massively (Slide 11). The latest advances in financial theory now conclusively show that this unattractive ‘negative’ asymmetry is what defines ‘risk’ for investors. The crucial point is that at low bond yields, bonds become as risky, or more risky, than equities (Slide 12). And this necessarily means that equities no longer need to deliver a superior return, a risk-premium, over the low bond yield (Slide 13). As bond yields decline this means equity valuations get an exponential boost because both components of the equity’s required return – the risk-free component and the risk-premium component – are collapsing simultaneously (Slide 14). But if bond yields rise substantially, the process would go into vicious reverse and equity valuations would fall off a cliff. Other risk-assets too, and bear in mind that if we include real estate – as we should – global risk-assets are worth $400 trillion, five times the size of the global economy!   Our research shows that the point of vulnerability is if the global 10-year bond yield approaches 2 percent, which is about 50 basis points above where it stands right now. And that, to me, is by far the biggest risk to the global economy. So to summarise, the biggest risks to the global economy are: China’s credit cycle; trade imbalances and technological disruption and their associated populist backlash; and the biggest risk is higher bond yields (Slide 15). In the near future I think alarm bells should start to ring if China’s credit cycle has tipped into a down-oscillation and/or the global 10-year bond yield is 50 bps higher. Don’t worry, the alarm bells are not ringing right now but they might be later this year. Finally, given the title you gave me, this presentation has necessarily focussed on the key risks. But I don’t want you to get too negative. I also have another presentation called ‘The Biggest Positives For The Global Economy Are…’ And for balance, I hope you invite me to present that next time! Thank you. How Do The Risks Translate Into A European Investment Context? Risk 1: China’s Credit Cycle, is highly relevant to European investors, for two reasons. First, the European economy is very open, meaning that exports make a substantial contribution to GDP growth. This is especially true in Europe’s engine economy, Germany, but it is also important for other major economies like Sweden. And it is evidenced in large trade surpluses as, for example, illustrated in Slide 5. Therefore, whenever China’s credit cycle enters a down-oscillation, as it did last year, Germany cannot escape the nasty chill coming through its all-important net export channel. Second, the European equity market is over-exposed to global growth sensitive sectors and companies – specifically, Industrials, Materials, and Financials. These sectors tend to have a very high operational gearing to global growth. Meaning that a small change in global growth has a disproportionate effect on these companies’ profits and share price performance. The upshot is that in a credit cycle up-oscillation, Europe’s global-growth sensitive stock markets and sectors benefit from a sharp burst of outperformance. The opposite applies in a credit cycle down-oscillation. It follows that if China’s credit cycle is due to tip into a down-oscillation later this year, it would be time to close our successful relative overweighting to European equities and to the global growth sensitive cyclical sectors. Risk 2: Trade Imbalances, is also highly relevant to European investors, for the obvious reason that European economies – especially Germany – are running huge trade surpluses. This puts these economies squarely in the cross-hairs of a retaliatory salvo involving tariffs, trade barriers, or worse, an all-out trade war. Clearly, Europe’s ‘exporting champions’ are the most vulnerable to this risk. The issue is important for the exchange rate too. We showed conclusively that Europe’s trade imbalance is the consequence of the depressed euro. It follows that another way to correct this imbalance is via a stronger euro. In this sense, the fundamentals imply euro upside from here. Risk 3: Technological Disruption, manifests through disruption in the jobs market, the lack of feel good, and the ensuing backlash leading to populism and nationalism. This is particularly relevant to Europe because its collection of nations, each with its own political processes, provides more scope for a political tail-event. A lull in the major political-event cycle is a good thing for Europe. In this regard, the upcoming EU parliamentary elections is not a big risk given the EU parliament’s inability, by itself, to drive policy. The risk increases approaching a meaningful political event, and this includes the date of Brexit. Therefore, this risk is likely to rise somewhat towards the end of the year. Risk 4: Higher Bond Yields, is clearly very relevant to Europe because many of the core euro area bond yields are at their lower bound. This means that the negative asymmetry of returns has its maximum impact on, for example, German bunds. It follows that German bunds are a sell in the near-term. Nevertheless, the upside to yields is ultimately limited given the aforementioned vulnerability of risk-asset valuations to higher bond yields. Therefore, the better long-term strategy is to short German bunds relative to U.S. T-bonds. Finally, a 50 basis points rise in 10-year yields from current levels would be a trigger to flip to underweight European equities.  Fractal Trading System* Crude oil is at a technical reversal level. The best way to play this is on a hedged basis versus metals: short WTI, long LMEX. Set the profit target at 5 percent with a symmetrical stop-loss. In other trades, we are pleased to report long AUD/CNY achieved its profit target at which it was closed. This leaves five open positions. For any investment, excessive trend following and groupthink can reach a natural point of instability, at which point the established trend is highly likely to break down with or without an external catalyst. An early warning sign is the investment’s fractal dimension approaching its natural lower bound. Encouragingly, this trigger has consistently identified countertrend moves of various magnitudes across all asset classes. The post-June 9, 2016 fractal trading model rules are: When the fractal dimension approaches the lower limit after an investment has been in an established trend it is a potential trigger for a liquidity-triggered trend reversal. Therefore, open a countertrend position. The profit target is a one-third reversal of the preceding 13-week move. Apply a symmetrical stop-loss. Close the position at the profit target or stop-loss. Otherwise close the position after 13 weeks. Use the position size multiple to control risk. The position size will be smaller for more risky positions. * For more details please see the European Investment Strategy Special Report “Fractals, Liquidity & A Trading Model,” dated December 11, 2014, available at eis.bcaresearch.com Dhaval Joshi, Chief European Investment Strategist dhaval@bcaresearch.com Fractal Trading Model Recommendations Asset Allocation Equity Regional and Country Allocation Equity Sector Allocation Bond and Interest Rate Allocation Currency and Other Allocation Closed Fractal Trades Trades Closed Trades Asset Performance Currency & Bond Equity Sector Country Equity Indicators Bond Yields Chart I-1Indicators To Watch - Bond Yields Indicators To Watch - Bond Yields Indicators To Watch - Bond Yields Indicators To Watch - Bond Yields   Interest Rate Indicators To Watch - Interest Rate Expectations Indicators To Watch - Interest Rate Expectations   Indicators To Watch - Interest Rate Expectations Indicators To Watch - Interest Rate Expectations  

Highlights Q1/2019 Performance Breakdown: Our recommended model bond portfolio underperformed the custom benchmark index by -17bps in the first quarter of the year. Winners & Losers: The underperformance came from the government side of the portfolio (-40bps), where our below-benchmark duration stance was mainly implemented through underweight positions in long-ends of government bond yield curves. On the other side was a solid outperformance from spread product allocations (+23bps) after our tactical upgrade to global corporates in January. Scenario Analysis For The Next Six Months: An improving global growth backdrop, and benign monetary policy backdrop, should help generate an outperformance of the model bond portfolio – mostly through credit, but also through moderate bear-steepening of government bond yield curves. Feature For fixed income markets, the start of 2019 has been categorized by three main trends: falling bond yields, narrowing credit spreads, and slower global growth. Central bankers have been forced to shift to a much more dovish stance on monetary policy, in response to heightened uncertainties over the global economy, helping trigger rallies in both government bonds and credit. In this report, we review the performance of the BCA Global Fixed Income Strategy (GFIS) model bond portfolio during the surprisingly eventful first 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. Q1/2019 Model Portfolio Performance Breakdown: Overweight Credit Pays Off, Below-Benchmark Duration Does Not Chart of the WeekDuration Losses Offset Credit Gains In Q1/2019 Table 1GFIS Model Bond Portfolio Q1/2019 Overall Return Attribution   The total return for the GFIS model portfolio (hedged into U.S. dollars) in the first quarter was 3.1%, underperforming the custom benchmark index by -17bps (Chart of the Week).1 The bulk of the underperformance came from the government bond side of the portfolio (-40bps) - a function of both our below-benchmark duration tilt and underweight stance on sovereign bonds (Table 1). Of course, the flipside of that government bond underweight is a spread product overweight. The tactical upgrade to global corporate debt (favoring the U.S.) that we introduced back on January 15 helped boost the credit piece of the model bond portfolio, which outperformed the custom benchmark by +23bps. The tactical upgrade to global corporate debt (favoring the U.S.) that we introduced back on January 15 helped boost the credit piece of the model bond portfolio, which outperformed the custom benchmark by +23bps. 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 main individual sectors of the portfolio that drove the excess returns were the following: Biggest outperformers Overweight U.S. investment grade industrials (+11bps) Overweight U.S. high-yield Ba-rated (+10bps) Overweight U.S. high-yield B-rated (+8bps) Overweight U.S. investment grade financials (+5bps) Overweight Japanese government bonds with maturity of 7-10 years (+4bps) Biggest underperformers Underweight Japanese government bonds with maturity beyond 10+ years (-17bps) Underweight U.S. government bonds with maturity beyond 10+ years (-12bps) Underweight France government bonds with maturity beyond 10+ years (-8bps) Underweight Emerging Markets U.S. dollar denominated corporates (-7bps) Underweight U.S. government bonds with maturity of 7-10 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 Q1/2019. 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 Q1/2019 (red for underweight, blue for overweight, gray for neutral). It was a great quarter for global fixed income, as all countries and spread products generated positive total returns. Generally, our allocations did reasonably well. There were more blue bars than red bars on the left side of Chart 4 (i.e. more overweights than underweights where returns were higher), and vice versa on the right side (more underweights than overweights where returns were lower). Some of the hit to performance from below-benchmark duration is already starting to be recouped in the first weeks of Q2 as markets become more comfortable with early signs of improving global growth. The negative overall Q1/2019 result is obviously not satisfactory, but we are still pleased with the positive returns generated from the spread product side after we did our January upgrade. More importantly, some of the hit to performance from below-benchmark duration is already starting to be recouped in the first weeks of Q2 as markets become more comfortable with early signs of improving global growth, pushing bond yields higher. Bottom Line: Our recommended model bond portfolio underperformed the custom benchmark index in the first quarter of the year. The underperformance came from the government side of the portfolio, where our below-benchmark duration stance was mainly implemented through underweight positions on the long-ends of government bond yield curves. On the other side was a solid outperformance from spread product allocations after our tactical upgrade to global corporates in January. Future Drivers Of Portfolio Returns Chart 6Overall Portfolio Duration: Below-Benchmark Looking ahead, the performance of the model bond portfolio will benefit from two main factors: our below-benchmark duration bias and our overweight stance on global corporate debt (favoring the U.S.) versus government bonds. In terms of the specific high-level weightings in the model portfolio, we are maintaining our tactical overweight tilt, equal to seven percentage points, on spread product versus government debt (Chart 5). This reflects a more constructive view on global growth, which appears to be bottoming out after the sharp slowdown seen in 2018, to the benefit of corporate bond performance. That faster growth backdrop will also benefit our below-benchmark duration stance through a rebound in government bond yields. This should happen only slowly, however, as global central bankers are likely to keep their newly-dovish policy bias in place for some time until there are more decisive signs of accelerating growth AND inflation. We are maintaining our significant below-benchmark duration tilt (one year short of the custom benchmark), but we recognize that the underperformance from duration seen in Q1 will only be clawed back slowly over the next 3-6 months (Chart 6). As for country allocation, we continue to favor regions where tighter monetary policy is least likely (overweight Japan, the U.K., and Australia, neutral core Europe and Canada). We are staying underweight the U.S., however, as the market’s expectations for the Fed is too dovish, with -25bps of rate cuts now discounted over the next twelve months. We expect to make some changes to those country allocations over the next few months, however - most notably a potential downgrade in core Europe, and upgrade in Peripheral Europe, if the euro area stabilizes on the back of firmer global growth. We expect to make some changes to those country allocations over the next few months, however - most notably a potential downgrade in core Europe, and upgrade in Peripheral Europe, if the euro area stabilizes on the back of firmer global growth. The overall yield from the model bond portfolio is modestly above that of the benchmark (+7bps). That is admittedly a fairly small amount of positive carry (Chart 7) given the overweight credit position. It is a consequence of our below-benchmark duration stance, which is focused on underweights in longer, higher-yielding ends of government bond yield curves (i.e. we have a bear-steepening bias in the U.S., core Europe and even the very long-end in Japan). Chart 7Portfolio Yield: Small Positive Carry Chart 8Portfolio Risk Budget Usage: Cautious   Even though we have decent-sized overall tilts on global duration and spread product allocation, our estimated tracking error (excess volatility of the portfolio versus its benchmark) remains low (Chart 8). This is a function of some of the offsetting country and sector tilts within the overall allocations (i.e. more Japan than Germany, more Spain than Italy, more U.S. corporates than EM corporates). We remain comfortable maintaining a tracking error target range of between 40-60bps, well below our self-imposed 100bps ceiling, as our internal weightings are helping keep overall portfolio volatility at a modest level. 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.2 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). 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, are all driven by what we continue to believe will be the most important driver of market returns in 2019 – the path of U.S. monetary policy. Our Base Case: the Fed stays on hold, the U.S. dollar remains flat, oil prices rise by +10%, the VIX index hovers around 15, and there is a mild bear-steepening of the U.S. Treasury curve. This is the case of a pickup in U.S. and global growth that is strong enough to support higher commodity prices, but not intense enough to rapidly boost U.S. core inflation, allowing the Fed to keep rates unchanged. A Very Hawkish Fed: the Fed does a surprise +25bps rate hike in June or September, the U.S. dollar rises by +3%, oil prices increase +10%, the VIX index climbs to 25 and there is a sharp bear-flattening of the U.S. Treasury curve. This would occur if the U.S. economy reaccelerates alongside improved global growth, U.S. core inflation and inflation expectations move higher, and market volatility increases from a surprisingly hawkish Fed. A Very Dovish Fed: the Fed cuts the funds rate by -25bps, the U.S. dollar falls by -3%, oil prices decline -15%, the VIX index increases to 35 and there is a sharp bull steepening of the U.S. Treasury curve. This is a scenario where U.S./global growth momentum fades once again, leaving the Fed little choice but to ease monetary policy as market volatility surges alongside elevated recession risks. 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 all unchanged from our late portfolio review in early January (Chart 9). The U.S. Treasury yield changes, however, are more moderate than what we used three months ago (Chart 10). That reflects the Fed’s dovish turn since then, which limits the upside for yields from multiple Fed hikes in 2019. Chart 9Risk Factors Assumptions For The Scenario Analysis Chart 10U.S. Treasury Yield Assumptions For The Scenario Analysis     The model bond portfolio is expected to outperform the custom benchmark index by +43bps in our Base Case scenario. This comes from the relative outperformance of credit versus government bonds in an environment of slowly rising bond yields (below-benchmark duration), and tighter credit spreads (overweighting U.S. corporates). In the Very Hawkish Fed scenario, our model portfolio is projected to outperform the benchmark by +29bps. This comes mostly from below-benchmark duration, with more muted credit performance as spreads widen and volatility increases due to the unexpected Fed rate hike. In the Very Dovish Fed scenario, the model bond portfolio is expected to lag the benchmark by -49bps. Performance would get hit from both credit and duration, as government bond yields fall and credit spreads widen sharply against a backdrop of even slower global growth. The overall expected excess return of our model bond portfolio over the benchmark is positive, given that the scenario analysis produces positive excess returns in the Base Case and Very Hawkish Fed scenarios. While we do not place probabilities on our scenarios in this analysis, if we did, the Very Dovish Fed scenario would be far less likely than the Very Hawkish Fed scenario (by definition, the Base Case is our most likely outcome). Global growth is much more likely to rebound than decelerate further over the rest of 2019. Thus, the overall expected excess return of our model bond portfolio over the benchmark is positive, given that the scenario analysis produces positive excess returns in the Base Case and Very Hawkish Fed scenarios. Bottom Line: An improving global growth backdrop, and benign monetary policy backdrop, should help generate an outperformance of the model bond portfolio – mostly through credit, but also through moderate bear-steepening of government bond yield curves.   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 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 Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Special Report

Highlights Portfolio rebalancing is the process of realigning portfolio weights back to their strategic allocations. Frequent rebalancing is essentially a counter-cyclical, or value, strategy. In effect, investors buy low and sell high. Infrequent rebalancing is a momentum-factor investing strategy. Maximizing risk-adjusted return is the reason investors should rebalance, not maximizing return per se. We find that calendar, deviation, or a combination of both methods of rebalancing, can all improve risk-adjusted return compared to a non-rebalanced portfolio. Feature What Do We Mean By Rebalancing? The first step of portfolio construction is strategic asset allocation. Simply put, it is determining a set of asset weights that best suits the investor’s return target, risk appetite, capabilities, and other considerations. Once a portfolio is constructed, divergent returns among asset classes cause the weights of the portfolio to shift. Portfolio rebalancing is therefore, the process of realigning portfolio weights back to their strategic allocations. Chart 1Rebalancing Can Imply Style Rebalancing is a means of reducing portfolio risk rather than increasing returns, and is necessary to maintain the desired risk exposure over time. Frequent rebalancing can be viewed as value investing: a style in which investors “buy low and sell high” (Chart 1). Given the mean-reverting nature of asset performance, buying the undervalued asset and selling the overvalued should imply that future returns would be higher than past returns. Through this process, investors are hoping to obtain a “rebalancing premium”. It is crucial to recognize that rebalancing works best at inflection points. Hence, that premium is gained when the rebalancing frequency is similar to the frequency of the mean-reversion feature of assets. Rebalancing also allows a portfolio to be consistent with the investor’s risk appetite in order to avoid a particular asset class dominating. However, this is easier said than done. An investor’s intuition usually acts in the opposite direction, pushing him or her to follow momentum rather than cut back the weight of a “winning” asset. The question that this Special Report aims to answer is not whether investors should rebalance or not, but rather what kind of rebalancing they should do. We discuss three different conventional rebalancing methods that investors can use, illustrating the risk-return characteristics of a simple two-asset-class (60% equity/40% bonds) portfolio since 1973. In doing so, we rebalance the portfolio back to its 60/40 strategic weights. Rebalancing is a means of reducing portfolio risk rather than increasing returns, and is necessary to maintain the desired risk exposure over time. It is important to note that rebalancing is no free lunch. Costs vary depending on the method used. Costs include trading and transaction costs, operational costs (trade lags, labor, and time to monitor the portfolio), and tax costs (capital gains on appreciated assets). In this paper, we do not consider the operational and tax costs (as they differ from investor to investor). Rather, we examine portfolio returns given: (1) zero trading costs, and (2) a variable cost of 10 bps dependent on trade size. Additionally, frequent rebalancing can introduce “negative convexity”, a return profile in which large divergences in asset performance exceed the rebalancing premiums investors obtain.1 Throughout our explanations, we show two tables for each method: Table A illustrates the returns given zero costs, while Table B illustrates the returns given the variable costs. It is key to note however that there is no one-size-fits-all rebalancing method. The important thing to realize is that rebalancing, done correctly, must find an optimal balance between cost minimization and managing portfolio risk. As a benchmark, we examine how an unbalanced portfolio, which we will refer to as a “drift portfolio”, comprised of 60% equities and 40% bonds in 1973, would have evolved over the past 46 years. Given that equities outperform bonds over the long run due to their riskier nature, the drift portfolio ends with an 86% allocation to equities, and a maximum allocation of 87% over the period (Chart 2). Chart 3Broken Equity/Bond Correlation   Before describing how each methodology performed, we need to highlight a key point in understanding the results that follow: the equity/bond correlation underwent a step-change around 1998. Between 1975 and 1998, the correlation between equities and bonds averaged about 0.4. However, declining inflation expectations led to a reversal of this relationship. Since 1998, the equity/bond correlation averaged -0.3 (Chart 3, top panel). It is key to note however that there is no one-size-fits-all rebalancing method. The important thing to realize is that rebalancing, done correctly, must find an optimal balance between cost minimization and managing portfolio risk. How does this affect the results? A positive correlation between equities and bonds means that asset-class returns moved together, reducing the advantages of rebalancing. Therefore, between the start of our sample period, 1973, and 1998, rebalanced portfolios only slightly outperformed a non-rebalanced portfolio. It is crucial to recognize that rebalancing portfolios should continue to be most advantageous during times when asset returns exhibit negative correlation. Portfolio Rebalancing can take place in different ways2 (Table 1). Table 1Conventional Methods Of Rebalancing Rebalancing Methodologies Time-Only Rebalancing The most common rebalancing methodology used by investors is on a simple calendar basis. A survey conducted by the Financial Planning Association showed that 48%, 36%, and 14% of financial planners rebalance quarterly, annually, and monthly respectively; 1% of respondents said they rebalanced based on a client’s request.3 This form of rebalancing involves bringing the asset-class weights back to the agreed-upon benchmark at the end of a specified period. Periods can range from daily (which is rare) to multiple years. Several academic papers and practitioners call for investors to rebalance at least annually. For the purpose of this report, we look at monthly, quarterly, semi-annual, annual, and bi-annual rebalancing.4 Rebalancing not only increases return at the margin, but also reduces portfolio risk and hence improves risk-adjusted returns. The risk-adjusted return increases as the rebalancing frequency decreases. Bi-annual rebalancing had a risk-adjusted return of 1.016 versus 0.895 for a non-rebalanced portfolio and 0.985 for a monthly-rebalanced portfolio over our entire sample period (Tables 2A and 2B). All calendar-rebalancing dates outperformed a non-rebalanced portfolio on a risk-adjusted basis due to lower volatility. The same results persist even when costs are factored in. Rebalancing too frequently not only increased costs, but also limited upside potential. That is noticeable from the number of rebalancing events for a monthly-rebalanced portfolio versus an annually or a bi-annually rebalanced portfolio. Unsurprisingly, we found that all rebalanced portfolios on average underperformed the drift portfolio during equity bull markets, and outperformed in the period leading up to recessions and equity corrections (Chart 4). Given that stocks peak on average six to 12 months before a recession, the higher weighting in bonds at the start of a correction explains the outperformance of a frequently rebalanced portfolio versus a drift portfolio during recessions and equity market corrections. To put this into context, the drift portfolio’s equity weight at the time of the S&P 500’s peak in the dot-com bubble was 84%, versus an average of 61% across the rebalanced portfolios. Similarly, at the peak before the latest market selloff starting on October 3, 2018, the drift portfolio had an 87% equity allocation versus a 61% average allocation for the frequently rebalanced portfolios. Chart 5 shows that rebalancing reduces downside risk relative to a drift portfolio during downturns and recessions. Chart 4Calendar Rebalancing: Relative Performance Chart 5Calendar Rebalancing: Lower Drawdown Threshold-Only Rebalancing Threshold rebalancing allows asset-class weights to be readjusted back to their target weights once they deviate away by a certain percentage. This can be set in terms of either a percentage-point or a percent deviation. Given that, in this paper, we illustrate our findings using just a two-asset class portfolio with relatively large weights in each asset, percentage-point deviations are more appropriate. However, percent deviations should be used when a certain asset class has only a small weight within a portfolio, for example, a 20% deviation away from the 5% target weight of an asset class. A key benefit of threshold-only rebalancing over calendar rebalancing in a multi-asset portfolio is lower transaction costs. Unlike calendar-only rebalancing where all asset classes are brought back to target weights, only the assets that have moved away from benchmark by the set deviation have to be bought and sold. For example, in a five-asset class portfolio, it could be the case that only the best and worst performers have hit their thresholds and have to be adjusted, whereas the other asset classes do not. Tables 3A and 3B show the risk-return characteristics of rebalanced portfolios based on 1, 5, 10, and 20 percentage-point deviations. Similarly to calendar rebalancing, the wider the threshold, the better the risk-adjusted return. The rebalanced portfolio with a 20-percentage point threshold outperforms all other deviations on both a return and risk-adjusted basis. All rebalanced portfolios led to better risk-adjusted returns than the drift portfolio, even after costs are factored in. Also similar to calendar rebalancing, threshold deviation rebalancing also outperforms during recessions and market corrections (Charts 6 & 7). Chart 6Threshold Rebalancing: Relative Performance Chart 7Threshold Rebalancing: Lower Drawdown The table also illustrates that picking the right threshold is crucial. A threshold set too wide will miss all turning-points and hence turn into a drift portfolio. Whereas, thresholds set too narrow will produce only a small improvement in return at the expense of more rebalancing events, and therefore higher costs. Time-And-Threshold Rebalancing A time-and-threshold rebalancing combines the merits of both strategies. The portfolio is rebalanced only when an asset class has deviated from its target allocation by a set threshold on the date of rebalancing. Assuming, for example, monthly rebalancing with a 10% deviation, a portfolio would be rebalanced on the next monthly date only if it had deviated by more than 10 percentage points. Otherwise, the portfolio would not be rebalanced. This implies that two decisions have to be made: a threshold band and a rebalancing frequency. We present the results of this method in a slightly different way. In this case, we show each metric (annualized return (Tables 4A & 5A), annualized volatility (Tables 4B & 5B) and risk-adjusted return (Tables 4C & 5C)) separately under assumptions of both zero costs and variable costs.   The highest risk-adjusted return of 1.023 was achieved with quarterly rebalancing and a 20 percentage point deviation. This resulted in only three rebalancing events throughout the 46-year period. However, this was not as good as simply relying on a 20 percentage point threshold deviation. Investors wanting to keep a tighter control over their portfolio could use a tighter band with a more frequent rebalancing. As noted earlier, rebalancing is a way to maximize risk-adjusted return rather than maximize return. To simply maximize return, annual rebalancing with a 10-percentage point threshold, which had an annualized return of 9.80%, would be the best combination. However, that came at the expense of high volatility and a higher average equity allocation. Having fewer rebalancing events does not necessarily mean lower costs. In fact, we noted that the fewer the rebalancing events, the higher the annualized cost per trade5 (Tables 6 and 7). Given that our variable cost was dependent on trade size, a rebalancing method that relied on wider bands would incur higher costs per trade relative to narrower bands. Table 6Time-And-Threshold Rebalancing: Rebalancing Events Table 7Time-And-Threshold Rebalancing: Cost Per Trade (Bps) Beyond The Conventional Methods New rebalancing strategies have evolved that rely on different metrics. These include timing rebalancing events using tracking error or risk deviation, absolute momentum, or analyzing the stage of the economic cycle. A recent paper published by Northern Trust discussed the merits of risk-based tracking-error rebalancing as a superior method to traditional strategies. The paper concluded that risk-based tracking had outperformed most other rebalancing strategies while requiring fewer rebalancing events. Within the core strategies mentioned, several adjustments could be made to obtain better results from rebalancing events. Some argue that rebalancing back to a tolerance band, rather than to the precise allocation target, could improve risk-adjusted returns. That band is usually set at half of the deviation threshold band, but can vary at the investor’s discretion. Given costs that vary based on trade size, it might be cheaper for an investor to use tolerance bands. However, relying on such a method can easily rack up costs if the investor is going against momentum prior to its end, since relying on tolerance bands would require more frequent rebalancing. Bottom Line Rebalancing is a means of maximizing risk-adjusted return, rather than increasing absolute return. Rebalancing is no free lunch. Investors must take various associated costs into account before considering how and when to rebalance. The added benefit of rebalancing might seem small in annualized returns. However, on average, rebalancing led to an annualized decrease in volatility in excess of 1% over the 46-year period. It might be best for investors to use a time-and-threshold rebalancing to find a balance between cost minimization and maximizing risk-adjusted returns. Amr Hanafy, Research Associate amrh@bcaresearch.com   1 Nick Granger, Douglas Greenig, Campbell Harvey, Sandy Rattray, David Zou, "The Unexpected Costs of Rebalancing And How To Address Them," AHL Partners LLP, July 2014. 2 Colleen Janconetti, Francis Kinniry Jr., Yan Zilbering, "Best Practices For Portfolio Rebalancing," Vanguard, July 2010. 3 Financial Planning Association, Longboard, and Journal Of Financial Planning, “2017 Trends In Investing,” www.onefpa.org. 4 We assumed that monthly rebalancing occurs on the first trading day of every month, quarterly rebalancing occurs on the first trading day of January, April, July, afn_4nd October, semiannual rebalancing on the first trading day of January and July, and annual rebalancing on the first trading day of the year. 5 Calculated as the difference in annualized return between 10 bps cost assumptions and 0 cost assumption multiplied by the number of years within the sample period divided by the number of trades.  

Highlights Monetary Policy: The Fed is in no rush to tighten, and will remain on hold until inflation expectations or financial conditions give them a reason to resume hikes. Investors should take advantage by overweighting spread product while keeping portfolio duration low. Municipal Bonds: The best value in municipal bonds is found at the long-end of the Aaa-rated municipal bond curve. Lower-rated and shorter maturity munis are much less appealing. Investors should focus their municipal bond exposure on Aaa-rated debt with 20-year and 30-year maturities. Fed Balance Sheet: The Fed has now announced almost all the details of its balance sheet normalization plan. The Fed’s asset holdings will stop falling at the end of September, and we project that it will start buying securities again in 2020. Feature The minutes from the March FOMC meeting, released last week, were about as bullish for risk assets as anyone could have hoped. Not only did we learn that the Fed’s consensus forecast calls for economic growth to trough in Q1: Underlying economic fundamentals continued to support sustained expansion, and most participants indicated that they did not expect the recent weakness in spending to persist beyond the first quarter.1 But we also learned that, despite its economic optimism, the FOMC sees no reason to telegraph another rate hike any time soon: Chart 1Stay Overweight Corporate Bonds [A] majority of participants expected that the evolution of the economic outlook and risks to the outlook would likely warrant leaving the target range unchanged for the remainder of the year. The overall message couldn’t be clearer. The Fed is inclined to let the economy run for a while before it steps in to spoil the party. This supportive policy backdrop, coupled with our positive view of global growth,2 argues for investors to be overweight risk assets. Fortunately, even those who have so far been reluctant to add credit risk probably still have time to get in on the action. High-yield excess returns have only just made up the ground they lost near the end of last year, and investment grade corporates have another 46 bps to go (Chart 1). Further, only spreads from the highest rated credit tiers have tightened back to the target levels we set in February.3 Baa and junk-rated spreads still have ample room to tighten (Charts 2A & 2B). Specifically, The average Aaa-rated spread is currently 59 bps, 19 bps below our target. The average Aa-rated spread is currently 57 bps, exactly equal to our target. The average A-rated spread is currently 85 bps, 2 bps below our target. The average Baa-rated spread is currently 140 bps, 9 bps above our target. The average Ba-rated spread is currently 205 bps, 27 bps above our target. The average B-rated spread is currently 348 bps, 72 bps above our target. The average Caa-rated spread is currently 714 bps, 145 bps above our target. Chart 2AInvestment Grade Spread Targets Chart 2BHigh-Yield Spread Targets As a result, we recommend that investors avoid Aaa-, Aa- and A-rated credits, but overweight the remaining corporate credit tiers. Who’s Watching The Punch Bowl? Even though a hike is not imminent, at some point the Fed will lift rates again. For this reason, and because the market is currently priced for 20 bps of rate cuts over the next 12 months, we recommend that investors maintain below-benchmark portfolio duration. Investors should avoid Aaa-, Aa- and A-rated credits, but overweight the remaining corporate credit tiers. But how will the Fed decide when to take away the punch bowl? In a recent report we made the case that the two most important factors to monitor will be (i) inflation expectations and (ii) financial conditions.4 Last week’s FOMC minutes only strengthened our conviction in that view. The Fed On Inflation Expectations The March FOMC minutes showed that participants are concerned that inflation expectations have become un-anchored to the downside. In the Fed’s thinking, it must ensure that policy is accommodative enough to re-anchor inflation expectations. Otherwise, a Japanese-style scenario of permanent deflation could unfold. From the minutes:     Several participants observed that limited inflationary pressures during a period of historically low unemployment could be a sign that low inflation expectations were exerting downward pressure on inflation relative to the Committee’s 2 percent inflation target; Consistent with these observations, several participants noted that various indicators of inflation expectations had remained at the lower end of their historical range… In light of these considerations, some participants noted that the appropriate response of the federal funds rate to signs of labor market tightening could be modest provided that signs of inflation pressures continued to be limited. These concerns about low inflation expectations are not unfounded. Long-maturity TIPS breakeven inflation rates are well below the 2.3% - 2.5% range that has historically been consistent with “well anchored” expectations (Chart 3). The University of Michigan Survey of household inflation expectations is also well below pre-crisis levels (Chart 3, bottom panel). We expect monthly core CPI will print above 1.8% more often than not going forward. Our sense is that expectations are depressed because many years of low inflation have convinced markets that the Fed cannot sustainably hit its 2% target. In fact, our Adaptive Expectations Model – a model driven purely by measures of actual inflation – does a good job explaining movements in the 10-year TIPS breakeven inflation rate (Chart 4).5 At present, our model shows that the 10-year breakeven is close to fair value. Although we expect the fair value reading from our model to creep slowly higher over time. Chart 3First Battleground: Inflation Expectations Chart 4Adaptive Expectations Model The most important independent variable in our model is trailing 10-year core CPI inflation, which is currently running at an annualized 1.8% clip. This means that as long as monthly core CPI prints above 1.8% (annualized), it will send our model’s fair value reading higher over time. While core CPI has printed below that threshold in each of the past two months, we expect it will more often than not exceed it going forward. Notice that while year-over-year core CPI has rolled over, trimmed mean CPI has increased and median CPI just made a new cycle high (Chart 5). Meanwhile, small businesses continue to report an elevated rate of price increases and ISM prices paid surveys recently ticked up, after having fallen sharply earlier this year (Chart 6). Chart 5Encouraging Inflation Readings... Chart 6...Alongside Continued Price Pressures The Fed On Financial Conditions The Fed didn’t have much to say about financial conditions at the March 2019 meeting. In fact, looking through the minutes we could only locate the following relevant passage: A few participants observed that the appropriate path for policy, insofar as it implied lower interest rates for longer periods of time, could lead to greater financial stability risks. The lack of references to financial conditions shouldn’t be too surprising. Financial conditions aren’t nearly as accommodative as they were last autumn, and hence are currently much less of a policy concern (Chart 7): Chart 7Second Battleground: Financial Conditions The financial conditions component of our Fed Monitor is at 0.5. It was more than one standard deviation easier than average only a few months ago (Chart 7, top panel). The average junk index spread is still 46 bps above its 2018 low (Chart 7, panel 2). The GZ Excess Corporate Bond Risk Premium, an estimate of the excess spread in corporate bonds after accounting for expected default risk, still hasn’t recovered after widening sharply near the end of last year (Chart 7, panel 3).6 At 16.8, the S&P 500 Forward P/E ratio is almost back to its October level of 17 (Chart 7, bottom panel). Now consider that last year, when financial conditions were much more accommodative, the Fed was much more concerned. Fed Governor Lael Brainard and Chairman Jerome Powell both warned that signs of economic overheating could show up in financial markets before they show up in price inflation. Also, the minutes from the September 2018 FOMC meeting reveal that participants were willing to use the risk of “financial imbalances” as justification for tighter policy. A few participants expected that policy would need to become modestly restrictive for a time and a number judged that it would be necessary to temporarily raise the federal funds rate above their assessments of its longer-run level in order to reduce the risk of sustained overshooting of the Committee’s 2 percent inflation objective or the risk posed by significant financial imbalances.7 Bottom Line: The Fed is in no rush to tighten, and will remain on hold until inflation expectations or financial conditions give them a reason to resume hikes. Investors should take advantage by overweighting spread product while keeping portfolio duration low. Extend Maturity In Municipal Bonds Chart 8Municipal / Treasury Yield Ratios We continue to recommend that investors hold an overweight allocation to tax-exempt municipal bonds. Not only does the sector tend to outperform during the mid-to-late innings of the cycle,8 but value also remains attractive, with one key caveat: The best value in the municipal bond space is found at the long-end of the Aaa curve. The Value In Aaa Munis Chart 8 shows yield ratios for different maturities of Aaa-rated municipal debt relative to Treasuries. Notice that the 2-year and 5-year yield ratios, at 65% and 70% respectively, are close to one standard deviation below average pre-crisis levels. In fact, the all-time low for the 2-year Muni / Treasury yield ratio is 61%, only 4% below the current level. The all-time low for the 5-year yield ratio is 66%, also only 4% below the current level. The 10-year yield ratio looks almost as expensive as the 2-year and 5-year. At 76%, it is also close to one standard deviation below its average pre-crisis level. It is also only 6% above its all-time low. The real value in Aaa municipal bonds is found at the very long-end of the curve, in the 20-year and 30-year maturities where yield ratios, at 92% and 94% respectively, remain well above average pre-crisis levels (Chart 8, bottom two panels). While yield ratios out to the 10-year maturity point likely don’t have much room to compress, they could still look enticing depending on an investor’s tax situation. For example, a 76% 10-year Muni / Treasury yield ratio means that an investor facing an effective tax rate above 24% would still earn a positive after-tax yield pick-up in the municipal bond relative to the 10-year Treasury. The Value In Lower-Rated Munis Table 1Municipal Revenue Bonds / U.S. Credit Index Yield Ratios When we move outside the Aaa-rated municipal bond space we find that relative value starts to evaporate. Table 1 shows yield ratios between different municipal revenue bonds and the U.S. Credit index. We did our best to match the duration and credit rating of the different muni sectors as closely as possible. The table shows that the highest available Muni / Credit yield ratio is for 20-year A-rated munis, and even that yield ratio is only 73%. This means that an investor would need an effective tax rate above 27% to earn a positive after-tax yield pick-up relative to the U.S. Credit index. In other words, investors can add a fair amount of value by swapping Aaa-rated munis into their portfolios in place of Treasuries, especially at the long-end of the curve. There is much less incremental value to be gained from replacing corporate credit with lower-rated municipal debt. The Yield Ratio Curve Chart 9A Supportive Environment For Munis Our research shows that the yield ratio advantage at the long-end of the Aaa-rated muni curve tends to be greatest when the fundamental credit back-drop is supportive and municipal ratings upgrades are far outpacing downgrades (Chart 9). Conversely, when downgrades increase, yield ratios usually widen at the short-end of the curve relative to the long-end. At present, the muni ratings back-drop looks fairly supportive. While state & local government interest coverage dipped in Q4 (Chart 9, panel 2), it remains positive and should rebound as tax receipts move back to levels that are more consistent with the trend in nominal income growth (Chart 9, bottom panel). Periods of negative interest coverage tend to precede downgrade spikes. Under normal circumstances, a positive ratings outlook would suggest that yield ratios should fall more at the short-end of the curve than at the long-end, but there is very little chance that short-maturity yield ratios can compress further from current levels. Instead, it makes sense for investors to camp out at the long-end of the Aaa muni curve. Not only is the yield pick-up greater, but long-maturity yield ratios should better weather the storm when the cycle eventually turns. Bottom Line: The best value in municipal bonds is found at the long-end of the Aaa-rated municipal bond curve. Lower-rated and shorter maturity munis are much less appealing. Investors should focus their municipal bond exposure on Aaa-rated debt with 20-year and 30-year maturities. Fed Balance Sheet Normalization Almost Complete The Fed also presented a much more detailed plan for balance sheet normalization at the March FOMC meeting. To summarize the details: The Fed will continue to allow assets to passively run off its balance sheet until the end of September. Beginning in May, the Fed will reduce the monthly cap on Treasury redemptions from $30 billion to $15 billion. This means that if $16 billion of the Fed’s Treasury holdings mature in May, $15 billion will be allowed to run off and $1 billion will be reinvested. The current monthly cap of $20 billion for MBS remains unchanged. After September, the Fed will keep its overall assets constant but will continue to allow its MBS holdings to run down. It will reinvest the proceeds from MBS run-off into Treasuries. After September, even though the Fed will keep the asset side of its balance sheet constant, the supply of bank reserves will continue to shrink because the Fed’s other non-reserve liabilities – mostly currency in circulation – will continue to grow. Eventually, reserves will shrink to a level that the Fed deems optimal for the future implementation of monetary policy. It will then start to increase its asset holdings by purchasing Treasury securities. To implement this policy the Fed will likely announce a “minimum operating level” of desired reserve supply and then buy enough Treasuries to ensure that reserves stay above that level. The Fed has not announced which maturities it will target when it re-starts Treasury purchases. In our view, there are only two remaining questions when it comes to the Fed’s balance sheet policy. What Treasury maturities will it purchase going forward? And, when will it start buying Treasuries again? The Treasury’s cash holdings will continue to decline until the fall, putting upward pressure on the supply of bank reserves. On the first question, we will have to wait for an official announcement. Though in our view the Fed will choose a policy that reduces the risk that it will be perceived to be easing or tightening monetary policy through its purchases. This could be achieved by either concentrating its purchases in T-bills, or by targeting maturities in proportion to the Treasury department’s issuance schedule. The second question comes down to estimating the minimum reserve supply that will ensure banks are fully satiated, so that they don’t start competing for scarce reserve balances, driving up overnight rates in the process. While that equilibrium reserve number is unknown, the New York Fed’s most recent Survey of Primary Dealers shows that the 25th and 75th percentile of dealer estimates range from $1.1 trillion to $1.3 trillion. With those figures in mind, we can turn to the simplified Fed balance sheet shown in Table 2. The current balance sheet is shown along with what the balance sheet will look like when run off stops at the end of September. Table 2Simplified Fed Balance Sheet Projections To forecast the Fed’s balance sheet we assume that MBS runs off at a pace of $15 billion per month and that currency-in-circulation grows at an annual rate of 5%. We also estimate a range of possible values for the Treasury department’s General Account. This is the account where the Treasury keeps its cash holdings, which currently total $246 billion. Because the Treasury is currently engaged in extraordinary measures to prevent the U.S. from breaching the debt ceiling, this cash balance will almost certainly decline between now and when the debt ceiling is raised in the fall. After the debt ceiling is raised, the Treasury will probably start to re-build its cash balance. All else equal, a decline in the Treasury’s cash holdings puts upward pressure on the supply of bank reserves, while an increase in the Treasury’s cash holdings causes the supply of bank reserves to fall. According to Table 2, the supply of bank reserves will be between $1.42 trillion and $1.66 trillion by the end of September, still above most estimates of its equilibrium level. The table also shows that reserves will then shrink to between $1.35 trillion and $1.60 trillion by June 2020 and to between $1.31 trillion and $1.55 trillion by the end of 2020. Based on those figures and the dealer estimates, the Fed can probably keep its asset holdings constant through the end of 2020 without losing control of the policy rate or causing a disruption in money markets. However, we expect the Fed will err on the side of caution and start purchasing Treasuries again much earlier, possibly in the first half of 2020. The reason for the Fed to act quickly is that it faces asymmetric risks. The Fed risks losing control of the policy rate if it allows reserves to fall too far, but there is no real downside to keeping the balance sheet “too large”. In any event, the Fed has already demonstrated that it has the tools to conduct monetary policy with a large balance sheet. Bottom Line: The Fed has now announced almost all the details of its balance sheet normalization policy. The Fed’s asset holdings will stop falling at the end of September, and we project that it will start buying securities again in 2020. Ryan Swift, U.S. Bond Strategist rswift@bcaresearch.com Footnotes 1 https://www.federalreserve.gov/monetarypolicy/files/fomcminutes20190320.pdf 2 Please see U.S. Bond Strategy Weekly Report, “Bond Kitchen”, dated April 9, 2019, available at usbs.bcaresearch.com 3 We moved to overweight corporate bonds (both investment grade and high-yield) in in the U.S. Bond Strategy Weekly Report, “Buy Corporate Credit”, dated January 15, 2019, available at usbs.bcaresearch.com. The rationale for our spread targets is found in U.S. Bond Strategy Weekly Report, “The Value In Corporate Bonds”, dated February 19 , 2019, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, “The New Battleground For Monetary Policy”, dated March 26, 2019, available at usbs.bcaresearch.com 5 For further details on our Adaptive Expectations Model please see U.S. Bond Strategy Weekly Report, “Adaptive Expectations In The TIPS Market”, dated November 20, 2018, available at usbs.bcaresearch.com 6 The Gilchrist and Zakrajsek (GZ) Excess Bond Premium is a measure of the excess spread available in a sample of nonfinancial corporate bonds after removing a bottom-up estimate of expected default losses for each security. Default losses are estimated based on the Merton Default model using each firm’s market value of equity and face value of debt. https://www.federalreserve.gov/econresdata/notes/feds-notes/2016/files/…; 7  https://www.federalreserve.gov/monetarypolicy/files/fomcminutes20180926.pdf 8 Please see U.S. Bond Strategy Special Report, “2019 Key Views: Implications For U.S. Fixed Income”, dated December 11, 2018, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Highlights The first quarter is in the books, … : Risk may have been out in the fourth quarter, but it is squarely back in fashion so far this year, with equities and high yield posting gaudy first-quarter returns. … and events have compelled us to modify our high-conviction Fed call, … : There may yet be another four or more rate hikes, but they’re not going to occur this year. … but we’re still confident in our asset-allocation recommendations, … : The Fed may no longer be a menacing presence, but that doesn’t mean Treasuries and longer-maturity bonds are going to have it easy from here. … which should benefit from a more accommodative monetary policy outlook: Conditions remain favorable for equities and spread product, and unfavorable for Treasuries, even if the underlying drivers have shifted. Feature Table 1Whipsaw Newton’s Third Law holds that for every action there is an equal and opposite reaction. Markets have been busy supporting the theorem, as the fourth quarter’s sharp selloff has been nearly erased by the potent first-quarter rally (Table 1). Risk assets have been on a rollercoaster ride, though our economic outlook has been more or less unchanged. We chalked up the fourth quarter’s selloff to fears that the Fed was threatening the expansion. Conversely, the first quarter’s snapback likely owed quite a bit to the Fed’s pivot. By shifting its emphasis from trying to prevent inflation from getting away on the upside to trying to keep inflation expectations from falling too far, the Fed has gone from removing the punch bowl to promising to keep it full. In financial markets, risk assets should be the biggest relative beneficiaries. The Fed’s turn thwarted our more-hikes-than-expected call, at least in the near term. That surprise has been compounded by the administration’s seeming intent to pack the board of governors with nominees chosen solely on the basis of their uber-dovishness, and has inspired us to reflect on our calls. We like to share our reflections, as well as the internal BCA discussions and the client questions that shed light on our views. This week’s report examines some of the most important issues on our minds, and the minds of our colleagues and clients. Q: What does the Fed do from here? The quarterly summary of economic projections compiles FOMC meeting participants’ expectations for the likely path of key economic indicators (real GDP growth, unemployment and inflation) and monetary policy. The latest release revealed that Fed governors and regional presidents sharply dialed back their rate hike expectations between the December meeting and the March meeting (Chart 1). The median participant lopped 50 basis points (“bps”) off of his/her year-end 2019 and terminal fed funds rate projections, calling for no hikes in 2019 and just one more for the current cycle, in 2020. The rationale is a bit of a mystery, as the median participant’s estimates of GDP and inflation only came down modestly, and his/her unemployment rate estimates only rose modestly. It made sense for the Fed to turn away from the gradual pace of hikes it pursued in 2017 and 2018 in response to the sharp tightening in financial conditions brought on by the fourth-quarter selloff. The ensuing rallies in equities and high-yield bonds have undone much of that tightening, however. From a data perspective, it seems the Fed is mostly holding off to see how the outlook for the rest of the world evolves. The minutes of the March meeting, released last week, suggested that there may be more nuance to the Fed’s embrace of patience than markets initially perceived. The money markets had been calling for a 25-bps cut in the fed funds rate, to 2.25%, by the end of 2020; following the March meeting, they swiftly moved to price in a high likelihood of a second cut, to 2% (Chart 2). That outlook does not exactly accord with the committee’s more measured take: “Several participants observed that the [‘patient’] characterization … would need to be reviewed regularly[.] … A couple of participants noted that the ‘patient’ characterization should not be seen as limiting the Committee’s options[.] … Several participants noted that their views of the appropriate target range for the federal funds rate could shift in either direction[.] … Some participants indicated that if the economy evolved as they currently expected, … they would likely judge it appropriate to raise the target range … modestly later this year[.]” Chart 2... To Keeping It Full We continue to believe that the Phillips Curve is alive and well inside the Fed’s policy framework. The inverse relationship between inflation and unemployment is embedded in its macroeconomic models, and will compel the Fed to tighten policy in response to an unemployment rate that is nosing around 50-year lows (Chart 3). With the committee seemingly willing to let inflation get a bit of a head start before it tightens policy, it may well have to hike faster, and establish a higher terminal rate, than it otherwise would have if it had continued to follow a steady course. We believe the tightening cycle has been postponed rather than truncated, contrary to the money market’s view. Chart 3Sixties Flashback Bottom Line: The Fed is not going to take the fed funds rate to 3.25 - 3.5% by year end, as we expected late last year. We still believe the terminal rate is in that neighborhood, however, and the longer the Fed cools its heels, the greater the potential that it could exceed our estimate. Q: What is the outlook for the rest of the world? The March minutes revealed that conditions in the rest of the world continue to influence the Fed’s policy decisions. The slowdown in China, the uncertain outcomes of ongoing trade talks and Britain’s separation from the EU shadow the outlook in emerging economies and the major non-U.S. developed economies. The outlook for China, other emerging markets, and Europe have been a spirited subject of discussion within BCA. With a majority of the managing editors perceiving the signs of some green shoots, we upgraded Chinese equities to overweight from equal weight, and European and EM equities to equal weight from underweight, at our monthly View Meeting last week. An end to China’s deleveraging campaign may be all the rest of the world needs to show a little more life. Chart 4As China Goes China is a critical influence on our global view. We expect that policymakers have already begun de-emphasizing their deleveraging campaign, as suggested by March’s credit data, released Friday, and will encourage lenders to lend. No one at BCA expects a stimulus campaign on the order of the massive 2008 and 2016 efforts, but the general view is that policymakers can take steps to end the deceleration in China’s growth, since it was rooted in their deleveraging drive. The deceleration weighed on trade and manufacturing activity around the world (Chart 4), and may have been the catalyst for the global mini-slowdown. The rest of the world should benefit from the easing in financial conditions driven by the global equity rally. The decline in bond yields has also helped ease financial conditions, and the nearly unanimous dovishness of major-economy central banks may provide investors and consumers with additional comfort. The key issue for the U.S. economy, and U.S.-oriented investors, is whether or not the other major economies will slow enough to cool off the U.S. at a time when its fiscal impulse is slowing. We have a sense that China and Europe are beginning to turn, and we do not expect spillovers to drag on U.S. growth, but continued rallies in U.S. risk assets probably require some sort of revival beyond its shores. Q: How do corporate profits look? Is the consensus overly optimistic? The corporate profit outlook is getting less ambitious by the day. Over the last three months, consensus expectations for first quarter S&P 500 share-weighted earnings have fallen by 6.5%, as analysts downwardly revised their year-over-year growth projections from +3.5% to -2.2%. Management teams seek to under-promise and over-deliver, and do their best to guide analyst expectations to a level their companies can exceed. Since 1994, according to Thomson Reuters, about two-thirds of companies have reported earnings that beat estimates. On average over that stretch, companies have beaten estimates by a margin of 3.2%. We are therefore inclined to take the projected earnings contraction with a grain of salt. Corporations seem to have lowered the bar to a level they should be able to clear without too much trouble. Chart 5Wages Aren't Yet Pressuring Margins ... We are further inclined to question the projected 2.2% contraction in earnings, given that revenues are projected to grow by 5% in the quarter. The disparity implies margin contraction of close to 7%. Compensation is the largest component of corporate expenses, with the remainder roughly split between interest expense and other input costs. The other meaningful input is the dollar, which should most often exhibit an inverse relationship with margins. Real unit labor costs is the compensation series that most directly impacts profit margins, and it has been contracting on a year-over-year basis, augmenting margins (Chart 5). It will continue to do so as long as nominal wage growth lags inflation and productivity gains. BBB-rated corporate yields were materially higher in the first quarter than they were a year ago, and may have taken a modest bite out of margins, but they’re now back to where they were then and cannot explain the projected 7-ppt margin haircut by themselves (Chart 6). Producer prices grew just 2.2% on a year-over-year basis, slightly ahead of consumer prices (Chart 7), suggesting that margins only slightly narrowed from the disparity between input costs and selling costs. Chart 6... And Interest Rates Aren't Anymore Chart 7Input Costs Are Manageable The broad trade-weighted dollar gained 6% from 1Q18 to 1Q19. Assuming corporations lower prices to defend market share against foreign competitors, profit margins should fall when the dollar rises. Dollar appreciation likely exerted some incremental pressure on margins, but the internal model we’ve previously referenced pegs the EPS impact of a 10% rise in the dollar at 2.5%, far too small for a 6% rise in the dollar to drive a 7-ppt fall in margins. If the revenue estimates are accurate, it seems to us that management must be sandbagging its earnings guidance to some degree. The 10-year Treasury yield will have a harder time falling further now that the Fed is already awfully dovish. Q: Are you having any second thoughts about your duration recommendation? Our below-benchmark duration call was largely founded on our expectation that the Fed was going to surprise complacent markets by hiking more than they expected. It instead surprised dovishly, and the OIS curve responded by pricing in an additional rate cut by the end of next year. The 10-year Treasury yield melted, in accordance with our U.S. Bond Strategy service’s golden rule1 (Chart 8). Chart 8The Golden Rule The surest way to mess up a Fed call is to allow what one thinks the Fed should do to intrude on one’s assessment of what the Fed will do. We did not fall into that trap: our view that the Phillips Curve exerts considerable influence over the Fed and other central banks is founded in the observation that virtually every mainstream macroeconomic model incorporates an inverse relationship between inflation and unemployment. As noted above, we see the Fed’s hiking campaign as extended rather than ended. We believe pausing the hiking campaign will extend the expansion and allow the economy to build up more momentum. More momentum would merit higher real rates, and we also expect it would promote inflation pressures given that the output gap is already closed. We were admittedly on the wrong side as the 10-year Treasury yield fell from 3.25% to 2.4%, but still lower yields would be incompatible with our constructive view of the U.S. economy. With much of the drag on Treasury yields seeming to have come from overseas, it’s also important to note that lower major-economy yields would be incompatible with our house view that the global economy is on the cusp of rebounding (Chart 9). Chart 9Yields Rise When Green Shoots Appear Bottom Line: We missed the slide in the 10-year Treasury yield because we failed to foresee the Fed’s pivot, and because we may have focused too much on U.S., rather than global, conditions. We do not see yields falling much further, however, now that the Fed’s capacity for dovish surprises is spent, and green shoots are starting to appear in China and Europe. Q: How was the Final Four? Fantastic, and we recommend gathering some old college friends and making the trip to cheer on your alma mater should it qualify. Bring your kids if they’re old enough. If your school wins it all, you’ll share lifelong memories of the sort the Virginia alumni who attended the games will cherish. We’ll always have Minneapolis. Go ‘Hoos!   Doug Peta, CFA Chief U.S. Investment Strategist dougp@bcaresearch.com     Footnotes 1      Treasuries beat cash when the Fed hikes less than the money market expects, and lag cash when it hikes more than expected. Please see the U.S. Bond Strategy Special Report, “The Golden Rule Of Bond Investing,” published July 24, 2018. Available at usbs.bcaresearch.com.

Highlights 10-Year Yield: In this week’s report we run through different macro factors that could be used to create a macroeconomic model of the 10-year Treasury yield, and describe the current outlook for each one. On balance, the indicators suggest that the 10-year Treasury yield is near its floor. Global Growth: Leading indicators have hooked up recently, suggesting that the Global Manufacturing PMI – a key driver of the 10-year Treasury yield – may rise in the coming months. Wages: Average hourly earnings softened in March, but survey measures suggest that wage growth remains in an uptrend. We show that rising wages have put considerable upward pressure on the 10-year yield in recent years, and should continue to do so going forward. Sentiment: The depressed Economic Surprise Index suggests that investor economic sentiment is downbeat. This means that the bar for positive data surprises (and higher bond yields) is relatively low. Feature Chart 1CRB/Gold Ratio On The Rise Treasury yields stabilized during the past week, and investors are trying to figure out whether the next big move will be higher or lower. We’re on the record as predicting that yields will eventually head higher, and have flagged the CRB Raw Industrials / Gold ratio as an important indicator to watch to time the next big move.1  Encouragingly, this indicator has risen during the past few weeks (Chart 1). Though the message from the CRB/Gold index is promising, the outlook for the 10-year Treasury yield remains uncertain. To shed some light on this important investment question, in this week’s report we run through different macroeconomic indicators that could be used to create a model of the 10-year Treasury yield. By performing this exercise out in the open, our goal is to present readers with a good way to think about the linkages between the economy and the 10-year Treasury yield. Recipe For A 10-Year Treasury Yield Model Ingredient #1: Growth Factors The first logical factor to include in any model of the 10-year Treasury yield is some measure of economic growth. We have found that the Global Manufacturing PMI is often highly correlated with the 10-year yield (Chart 2). Interestingly, the manufacturing PMI correlates more strongly with the 10-year yield than do the services or composite (manufacturing + services) PMIs. The Global PMI also correlates more strongly with the U.S. 10-year yield than does the U.S. PMI. It only takes a quick glance at the Global Manufacturing PMI to see why the 10-year Treasury yield fell this year. The Global PMI has been in a sharp downtrend for some time, driven mostly by the Euro Area and China. U.S. PMIs have also weakened in recent months, though they remain above levels seen in Europe and China. Another global growth indicator that correlates tightly with the 10-year Treasury yield is investor sentiment toward the U.S. dollar (Chart 3). Since the dollar is a countercyclical currency that appreciates when global growth slows and depreciates when it quickens, we observe that the 10-year Treasury yield tends to be lower when investors are extremely bullish on the U.S. dollar and higher when they are more bearish on the dollar. Chart 2Growth Factor Ingredient 1: Global Manufacturing PMI Chart 3Growth Factor Ingredient 2: Dollar Bullish Sentiment     Notice in Charts 2 and 3 that the Global Manufacturing PMI and dollar bullish sentiment are both close to levels seen near the 10-year yield’s mid-2016 trough. At 50.6, the PMI is only slightly above its 2016 low of 49.9. Meanwhile, dollar bullish sentiment is currently 79%. It maxed out at 82% in 2016. Interestingly, despite the fact that our economic growth indicators paint a similar growth back-drop as 2016, the 10-year yield remains well above its mid-2016 low of 1.37%. Logically, we must conclude that some other “non-growth” factor is propping yields up (more on this below). The 10-year Treasury yield tends to be lower when investors are extremely bullish on the U.S. dollar and higher when they are more bearish on the dollar.  Looking ahead, we remain optimistic that the most important global growth indicators (Global Manufacturing PMI and dollar bullish sentiment) will soon reverse course, as some leading global growth indicators have recently turned a corner. We already saw that the CRB Raw Industrials index has broken out (Chart 1). Additionally: Chart 4The Worst Is Behind Us? The Global ZEW Economic Sentiment index has risen in two consecutive months (Chart 4, top panel). Our Global LEI Diffusion Index shows that more than half of the countries in our sample now have improving leading economic indicators (Chart 4, panel 2). Our BCA Boom/Bust Indicator – an indicator based on the CRB index, Global Metals equities and U.S. unemployment claims – has also jumped (Chart 4, bottom panel). Ingredient #2: Output Gap As noted above, the 10-year Treasury yield looks too high relative to our preferred economic growth indicators. This could be because yields haven’t yet caught up to the deteriorating global economy, but more likely it is because our bond model is still missing some key ingredients. The next most obvious factor to incorporate into our model is some measure of the output gap. If an economy is operating at very close to its peak capacity, with a small output gap, then it doesn’t take much additional growth to spark inflation. Conversely, even rapid economic growth will not be inflationary if the output gap is large. As long as the central bank is expected to lean against rising inflation with higher interest rates, then some measure of the output gap should be included in our bond model. Unfortunately, appropriate output gap measures are difficult to find. We could rely on the CBO or IMF’s output gap estimates, but those are often subject to large ex-post revisions – not ideal if we want to create a bond model that is useful in real time. Since the Fed tends to lift rates when the output gap closes, another option would be to include the fed funds rate as an independent variable in our model. However, this is also not ideal since we would expect the macroeconomic data and the 10-year yield to lead changes in the policy rate. Some measure of inflation might be the best factor to include. However, we find that the correlation between different price inflation measures and the 10-year Treasury yield is incredibly unstable over time. This is likely because the Fed targets price inflation explicitly, making its correlation with bond yields less empirically apparent. Wage growth is the best “output gap” measure to include in a 10-year Treasury yield model.  In fact, our analysis reveals that wage growth is the best “output gap” measure to include in a 10-year Treasury yield model. Specifically, average hourly earnings from the monthly employment report. Not only does the fed funds rate respond – with a lag – to changes in average hourly earnings, but average hourly earnings also line up reasonably well with the 10-year yield over time (Chart 5). Looking at Chart 5, we can now clearly see why the 10-year yield is above its mid-2016 low, despite the poor readings from our growth indicators. Wages have risen sharply since mid-2016, indicating that the output gap has closed, and the Fed has hiked rates 8 times as a result. The obvious conclusion is that in the present situation, with a much smaller output gap than in 2016, it would require a Global Manufacturing PMI well below 50 to produce a 10-year yield near 2% or below. Going forward, we see the uptrend in wage growth continuing for some time. The proportion of workers quitting their jobs each month, a signal of worker bargaining power, remains very high relative to history. Meanwhile, many more households continue to describe jobs as “plentiful” as opposed to “hard to get” (Chart 6). Chart 5Output Gap Ingredient: Average Hourly Earnings Chart 6More Room For Wages To Grow Ingredient #3: Policy Uncertainty The third ingredient we’ll add to our 10-year Treasury yield model is a measure of policy uncertainty. Specifically, the index of Global Economic Policy Uncertainty created by Baker, Bloom and Davis.2  Investors often flock to the safety of U.S. Treasuries in times of economic distress. But Treasuries can also benefit from flight-to-quality flows during periods of stable economic growth but heightened political turmoil. In other words, elevated political uncertainty can make investors fear a downturn in the future, and drive a flight into the safety of U.S. Treasuries. The Global Economic Policy Uncertainty index also shows a relatively strong correlation with the 10-year Treasury yield over time (Chart 7). Chart 7Policy Uncertainty Ingredient: Global Economic Policy Uncertainty Index Looking more closely at Chart 7, we see that global policy uncertainty is currently as high as it was in mid-2016, when the 10-year Treasury yield hit its cycle low. This lines up pretty well with intuition, since investors are understandably quite nervous about the state of Brexit negotiations and U.S./China trade relations. In that context, it is reasonable to expect that some geopolitical risk premium is currently priced into the 10-year Treasury yield, though a smaller output gap than in 2016 is preventing the 10-year yield from reaching mid-2016 levels. Going forward, though political uncertainty will probably stay elevated compared to history. It seems increasingly likely that a “hard Brexit” will be avoided and that President Trump will seek some sort of agreement with China in advance of the 2020 U.S. election.3 The political risk premium in 10-year notes could unwind somewhat in the coming months. Ingredient #4: Sentiment The fourth and final ingredient we’ll add to our 10-year Treasury yield model is a component related to investor sentiment. Our favorite being the U.S. Economic Surprise Index. Chart 8Sentiment Ingredient: Economic Surprise Index Investors don’t often think of the Surprise index as a sentiment indicator, but in fact that’s exactly what it is. It measures whether the economic data exceeded or fell short of expectations during the past 30 days, a measurement that is heavily influenced by whether investor expectations are optimistic or pessimistic. When economic expectations are extremely downbeat it doesn’t take much good news to generate a positive surprise, and vice-versa. Also, investor expectations are influenced in one direction or the other by whether the recent economic data are positive or negative. This behavioral dynamic causes the Economic Surprise Index to be a mean-reverting series, one that we can even describe with a simple auto-regressive model, as shown in Chart 8. More importantly, we have found that the Economic Surprise Index is tightly correlated with the change in the 10-year Treasury yield. A given month that ends with the Surprise index above zero is usually a month when the 10-year Treasury yield increased, and vice-versa (Chart 9). This correlation also holds relatively well over 3-month and 6-month horizons (Charts 10 & 11), but breaks down beyond that.4   The U.S. data surprise index is deeply negative at present, and has been for several weeks. But the longer the data continue to disappoint, the more downbeat investor expectations become and the more likely it is that the surprise index will rise in the future. Right now, our simple auto-regressive model projects that the surprise index will be slightly higher in one month’s time, though still deeply negative. Nevertheless, the Surprise index suggests we are approaching a turning point in investor sentiment. Mix Well, Cover, Stir Occasionally We’ve now presented what, in our view, is a fairly complete list of factors that should be included in a macroeconomic model of the 10-year Treasury yield. Importantly, each factor complements the other ones in the sense that they each capture a different element of the economic landscape. At this stage, it would be nice to weight all of the factors together and arrive at a fair value estimate for the 10-year yield. Unfortunately, we won’t be performing that exercise in this report (we may do so in the future). The key challenge in combining all of the indicators together is that the sensitivity of the 10-year yield to each of the above factors changes over time. For example, there are periods when policy uncertainty appears to be a very significant driver of the 10-year yield, and other times when it appears to not matter much at all. The macro indicators listed in this report generally signal that the 10-year yield is near its trough. While it is often useful to boil all of the important drivers down into a point estimate of the 10-year yield, such an exercise can also create problems if it causes us to zero-in on the model’s output and avoid thinking critically about what the different macro inputs are telling us. As of today, we think the macro indicators listed above generally signal that the 10-year yield is near its trough. Leading global growth indicators have hooked up, suggesting that the Global Manufacturing PMI will improve during the next few months and that bullish dollar sentiment could soften. Survey indicators suggest that the labor market remains tight, and that wage growth will stay in an uptrend. Policy uncertainty will probably continue to apply some downward pressure to yields, but a long Brexit extension and/or trade agreement between the U.S. and China could cause that impact to wane in the next few months. Economic sentiment is likely quite depressed, meaning that the bar for positive surprises is low. All in all, our investment strategy is unchanged. We recommend that investors maintain below-benchmark duration in U.S. bond portfolios, while focusing short positions on the 5-year and 7-year maturities.   Ryan Swift, U.S. Bond Strategist rswift@bcaresearch.com Footnotes 1 The rationale for tracking the CRB/Gold ratio can be found in U.S. Bond Strategy Weekly Report, “The Search For Aaa Spread”, dated March 12, 2019, available at usbs.bcaresearch.com 2 www.policyuncertainty.com 3 Please see Global Investment Strategy Quarterly Outlook, “From Dead Zone To End Zone”, dated March 29, 2019, available at gis.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, “How Much Higher For Yields?”, dated October 31, 2017, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification