Fixed Income

Highlights Clients frequently ask us what long-term returns they should assume when constructing strategic portfolios. In this report, we use a range of methodologies to arrive at reasonable return assumptions for bonds, equities, alternative assets, and currencies on a 10-15 year investment horizon. We conclude that global bonds are likely to return around 1.5% in nominal terms (compared to 5.3% over the past 20 years), and global equities 4.6% (compared to 6.1%). Alternative assets look rather more attractive with, for example, private equity projected to return 9% and real estate 7.1%. Nonetheless, the typical pension fund portfolio, consisting of 50% equities, 30% fixed income, and 20% alts, will be unable to achieve its return target (still typically 7% or higher). Feature Pension plan sponsors and wealth managers need realistic assumptions about the likely returns from different assets in order to construct strategic portfolios, for example when calculating the efficient frontier using a mean-variance optimizer (MVO). Using historical data is the simplest way to do this, but can be very misleading: for example, global bonds have delivered an annual nominal return of 5.3% over the past 20 years but, with bond yields currently so low, it is almost mathematically impossible for them to return anything close to that over coming years (our estimate for future returns is 1.5%). This Special Report is our attempt to produce long-run return assumptions for strategic portfolios, something that GAA clients frequently ask us for. We want to emphasize that these are reasonable assumptions, not forecasts. The value of forecasting the world economy over the next decade or more is questionable. Consider if we had carried out this exercise in 2002: how likely is it that we would have predicted the rise and fall of emerging markets, the U.S. housing crisis, and the subsequent "secular stagnation"? Our analysis, therefore, is mostly based on the philosophy that long-run historical relationships (for example, credit spreads, or the excess return of small cap stocks) are fairly constant, and that most variables (profit margins, valuation, productivity) mean revert over the long term. Our time horizon is 10-15 years. We chose this - rather than the five or seven years that is perhaps more common in such analyses - because it is closer to the investment horizon of pension funds and most individual investors. It also allows us to avoid making a call on where we are currently in the cycle, and how long the next recession and expansion will last. It is likely we are close to the peak of the current economic expansion and equity bull market (the "X" on Chart 1): choosing a shorter time horizon would mean making judgements about the timing of the cycle. Conceptually, we prefer to forecast the trend line on the chart. Chart 1Stylized Trend Versus Cyclical Movements Our assumptions are inevitably approximate. In many cases (particularly for equity returns), we use multiple methodologies and take the average result. Does it matter that the estimation error of our assumptions is likely to be large? Most academic evidence finds not.1 The reason is that, for closely correlated assets, errors in the return estimates (and therefore the optimal weights in a portfolio) will not greatly affect a portfolio's risk and return; while, for assets that are very different, errors in the estimates will not have much effect on the optimal portfolio weights. Rough estimates, therefore, are sufficient for portfolio construction purposes. In any case, using common-sense projections is better than unrealistic historical averages, and investors do need some assumptions to work with when constructing portfolios. How To Forecast Economic Growth A key input (especially when considering earnings growth, which is one factor driving equity returns) is the likely rate of economic growth in various countries and regions over our time horizon. Our simplified way of deriving this is to assume that GDP growth is a factor of (1) demographics (specifically, the growth in the population of working age), and (2) productivity growth. (We assume that capital intensity is steady.) For the demographic assumptions, we use the United Nations' median forecast of the annual growth in population aged 25-64 between 2015 and 2030 (Table 1). Productivity growth is harder to estimate. Productivity has been poor in recent years compared to history (Chart 2). There is significant uncertainty about whether this is caused by cyclical factors (the Great Recession, for example) or structural factors (the end of positive effects from the IT revolution etc.), and whether a potential new wave of technology (artificial intelligence, self-driving vehicles) will raise productivity in future. Table 1Demographic Assumptions Chart 2Productivity Growth Our approach is to assume that productivity in the U.S. will return to its 40-year average, and that productivity growth in the main European economies will be 50 bp lower than the U.S. and in Japan 80 bp lower (in line with recent averages). The estimate is harder for emerging markets, so we use two scenarios: one in which structural reforms, particularly in China, bring productivity growth back up to the average of the past 10 years, 3.5%; and a second scenario in which governments fail to reform, and therefore productivity growth continues to fall to only 1%. For inflation, we assume that central banks over the long-term largely achieve their current inflation goals. The results of our assumptions for GDP growth are shown in Table 2. Table 3 shows the summary of our results: the 10-15 year return assumptions for all the assets in our analysis. We also show historic returns and volatility for comparison (for the past 20 years, where data is available). Below, we describe in detail how we arrived at these numbers. Table 2GDP Growth Assumptions Table 3BCA Assumed Returns All our results are shown in nominal terms and in local currencies. While strictly speaking, it might be theoretically better to estimate real returns, in practice most investors and advisers tend to work on a nominal basis. Moreover, since we have made assumptions for inflation in each region, it is simple to translate our nominal returns into real ones. There is also a trade-off between inflation and currency movements (and interest rates). At the end of the report, we consider the impact of relative inflation rates on currency returns, allowing investors to work the returns back into their own currencies. 1. Fixed income We start from a base that is known: the return on long-term government bonds. If an investor today buys a 10-year U.S. Treasury bond, his or her annual nominal return over the next 10 years will almost certainly be 2.3% (today's yield). The only uncertainties come from (1) reinvesting coupons at the future rate of interest, but the impact of this is small, and (2) the (presumably minimal) risk of a U.S. government default. Of course, investors do not own just 10-year bonds, and indeed the average duration of U.S. Treasuries is currently 5.7 years. But changes in interest rates make relatively little difference to future returns: a rise in interest rates causes a capital loss but a higher yield on rolled-over positions after bonds mature (though, admittedly, the convexity effect is greater when rates are low, as they are now). Even if interest rates were to double over the next decade, the return from U.S. Treasuries would fall only to around 1.5% and, if interest rates fell to 0%, the return would be only about 3%. Moreover, the effect diminishes over time as more bonds are redeemed at par. Empirically, we can see that there is a strong correlation between starting yield on 10-year bonds and long-term returns from U.S. Treasuries (Chart 3). Chart 3Government Bond Returns Driven By The Starting Yield For our cash assumption, we first calculate a proxy for the current cash yield using the average spread between 10-year government bonds and three-month bills over a long-run history (using data from Dimson, Marsh and Staunton which goes back to 1900 and covers a range of countries, Table 4).2 While it is true that the yield curve steepens and flatten along with the cycle, the average yield curve shape should be a good proxy for long-term future expected returns. Of course, this assumes that the term premium comes back. It may not if bonds now are a good hedge against recession risk. However, we also need to take into account that interest rates and inflation are likely to change over the next 10-15 years. We assume that both will rise to an equilibrium level over that time. Our assumption is that central banks will get close to hitting their inflation targets (in the U.S., 2% on PCE inflation, which translates into 2.5% on CPI; in Europe, "around but below 2%"; and in Japan, 2%). For the equilibrium real rate, we take BCA's current estimate (Chart 4) and assume a small rise over the next decade as some of the after-effects of the Great Recession and secular stagnation wear off: to 0.4% in the U.S., -0.1% in the euro area, and -0.2% in Japan. Table 4Historic Spread Government Bonds To Bills (1900-2016) Chart 4Current Equilibrium Real Rates Our calculation of the return from cash over the 10-15 year horizon is based on a steady rise from the current cash return to that implied by the inflation and equilibrium real rate assumptions (Table 5). Table 5Calculation Of Assumption For Cash Return For other fixed-income instruments, we make the following assumptions: Government bonds. We assume that the spread between 10-year and 7-year bonds and 3-month bills will be similar to the historical average (Chart 5), and calculate the return from the government bond index based on this and our estimate for 10-year returns, adjusted by the duration of outstanding bonds in the index: 5.7 years for the U.S., 7.1 for Europe and 8.6 for Japan. For U.S. investment-grade and high-yield corporate bonds, we take the average spread, default rate, and recovery rate in history (Table 6). Obviously, spreads and default rates, especially for high-yield bonds, also jump around massively over the cycle (Chart 6), but we think it is reasonable to assume in our long-term projections that they revert to the mean. Reliable data for European and Japanese credit has a short history but, over the past 10 years, spreads and default rates have been similar to the U.S., so we use the U.S. assumptions for these markets too. Chart 5Yield Curves Table 6U.S. Corporate Credit Assumptions Chart 6Credit Spreads And Default Rates Move With The Cycle Government-related bonds and securitized bonds (MBS, ABS etc.) are an important part of the Barclay's Aggregate Bond indexes: in the U.S., for example, securitized bonds comprise 31% of the index, and government-related ones 7%; in Europe, the weights are 8% and 17% respectively. For our projections of government-related bonds, we assume historic average spreads will continue (Table 7). For securitized bonds, we assume that the historic average spread in the U.S. will continue, and will be the same in Europe and Japan (where historic data is less readily available). Inflation-linked bonds. We assume that the average real yield of the past 10 years, 0%, will continue in future (Chart 7). Table 7Spreads Over Government Bonds Chart 7Real Yield On U.S. TIPs 2. Equities There are a number of ways to think about forward equity returns, all with a high degree of uncertainty. These could be based on starting valuations (but which valuation measure to use?); related to likely earnings growth in future years (hard to forecast); or based on a reversion to the mean of valuations and profits. We decided to take a range of different measures, and average the results. In practice, the results are similar, except for emerging markets (see below for more on EM). Table 8 summarizes the equity return calculations. Table 8Equity Return Calculations AVERAGE EQUITY The thinking behind the six measures we use is as follows. Equity risk premium (ERP). The most obvious methodology: historically, over the long run equities have returned more than government bonds. But which risk premium to use? Dimson, Marsh and Staunton's work includes the excess performance of equities over bonds since 1900 for a range of countries (Table 9). We decided not to choose a different ERP for each developed region, as the historical data would suggest, since it is difficult to argue that the U.S. is likely to be riskier in future than Europe and since, for parts of this history, Japan and the U.S. were essentially emerging markets. We, therefore, take a rounded average of world ERP over the past 116 years, 3.5%. For emerging markets, we multiply this by the average beta of EM relative to global equities over the past 30 years, 1.2, to give an ERP of 4.2%. Growth model. Think of a Gordon Growth Model, which defines the return from equities as the starting dividend yield plus future earnings growth (strictly speaking, dividend growth; we are assuming that the payout ratio will stay constant). We need to make a couple of adjustments to this. First, earnings growth has historically been correlated to nominal GDP growth but has lagged it - in the U.S. by 1.5 percentage points in the period 1918-2016 - although, since 1981, earnings have grown significantly faster than GDP (Chart 8). For the future, we assume that the long-run lag returns. Second, we need to add share buybacks to the dividend yield since, in some countries, such as the U.S., for tax reasons companies prefer to buy back shares rather than increase dividends. However, we should do this on a net basis since equity holders are penalized by companies that issue new shares. In the U.S. net equity withdrawal has been 0.3% over the past 10 years, but in both Europe and Japan, annual net new equity issuance has averaged 1.6% (Chart 9). In EM, the dilution has been even more extreme, averaging 6% over the past 10 years (and much more over the past 25 years). We subtract this dilution from future returns. Table 9Equity Excess Return Over Bonds Chart 8U.S. EPS Growth Versus Nominal GDP Growth Chart 9Net Equity Issuance Growth plus reversion to the mean. This takes the Gordon Growth Model but adds to it an assumption that PE multiples and profit margins revert to the historical mean. We again use dividend yield adjusted by net equity issuance. We assume that the current trailing PE and profit margin revert to the average since 1980 (see Table 8 above for the data) over the next 10 years. In the U.S., PE and margins are currently somewhat higher than history, but this is less the case in Europe or Japan (Charts 10 and 11). Additionally, assuming that the mean reversion happens over 10 years means that the effect on annual returns is not especially large, even for the U.S. Chart 10Net Profit Margin Chart 11Trailing PE History Earnings yield (EY). The simplest of the three valuation measures we use, the assumption is that companies reward shareholders either by paying them a dividend this year, or by reinvesting retained earnings to pay dividends in future. If you assume (admittedly a rash assumption) that the future return on investment will be similar to the current return on investment, it should be immaterial how the company pays out to shareholders. Therefore, the trailing earnings yield (1/PE ratio) should be a good proxy for future returns. Empirically, the relationship between earnings yield and 10-year future returns has been quite strong (Chart 12). However, returns have been somewhat higher on average than the EY would indicate (between 1900 and 2006, 9.7% versus an average EY of 7.5%) mainly because of rising PE multiples since 1980 (Chart 13). We think it unlikely that valuations will continue to rise, and so the EY should be a reasonable guide to future returns. Chart 12Earnings Yield And 10-Year Future Returns Chart 13Trailing Price/Earnings Multiple S&P500 Shiller PE. The cyclically-adjusted price/earnings ratio (CAPE, or Shiller PE) - the current share price divided by the 10 year average of historic inflation-adjusted earnings - has historically had a good correlation with future long-term returns (Chart 14). A regression model of this indicates that the current Shiller PE points to long-run forward returns for the U.S. of 4.9%, for Japan 3.6%, Europe 8.5% and EM 10.8%. Valuation composite. The Shiller PE has some flaws, for example in using a fixed 10-year period for earnings when the length of cycles varies. It has not necessarily mean-reverted in history (perhaps because of long-term trends in interest rates, which it doesn't take into account). It may be more reasonable, then, to use a mixture of different valuation metrics. BCA's Composite Valuation Indicator has had a good correlation with long-run future returns (Chart 15).3 A regression model of this indicator against 15-year returns currently points to returns from the U.S. of 5.2%, Europe of 4.1%, Japan 5.1% and EM 11.0%. Small-cap stocks. We take the 2.4% excess annual return of small cap stocks over large caps in the U.S. for 1926-2016, as calculated by Dimson, Marsh & Staunton. Chart 14Shiller PE Versus ##br##15-Year Equity Return Chart 15Composite Valuation Measure Versus ##br##Long-Run Future Returns Emerging Markets The return assumption for emerging market equity returns has a much higher degree of uncertainty. On our three valuation measures, EM equities look attractive: the average return expectation of the three valuation indicators points to an annual return of 9.4%. However, the growth outlook is murky: as described above, a wave of structural reform in emerging markets, especially China, would be necessary to keep productivity - and, therefore, earnings growth - up, in order for returns to be as good as the current valuation level suggests. Another worry is the degree of equity dilution: it has averaged 6% a year over the past 10 years, and is unlikely to fall much unless corporate governance improves significantly. The range of expected returns derived from our various methodologies, therefore, varies from -1% to +11% a year. Moreover, as described in the currency section below, investors should expect a depreciation in some EM currencies over the next decade, which will also eat into returns. However, due to the influence of China, where the currency is projected to appreciate almost 2% a year against the USD, the EM equity index will see an overall boost to USD-based returns due to the currency effect. 3. Alternative Assets We consider the likely future returns for nine of the 10 alternative assets that Global Asset Allocation regularly covers (we omit wine, which is hard to value on the basis of fundamental macro factors and, anyway, is owned by few institutional investors).4 Alts are harder to forecast than public securities since data is less easily available (and may be only quarterly and based on estimated values), and since some alternative assets have not existed in their current form for very long (venture capital, for example). Moreover, alternative assets tend to have non-normal returns with skewed distributions. Table 10 shows the historical returns and volatility of the nine alternative asset classes both over the longest period for which we have data, and since 1997, when we have data for all of them. Table 10Returns And Volatility For Alternative Assets We, therefore, take a more ad hoc approach, projecting each asset class differently. Generally, we assume that future returns will look similar to historical ones. Specifically, the assumptions we use are as follows. Hedge funds. We assume a return of cash + 3.5%. Hedge fund returns have trended down over time (Chart 16), as more entrants have arbitraged away alpha. We choose to use the average return over cash of the past 10 years, 3.5% (net of fees). It is unlikely that hedge funds returns will rise back anywhere close to earlier levels, for example that of the 1990s when they returned cash +14%. Chart 16Hedge Fund Historic Returns U.S. Direct real estate. We find reasonably good results (R2 = 24%) from regressing U.S. nominal GDP growth against real estate returns. The regression equation is 1.25 x nominal GDP growth + 1.9%. Conceptually, this probably represents a cap rate plus growth of capital values slightly higher than economic growth due to supply shortages in certain key locations. We project real estate to return 7.2% annually. One risk to this assumption, however, is that commercial real estate prices are already above the previous peak from 2007; high valuations may dampen future returns. U.S. REITs. We find only weak correlations with direct real estate investment, although REITs have outperformed real estate over time (perhaps because of the inbuilt leverage of REITs). Over time, REITs have become increasingly correlated with equities. We, therefore, use a regression against U.S. equity returns (R2 = 42%), with REIT returns 0.49 x equity returns + 7.7%. This indicates 10.1% annual return from REITs in the long run. U.S. Private equity (PE). In the past, returns from private equity have been 5 or 6 percentage points higher than from public equities. This is most likely due to their higher leverage, bias towards small-cap companies, and stronger shareholder control over the companies they invest in; it can also be thought of as an illiquidity premium. However, it seems likely that excess returns will be lower in future given the bigger size of the PE industry now and relatively high valuations currently. Moreover, the PE industry currently has almost USD 1 Trn in dry power (uninvested capital), a sign that investment opportunities are limited. We assume, therefore, a slightly lower premium over public equities in future of 4 ppts. This results in a total annual return of 9.5%. U.S. Venture capital (VC). Historically (using data since 1986) VC returns have been 0.6 ppts higher than for PE (probably representing a premium for greater risk and smaller size of the companies invested in). We assume 0.5 ppt higher return in future. This leads to a return assumption of 10%. U.S. Structured products. As discussed in the fixed income section above, we use the 20-year average spread over the aggregate bond index of 0.7 ppt. Total assumed return, therefore, is 3.3%. U.S. Farmland. The value of farmland has risen by an average of 4.4% a year since 1920, a period which included five agricultural cycles. We assume that the value of land will continue to rise at the same rate. We think this is a reasonable assumption since, although nominal GDP growth in the U.S. may be lower in future than in the past, global demand for food is likely to continue to grow rapidly. The total return from investment in farm land, using a regression, produces: growth of farm land value x 1.81 + 0.64% = 8.6%. Chart 17Long-Term Commodity Prices U.S. Timberland is more defensive than farmland since trees can be stored "on the stump" and don't need to be harvested each year in the way that crops do even when prices are unattractive. Historically, timberland has returned about 1 ppt less a year than farmland, and we assume that this will continue. Commodities move in long-run cycles, with a commodity super-cycle of around 10 years, in which prices rise by 3-4x, followed by a bear market of 20 or 30 years in which they fall or stagnate (Chart 17). This is driven by a build-up of excess supply, because of the capex done during the super-cycle, and often by a structural shift on the demand side too. We see no reason why this pattern should change, with China's re-engineering of its economy away from dependence on infrastructure spending likely to be a particularly important factor over the next decade. We assume that commodity prices will, over the current bear market (now about five years old), fall by the same amount and over the same number of years as the average of previous bear markets since the 19th century. This means they have 16% further to fall over 200 months, giving a return of -1% a year. 4. Currencies Most investors are unable or unwilling to fully hedge currency exposure over very long periods. So, a consideration of how returns from different countries' assets might be affected by relative currency movements over the next 10-15 years is an important element in calculating likely returns. Fortunately, for developed market currencies at least, there is a simple, and historically fairly reliable, way to make assumptions of currency movements: reversion to purchasing power parity. As shown in Chart 18, major currencies have fairly consistently reverted to their PPP over the long run. So we can forecast likely future currency movements as a combination of 1) how far away the currency is currently from PPP against the U.S. dollar, and 2) the likely change in the PPP over the period. The latter we calculate from the IMF's forecasts of relative consumer inflation between each country and the U.S. (the IMF makes this forecast only for the next five years, but we assume that the differential continues at the same rate after 2022). Table 11 shows that most major currencies are expected to rise against the U.S. dollar over the coming decade or so. Except for Australia, they are likely to have slightly lower inflation. And - again with the exception of Australia - they all look a little undervalued currently relative to the USD. Table 11Assumed Annual Change Versus U.S. Dollar Over Next 10-15 Years Unfortunately, this approach does not work for EM currencies. They have historically traded at a level consistently well below PPP. This is mainly because, while tradable goods prices tend to be driven by international prices movements and relative unit labor costs, local services prices (which cannot be arbitraged across borders) do not. Also, inflation in emerging markets has historically been much higher than in the U.S. (Chart 19), meaning that their PPP has shifted significantly lower over time. However, China's inflation is now not dissimilar to that of the U.S. (the IMF forecasts it will be only 50 basis points a year higher over the coming five years). And China has shown some tendency for the currency to move towards PPP - 20 years ago the RMB was 190% below PPP; now it is "only" 97% below. Chart 18Reversion To PPP Chart 19U.S. And Emerging Market Inflation We, therefore, take an alternative approach to estimating currency returns for EM economies. We run a regression analysis of the annual change in each country's exchange rate versus the U.S. dollar against its CPI inflation relative to the U.S. We find mostly acceptable r-squared scores (ranging from 57% for Turkey to 1% for Taiwan). For most countries, the intercept is positive (suggesting the currency is trending over time towards PPP) and the coefficient for CPI is, as expected, negative (Table 12). Table 12Calculations For EM Currency Moves A number of EM currencies, on this analysis, would be expected to depreciate against the U.S. dollar over coming years, including Indonesia, Mexico and Turkey. But, weighting the countries by their weights in the MSCI ACWI index, on average the EM universe would be expected to see a currency appreciation against the U.S. dollar of around 2% a year. This is largely due to the influence of China, which has a 29% weight in the EM index. This would be a much better result than the past 10 years when, for example, the Brazilian real has depreciated by 12% a year, the Indonesian rupiah by 16% and the Turkish lira by 37%. This could be because the IMF forecasts of future inflation (4.9% for India, 4.5% for Brazil and 4.1% for Russia), are too optimistic. They are certainly much better than these countries have achieved in the past 10 years (8.0% in India, 6.2% in Brazil, and 9.2% in Russia). Conclusion Arriving at assumptions for future returns is as much an art as a science. Our analysis is based principally on the concept that the future will be similar to long-term history (but not necessarily to the history of the past 30 years, which in many ways were abnormal for financial markets with, for example, a continuous decline in interest rates and inflation). Obviously, therefore, a very different macro environment over the next 10-15 years (for example, one in which inflation spiked, or secular stagnation deepened) would produce a very different results for economic growth and interest rates. However, it will be clear from our analysis that a great deal of the long-term return for equities and bonds is derived from the valuation at the start. Given that current valuations in almost all asset classes are expensive relative to history, this implies that future portfolio returns will be poor compared to recent, and long-term, history. Based on our return assumptions, a typical global portfolio (with 50% equities, 30% bonds, and 20% alternatives) will produce a nominal return of only 4.1% a year over the next decade or so, and a similar U.S. portfolio only 4.6%. This compares to 6.3% and 7.0% over the past 20 years. For pension funds which assume an 7.5% or 8% annual return (as many in the U.S. do), or individual investors planning their retirement on the basis of, say, a 5% annual real return, that outcome would come as a nasty shock. Garry Evans, Senior Vice President Global Asset Allocation garry@bcaresearch.com 1 For the best summary of the evidence on this, please see A Practitioner's Guide To Asset Allocation, by William Kinlaw, Mark Kritzman and David Turkington, Wiley 2017. 2 Please see Credit Suisse Global Investment Returns Yearbook 2017 by Elroy Dimson, Paul Marsh and Mike Staunton, February 2017 3 BCA's Composite Valuation Indicator comprises, for the U.S.: market value of equities / non-financial gross value added adjusted for foreign revenues, trailing PE, Shiller PE, and price to sales. And for other regions: divided yield, market Cap/GDP, trailing PE, price to book, forward PE, price to cash flow, price to sales, and enterprise value/total assets. 4 Please see Global Asset Allocation Special Report, "Alternative Assets: More Important Than Ever", dated 11 March 2016, available at gaa.bcaresearch.com Appendix Correlation Matrix

Highlights Question 1: Why is U.S. inflation still so low? Question 2: How important is the upcoming change in Fed leadership? Question 3: What are the implications of the U.S. tax cuts? Question 4: What is the outlook for the ECB next year, and how will this impact the U.S. dollar versus the euro? Question 5: Are markets underestimating the potential impact from slower growth of central bank balance sheets? Question 6: How much longer can this powerful rally in Emerging Markets continue? Question 7: What are other investors worried about? Feature I have just returned from an extended two-week trip visiting clients in the Asia-Pacific region. The meetings were all very well attended, with even many non-dedicated fixed income investors turning up to ask tough questions about global bond markets. My impression was that given the powerful returns earned in virtually all risk assets this year (equities, credit, Emerging Markets), our clients are growing more concerned about the potential risks from tighter global monetary policy and rising interest rates than they have been for some time. Oddly enough, this is despite not fearing either a serious rise in inflation or a major growth slowdown next year. If such a thing as "nervous complacency" can exist, it seemed widely evident in most of my meetings. This week, I am taking a more personal tone than in a typical Global Fixed Income Strategy Weekly Report to summarize the key client questions from ten days of meetings, spread across six cities in five countries on two continents. Why is U.S. inflation still so low? Chart 1Tightest Global Labor Market##BR##Since The Mid-2000s Almost all of the meetings began with a discussion of the current situation in the U.S., particularly the lack of inflation. The current BCA view that U.S. inflation will accelerate in 2018 was met with some skepticism, particularly when framed in the context of the uncertain unemployment/inflation trade-off. In one meeting, outright laughter actually broke out when the term "Phillips Curve" was mentioned! Clearly, the burden of proof is on the inflation data itself. On that note, I presented several of the best BCA charts from recent months that show how the backdrop is ripe for a turnaround in global inflation. Clients were impressed when shown that nearly ¾ of the countries in the OECD had unemployment rates below the full-employment NAIRU, a level not seen since the period of strong coordinated global growth and rising inflation in the mid-2000s (Chart 1). Yet when I then presented a chart showing the actual inflation/unemployment data in the U.S. over the past 20 years, with a clear "kinked" Phillips Curve and the latest data point well on the steeper portion of that curve (Chart 2), the majority of clients were less convinced. The most cited reason was that the U.S. inflation data simply did not accelerate in 2017 when it was supposed to given the steady fall in unemployment over the preceding few years. Perhaps most surprising was that, rather than believe that the NAIRU rate may simply be lower now than in past business cycles, so many people that I met were willing to discard the entire Phillips Curve concept as a useful framework to forecast inflation. When presented with charts showing non-Phillips Curve reasons to expect higher inflation, however, there was far less skepticism. Perhaps the most compelling chart showed the typical 18-month lag between U.S. economic growth and the "momentum" of U.S. inflation (Chart 3). Upon seeing this, clients were more convinced that inflation would pick up next year in response to the current U.S. growth upturn. Chart 2U.S. Economy Has Moved Into##BR##The "Steep" Part Of The Phillips Curve Chart 3Inflation Typically Follows Economic Growth With A Long Lag I was also able to break down some of the skepticism on the U.S. inflation outlook even more after discussing the bullish oil forecast from our colleagues at Commodity & Energy Strategy. Admittedly, their view that the benchmark Brent oil price will average $65/bbl in 2018 sounds far less dramatic given that the current spot price has risen to nearly that level in the aftermath of the recent political turmoil in Saudi Arabia. Yet clients did appreciate that our bullish oil call was driven more by a view of improving global oil demand and continued production discipline by oil producers (especially for the so-called "OPEC 2.0" nations of Russia and Saudi Arabia). When shown our chart describing how oil prices persistently in the mid-$60s next would put some upward pressure on the inflation expectations component of global bond yields (Chart 4), there was virtually no disagreement from any clients that I met. There was a bit more pushback on the view that, if the BCA forecast of higher U.S. inflation and rising oil prices in 2018 comes to fruition, there is room for a substantial rise in U.S. Treasury yields from current levels. When presented a chart showing that market-based inflation expectations (both using TIPS breakevens and CPI swaps) could rise by 50-60bps just to get back to levels consistent with the Fed's inflation target (Chart 5), most clients politely nodded and basically said "show me the actual inflation first." Although there was widespread agreement with our view that it would take that kind of move in inflation expectations to prompt the Fed to fully deliver on the 100bps of rate hikes it is currently projecting to occur over the next year. Chart 4A Boost To Inflation Expectations##BR##From Higher Oil In 2018 Chart 5The Normalization Of U.S. Inflation##BR##Expectations Will Continue How important is the upcoming change in Fed leadership? The vast majority of clients that I met asked about the BCA view on the nomination of Jerome Powell as the new Fed Chair, replacing Janet Yellen. My impression was that there was not a lot of concern over the potential for serious alterations to the future path of U.S. monetary policy under new leadership. Yet it was still potentially a big enough change to ask questions about it. Most clients agreed with the BCA view that a Fed Chair Powell will not act much differently than Yellen. His voting history has aligned with hers and, by his own admission, he is a very data dependent central banker given that he is not a formally-trained economist. Only by knowing the ins and outs of the data has he been able to debate successfully with the Ph.D economists on the FOMC. Powell will likely be a data-driven Fed Chair that would not look to hike rates without higher inflation (and vice versa). Chart 6A Communications Problem##BR##For Jerome Powell? One point that I raised in all the meetings was that the Fed's communication strategy on future rate increases is the more worrisome issue for financial markets at the moment. The U.S. money market curve is still priced for only 50bps of rate increases over the next year, while the Fed "dots" are signaling 100bps of hikes. We think the Fed will deliver on its projections, which is one of the reasons we are recommending a below-benchmark duration stance in the U.S. (the upside in inflation expectations is the other reason). More importantly, the Fed's so-called "terminal rate" projection is at 2.75%, while our proxy for the market pricing of that rate - the 5-year U.S. Overnight Index Swap rate, 5-years forward - is hovering just above 2% (Chart 6). The persistent disagreement between the market and the Fed over the appropriate level of the terminal rate will become a problem later in 2018 if the Fed does indeed raise the funds rate to over 2% and continues to signal that more rate hikes will come to get the funds rate up to "neutral" (the terminal rate). If the Fed is not able to change the market's mind about the appropriate neutral level of the funds rate, then a move to the Fed's estimated terminal rate of 2.75% would push U.S. monetary policy into what will would be perceived a restrictive stance. This would have implications for the shape of the U.S. Treasury curve (a lot flatter) and for future growth expectations (a lot slower) heading into 2019. My impression from my meetings was that this possibility - that the Fed could engineer what would look to the markets like a policy mistake simply by sticking to its forecasts - was not at the forefront of clients' thinking at the moment. Yet there was no disagreement with the logic of how that could play out. The new Fed leadership under Jerome Powell may have its hands full clearly explaining their policy decisions in 2018, which could create some turbulence in global financial markets later in the year. What are the implications of the U.S. tax cuts? The details of the tax plans from the U.S. House of Representatives and the U.S. Senate were a very hot topic in all of my client meetings. Considering all the ideas being proposed, from cuts in corporate tax rates to changes in the tax treatment of debt interest costs to removing the disincentive to repatriate profits earned abroad, it is no surprise that both equity and fixed income clients had a lot of questions on future U.S. tax policy. It is difficult right now to judge the net impact of the tax changes, as not all of the proposals in the two Congressional tax plans will likely be implemented. There will be plenty of horse trading between the Republicans and Democrats (and between the Republicans themselves) before the final tax deal is done. Yet there was a lot of concern among clients in my meetings over the likelihood that the tax cuts will be implemented at all. After seeing President Trump lose the battle on health care reform earlier this year, many clients were worried that a repeat could happen for the Trump tax cut agenda. This would have negative implications for U.S. equity markets, the U.S. dollar and future Fed policy moves. I explained the views from our colleagues at Geopolitical Strategy, who strongly believe that a tax cut will eventually pass (likely in early 2018) given the need for Congressional Republicans to have something positive to present to voters heading into the 2018 U.S. midterm elections. The tax cuts will have a moderate stimulative effect on the U.S. economy that the markets were not yet fully discounting. I also presented the chart from Global Fixed Income Strategy showing that wider U.S. budget deficits usually coincide with a steeper U.S. Treasury curve, almost always because the U.S. economy is slowing down, prompting looser fiscal policy and also Fed rate cuts (Chart 7). This time is different, however, since the Trump tax cuts will be stimulating an economy currently at full employment (middle panel). This has the potential to trigger more inflation through faster economic growth and even tighter labor markets which could prompt the Fed to move more aggressively on interest rate increases next year and eventually flatten the UST curve (bottom panel). Chart 7A Full-Employment Fiscal Stimulus Will Bear-Steepen The UST Curve The idea of a "steeper, then flatter" Treasury yield curve in response to U.S. fiscal policy stimulus generated a lot of discussion in my meetings. Some even noted that the recent flattening of the curve was a sign that the markets were discounting a lower probability of a tax deal being reached in D.C. I described the flat curve as a consequence of inflation expectations remaining too low, as the Treasury curve was much flatter than implied by the low level of the real fed funds rate, which is one of the most reliable relationships in the bond markets (higher real rates = a flatter curve, and vice versa). My conclusion from these meetings (and from the current market pricing) is that clients are a bit skeptical that a tax deal will be reached. This suggests there is room for bond yields to rise, and the Treasury curve to bear-steepen, if our political strategists are right and the tax cuts will happen. What is the outlook for the ECB next year, and how will this impact the U.S. dollar versus the euro? While most of the questions in my meetings focused on the U.S. outlook, several clients asked about the next move from the European Central Bank (ECB). This was both from a fixed income perspective and, perhaps even more importantly, with an eye on the future direction of the euro versus the U.S. dollar. I made the straightforward argument that with Euro Area economic growth showing strong momentum that is unlikely to slow much in 2018, and with headline Euro Area inflation likely to surprise to the upside based on our bullish oil call (Chart 8), the ECB would likely be forced to signal a tapering of its asset purchase program to zero by the end of next year. The oil view was especially important, as the ECB is expecting a slowing of headline Euro Area inflation to 1% in early 2018 based on the base effects from comparisons to the rise in oil prices seen in early 2017. If our house view on oil prices plays out, then there is potential for inflation to catch the ECB by surprise in 2018. The key will be how core inflation plays out as oil prices rise further. Core Euro Area inflation has dipped lower in recent months, even as wage growth has accelerated (bottom panel). Given tightening Euro Area labor markets, and robust domestic demand, the recent dip in core inflation is likely to bottom out sometime in the first few months of 2018. But until that happens, there is more potential for higher U.S. bond yields through faster increases in inflation expectations and Fed rate hikes (Chart 9). This will support a higher U.S. dollar versus the euro through wider interest rate differentials (bottom panel). Chart 8ECB Will Fully Taper##BR##By The End Of 2018 Chart 9UST-Bund Spread Will Widen Next Year,##BR##Supporting The USD Clients were generally in agreement with that view on relative interest rates, but the views on the direction of EUR/USD were far more mixed. My impression is that if the Fed delivers the rate hikes that we expect in 2018, EUR/USD has room to move lower as investors were not prepared for this. Are markets underestimating the potential impact from slower growth of central bank balance sheets? I received many questions on the potential impact of central banks either shrinking balance sheets (the Fed) or slowing their expansion (the ECB and Bank of Japan). The chart showing how the growth in central bank money printing since 2015 (when the ECB began buying bonds) has correlated strongly with the bull markets in virtually all global risk assets garnered a lot of attention (Chart 10). This was especially true when I showed the chart that converted the level of the major central bank balance sheets to a growth rate and plotted that versus the returns on global equities and credit markets (Chart 11). The implication - expect lower returns on global equity markets, and MUCH lower returns from corporate bond markets next year. Chart 10CB Liquidity Has Supported Risk Assets... Chart 11...But That Tailwind Will Fade Next Year On this point, there was almost no disagreement from clients. There is widespread awareness that this era of puny interest rates, spurred on by central banks buying up huge quantities of government bonds and other financial assets, was forcing investors to take on far more risk in their portfolios to achieve acceptable returns. The key is when this will all turn around. Clients were generally in agreement with my view that the final leg of this liquidity-driven global bull market in risk assets will best be played through equity markets over corporate credit. These stable, earnings-driven rallies seen in global equity markets have not yet reached a "blowoff" phase that would suggest a larger correction is imminent. Perhaps it will take a final asset allocation decision to move more money out of bonds into equities to trigger that final run-up in equity prices before tighter monetary policies and slower growth expectations begin to damage returns later in 2018 into 2019. How much longer can this powerful rally in Emerging Markets continue? This is a topic that generated a healthy amount of debate in my meetings, particularly given the bearish views on Emerging Market (EM) assets from my colleagues at Emerging Markets Strategy. Here again, clients were generally looking at EM as a way to achieve acceptable returns in their portfolios while also participating in the global economic upturn through growth-sensitive assets. The previous chart showing the impact of diminished central bank liquidity on EM credit markets got some clients a bit nervous about the outlook for EM markets. What also spooked them were the charts from our EM strategists showing accelerating Chinese inflation (Chart 12) and slowing Chinese money growth. There is obviously a connection between the two, as China's policymakers are being forced to tighten monetary policy, and clamp down on excess credit creation, in response to accelerating inflation and very high debt levels. The chart showing how our "China M3 Impulse" had turned negative this year and was pointing to slower growth in industrial metals prices and China capital goods imports (Chart 13) was particularly unnerving for even the most bullish of EM clients. Chart 12This Is Why China Is Tightening Monetary Policy Chart 13Prepare For Slower Chinese Growth In 2018 My impression is that the clients I met were fully loaded up on EM assets but were comfortable holding those positions based on expectations of solid Chinese economic growth and continued inflows into EM assets from yield-starved global investors. If BCA's view that Chinese growth will slow next year comes to fruition, combined with rising U.S. interest rates and a stronger U.S. dollar as the Fed tightens more than currently discounted by the markets, then there is potential for outflows from EM markets to accelerate, to the detriment of EM returns. What are other investors worried about? This is a question that comes up a lot at BCA meetings, as clients are always curious as to what we are hearing from other investors. Perhaps this can be chalked up to a version of "confirmation bias", where investors like to hear that their own views are shared by others in the markets. In my meetings over the past two weeks, however, I got the sense that clients are heavily exposed to risk assets, which have performed beyond their expectations, and are growing more worried about how things can go wrong. Like an end to the current low volatility regime, for example. Given the BCA views on the likelihood of global inflation increasing next year, triggering a more hawkish response from policymakers, I noted that I did not believe that clients were prepared for that outcome. This suggests that the beginning of the end of the current low volatility regime, which is seen across all asset classes (Chart 14), will occur through a pickup in bond volatility. This will take place from a rise in inflation expectations first, and a rise in policy rate expectations later. My advice to clients was that if realized bond volatility picks up, this is the signal to reduce exposure to credit and equity markets. We anticipate making such a recommendation sometime during 2018. Chart 14The Low Market Volatility Backdrop Will End When Bond Volatility Rises Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Recommendations Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Highlights A Quick Primer: Convertible bonds have a risk/reward profile that falls somewhere between B-rated and Caa-rated high-yield bonds. The key difference is that convertible bonds are less exposed to credit spreads than junk bonds and more exposed to the equity market. Performance Vs. Junk: Convertibles tend to outperform junk bonds in the late stages of bull markets. Junk spreads tend to trough prior to the peak in equities, and rising Treasury yields also favor convertibles. The Value Proposition: Convertible bonds appear somewhat cheap relative to equities, but equities are fairly valued compared to junk bonds and convertible bonds currently offer an unattractive investment premium. Valuation is by no means compelling, but it should not prevent convertibles from outperforming junk bonds in the late stages of the recovery. Feature From time to time we are asked whether convertible bonds have a place in U.S. fixed income portfolios, and this is the question we consider in today's Special Report. Being a hybrid credit/equity investment, it is clear that convertibles are a fairly high-risk proposition for U.S. bond funds. In fact, we find that the historical risk/reward profile of the sector falls somewhere between B-rated and Caa-rated High-Yield corporate debt (Chart 1). So while not for every investor, we do find compelling reasons for why, in the current macro environment, U.S. bond funds should consider replacing some high-yield debt with an allocation to convertible bonds. Chart 1Fixed Income Universe - Risk Vs. Return* (1990s To Present) This report proceeds as follows. The first section provides a quick primer on convertible bonds, with a focus on the risk factors that drive the sector's performance. The second section takes a look at the historical performance of convertibles in different macro environments. The last section considers valuation in the sector. A Quick Primer Convertible bonds are much like traditional corporate bonds in that they pay periodic fixed coupons and have a maturity date. However, convertibles also give bondholders the option to convert the bond into a pre-specified number of shares of common stock. Investors obviously pay a premium for this option but have the potential to realize large gains if the firm's stock price rises. Convertibles are typically issued by smaller firms with low credit ratings. Fifty percent of the face value of the Bloomberg Barclays U.S. Convertibles index comes from small-cap and mid-cap firms. In addition, 58% of the Convertibles index face value comes from non-rated firms. For comparison, only 0.22% of the Bloomberg Barclays High-Yield bond index is un-rated. Technology firms make up 35% of the Convertibles index, which should not be surprising since the convertibles market appeals to small issuers with large financing demands who may have limited access to traditional debt and equity markets. The Convertibles index also has a large concentration in Healthcare & Pharmaceuticals (19%) and Financials (15%). In contrast, the largest sectors in the High-Yield bond index are Communications (20%), Consumer Cyclicals (15%) and Energy (14%). Chart 2Convertible Bond Value Illustrated The value of a convertible bond is illustrated in Chart 2. This chart shows how the price of a convertible bond (the line labeled "Convertible Bond Value Curve") evolves relative to the firm's stock price (shown on the x-axis). It also helps define a few key terms: Parity: Also called the "conversion value" of the security. This represents the value of the security if the conversion option is exercised. It is the 45 degree line passing through the origin of Chart 2. Conversion Premium: This is the difference between the price of the convertible bond and its parity value. In other words, it is the extra price an investor must pay for a convertible bond relative to simply buying the firm's equity. It can be thought of as the cost of downside protection for an equity investor. Chart 2 shows that this cost is larger when the firm's equity price is low. Investment Value: Also called the "bond floor". This is the value of the security if the conversion option is never exercised. It is the thick dashed line in Chart 2. Investment Premium: This is the difference between the price of the convertible bond and its investment value. In other words, it is the extra price an investor must pay for a convertible bond relative to simply buying a corporate bond. It can be thought of as the cost a convertible bond investor must pay to get potential equity exposure. Chart 2 shows that this cost is greater when the firm's stock price is high. We can use the concepts of conversion premium and investment premium to define three types of convertible bonds. A convertible bond is called "credit sensitive" when its underlying stock price is low. This type of convertible bond is valued very similarly to a corporate bond because there is only a low chance that the conversion option will be exercised. It therefore has a large conversion premium and a small investment premium. Conversely, an "equity sensitive" convertible bond is valued similarly to an equity. This is a convertible bond with an elevated underlying stock price, one that would make it beneficial to exercise the conversion option. These securities have high investment premiums and very low conversion premiums. "Balanced" convertible bonds fall between the other two categories. The stock price has not quite reached the value that makes the conversion option attractive, but it is close enough that the security trades at a significant investment premium. Risk Factors & Performance Drivers Based on the historical risk/reward relationship shown in Chart 1, it is clear that if convertible bonds should be included in U.S. fixed income portfolios it is in place of B-rated and Caa-rated junk bonds. We must therefore consider what factors determine the relative performance between high-yield bonds and convertibles. Chart 3Risk Decomposition, 1999-Present To do that we performed regressions of monthly total returns for both an index of convertible bonds and an index of B-rated and Caa-rated high-yield bonds on three risk factors: Interest rate risk: proxied by the yield on the Treasury Master index Credit risk: proxied by the change in the high-yield bond spread Equity risk: proxied by monthly returns from the S&P 600 The top panel of Chart 3 shows the betas from the regressions. It shows that junk bonds have greater sensitivity to interest rate risk and credit risk, but are less sensitive to equity risk. However, to perform a complete risk attribution we need to consider both the sensitivity of returns to each risk factor and the volatilities of the risk factors themselves. The second panel of Chart 3 shows the variance of each of the three risk factors and the covariance between each pair of risk factors. With that information we are able to calculate the exposure of both junk bond and convertible bond returns to each risk factor by multiplying the squared beta by the variance of each risk factor.1 The results of this complete risk accounting are presented in the bottom panel of Chart 3. It shows that both convertible bond and junk bond total returns have small exposures to interest rate risk, with junk bonds having a somewhat larger exposure. However, the big difference between the two assets is that junk bond returns are mostly determined by credit risk while convertible bond returns are mostly determined by equity risk. This means that we can boil the decision of whether to invest in junk bonds or convertible bonds down to the question of whether junk spreads are likely to outperform equities. It is this question that we address in the next section. Convertible Bond Performance: Equities Vs. Credit Chart 4 shows that the history of relative total returns between convertible and high-yield bonds has gone through five distinct phases since the early 1990s. Chart 4Five Phases Of Convertible Bond Performance In the first phase, which culminated in the 1999/2000 tech boom, convertibles significantly outperformed junk. This is because junk spreads widened while equities performed exceptionally well. The fact that convertible bonds were a popular financing choice for small-cap tech companies also helped the sector's returns in 1999. The second phase, between 2000 and 2005, was characterized by falling Treasury yields and tighter credit spreads. Equities also performed well during this period, but not by enough to offset the impact of falling yields and tighter spreads. Junk bonds outperformed convertibles. Convertibles then outperformed junk from 2005 until the peak of the stock market in 2007. Junk spreads widened prior to the peak in the stock market, and this caused convertibles to outperform junk. Junk outperformed convertibles from 2007 until the mid-2012 trough in Treasury yields. This period was somewhat unique in that both equities and junk spreads were relatively flat. It was the greater impact of falling Treasury yields on junk bonds that drove the relative performance. Finally, convertibles have outperformed sharply since 2012, due to much higher equity prices. Junk spreads are also tighter but did experience a large widening in 2014/15. Higher Treasury yields during this time have also favored convertibles over junk. The Outlook For Convertibles The above framework gives us a way to qualitatively assess whether convertibles are likely to outperform junk going forward. First, with the Fed likely still not passed the mid-point of its rate hike cycle, Treasury yields will probably continue to rise. This will favor convertible bonds over junk bonds. Second, we have previously shown that junk bond spreads are fast approaching historically tight levels.2 In fact, we can calculate that B-rated junk spreads can only tighten another 180 bps before reaching all-time expensive valuations and Caa-rated spreads can only tighten another 375 bps. This represents only four months and five months of average monthly spread tightening, respectively. Further, junk spreads have a history of starting to widen before equity bear markets. The 2008 and 2002 bear markets being the most obvious examples (Chart 4, bottom 2 panels). If this pattern is repeated it will help convertibles outperform junk in the late stages of the current recovery. On the equity side, although valuations are also expensive, prices should continue to rise as long as inflation stays low enough for the Fed to maintain an accommodative policy stance. Our U.S. Investment Strategy service also thinks that small cap equities will outperform large caps in the coming months.3 As discussed above, convertible bonds are mostly issued by small cap firms. Our Cyclical Capitalization Indicator is above the zero line, indicating a favorable macro environment for small caps, and our valuation indicator shows that small caps are relatively cheap (Chart 5). However, arguably more important is that small caps remain a strong high-beta equity play (Chart 5, bottom panel). As long as the equity bull market is maintained, then small caps should outperform. One final factor that should drive the relative outperformance of small cap equities is the potential for Congress to pass tax legislation in the first half of next year. The relative small cap trade has been strongly correlated with other trades that are likely to benefit from proposed tax legislation and right now the market is probably assigning too low a probability to the chance that such legislation will pass (Chart 6). The biggest near-term risk for convertible bond returns relative to junk stems from the Technology sector. Chart 7 shows that divergences in performance between convertible bond sectors and high-yield bond sectors tend to be short lived. It also shows that Technology convertible bonds have strongly outperformed their high-yield counterparts in recent months. It is almost certain that convertible Technology bonds are due for a period of underperformance, much like what happened to Healthcare & Pharmaceutical convertibles in 2014/15 (Chart 7, bottom panel). Given the size of the convertible Technology sector, any period of sector-specific underperformance would also exert a meaningful drag on the overall index. Chart 5Small Caps Poised For Outperformance Chart 6The Trump Trades Are Back On Chart 7Tax Cuts Will Help Small Caps In general, traditional late-cycle dynamics - rising Treasury yields and junk spreads widening before equities sell off - should favor convertible bonds over junk between now and the end of the economic recovery. We expect this will also be the case in the current cycle, although the recent outperformance in the technology sector makes us cautious in the near-term. The best strategy is probably to replace some low-rated high-yield bonds with convertibles, while avoiding the technology sector, and then subbing out even more junk for convertibles once Technology convertibles have come back down to earth. Bottom Line: Convertibles tend to outperform junk bonds in the late stages of bull markets. Junk spreads tend to trough prior to the peak in equities, and rising Treasury yields also favor convertibles. This pattern will likely be repeated in the current cycle, although lofty valuations in the Technology sector make us cautious in the near-term. The Value Proposition The final question worth considering is whether convertible bonds appear fairly valued relative to history and investment alternatives. Specifically, we consider valuation from three different perspectives: Are equities fairly valued relative to high-yield bonds? If equities are cheap relative to high-yield bonds then it is reasonable to expect that convertible bonds are also cheap. Are convertible bonds fairly valued relative to equities? Are convertible bonds fairly valued relative to themselves? To answer the first question we use the BCA S&P 500 Valuation Indicator as our measure of equity valuation (Chart 8, top panel) and the 12-month high-yield breakeven spread as our measure of junk bond valuation (Chart 8, bottom panel). Both indicators are at relatively expensive levels. According to our chosen valuation metrics, equity valuation was only more expensive than it is today during the late 1990s tech boom, but in that period it reached much more expensive levels before changing course. In contrast, there have been several periods when junk breakeven spreads have been tighter, but even the all-time low is not that far below current levels. To create a true relative value indicator we standardized both our equity valuation indicator and the 12-month breakeven junk spread, and then subtracted one from the other. The result is shown in the middle panel of Chart 9 and it suggests that equities look a tad cheap compared to junk. However, relative valuation is nowhere near an extreme, and a more reasonable conclusion would be that equities and junk bonds appear fairly valued relative to each other. Chart 8Equity, Convertible & Junk Valuation I Chart 9Equity, Convertible & Junk Valuation II To answer the second question, whether convertible bonds look cheap relative to equities, we compare the parity (or conversion value) of the Convertible Bond index to our equity valuation indicator. The parity of the Convertible index is shown in the middle panel of Chart 8, and it shows that convertible bonds are quite expensive, but not as expensive as they were between 2013 and 2015. In contrast, our equity valuation indicator is at its most expensive levels of the current cycle. Once again, we standardize both our equity valuation indicator and the Convertible Bond index parity and take the difference. The result is shown in the bottom panel of Chart 9, and unsurprisingly, it suggests that convertible bonds appear relatively cheap compared to equities. To answer the final question, how convertibles are valued relative to themselves, we turn to Charts 10 and 11. Chart 10 shows the conversion premium of the Convertible Bond index relative to its parity. In essence, this chart shows how much extra an investor is being charged for convertible bonds relative to a similar equity portfolio. For a given parity level, a low conversion premium would suggest a more attractive valuation, and vice-versa. At the moment, Chart 10 shows that the index conversion premium is almost exactly in line with its expected value given the level of parity. Chart 10Conversion Premium Vs. Parity Chart 11Investment Premium Vs. Parity Chart 11 shows the Convertible Bond index's investment premium relative to parity. This chart shows how much extra an investor must pay for convertible bonds relative to a similar basket of corporate bonds. Once again, a low investment premium would suggest a more attractive valuation, but at the moment the investment premium appears quite elevated. For the given parity level, investors are paying a bit too much for convertibles relative to a similar basket of corporate bonds. In summary, while convertible bonds do appear somewhat cheap relative to equities, on balance, there is not much of a pure valuation case for the asset class. Equities appear fairly valued relative to junk bonds, and convertibles are trading at an elevated investment premium. We do not think that valuation will be a significant headwind to the typical late-cycle outperformance of convertibles versus junk. Bottom Line: Convertible bonds appear somewhat cheap relative to equities, but equities are fairly valued compared to junk bonds and convertible bonds currently offer an unattractive investment premium. Valuation is by no means compelling, but it should not prevent convertibles from outperforming junk bonds in the late stages of the recovery. Alex Wang, CFA, Research Analyst alexw@bcaresearch.com Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Jeremie Peloso, Research Assistant jeremiep@bcaresearch.com 1 We also use beta weights to attribute the covariance risk to each individual risk factor and calculate the unexplained risk as the variance of the bond returns less the risk exposure from each of our three factors. 2 Please see U.S. Bond Strategy Weekly Report, "The Fed Will Fall Behind The Curve", dated October 24, 2017, available at usbs.bcaresearch.com 3 Please see U.S. Investment Strategy Weekly Report, "Small Cap Surge", dated October 9, 2017, available at usis.bcaresearch.com

Highlights Chart 1Fed Must Fall Behind The Curve Jerome Powell will assume the Fed Chairmanship at a critical juncture for monetary policy. Core PCE inflation is still well below the Fed's 2% target, and yet, the slope of the 2/10 Treasury curve is a mere 71 bps (Chart 1). Such a flat yield curve alongside such low inflation suggests that the market believes the Fed will tighten the yield curve into inversion before inflation even regains the Fed's target. That would be an unprecedented policy mistake that the new Chairman will seek to avoid at all costs. This means either inflation will soon rise, justifying the FOMC's median rate hike projections, or inflation will stay low and the Fed will be forced to take a dovish turn. Either way the Fed must "fall behind the curve" and start chasing inflation higher. The act of falling behind the inflation curve means that long-maturity TIPS breakevens are likely to widen, the yield curve will steepen and the policy back-drop will stay accommodative for spread product. We recommend positioning for all three of these outcomes. Feature Investment Grade: Overweight Chart 2Investment Grade Market Overview Investment grade corporate bonds outperformed the duration-equivalent Treasury index by 52 basis points in October, bringing year-to-date excess returns up to 288 bps. The average index option-adjusted spread tightened 6 bps on the month, and now sits at 97 bps. Two weeks ago we noted that there is simply not much room for investment grade corporate spreads to tighten.1 Looking at 12-month breakeven spreads shown as a percentile rank relative to history, we see that A-rated paper has only been more expensive than it is today 7% of the time. Baa-rated paper has been more expensive only 9% of the time (Chart 2).2 Further, we calculate that at current duration levels Baa-rated option-adjusted spreads can only tighten another 36 bps before the sector is more expensive than it has ever been. Similarly, A-rated spreads can tighten another 14 bps, Aa-rated spreads another 17 bps and Aaa-rated spreads another 7 bps. All this to say that corporate bonds are essentially a carry trade at this stage of the cycle. The important question is how much longer we can pick up the carry before a period of significant spread widening. With low inflation keeping monetary policy accommodative and accelerating profit growth putting downward pressure on leverage (bottom 2 panels), the carry trade appears safe for now (Table 3). Table 3ACorporate Sector Relative Valuation And Recommended Allocation* Table 3B Corporate Sector Risk Vs. Reward* High-Yield: Overweight Chart 3High-Yield Market Overview High-Yield outperformed the duration-equivalent Treasury index by 51 basis points in October, bringing year-to-date excess returns up to 580 bps. The index option-adjusted spread (OAS) tightened 9 bps on the month, and currently sits at 339 bps. Based on our current forecast for default losses we calculate that, if junk spreads remain flat, high-yield excess returns will be 230 bps for the next 12 months. If spreads tighten by 100 bps we should expect excess returns of 606 bps, and if spreads widen by 100 bps we should expect excess returns of -145 bps (Chart 3). Given that the OAS for the high-yield index can only tighten another 139 bps before it reaches all-time expensive valuations, 606 bps is a fairly optimistic excess return projection. But equally, with inflation pressures still muted and monetary policy still accommodative, more than 100 bps of spread widening is also unlikely. Our base case forecast is that high-yield excess returns will be between 2% and 5% (annualized) on a 6-12 month investment horizon.3 In a recent report we noted that high-yield generally looks more attractive than investment grade after adjusting for differences in spread volatility between the two sectors.4 Specifically, we calculate that it will take 39 days of average spread tightening before B-rated bonds reach all-time expensive levels. The same calculation shows it will take 19 days for A-rated debt. MBS: Neutral Chart 4MBS Market Overview Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 4 basis points in October, bringing year-to-date excess returns up to 31 bps. The conventional 30-year zero-volatility MBS spread was roughly flat on the month, as was the option-adjusted spread (OAS) and the compensation for prepayment risk (option cost). Last month we upgraded Agency MBS from underweight to neutral, noting that OAS have become significantly more attractive during the past year, particularly relative to corporate credit (Chart 4). The spread widening likely resulted from the market pricing-in the impact of the Fed's balance sheet run-off. Now that the run-off has begun, and its future pace has been well telegraphed, its impact has probably also been fully priced. While OAS is the correct measure of MBS carry because it adjusts for expected losses due to prepayments, it is the change in the nominal spread that determines capital gains and losses. With that in mind, it is difficult to see a catalyst for significantly wider nominal MBS spreads on a 6-12 month horizon. The two factors that correlate most closely with nominal MBS spreads - credit spreads and mortgage refinancings - are likely to stay depressed (bottom panel). Higher mortgage rates would obviously prevent refinancings from rising. But we showed in a recent report that even if rates move lower the coupon and age distribution of outstanding mortgages has made refi activity much less sensitive to rates than in the past.5 Government-Related: Underweight Chart 5Government-Related Market Overview The Government-Related index outperformed the duration-equivalent Treasury index by 12 basis points in October, bringing year-to-date excess returns up to 193 bps. Sovereign bonds underperformed the Treasury benchmark by 5 bps on the month. Foreign and Domestic Agency bonds outperformed by 2 bps and 9 bps, respectively. Supranationals outperformed by 4 bps. The underperformance in Sovereigns was concentrated in Mexican debt, which sold off as the White House took a hard line on NAFTA negotiations. Local Authority bonds outperformed by 62 bps in October, bringing year-to-date excess returns up to 367 bps (Chart 5). Excess returns for Local Authority debt - mostly taxable municipal debt and USD-denominated Canadian provincial debt - have exceeded excess returns from Baa-rated corporate debt so far this year, despite the sector's average credit rating of Aa3/A1. In a recent report we looked at whether USD-denominated Emerging Market Sovereign debt is an attractive alternative to U.S. high-yield corporates.6 We observed that hard currency EM sovereigns and similarly rated U.S. corporate bonds offer almost exactly the same breakeven spread, and also that EM Sovereigns have been getting comparatively cheaper since early last year. Further, we observed that periods when EM Sovereigns outperform U.S. corporates tend to coincide with falling U.S. rate hike expectations, as measured by our 24-month fed funds discounter. At present, our 24-month discounter is at 74 bps, meaning the market expects less than three Fed hikes during the next two years. We anticipate a better opportunity to move into EM Sovereigns once U.S. rate hike expectations have adjusted higher. Municipal Bonds: Underweight Chart 6Municipal Market Overview Municipal bonds outperformed the duration-equivalent Treasury index by 43 basis points in October (before adjusting for the tax advantage). Munis have outperformed the Treasury benchmark by 251 bps, year-to-date. The average Municipal / Treasury (M/T) yield ratio edged down in October and currently sits at 87%, still extremely tight relative to its post-crisis trading range. M/T yield ratios look much more attractive at the long-end of the curve (Chart 6), and we continue to recommend that investors extend maturity within their municipal bond allocations. Congress released its first draft of proposed tax legislation last week, and while it will certainly undergo some changes in the coming months, it appears as though it will not be very negative for municipal bondholders. Crucially, the top marginal personal tax rate remains unchanged at 39.6% and demand for munis should benefit from the removal of other deductions. A reduction of the corporate tax rate to 20% remains a risk, but that will likely be revised higher as the bill is re-written. Fundamentally, state & local government health improved sharply in Q3, with net borrowing likely falling to $157 billion from $211 billion in Q2, assuming that corporate tax revenues are unchanged (Chart 6).7 The rate of growth in state & local tax revenues now exceeds expenditures and that should put further downward pressure on borrowing in the coming quarters. However, a decline in state & local government borrowing is already reflected in historically tight M/T yield ratios. Treasury Curve: Favor 5-Year Bullet Over 2/10 Barbell Chart 7Treasury Yield Curve Overview The Treasury curve bear-flattened in October alongside a sharp move higher in the expected pace of Fed rate hikes (Chart 7). The 2/10 Treasury slope flattened 8 bps and the 5/30 slope flattened 7 bps. The upward adjustment in rate hike expectations benefited our recommendation to short the July 2018 fed funds futures contract. That trade is now 13 bps in the money since it was initiated on July 10. Further, the July 2018 contract is still discounting fewer than two rate hikes between now and next July. If two more hikes are delivered by July our trade will earn an additional 5 bps. If three more hikes are delivered it will earn an additional 31 bps. In a recent report we discussed why the Fed must soon "fall behind the curve" on inflation and allow the yield curve to steepen.8 Essentially, unless the Fed starts to chase inflation higher it will soon invert the yield curve without having met its inflation goal. That would be a severe policy mistake. This means that either inflation must start to rise, or the Fed must slow its pace of rate hikes. Both scenarios lead to a steeper yield curve. We continue to position for a steeper curve via a long position in the 5-year bullet versus a short position in the 2/10 barbell. At the moment our model shows the 5-year bullet trading roughly in-line with its fair value, or alternatively that the 2/5/10 butterfly spread is priced for an unchanged 2/10 slope on a 6-month horizon.9 TIPS: Overweight Chart 8TIPS Market Overview TIPS outperformed the duration-equivalent nominal Treasury index by 33 basis points in October, bringing year-to-date excess returns up to -99 bps. The 10-year TIPS breakeven inflation rate rose 4 bps on the month but, at 1.86%, it remains well below its pre-crisis trading range of 2.4% to 2.5%. As was pointed out on the front page of this report, the Fed must "fall behind the curve" on inflation if it wants to avoid a policy mistake. Our expectation is that this will occur because inflation will move higher in the coming months. The 6-month rate of change in trimmed mean PCE has already bounced off its lows (Chart 8) and pipeline measures of inflation are soaring (panels 3 & 4). However, even if inflation remains stubbornly low, we think any downside in long-maturity TIPS breakeven rates will prove fleeting. We are approaching an inflection point where if inflation does not rise the Fed will have to adopt a much more dovish policy stance. This should limit any downside in long-dated breakevens. As long as the Fed can maintain interest rates low enough for realized inflation to eventually recover to its target, then we anticipate that long-maturity TIPS breakeven rates will settle into a range between 2.4% and 2.5% by the time that occurs. According to our model, the 10-year TIPS breakeven inflation rate is currently trading in-line with other financial market variables - oil, the trade-weighted dollar and the stock-to-bond total return ratio (panel 2). ABS: Neutral Chart 9ABS Market Overview Asset-Backed Securities outperformed the duration-equivalent Treasury index by 13 basis points in October, bringing year-to-date excess returns up to 81 bps. Aaa-rated ABS outperformed the Treasury benchmark by 10 bps on the month, bringing year-to-date excess returns up to 71 bps. Non-Aaa ABS outperformed the benchmark by 32 bps, bringing year-to-date excess returns up to 176 bps. The index option-adjusted spread for Aaa-rated ABS tightened 5 bps in October and, at 33 bps, it remains well below its average pre-crisis trading range. We continue to favor credit cards over auto loans within Aaa-rated ABS, despite the modest additional spread pick-up available in autos (Chart 9). The main reason is that auto loan net losses have been trending steadily higher for several years while credit card charge-offs are still depressed (panel 4). However, even the credit card space is starting to see rising delinquency rates, albeit off a low base, and banks are tightening lending standards on both auto loans and cards (bottom panel). We expect that tight labor markets and solid income growth will prevent a surge in consumer delinquencies, but these are nonetheless troubling signals that bear monitoring. From a valuation perspective, with the 33 bps OAS offered from Aaa-rated Consumer ABS now only slightly higher than the 29 bps offered by Agency Residential MBS, we advocate a neutral allocation to consumer ABS. Further increases in delinquencies could warrant an eventual downgrade, stay tuned. Non-Agency CMBS: Underweight Chart 10CMBS Market Overview Non-agency Commercial Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 71 basis points in October, bringing year-to-date excess returns up to 182 bps. The index option-adjusted spread (OAS) for non-agency Aaa-rated CMBS tightened sharply in October, from 74 bps to 65 bps. At current levels it is now one standard deviation below its pre-crisis average (Chart 10). With spreads at such low levels in an environment of tightening commercial real estate (CRE) lending standards and falling CRE loan demand, we view the risk/reward trade-off in non-Agency CMBS as quite unfavorable. Agency CMBS: Overweight Agency CMBS outperformed the duration-equivalent Treasury index by 34 basis points in October, bringing year-to-date excess returns up to 96 bps. The index OAS for Agency CMBS tightened 6 bps on the month but, at 46 bps, the sector continues to offer an attractive spread pick-up relative to other low-risk spread product. The Aaa-rated consumer ABS OAS is only 33 bps, and the OAS on conventional 30-year Agency MBS is a mere 29 bps. Such an attractive spread pick-up in a sector that benefits from Agency backing is probably worth grabbing. Treasury Valuation Chart 11Treasury Fair Value Models The current reading from our 2-factor Treasury model (based on Global PMI and dollar sentiment) pegs fair value for the 10-year Treasury yield at 2.69% (Chart 11). Our 3-factor version of the model (not shown), which also incorporates the Global Economic Policy Uncertainty Index, places fair value at 2.67%. The Global Manufacturing PMI increased to 53.5 in October, its highest level in six-and-a-half years. Bullish sentiment toward the dollar also edged higher, but not by enough to prevent the fair value reading from our 2-factor Treasury model from climbing. Last month's fair value reading was 2.65%. The U.S. and Eurozone PMIs continued to trend up, while the Chinese PMI held flat. The Japanese PMI ticked down from 52.9 to 52.8. Most importantly, of the 36 countries we track 34 now have PMIs above the 50 boom/bust line. The global economic recovery has become incredibly broad based, a bearish development for U.S. Treasury yields. For further details on our Treasury models please refer to U.S. Bond Strategy Weekly Report, "The Message From Our Treasury Models", dated October 11, 2016, available at usbs.bcaresearch.com At the time of publication the 10-year Treasury yield was 2.33%. 1 Please see U.S. Bond Strategy Weekly Report, "The Fed Will Fall Behind The Curve", dated October 24, 2017, available at usbs.bcaresearch.com 2 We use breakeven spreads to adjust for the changing duration of the index over time. We calculate the 12-month breakeven spread as option-adjusted spread divided by duration. We ignore the impact of convexity. 3 Please see U.S. Bond Strategy Weekly Report, "Living With The Carry Trade", dated October 17, 2017, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, "The Fed Will Fall Behind The Curve", dated October 24, 2017, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, "Dollar Watching: Yet Another Update", dated October 10, 2017, available at usbs.bcaresearch.com 6 Please see U.S. Bond Strategy Weekly Report, "Living With The Carry Trade", dated October 17, 2017, available at usbs.bcaresearch.com 7 Please see U.S. Bond Strategy Weekly Report, "How Much Higher For Yields?", dated October 31, 2017, available at usbs.bcaresearch.com 8 Please see U.S. Bond Strategy Weekly Report, "The Fed Will Fall Behind The Curve", dated October 24, 2017, available at usbs.bcaresearch.com 9 For further details on our model please see U.S. Bond Strategy Special Report, "Bullets, Barbells And Butterflies", dated July 25, 2017, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification Corporate Sector Relative Valuation And Recommended Allocation Total Return Comparison: 7-Year Bullet Versus 2-20 Barbell (6-Month Investment Horizon)

Highlights London house prices have dropped 7% since the U.K. Government started the formal process of Brexit seven months ago. Stay underweight U.K. real estate and consumer services versus German real estate and consumer services. The global bond yield mini-cycle is driving asset allocation, sector allocation, value/growth allocation and country allocation. We are more than half way through the current mini-upswing in global bond yields. Look for opportunities to cut back overall portfolio cyclicality towards the end of the year. Feature London house prices have dropped 7% since the U.K. Government started the formal process of Brexit seven months ago (Chart of the Week). The average London home is now worth £584,000,1 down from £628,000. Moreover, our leading indicator for London house prices which compares the number of new viewings (demand) with the number of new listings (supply) suggests no imminent end to the sharpest price decline since the 2008 financial crisis (Chart I-2). Chart I-1Brexit Begins To Bite In London Chart I-2The Sharpest Decline In London House Prices Since 2008... Unsurprisingly, the many uncertainties surrounding the unfolding Brexit process are having a much greater impact on the London housing market than on the U.K. housing market as a whole. Outside London, the housing market is broadly flat-lining (Chart I-3). The average U.K. home outside London is now worth £256,500, modestly down from £260,000. Chart I-3 ...But Outside London, Prices Are Flat-Lining U.K. Households Squeezed We are writing ahead of the Bank of England monetary policy meeting, at which the BoE may deliver its first interest rate hike since July 2007. But hike or no hike, we can confidently say one thing: U.K. households will be squeezed. If the BoE does hike the base rate in an attempt to counter overshooting inflation, it could tip the precariously flat-lining housing market outside London into a downturn - as this market is much more exposed to mortgage affordability than it is to Brexit uncertainties. Alternatively, if the BoE does not hike the base rate, the boost to sterling from recent hawkish rhetoric will be priced out, and the pound will come under renewed downward pressure. This would keep U.K. inflation elevated, and further choke U.K. households' real incomes. Absent the post Brexit vote slump in the pound, U.K. inflation would be substantially lower than it is (Chart I-4 and Chart I-5). So the pound's weakness explains why the U.K. is one of the few major economies where inflation is running well north of 2%. Unfortunately for U.K. households, nominal wage inflation has not followed price inflation higher. And as we explained in Why Robots Will Kill Middle Incomes,2 nor is it likely to in the near future. Chart I-4The Weaker Pound Lifted U.K. Headline Inflation... Chart I-5...And U.K. Core Inflation But doesn't textbook economic theory say that the pound's weakness should make U.K. exports more competitive - thereby boosting the net export contribution to economic growth? Yes, the theory does say that a currency devaluation should allow firms to trade in markets that were previously unprofitable to them. However, to trade in these newly profitable markets, firms first need to invest - for example, in marketing and distribution. The trouble is that, post-Brexit, many of the newly profitable markets may be unavailable, or come with heavy tariffs. So firms will hold off making the necessary investments, unless the currency devaluation is massive. But in this case, the corresponding surge in inflation and choke on households' real incomes would also be massive. In summary, U.K. consumer spending faces a continued squeeze. If the BoE delivers a rate hike, household borrowing is likely to fade as a driver of spending. But if the BoE does not deliver the rate hike, the pound will once again weaken, keeping inflation elevated and weighing on real incomes. Stay underweight U.K. consumer services versus German consumer services (Chart I-6). And stay underweight U.K. real estate versus German real estate - expressed either through direct real estate exposure or through real estate equities (Chart I-7). Chart I-6U.K. Consumer Services Equities Are Underperforming Chart I-7U.K. Real Estate Equities Are Underperforming Investment Reductionism Illustrated Turning to markets more generally, it is crucial to understand that most of the moves in most financial markets reduce to a very small number of over-arching macro drivers. We call this very important principle Investment Reductionism. Investment Reductionism emerges from two guiding philosophies: Occam's Razor - which says that when there are competing explanations for the same effect, the simplest explanation is usually the best; and the Pareto Principle (the 80:20 rule) - which says that a small minority of causes usually explain a large majority of effects. The upshot of Investment Reductionism is that the seeming complexity of asset allocation, sector selection, the choice between value or growth, and country allocation usually reduces to something much simpler. Let's illustrate this. The global 6-month credit impulse leads the cyclical direction of the global bond yield, and thereby determines asset allocation (Chart I-8). The direction of the global bond yield drives sector selection: for example Banks versus Healthcare. This is because higher bond yields imply higher net interest margins for banks as well as an improving growth outlook, favouring cyclicals over defensives. And vice-versa (Chart I-9). Chart I-8Investment Reductionism Step 1: ##br##The Global Credit Impulse Leads The Bond Yield Cycle Chart I-9Step 2: The Bond Yield Drives ##br##Sector Performance Banks versus Healthcare determines the European Value versus Growth decision. This is because in Europe, Banks and Healthcare are the dominant value sector and growth sector respectively (Chart I-10). Banks versus Healthcare also determines the country allocation between, say, Italy's MIB - which is bank heavy - and Denmark's OMX - which is healthcare heavy (Chart I-11). Chart I-10Step 3: Sector Performance Drives Value ##br##Vs. Growth Chart I-11Step 4: Sector Performance Drives ##br##Country Performance Therefore, the important lesson from Investment Reductionism is to ignore the hundreds of things that matter little, and to focus on the very small number of things that matter a lot. And one of the things that matters a lot is the global bond yield mini-cycle. Where Are We In The Bond Yield Mini-Cycle? Empirically, the acceleration and deceleration of global bank credit flows - as measured in the global credit impulse - exhibits a remarkably regular wave like pattern, with each half-cycle lasting about 8 months (Chart I-12). The global bond yield shows a similarly regular wave like pattern with each half-cycle also averaging about 8 months (Chart I-13). Chart I-12The Global Credit Impulse Has Also Shown A Regular Wave Like Pattern Chart I-13The Global Bond Yield Has Shown A Regular Wave Like Pattern It is not a coincidence that the bank credit impulse and bond yield exhibit near identical half-cycle lengths. The global credit impulse and global bond yield are inextricably embraced in a perpetual mini-cycle. A stronger credit impulse boosts economic growth. In response to the stronger economic data, the bond yield rises, which slows credit growth. A weaker credit impulse weighs down economic growth. In response to the weaker economic data, the bond yield declines, which re-accelerates credit growth. Go back to step 1 and repeat ad perpetuam. At this moment, from an investment perspective, there are three points worth making: first, bond yield mini-upswings tend to occur mostly within the credit impulse upswing; second, credit impulse mini-upswings have a consistent duration lasting about 8 months; and third, the current mini-upswing started in May. What does this mean for investment strategy? It means that we are more than half-way through the current mini-upswing which we would expect to end around January/February. And at some point early next year we are likely to enter a mini-downswing. So it is slightly premature to cut back cyclical exposure right now. But we would certainly consider opportunities as we move to the end of the year - especially if our now tried and tested fractal timing indicators signal that the price action in specific investments has reached a technical tipping point. Stay tuned. Dhaval Joshi, Senior Vice President Chief European Investment Strategist dhaval@bcaresearch.com 1 Source: LSL Acadata 2 Please see the European Investment Strategy Special Report "Why Robots Will Kill Middle Incomes", dated August 10 2017 available at eis.bcaresearch.com. Fractal Trading Model* This week, our model suggests that the New Zealand dollar is oversold and ripe for a technical rebound. The recommended trade is long NZD/USD with a profit target/stop loss set at 3%. In other trades, long Canada 10-year bond/short German 10-year bund achieved its profit target while short Norway/long Switzerland hit its stop loss. This leaves five open trades. 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. Chart I-14 * 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. Recommendations Equities Bond & Interest Rates Currency & Other Positions Closed Fractal Trades Trades Closed Trades Asset Performance Currency & Bond Equity Sector Country Equity Indicators Bond Yields Chart II-1Indicators To Watch - Bond Yields Chart II-2Indicators To Watch - Bond Yields Chart II-3Indicators To Watch - Bond Yields Chart II-4Indicators To Watch - Bond Yields Interest Rate Chart II-5Indicators To Watch - Interest Rate Expectations Chart II-6Indicators To Watch - Interest Rate Expectations Chart II-7Indicators To Watch - Interest Rate Expectations Chart II-8Indicators To Watch - Interest Rate Expectations

Highlights Emerging Market (EM) hard currency debt, both sovereign and corporate, has consistently outperformed the broad global bond index. However, investors should steer clear of always maintaining maximum overweights to EM given its weak volatility reduction benefits and a much higher-than normal tendency of experiencing outsized, negative returns. Our long-term analysis suggests a structural 5% allocation offers the best risk/reward potential. The Fed is still in the early stages of rate normalization. At this point in the Fed policy cycle, where the Fed is hiking rates but monetary conditions are still stimulative, EM hard currency debt has historically performed well both on a relative and absolute basis. Looking ahead, EM returns should begin to suffer in latter half of 2018 as the Fed moves to more restrictive policy stance. While global growth will remain supportive of EM credit next year, renewed U.S. dollar strength and a re-convergence to the downside with commodity prices present considerable headwinds. Maintain an underweight stance on EM hard currency debt. Favor DM spread product due to more supportive relative growth trends and valuations. Feature Emerging market (EM) sovereign and corporate debt returns have surged in 2017, returning 9.4% and 7.5%, respectively (Chart 1). Investor interest has been renewed, with the latest IMF Financial Stability Report indicating that non-resident inflows of portfolio capital to EM countries have recovered since early 2016 and reached $205 billion for 2017 through August. Against a backdrop of above-trend global economic growth, monetary policy settings from the major central banks that are still accommodative, and some diminished risks from the world's geopolitical hotspots, the current uptrend for EM debt performance could continue. Nevertheless, we urge caution. We moved to a moderate underweight stance on EM hard currency debt back in August, while at the same time increasing our current recommended overweight to U.S. investment grade (IG) corporate debt on the other side of the trade.1 Even with synchronized global growth boosting both EM export demand and industrial commodity prices, we prefer U.S. credit exposure over EM at this point in the cycle, for several reasons: The massive flow-driven EM rally has resulted in not only outsized returns but stretched valuations, with EM debt spreads now back to post-2008-crisis low (or even through those levels for EM hard currency corporates) without any major improvement in EM fundamentals; The previously reliable correlation between EM debt and commodity prices, a long-time driver of EM performance, has broken down, bullishly, for EM - potentially another sign of flow-driven overvaluation; Growing uncertainty over the near-term China growth outlook raises risks on further gains in industrial commodity demand and EM exports; The USD will appreciate once again on the back of additional Fed interest rate hikes beyond levels currently discounted by markets, which could trigger some reversal of the sharp inflows into EM seen this year. Over a strategic horizon, however, it remains difficult to argue against owning a core structural allocation of EM hard currency debt within global fixed income portfolios, given the higher yields that are typically on offer and the fairly consistent historical outperformance over Developed Market (DM) debt. Although the benefits of EM in a portfolio context are slightly overstated given its skewed risk profile (i.e. fat negative tails) and high correlation with DM spread product, specifically U.S. high-yield corporates (Chart 2). Chart 1How Much Longer Can This Rally Last? Chart 2EM Debt Offers Little Diversification Benefits In this Special Report, we examine the long-term role of EM hard currency debt within a fixed-income portfolio, and re-iterate our case for being underweight EM debt on a cyclical basis. The Long-Run Case For Owning EM Debt: A Moderate Core Allocation Makes Sense It is not a stretch to say that EM debt has become the most important part of global bond portfolios in the 21st century. Having a significant EM allocation at the right time can make a bond manager's year, while having it at the wrong time can end a bond manager's career. But what is the "right" allocation to optimize the long-run contribution to returns in a global fixed income portfolio? To answer this question, we took a look at the historical performance of a global bond portfolio that consisted of both DM and EM debt (sovereign and corporate), looking for the combination that would maximize the risk-adjusted return of the portfolio. In our analysis, we ran calculations for two different time periods as the available index data for EM sovereign debt goes back to 1994, while EM corporate debt indices begin in 2002. For DM debt, we used a single index - the Bloomberg Barclays Global Aggregate - as this has a long history and is a common benchmark used by global bond managers that includes both DM sovereign and corporate debt. Though the sample size of our combined global portfolio is limited due to the shorter history of the EM corporates asset class, the findings generally align with our intuition. On a standalone basis, modern portfolio theory proposes that an individual asset should be included within a portfolio if its excess return divided by its standard deviation is higher than the excess return of the portfolio divided by the portfolio's standard deviation, multiplied by the correlation between the portfolio and the asset. Though the correlation to the DM portfolio from 2004 was fairly high for both assets at over 0.6, when we applied this formula, both EM sovereign and corporate debt warranted an allocation in a standard global fixed-income portfolio. EM sovereign debt scored higher, by offering a considerably better Sharpe ratio with only a minimally higher correlation to DM fixed income. While EM hard currency debt has fairly consistently outperformed the DM benchmark on a 12-month rolling basis, investors must be careful not to simply maintain large positions at all times. Obviously, the majority of fixed-income investors have volatility constraints that impose limits on credit allocations. Additionally, apart from simple volatility measures, EM debt has a "hidden" risk profile when looking at the higher moments of return distributions. Table 1EM Debt Returns Are##BR##Negatively Skewed Both EM sovereign and corporate credit historical returns have exhibited significant negative skewness and excess kurtosis, indicating a much higher-than-normal tendency of experiencing outsized, negative returns (Table 1). This is confirmed through Historical Value-at-Risk (VaR) analysis, where the 5% worst returns far eclipsed those of DM investment grade and government debt. Nevertheless, it is important to view EM from a holistic perspective. For example, an asset with a high standard deviation may be less desirable as a standalone investment, but can be highly beneficial if it enhances overall the returns of a portfolio while also reducing its volatility. We tested these "portfolio effects" of EM debt by creating 21 hypothetical portfolios. We began with a DM-only portfolio (consisting of the Global Aggregate index) and increased the weighting toward EM debt by one percentage point in each portfolio, with the last portfolio having a 20% weighting toward EM. The breakdown within EM was 62% corporates and 38% sovereigns based on the market capitalizations of the relevant benchmark indices. Our calculations indicate that the highest portfolio Sharpe ratio was achieved with a 5% EM debt allocation, which also happens to be the "neutral" weighting of EM debt in the BCA Global Fixed Income Strategy model portfolio benchmark index (Chart 3).2 Global bond investors should hover around this weighting on EM hard currency debt, absent a high conviction view on EM. Chart 3The Optimal EM Hard Currency Debt Allocation Is 5% So while the data suggests that EM hard currency debt warrants a long-term allocation, its beneficial impact on a fixed-income portfolio is at least slightly exaggerated. Portfolio managers are typically seeking out assets that can both improve return and decrease overall volatility, thereby increasing the efficiency of their portfolios. This was not the case with EM debt. In our study, increasing the EM allocation consistently raised both returns and volatility. Chart 4EM/DM Correlations Should Decline In 2018 This lack of diversification benefit is a result of the high correlation between EM hard currency debt and DM fixed income. Currently, the correlation between EM and DM (the Global Aggregate) is 0.90, near the upper end of its range, indicating that diversification benefits over the last year were essentially non-existent (Chart 4). Nevertheless, this relationship clearly exhibits a mean reversion tendency. That EM/DM correlation in recent years has been itself correlated to global growth and monetary policy changes. As we show in Chart 4, our diffusion index of OECD Leading Economic Indicators (LEI) - the number of countries with a rising LEI relative to those with a declining LEI - does tend to lead the EM/DM correlation and is currently pointing to a lower correlation as global growth becomes a little less synchronized in 2018. The same goes for the growth rate of major central bank balance sheets which is already slowing and will decelerate even more in 2018 on the back of a diminished pace of bond buying by the ECB and the Fed runoff of maturing bonds on its balance sheet. The conclusion is this - the EM/DM correlation should decline in 2018 but, as we discuss below, we think that happens through relative underperformance of EM credit. Bottom Line: EM hard currency debt, both sovereign and corporate, has consistently outperformed the broad global index. However, investors should steer clear of always maintaining maximum overweights given its weak volatility reduction benefits and a much higher-than normal tendency of experiencing outsized, negative returns. Our long-term analysis suggests a structural 5% allocation offers the best risk/reward potential. The Shorter-Run Case For Owning EM Debt: Will Macro Drivers Remain Supportive? So far in 2017, EM sovereign and corporate debt have been beneficiaries of robust global growth, a declining USD and a decoupling from a broader index of commodity prices. While we expect global growth will remain strong over the medium term, our outlook for the USD is still bullish and there is a risk that commodity prices and EM debt performance re-converge to the downside. Global growth will remain strong. Outside of a major global growth slowdown, which we currently view as a low probability event, a mass flight out of EM assets anytime soon is highly unlikely. Indicators such as the global PMI index, industrial production growth and the OECD leading economic indicator are all booming (Chart 5). Inflation will head higher on the back of rising oil prices, but the increase is likely to be gradual. Importantly, this is happening alongside global monetary conditions that remain generally accommodative, even with the Fed in a tightening cycle. Credit, both DM & EM, has historically performed well against this backdrop, as we discuss in the next section of this report. A renewed upleg in the USD bull market is already underway. The correlation between EM currencies and EM debt performance has recovered after breaking down during 2013-15 (Chart 6). Year-to-date, EM currency strength - the flipside of the weaker U.S. dollar - has been a major driver of EM relative performance. Using the IMF's measure real effective exchange rates based on unit labor costs, the U.S. dollar is fairly valued.3 Neutral valuations suggest that directional market indicators are driving currency movements. As the EM business cycle slows and the Fed ramps up its rate hikes in response to rising inflation, the USD cyclical bull market should resume. Chart 5Robust Global Growth##BR##Is Supportive For EM Chart 6Can EM Ignore Another##BR##Round Of USD Strength? The de-coupling between EM debt and commodity price movements is unsustainable. EM debt has experienced a strong rally since 2016 with only a moderate rise in commodity prices compared to past periods of EM strength. We view this decoupling to be temporary (Chart 7). Many sovereign EM issuers are commodity producers, suggesting that this divergence is unsustainable. EM sovereign and corporate debt will not be able to continue their massive rallies if commodity prices relapse. We maintain a bullish view on oil prices, but there are signals that base metal prices are at risk over the next 6-12 months. Chinese monetary authorities have tightened policy and the resulting sharp slowdown in money supply growth is a worrisome sign for Chinese demand for commodities (Chart 8).4 Chart 7EM-Commodity Divergence##BR##Is Unsustainable Chart 8China Downside Risks For##BR##Industrial Commodity Prices Bottom Line: While global growth will remain supportive of EM credit, currency weakness and a re-convergence with commodity prices present considerable headwinds. EM Debt Performance & The Fed Policy Cycle Chart 9The Fed Policy Cycle As more central banks are shifting to a tightening bias, investors are becoming increasingly concerned over policy normalization and its potential impact on credit market performance. Given the strong historical linkages between EM debt performance and Fed policy changes, the current U.S. tightening cycle looms as a major potential problem for EM assets. We have found it most useful to think about changes in Fed monetary policy and asset market performance in terms of breaking up the Fed policy into four distinct phases (Chart 9).5 These are characterized by both the level of interest rates (whether they are above or below "equilibrium") and the direction of policy changes (whether the Fed is raising or cutting rates):6 Phase 1 - the Fed is hiking while the fed funds rate is below equilibrium (i.e. monetary conditions are stimulative). Phase 2 - the Fed is hiking or keeping policy on hold while the fed funds rate is above equilibrium (i.e. monetary conditions are restrictive). Phase 3 - the Fed is cutting while the fed funds rate is above equilibrium (i.e. monetary conditions are restrictive). Phase 4 - the Fed is cutting rates while the fed funds rate is below equilibrium (i.e. monetary conditions are stimulative). For EM sovereign debt where we have index data going back to 1994, there have been four episodes of Phase 1 and three episodes of the other phases. For EM corporate debt, where the index data begins in 2002, there have been two episodes of Phases 1 and 4 and only one occurrence of Phases 2 and 3. We present the excess returns of EM debt relative to other major fixed income classes by phase in Table 2. In the limited sample, EM sovereign debt and corporate debt consistently outperformed the Global Aggregate index and most individual bond classes. However, relative to DM high-yield debt, which has the most comparable risk profile, EM sovereign bonds underperformed in Phase 1 and EM corporate debt underperformed in all phases. Table 2Relative EM Debt Performance Worsens As Fed Policy Tightens Excess returns for both EM debt classes were highest in Phase 4, where the central bank is easing while conditions are stimulative. Similar to other risk assets, EM debt also outperformed in Phase 1, where the central bank is tightening while rates are below equilibrium. This makes sense, as the early stages of monetary tightening typically occur in conjunction with stable, above-trend growth. Liquidity conditions are still stimulative in Phase 1, which provides a substantial tailwind for spread product performance. On the other end of the spectrum, EM debt excess returns were relatively low during Phase 2 and Phase 3, and even negative in the case of EM corporate debt for Phase 3. Surprisingly, EM debt has been less affected by the direction of U.S. interest rates than what we would have expected. Monetary easing in Phase 3 was not enough to substantially boost EM relative returns and tightening in Phase 1 did not derail growth or lift the USD enough for EM debt to underperform. In fact, because EM debt still offers robust excess returns during Phase 1 when the central bank is tightening, while also suffering during Phase 3 during central bank easing, we can conclude that the level of policy rates relative to equilibrium has a greater impact on returns than the direction of rates. The severity of the Global Financial Crisis and the relatively subdued pace of recovery for both growth and inflation led to one of the longest Phase 4s in history. Given the low level of starting yields, indicating a large gap to equilibrium, and the 'gradual' pace of normalization, the current Phase 1 should also last longer than it typically has. This bodes well for all credit sectors, including EM sovereign and corporate debt, if history is any guide. However, there are still reasons to be concerned about the impact of U.S. monetary policy on EM assets next year. If the Fed follows through with the interest rate hikes it is currently projecting - another 100bps in total by the end of 2018 - the funds rate will be much closer to equilibrium. If the U.S. dollar rallies alongside that Fed tightening, as we expect, overall U.S. monetary conditions could end up being much closer to a restrictive level than implied by strictly looking at our Fed Policy Cycle (which only looks at the funds rate to determine monetary conditions). Also, the equilibrium funds rate may now be lower than the levels we are assuming in the Fed Policy Cycle framework, suggesting that policy could turn restrictive more quickly in the current tightening cycle. Bottom Line: The Fed is still in the early stages of rate normalization. At this point in the Fed policy cycle, where the Fed is hiking rates but monetary conditions are still stimulative, EM hard currency debt has historically performed well both on a relative and absolute basis. Looking ahead, EM returns should begin to suffer in latter half of 2018 as the Fed moves to more restrictive policy stance. Another Reason For Caution: Our EM Corporate Health Monitor The BCA EM Corporate Health Monitor (CHM) is a directional indicator aimed at modeling the path of EM corporate spread movements. Financial data from 220 emerging market companies in over 30 countries is aggregated. Only firms that issue USD-denominated bonds are included, with banks and other financials also omitted in a similar fashion to the CHMs we have constructed for DM corporates. The indicator is made up of four financial ratios: profit margins, free cash flow to total debt, liquidity and leverage. Unlike the DM CHMs, the ratios are not equally weighted in the construction of the EM CHM. Profit margins and free cash flow to debt combined represent 75% of the EM CHM. The latest available reading is from Q2 2017, showing a large decrease, with the indicator now only barely in 'Improving Health' territory (Chart 10). This has occurred in tandem with EM corporate spreads narrowing to post-crisis lows, leaving EM debt at potentially overvalued levels on a fundamental basis. While this slowdown in the EM CHM is not yet a cause for concern, if this became an extended trend of financial health deterioration, the divergence with EM corporate debt performance would be unsustainable and leave EM corporates highly vulnerable to a correction. Chart 10The BCA EM Corporate Health Monitor Has Rolled Over EM Corporate Health Monitor Is Sending A 'Sell' Signal Bottom Line: Our EM Corporate Health Monitor has declined drastically and is barely in 'Improving Health' territory. This alone is not cause for concern yet, but further deterioration in our Monitor combined with additional credit spread narrowing would be a worrisome divergence. Investment Implications Emerging market debt is facing conflicting forces. While continued robust global growth and accommodative monetary policy provide a substantial tailwind for credit performance, extended valuations, the turn in the USD and a potentially worsening commodities outlook present difficult hurdles for EM to overcome. Given the mixed messages, we prefer owning cyclical credit exposure through DM corporate debt, particularly U.S. investment grade. EM debt yields have collapsed and are expensive relative to DM investment grade debt (Chart 11). Combined with a higher risk profile in EM, elevated valuations indicate that EM sovereign and corporate debt are vulnerable to larger corrections. From a return perspective, the difference in the corporate option-adjusted spreads (OAS) has been an excellent leading indicator for relative total returns (Chart 12). This differential indicates that there is considerable relative upside potential for U.S. investment grade over EM hard currency debt. Additionally, while global growth should support credit-related plays, relative growth dynamics are more supportive of U.S. investment grade because the next phase of the global growth upturn will be driven by DM countries and not EM. The difference between the manufacturing PMIs in the U.S. and EM has historically been a good directional indicator for the spread between U.S. corporate bond spreads and EM debt spreads (Chart 13). The gap between the relative manufacturing PMI readings is at a post-crisis high, and could widen further if EM economies suffer on the back of any pullback in Chinese growth in 2018. Chart 11EM Yields & Spreads Look Full Valued Chart 12Favor U.S. IG Over EM Corporates... Chart 13...Because Of Stronger U.S. Growth What are the risks to our view? Our recommended position would suffer in the event that inflation in the U.S. slows, keeping the Fed on hold and maintaining this year's USD downtrend. Also, if China were to ease up on its policy tightening, industrial commodity prices could strengthen once again. Under these scenarios, EM hard currency debt would likely outperform DM spread product. Bottom Line: Maintain moderate underweight positions in EM hard currency debt. Favor DM spread product (especially U.S. investment grade corporates) due to more supportive relative valuations and growth trends. Patrick Trinh, Associate Editor Global Fixed Income Strategy patrick@bcaresearch.com Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 This “EM versus IG” trade was implemented in both our Emerging Markets Strategy and Global Fixed Income Strategy services. Please see BCA Emerging Markets Strategy Weekly Report, “EM: The Focus Is On Profits”, dated August 16th 2017, available at ems.bcaresearch.com, as well as the BCA Global Fixed Income Strategy Weekly Report, “A Lack Of Leadership”, dated August 22nd 2017, available at gfis.bcaresearch.com. 2 The weighting to EM debt in the Global Fixed Income Strategy model bond portfolio benchmark is based on market capitalizations of all the fixed income sectors we wanted to have in the benchmark, which includes non-investment grade debt like global high-yield corporates. It is reassuring to see that our benchmark weighting is also the desired weighting from a long-run portfolio optimization perspective. 3 Please see BCA Emerging Markets Strategy Weekly Report, "Is The Dollar Expensive, And Are EM Currencies Cheap?", dated October 11th, 2017, available at ems.bcaresearch.com. 4 Please see the joint BCA Global Asset Allocation/Emerging Markets Strategy Special Report, "Global Equity Allocation: The Underwhelming Case For EM", dated August 9th 2017, available at ems.bcaresearch.com & gaa.bcaresearch.com. 5 Please see BCA U.S. Bond Strategy Special Report, "Bonds And The Fed Funds Rate Cycle", dated May 27th 2014, available at usbs.bcaresearch.com. 6 The equilibrium policy rate is a BCA calculation based on long-run real potential GDP growth and long run inflation expectations.

Highlights Emerging Market (EM) hard currency debt, both sovereign and corporate, has consistently outperformed the broad global bond index. However, investors should steer clear of always maintaining maximum overweights to EM given its weak volatility reduction benefits and a much higher-than normal tendency of experiencing outsized, negative returns. Our long-term analysis suggests a structural 5% allocation offers the best risk/reward potential. The Fed is still in the early stages of rate normalization. At this point in the Fed policy cycle, where the Fed is hiking rates but monetary conditions are still stimulative, EM hard currency debt has historically performed well both on a relative and absolute basis. Looking ahead, EM returns should begin to suffer in latter half of 2018 as the Fed moves to more restrictive policy stance. While global growth will remain supportive of EM credit next year, renewed U.S. dollar strength and a re-convergence to the downside with commodity prices present considerable headwinds. Maintain an underweight stance on EM hard currency debt. Favor DM spread product due to more supportive relative growth trends and valuations. Feature Emerging market (EM) sovereign and corporate debt returns have surged in 2017, returning 9.4% and 7.5%, respectively (Chart 1). Investor interest has been renewed, with the latest IMF Financial Stability Report indicating that non-resident inflows of portfolio capital to EM countries have recovered since early 2016 and reached $205 billion for 2017 through August. Against a backdrop of above-trend global economic growth, monetary policy settings from the major central banks that are still accommodative, and some diminished risks from the world's geopolitical hotspots, the current uptrend for EM debt performance could continue. Nevertheless, we urge caution. We moved to a moderate underweight stance on EM hard currency debt back in August, while at the same time increasing our current recommended overweight to U.S. investment grade (IG) corporate debt on the other side of the trade.1 Even with synchronized global growth boosting both EM export demand and industrial commodity prices, we prefer U.S. credit exposure over EM at this point in the cycle, for several reasons: The massive flow-driven EM rally has resulted in not only outsized returns but stretched valuations, with EM debt spreads now back to post-2008-crisis low (or even through those levels for EM hard currency corporates) without any major improvement in EM fundamentals; The previously reliable correlation between EM debt and commodity prices, a long-time driver of EM performance, has broken down, bullishly, for EM - potentially another sign of flow-driven overvaluation; Growing uncertainty over the near-term China growth outlook raises risks on further gains in industrial commodity demand and EM exports; The USD will appreciate once again on the back of additional Fed interest rate hikes beyond levels currently discounted by markets, which could trigger some reversal of the sharp inflows into EM seen this year. Over a strategic horizon, however, it remains difficult to argue against owning a core structural allocation of EM hard currency debt within global fixed income portfolios, given the higher yields that are typically on offer and the fairly consistent historical outperformance over Developed Market (DM) debt. Although the benefits of EM in a portfolio context are slightly overstated given its skewed risk profile (i.e. fat negative tails) and high correlation with DM spread product, specifically U.S. high-yield corporates (Chart 2). Chart 1How Much Longer Can This Rally Last? Chart 2EM Debt Offers Little Diversification Benefits In this Special Report, we examine the long-term role of EM hard currency debt within a fixed-income portfolio, and re-iterate our case for being underweight EM debt on a cyclical basis. The Long-Run Case For Owning EM Debt: A Moderate Core Allocation Makes Sense It is not a stretch to say that EM debt has become the most important part of global bond portfolios in the 21st century. Having a significant EM allocation at the right time can make a bond manager's year, while having it at the wrong time can end a bond manager's career. But what is the "right" allocation to optimize the long-run contribution to returns in a global fixed income portfolio? To answer this question, we took a look at the historical performance of a global bond portfolio that consisted of both DM and EM debt (sovereign and corporate), looking for the combination that would maximize the risk-adjusted return of the portfolio. In our analysis, we ran calculations for two different time periods as the available index data for EM sovereign debt goes back to 1994, while EM corporate debt indices begin in 2002. For DM debt, we used a single index - the Bloomberg Barclays Global Aggregate - as this has a long history and is a common benchmark used by global bond managers that includes both DM sovereign and corporate debt. Though the sample size of our combined global portfolio is limited due to the shorter history of the EM corporates asset class, the findings generally align with our intuition. On a standalone basis, modern portfolio theory proposes that an individual asset should be included within a portfolio if its excess return divided by its standard deviation is higher than the excess return of the portfolio divided by the portfolio's standard deviation, multiplied by the correlation between the portfolio and the asset. Though the correlation to the DM portfolio from 2004 was fairly high for both assets at over 0.6, when we applied this formula, both EM sovereign and corporate debt warranted an allocation in a standard global fixed-income portfolio. EM sovereign debt scored higher, by offering a considerably better Sharpe ratio with only a minimally higher correlation to DM fixed income. While EM hard currency debt has fairly consistently outperformed the DM benchmark on a 12-month rolling basis, investors must be careful not to simply maintain large positions at all times. Obviously, the majority of fixed-income investors have volatility constraints that impose limits on credit allocations. Additionally, apart from simple volatility measures, EM debt has a "hidden" risk profile when looking at the higher moments of return distributions. Table 1EM Debt Returns Are##BR##Negatively Skewed Both EM sovereign and corporate credit historical returns have exhibited significant negative skewness and excess kurtosis, indicating a much higher-than-normal tendency of experiencing outsized, negative returns (Table 1). This is confirmed through Historical Value-at-Risk (VaR) analysis, where the 5% worst returns far eclipsed those of DM investment grade and government debt. Nevertheless, it is important to view EM from a holistic perspective. For example, an asset with a high standard deviation may be less desirable as a standalone investment, but can be highly beneficial if it enhances overall the returns of a portfolio while also reducing its volatility. We tested these "portfolio effects" of EM debt by creating 21 hypothetical portfolios. We began with a DM-only portfolio (consisting of the Global Aggregate index) and increased the weighting toward EM debt by one percentage point in each portfolio, with the last portfolio having a 20% weighting toward EM. The breakdown within EM was 62% corporates and 38% sovereigns based on the market capitalizations of the relevant benchmark indices. Our calculations indicate that the highest portfolio Sharpe ratio was achieved with a 5% EM debt allocation, which also happens to be the "neutral" weighting of EM debt in the BCA Global Fixed Income Strategy model portfolio benchmark index (Chart 3).2 Global bond investors should hover around this weighting on EM hard currency debt, absent a high conviction view on EM. Chart 3The Optimal EM Hard Currency Debt Allocation Is 5% So while the data suggests that EM hard currency debt warrants a long-term allocation, its beneficial impact on a fixed-income portfolio is at least slightly exaggerated. Portfolio managers are typically seeking out assets that can both improve return and decrease overall volatility, thereby increasing the efficiency of their portfolios. This was not the case with EM debt. In our study, increasing the EM allocation consistently raised both returns and volatility. Chart 4EM/DM Correlations Should Decline In 2018 This lack of diversification benefit is a result of the high correlation between EM hard currency debt and DM fixed income. Currently, the correlation between EM and DM (the Global Aggregate) is 0.90, near the upper end of its range, indicating that diversification benefits over the last year were essentially non-existent (Chart 4). Nevertheless, this relationship clearly exhibits a mean reversion tendency. That EM/DM correlation in recent years has been itself correlated to global growth and monetary policy changes. As we show in Chart 4, our diffusion index of OECD Leading Economic Indicators (LEI) - the number of countries with a rising LEI relative to those with a declining LEI - does tend to lead the EM/DM correlation and is currently pointing to a lower correlation as global growth becomes a little less synchronized in 2018. The same goes for the growth rate of major central bank balance sheets which is already slowing and will decelerate even more in 2018 on the back of a diminished pace of bond buying by the ECB and the Fed runoff of maturing bonds on its balance sheet. The conclusion is this - the EM/DM correlation should decline in 2018 but, as we discuss below, we think that happens through relative underperformance of EM credit. Bottom Line: EM hard currency debt, both sovereign and corporate, has consistently outperformed the broad global index. However, investors should steer clear of always maintaining maximum overweights given its weak volatility reduction benefits and a much higher-than normal tendency of experiencing outsized, negative returns. Our long-term analysis suggests a structural 5% allocation offers the best risk/reward potential. The Shorter-Run Case For Owning EM Debt: Will Macro Drivers Remain Supportive? So far in 2017, EM sovereign and corporate debt have been beneficiaries of robust global growth, a declining USD and a decoupling from a broader index of commodity prices. While we expect global growth will remain strong over the medium term, our outlook for the USD is still bullish and there is a risk that commodity prices and EM debt performance re-converge to the downside. Global growth will remain strong. Outside of a major global growth slowdown, which we currently view as a low probability event, a mass flight out of EM assets anytime soon is highly unlikely. Indicators such as the global PMI index, industrial production growth and the OECD leading economic indicator are all booming (Chart 5). Inflation will head higher on the back of rising oil prices, but the increase is likely to be gradual. Importantly, this is happening alongside global monetary conditions that remain generally accommodative, even with the Fed in a tightening cycle. Credit, both DM & EM, has historically performed well against this backdrop, as we discuss in the next section of this report. A renewed upleg in the USD bull market is already underway. The correlation between EM currencies and EM debt performance has recovered after breaking down during 2013-15 (Chart 6). Year-to-date, EM currency strength - the flipside of the weaker U.S. dollar - has been a major driver of EM relative performance. Using the IMF's measure real effective exchange rates based on unit labor costs, the U.S. dollar is fairly valued.3 Neutral valuations suggest that directional market indicators are driving currency movements. As the EM business cycle slows and the Fed ramps up its rate hikes in response to rising inflation, the USD cyclical bull market should resume. Chart 5Robust Global Growth##BR##Is Supportive For EM Chart 6Can EM Ignore Another##BR##Round Of USD Strength? The de-coupling between EM debt and commodity price movements is unsustainable. EM debt has experienced a strong rally since 2016 with only a moderate rise in commodity prices compared to past periods of EM strength. We view this decoupling to be temporary (Chart 7). Many sovereign EM issuers are commodity producers, suggesting that this divergence is unsustainable. EM sovereign and corporate debt will not be able to continue their massive rallies if commodity prices relapse. We maintain a bullish view on oil prices, but there are signals that base metal prices are at risk over the next 6-12 months. Chinese monetary authorities have tightened policy and the resulting sharp slowdown in money supply growth is a worrisome sign for Chinese demand for commodities (Chart 8).4 Chart 7EM-Commodity Divergence##BR##Is Unsustainable Chart 8China Downside Risks For##BR##Industrial Commodity Prices Bottom Line: While global growth will remain supportive of EM credit, currency weakness and a re-convergence with commodity prices present considerable headwinds. EM Debt Performance & The Fed Policy Cycle Chart 9The Fed Policy Cycle As more central banks are shifting to a tightening bias, investors are becoming increasingly concerned over policy normalization and its potential impact on credit market performance. Given the strong historical linkages between EM debt performance and Fed policy changes, the current U.S. tightening cycle looms as a major potential problem for EM assets. We have found it most useful to think about changes in Fed monetary policy and asset market performance in terms of breaking up the Fed policy into four distinct phases (Chart 9).5 These are characterized by both the level of interest rates (whether they are above or below "equilibrium") and the direction of policy changes (whether the Fed is raising or cutting rates):6 Phase 1 - the Fed is hiking while the fed funds rate is below equilibrium (i.e. monetary conditions are stimulative). Phase 2 - the Fed is hiking or keeping policy on hold while the fed funds rate is above equilibrium (i.e. monetary conditions are restrictive). Phase 3 - the Fed is cutting while the fed funds rate is above equilibrium (i.e. monetary conditions are restrictive). Phase 4 - the Fed is cutting rates while the fed funds rate is below equilibrium (i.e. monetary conditions are stimulative). For EM sovereign debt where we have index data going back to 1994, there have been four episodes of Phase 1 and three episodes of the other phases. For EM corporate debt, where the index data begins in 2002, there have been two episodes of Phases 1 and 4 and only one occurrence of Phases 2 and 3. We present the excess returns of EM debt relative to other major fixed income classes by phase in Table 2. In the limited sample, EM sovereign debt and corporate debt consistently outperformed the Global Aggregate index and most individual bond classes. However, relative to DM high-yield debt, which has the most comparable risk profile, EM sovereign bonds underperformed in Phase 1 and EM corporate debt underperformed in all phases. Table 2Relative EM Debt Performance Worsens As Fed Policy Tightens Excess returns for both EM debt classes were highest in Phase 4, where the central bank is easing while conditions are stimulative. Similar to other risk assets, EM debt also outperformed in Phase 1, where the central bank is tightening while rates are below equilibrium. This makes sense, as the early stages of monetary tightening typically occur in conjunction with stable, above-trend growth. Liquidity conditions are still stimulative in Phase 1, which provides a substantial tailwind for spread product performance. On the other end of the spectrum, EM debt excess returns were relatively low during Phase 2 and Phase 3, and even negative in the case of EM corporate debt for Phase 3. Surprisingly, EM debt has been less affected by the direction of U.S. interest rates than what we would have expected. Monetary easing in Phase 3 was not enough to substantially boost EM relative returns and tightening in Phase 1 did not derail growth or lift the USD enough for EM debt to underperform. In fact, because EM debt still offers robust excess returns during Phase 1 when the central bank is tightening, while also suffering during Phase 3 during central bank easing, we can conclude that the level of policy rates relative to equilibrium has a greater impact on returns than the direction of rates. The severity of the Global Financial Crisis and the relatively subdued pace of recovery for both growth and inflation led to one of the longest Phase 4s in history. Given the low level of starting yields, indicating a large gap to equilibrium, and the 'gradual' pace of normalization, the current Phase 1 should also last longer than it typically has. This bodes well for all credit sectors, including EM sovereign and corporate debt, if history is any guide. However, there are still reasons to be concerned about the impact of U.S. monetary policy on EM assets next year. If the Fed follows through with the interest rate hikes it is currently projecting - another 100bps in total by the end of 2018 - the funds rate will be much closer to equilibrium. If the U.S. dollar rallies alongside that Fed tightening, as we expect, overall U.S. monetary conditions could end up being much closer to a restrictive level than implied by strictly looking at our Fed Policy Cycle (which only looks at the funds rate to determine monetary conditions). Also, the equilibrium funds rate may now be lower than the levels we are assuming in the Fed Policy Cycle framework, suggesting that policy could turn restrictive more quickly in the current tightening cycle. Bottom Line: The Fed is still in the early stages of rate normalization. At this point in the Fed policy cycle, where the Fed is hiking rates but monetary conditions are still stimulative, EM hard currency debt has historically performed well both on a relative and absolute basis. Looking ahead, EM returns should begin to suffer in latter half of 2018 as the Fed moves to more restrictive policy stance. Another Reason For Caution: Our EM Corporate Health Monitor The BCA EM Corporate Health Monitor (CHM) is a directional indicator aimed at modeling the path of EM corporate spread movements. Financial data from 220 emerging market companies in over 30 countries is aggregated. Only firms that issue USD-denominated bonds are included, with banks and other financials also omitted in a similar fashion to the CHMs we have constructed for DM corporates. The indicator is made up of four financial ratios: profit margins, free cash flow to total debt, liquidity and leverage. Unlike the DM CHMs, the ratios are not equally weighted in the construction of the EM CHM. Profit margins and free cash flow to debt combined represent 75% of the EM CHM. The latest available reading is from Q2 2017, showing a large decrease, with the indicator now only barely in 'Improving Health' territory (Chart 10). This has occurred in tandem with EM corporate spreads narrowing to post-crisis lows, leaving EM debt at potentially overvalued levels on a fundamental basis. While this slowdown in the EM CHM is not yet a cause for concern, if this became an extended trend of financial health deterioration, the divergence with EM corporate debt performance would be unsustainable and leave EM corporates highly vulnerable to a correction. Chart 10The BCA EM Corporate Health Monitor Has Rolled Over EM Corporate Health Monitor Is Sending A 'Sell' Signal Bottom Line: Our EM Corporate Health Monitor has declined drastically and is barely in 'Improving Health' territory. This alone is not cause for concern yet, but further deterioration in our Monitor combined with additional credit spread narrowing would be a worrisome divergence. Investment Implications Emerging market debt is facing conflicting forces. While continued robust global growth and accommodative monetary policy provide a substantial tailwind for credit performance, extended valuations, the turn in the USD and a potentially worsening commodities outlook present difficult hurdles for EM to overcome. Given the mixed messages, we prefer owning cyclical credit exposure through DM corporate debt, particularly U.S. investment grade. EM debt yields have collapsed and are expensive relative to DM investment grade debt (Chart 11). Combined with a higher risk profile in EM, elevated valuations indicate that EM sovereign and corporate debt are vulnerable to larger corrections. From a return perspective, the difference in the corporate option-adjusted spreads (OAS) has been an excellent leading indicator for relative total returns (Chart 12). This differential indicates that there is considerable relative upside potential for U.S. investment grade over EM hard currency debt. Additionally, while global growth should support credit-related plays, relative growth dynamics are more supportive of U.S. investment grade because the next phase of the global growth upturn will be driven by DM countries and not EM. The difference between the manufacturing PMIs in the U.S. and EM has historically been a good directional indicator for the spread between U.S. corporate bond spreads and EM debt spreads (Chart 13). The gap between the relative manufacturing PMI readings is at a post-crisis high, and could widen further if EM economies suffer on the back of any pullback in Chinese growth in 2018. Chart 11EM Yields & Spreads Look Full Valued Chart 12Favor U.S. IG Over EM Corporates... Chart 13...Because Of Stronger U.S. Growth What are the risks to our view? Our recommended position would suffer in the event that inflation in the U.S. slows, keeping the Fed on hold and maintaining this year's USD downtrend. Also, if China were to ease up on its policy tightening, industrial commodity prices could strengthen once again. Under these scenarios, EM hard currency debt would likely outperform DM spread product. Bottom Line: Maintain moderate underweight positions in EM hard currency debt. Favor DM spread product (especially U.S. investment grade corporates) due to more supportive relative valuations and growth trends. Patrick Trinh, Associate Editor Global Fixed Income Strategy patrick@bcaresearch.com Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 This “EM versus IG” trade was implemented in both our Emerging Markets Strategy and Global Fixed Income Strategy services. Please see BCA Emerging Markets Strategy Weekly Report, “EM: The Focus Is On Profits”, dated August 16th 2017, available at ems.bcaresearch.com, as well as the BCA Global Fixed Income Strategy Weekly Report, “A Lack Of Leadership”, dated August 22nd 2017, available at gfis.bcaresearch.com. 2 The weighting to EM debt in the Global Fixed Income Strategy model bond portfolio benchmark is based on market capitalizations of all the fixed income sectors we wanted to have in the benchmark, which includes non-investment grade debt like global high-yield corporates. It is reassuring to see that our benchmark weighting is also the desired weighting from a long-run portfolio optimization perspective. 3 Please see BCA Emerging Markets Strategy Weekly Report, "Is The Dollar Expensive, And Are EM Currencies Cheap?", dated October 11th, 2017, available at ems.bcaresearch.com. 4 Please see the joint BCA Global Asset Allocation/Emerging Markets Strategy Special Report, "Global Equity Allocation: The Underwhelming Case For EM", dated August 9th 2017, available at ems.bcaresearch.com & gaa.bcaresearch.com. 5 Please see BCA U.S. Bond Strategy Special Report, "Bonds And The Fed Funds Rate Cycle", dated May 27th 2014, available at usbs.bcaresearch.com. 6 The equilibrium policy rate is a BCA calculation based on long-run real potential GDP growth and long run inflation expectations.

Dear Clients, Please note there was an error in the Recommend Asset Allocation table published on November 1, 2017. This has now been amended. We apologize for the confusion and any inconvenience it may have caused. Best Regards, Garry Evans Senior Vice President Global Asset Allocation Reflation Trade Returns Recommended Allocation The market mood has shifted remarkably quickly over the past couple of months. The probability of a December Fed rate hike has moved up from 20% in early September to close to 100%, pushing the 10-year Treasury bond yield from 2.0% to 2.4% and causing the trade-weighted U.S. dollar to appreciate by 2%, and Emerging Market equities to underperform. We expect this trend to continue. Global growth continues to surprise to the upside (Chart 1). The softness in U.S. inflation this year is likely to reverse over coming quarters - an argument supported by the New York Fed's new Underlying Inflation Gauge, which indicates that sustained movements in inflation continue to trend higher (Chart 2). This makes it likely that the Fed will move ahead with its forecast three rate hikes in 2018, which the market has not yet priced in (Chart 3) - the implied probability of this is only 10%. Consequently, rates have further to rise: our fair value for the U.S. 10-year Treasury yield currently is 2.7%. And the increasing gap between U.S. and euro zone interest rates suggests that the dollar can appreciate further (Chart 4). All this supports our view that risk assets (equities and corporate credit) should outperform over the next 12 months, with developed government bonds producing a negative return, and emerging markets lagging because of rising rates and the stronger dollar (and a possible slowdown in China, as it focuses on reforming its economy and cleaning up the debt situation). Chart 1Growth Surprising To The Upside Chart 2Underlying Inflation Still Trending Up Chart 3Market Expects Fed To Move Only Slowly Chart 4Rate Gap Suggests Dollar Appreciation The key question, though, is how long this positive scenario can continue. With stock market valuations expensive (Chart 5) and investors fully invested, though not yet euphoric (Chart 6), we are clearly in late cycle. Rising rates could put a dampener on growth. Chart 5 Equities Close To Extremely Overvalued Chart 6Investors Are Fully Invested, But Cautious We find the Fed policy cycle a useful tool for thinking about probable investment returns from different assets (Chart 7). The best quadrant for risk assets is when the Fed is easing and policy is easy (with the Fed Funds Rate below the neutral rate). Currently we are in the bottom-right quadrant (Fed tightening, but not yet in the tight zone), which also has produced attractive returns for equities and credit. But once the Fed Funds Rate (FFR) moves above the neutral rate, returns from risk assets are on average poor and, historically, recession often followed quite quickly. How much longer do we have before Fed policy moves into the top-right quadrant? The Fed's own estimate of the neutral rate, in real terms, is 0.3%. The current real FFR (using core PCE inflation, 1.3%, as the deflator) is -0.17 (Chart 8). This implies that it will take only two further Fed hikes to move into the tight zone, which could happen as soon as March. This is why the outlook for inflation is critical. If, as the Fed forecasts and we also expect, core PCE inflation rises to 2%, it will be another five hikes before policy turns tight - we are unlikely to get there until early 2019. Chart 7The Fed Policy Cycle Chart 8How Far From The Tight Zone? For now, therefore, we continue to recommend an overweight on risk assets and pro-cyclical portfolio tilts. Global monetary policy remains easy and we see no indicators that suggest growth is slowing or that the risk of recession over the next 12 months is rising. The risks to this optimistic scenario (a hawkish Fed, over-eager structural reform in China, provocation from North Korea) seem limited. But we also continue to warn of the possibility of a recession in 2019 or 2020 caused, as so often, by excessive Fed tightening. We see, therefore, the possibility of our turning more defensive somewhere in mid-2018. Equities: We prefer developed over emerging market equities. Rising interest rates and an appreciating dollar will be headwinds for EM. Moreover, Xi Jinping's speech at the Communist Party Congress hinted at supply side structural reforms, overcapacity reduction, and deleveraging efforts. A renewed reform effort could dampen Chinese growth somewhat which, as in 2013-15, would negatively impact EM equities (Chart 9). Within DM, we are overweight euro zone and Japanese equities, which are higher beta, have stronger earnings momentum, and benefit from looser monetary policy. Fixed Income: We expect bonds to underperform over coming quarters, as U.S. inflation picks up and the Fed moves raises rates in line with its "dots". Corporate credit still has some attractions, provided the economic expansion continues. U.S. sub-investment grade bonds, in particular, have an attractive default-adjusted yield, as long as a strong economy keeps the default rate over the next 12 months to the historically low 2% our model suggests (Chart 10). The pick-up in inflation we expect would mean inflation-linked bonds outperform nominal bonds. Chart 9Slowing China Would Hurt EM Equities Chart 10Junk Attractive If Defaults Stay This Low Currencies: The ECB delivered a dovish tapering last month, extending its asset purchases until at least September 2018 and emphasizing that its current low interest rates will continue "well past the horizon of our net asset purchases". Given this, and the gap between U.S. and euro zone interest rates (Chart 4), we expect moderate further euro weakness over coming months. The dollar is likely to appreciate even more against the yen. There are the first tentative signs of inflation emerging in Japan (Chart 11) which, combined with the Bank of Japan sticking to its 0% 10-year JGB target and rising global interest rates, could push the yen to 120 against the dollar over coming months. Commodities: BCA's energy strategists recently revised up their crude oil forecasts on the back of strong demand, a likely extension of the OPEC agreement until at least end-2018, and possible supply disruptions in Iraq, Venezuela and other troubled regions.1 They see inventories continuing to draw down until at least 2H 2018 (Chart 12). Accordingly, they forecast $65 a barrel for Brent and $63 for WTI and flag upside risk to those projections. The outlook for industrial and precious metals, however, is less positive. A stronger dollar and a shift in the growth drivers in China will depress prices for base metals. Rising real interest rates will hurt gold, although we still like precious metals as a long-term hedge. Chart 11First Signs Of Inflation In Japan? Chart 12Oil Inventory Drawdowns Support Higher Price Garry Evans, Senior Vice President Global Asset Allocation garry@bcaresearch.com 1 Please see Commodity & Energy Strategy Weekly Report "Oil Forecast Lifted As Market Tightens," dated 19 October 2017, available at ces.bcaresearch.com GAA Asset Allocation

Highlights Duration: Treasury yields likely have another 50-60 basis points of upside during this cycle, and at least for now their uptrend should not be constrained by unreasonably elevated economic expectations. Stay at below-benchmark duration. Economy & Inflation: GDP growth remains firmly entrenched above levels necessary to ensure that the unemployment rate continues to fall and inflation is pressured higher. Weakness in residential investment presents a risk to the view that above-trend growth will persist, but leading housing indicators suggest it will bounce back in the coming quarters. Municipal Bonds: State & Local government net borrowing declined in the third quarter, but the improvement is already reflected in historically tight Muni / Treasury yield ratios. Remain underweight municipal bonds. Feature Chart 1Discounting An Inflation Rebound In last week's report we pointed out that a flat yield curve is incompatible with core inflation so far below the Fed's target and that the bond market is fast approaching a day of reckoning where either inflation will rise quickly enough to justify the Fed's rate hike expectations, or those expectations will be revised lower.1 Meantime, the Treasury curve has been bear-steepening since early September, and the 37 basis point increase in the 10-year yield has been driven both by higher real yields and a higher cost of inflation compensation (Chart 1). This suggests that the market is pricing-in a rebound in inflation rather than a capitulation from the Fed. Yesterday's PCE inflation report didn't do much to justify those expectations, coming in at only 1.33% year-over-year, not far above the 1.30% reading from August. However, we have previously noted mounting evidence that we are past the worst of the inflation downtrend.2 This raises the question of how much higher Treasury yields can rise, and this is the question we tackle in this week's report. Data Surprises & Playing The Odds Longer run, the 10-year cost of inflation compensation (currently 1.88%) will likely settle into a range between 2.4% and 2.5% by the time that core inflation returns to the Fed's 2% target. Assuming that inflationary pressures are sufficiently strong for that outcome to be achieved with the Fed lifting rates at a pace of about 50 bps per year, then long-dated real yields should stay roughly flat. This means that the nominal 10-year Treasury yield can move another 50-60 bps higher before the end of the cycle. But in the meantime, depending on swings in the macroeconomic data, bonds could experience several playable rallies and sell-offs. Is there a way for us to get a handle on when those might occur? One way might be to examine the economic surprise index (ESI). This index tracks whether economic data are over- or under-shooting consensus expectations. In this way it is very much like a financial market price. It moves higher when the incoming data suggest a rosier outlook than is currently anticipated, but then falls once expectations become so bullish they can no longer be surpassed. This is exactly what happened at the beginning of the year when the 10-year Treasury yield peaked at 2.62% following an extended period of elevated data surprises (Chart 2). Chart 2Economic Surprises Are Mean Reverting More specifically, we observe that when the ESI ends a month above (below) the zero line, it is very likely that the 10-year Treasury yield increased (decreased) during that month (Chart 3). The same is also true for 3-month and 6-month investment horizons, although the correlation is less robust, particularly for values close to zero (Charts 4 & 5). It follows that if we know whether the economic data will surprise on the upside or on the downside in a given month, then we can predict whether Treasury yields will rise or fall. Chart 3Economic Surprise Index & ##br##1-Month Change In Yields Chart 4Economic Surprise Index & ##br##3-Month Change In Yields Chart 5Economic Surprise Index & ##br##6-Month Change In Yields Unfortunately that is not a very profound statement. It is similarly easy to decide how much to bet on a hand of blackjack if you already know what cards will be dealt. But while it is obviously impossible to predict whether data surprises will be positive or negative in a given month, much like a card counter in blackjack, a study of events that have just occurred can help us make inferences that tilt the odds in our favor. In other words, we know that the ESI is mean reverting. A long sequence of elevated readings means it is more likely to fall, and a long sequence of depressed readings means it is more likely to rise. We can even use an AR(3) model to quantify the extent of mean reversion in the index. Using monthly data we run a regression of the ESI on its three most recent lags and get the following result which explains 55% of the variation since 2003: Notice that the index is positively correlated to its reading from the prior month, but negatively correlated with its readings from two and three months ago. Let's now consider that the most recent reading from the ESI is 38.2. One month ago it was -7.9 and two months ago it was -23.1. Using our formula, our best prediction for where the surprise index will be next month is 39. This is still deep in positive territory, meaning that if the model is correct, Treasury yields will remain under upward pressure. More decisively, we conclude from our model that it is unlikely that investor expectations have become so elevated that markets are set up for disappointment. The Appendix to this report provides a reference table for different ranges of the surprise index based on the above formula. It can be used as a quick reference guide for predicting where the ESI is likely to fall next month based on its readings from the prior three months. Bottom Line: Treasury yields likely have another 50-60 basis points of upside during this cycle, and at least for now their uptrend should not be constrained by unreasonably elevated economic expectations. Stay at below-benchmark duration. Economy & Inflation No Signs Of A Slowdown Last week we learned that GDP grew at an annualized rate of 3.0% in the third quarter, well above the Fed's 1.8% estimate of trend. The number was boosted by strong contributions from inventory accumulation (+0.73%) and net exports (+0.41%), but even stripping out those more volatile components to focus on real final sales to domestic purchasers reveals that growth is firmly above trend (Chart 6). Above-trend GDP growth will ensure that the unemployment rate continues to decline, which in turn will ensure that inflation moves higher. The unemployment rate had come close to flattening off late last year as growth decelerated toward 1.8%, but has since started to fall more rapidly alongside the re-acceleration in GDP (Chart 6, bottom panel). In fact, we attribute this year's decline in inflation to last year's growth deceleration and expect inflation will soon follow GDP growth higher (Chart 7). Chart 6Growth Is Steady, And Well Above Trend Chart 7Inflation Lags Growth Considering the contributions from the more stable sources of growth, we observe the following (Chart 8): Consumer spending remains firm, still above its post-2010 average. Nonresidential investment is accelerating back toward its post-2010 average, following a period of weakness that was driven by the mid-2014 commodity price collapse. Leading capex indicators, such as new orders surveys, suggest the acceleration will continue. Residential investment is a source of concern. It had already decelerated to well below its post-2010 average even prior to the hurricanes that depressed its contribution to growth in Q3. We are not yet concerned that the weakness in residential investment will morph into a broader slowdown. In fact, it appears quite likely that residential investment will bounce back in the coming quarters. Growth in residential investment is correlated with changes in the inventory of outstanding homes (Chart 9). Typically, large slowdowns in residential investment are preceded by a big run-up in supply. But at the moment, supply continues to contract, whether or not we include the shadow inventory from properties that were foreclosed upon during the housing bust. This shadow inventory has mostly evaporated in any case (Chart 9, panel 3). Chart 8Housing Not Keeping Pace Chart 9Inventories Still Falling Further support for residential investment comes from homebuilder sentiment which remains very strong (Chart 9, bottom panel). Bottom Line: GDP growth remains firmly entrenched above levels necessary to ensure that the unemployment rate continues to fall and inflation is pressured higher. Weakness in residential investment presents a risk to the view that above-trend growth will persist, but leading housing indicators suggest it will bounce back in the coming quarters. An Improvement In State & Local Government Balance Sheets Assuming that corporate tax revenues were the same in Q3 as in Q2, we can estimate that state & local government net borrowing declined to $163 billion in the third quarter. This represents a substantial improvement from prior quarters, but one that has already been discounted in Municipal / Treasury (M/T) yield ratios (Chart 10). M/T yield ratios are extremely tight, even compared to average pre-crisis levels (Chart 11), and the unattractive valuation underscores our negative stance on the sector. However, at least for now, there are no signs of an imminent surge in state & local government net borrowing that could cause a credit premium to get priced into muni yields. Chart 10Less Borrowing Is In The Price Chart 11Muni / Treasury Yield Ratios Until last quarter, growth in state & local current expenditures had been running consistently above growth in current revenues (Chart 12). However, the weakness in current revenues was mostly attributable to a slowdown in transfers from the federal government. When we look at growth in state & local government tax revenues only, we find that it is substantially outpacing expenditure growth (Chart 12, panel 2). Chart 12Tax Revenue Growth Greater Than Expenditure Growth The acceleration in transfers from the federal government that started in mid-2014 reflected the expansion of Medicaid under the Affordable Care Act. Now that most eligible individuals have signed up, we would expect growth in federal transfer payments to level-off. Unless legislation is passed to further curb transfers from the federal government, state & local borrowing should continue its decline in the coming quarters. Bottom Line: State & Local government net borrowing declined in the third quarter, but the improvement is already reflected in historically tight Muni / Treasury yield ratios. Remain underweight municipal bonds. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com 1 Please see U.S. Bond Strategy Weekly Report, "The Fed Must Fall Behind The Curve", dated October 24, 2017, available at usbs.bcaresearch.com. 2 Please see U.S. Bond Strategy Weekly Report, "Living With The Carry Trade", dated October 17, 2017, available at usbs.bcaresearch.com. Appendix Table 1 Table 2 Table 3 Fixed Income Sector Performance Recommended Portfolio Specification

Highlights Risk assets are responding well to better data and rising rates. Q3 EPS results beating lowered expectations, but growth earnings will peak soon. The conditions are in place for robust capital spending. Financial assets are adhering to the post-Hurricane playbook, with a few notable exceptions. Feature Chart 1Risk Assets Higher Despite Higher Rates Risk assets rose last week for the 6th week in a row (Chart 1). A solid start to Q3 earnings season, more legislative progress on the GOP's tax plan and a narrowing of President Trump's choice for Fed Chair (Jerome Powell, John Taylor and incumbent Janet Yellen) all added to the positive backdrop. The 4 bps rise in the 10 year Treasury yield last week (and 37 bps since early September) was not an impediment to higher equity and oil prices, and gains for small caps and high yield bonds. The positive reaction likely reflected the fact that yields rose more because of increased growth expectations than higher inflation expectations. Despite the impact of Hurricanes Harvey and Irma, Q3 GDP posted an impressive 3% gain. The composition of the Q3 readings suggests an even stronger report in Q4 (Chart 2). At 2.3%, the year-over-year change in real GDP is close to the Fed's 2017 forecast (2.4%) and above the long run forecast (1.8%). The implication for investors is that because U.S. economic growth is faster than its long-term potential, the labor market is tightening and inflation is poised to move higher. Accordingly, market odds for a Fed hike in December are near 90% and participants expect 51 bps more hikes in the next 12 months (Chart 1, panel 3). BCA's view is that U.S. economic growth is set to accelerate in the coming quarters aided by a post hurricane rebound in housing. The Fed will raise rates in December and three more times next year as inflation returns to 2% and perhaps beyond. Corporate profit growth will peak in the next few quarters, but remain supportive of higher stock prices for now. The rise in the Economic Surprise Index will continue for another few months, and provide another lift for risk assets. A surge in capital spending adds to the upbeat tone. Chart 2GDP Growth Remains Below Average, But Above Fed's Long Run Target Capital Spending Blasts Off Business capital spending is on the upswing. The robust readings in September on core durable goods orders (7.8% year-over-year) and shipments reported last week were paybacks for the Hurricane-weakened August report. Nonetheless, the impressive soundings on the three -month change in both orders and shipments were not distorted by the storms. Moreover, the durable goods report was one of the latest in a series of data points brightening capex's outlook (Chart 3). Both BCA's real and nominal capex models, driven by surging capital goods orders along with elevated ISM readings and soaring sentiment on business spending, indicate strong investment in plant and equipment in the next few quarters. CEO confidence soared to a 13-year high in Q1 according to the latest Duke University/CFO Magazine Business Outlook, but retreated modestly in Q2 and Q3 (Chart 4). Surveys by the Conference Board and Business Roundtable show a similar pattern. Notably, readings on all three surveys have climbed since Trump's election in November 2016, but then retreated as his pro-business agenda stalled. The drop in sentiment reflects the lack of legislative progress in Washington (Chart 5). The dip in CEO sentiment in Q2 and Q3 is in sharp contrast with the easing of policy concerns in the Beige Book. Chart 3Bright Outlook For Capital Spending Chart 4Capital Spending Plans Upbeat Chart 5Managements Remain Upbeat The upbeat numbers in the regional Federal Reserve Banks' surveys of capital spending intentions further support rising capex spending in the next few quarters. The average readings from the New York, Philadelphia and Richmond Feds' capex survey plans are close to cycle highs, despite a modest pullback in the summer months. Moreover, the regional Feds' capex spending plans diffusion index hit an eight-year high in October (Chart 5, panel 3). Bottom Line: Stay overweight stocks versus bonds, and underweight duration. Rising capex will drive up GDP, employment and EPS in the coming quarters. Q3 Earnings Beating Lowered Expectations The Q3 earnings reporting season is off to a strong start, with both EPS and sales growth well ahead of consensus expectations as we forecast in our October 2 preview. Moreover, the counter-trend rally in profit margins is still in place. Just under 55% of companies have reported results so far, with 74% beating consensus EPS projections just above the long-term average of 55%. Furthermore, 67% have posted Q3 revenues that topped expectations, which exceeded the LT average of 69%. The surprise factor for Q3 stands at 5% for EPS and 2% for sales. These compare favorably with the average EPS (4.2%) and sales (1.2%) in the past five years. We anticipate the secular mean-reversion of margins to re-assert itself in the S&P data, perhaps beginning early in 2018. Nonetheless, initial results imply that Q2 will be another quarter of margin expansion. Average earnings growth (Q3 2017 versus Q3 2016) is solid at 7% with revenue growth at 5%. Strength in earnings and revenues is broad based (Table 1). Earnings per share increased in Q3 2017 versus Q3 2016 in eight of the 11 sectors. The 7.3% year-over-year drop in the financial sector is linked to the impact of the hurricanes on the insurance and reinsurance industries. Excluding those industries, financial EPS is up 4.7% from a year ago. EPS results are particularly stout in energy (164%), technology (18%) and healthcare (7%). Those sectors likewise experienced significant sales gains (16%, 9% and 5% respectively). Corporate managements are more focused on the message in Washington than on the President (Chart 6). Trump's name was mentioned just once in the Q3 earnings calls held through October 27, matching Q2's reporting period. CEOs and CFOs have cited Trump's name at least once in each earnings season since Q2 2016. The peak in mentions occurred immediately after Trump took office in early 2017. Table 1S&P 500:##BR##Q3 2017 Results* Chart 6Managements Focused On##BR##The Message Out Of DC In contrast, the words "tax" and "reform" have appeared 39 times thus far in Q3 conference calls, most often in a positive light. There were only five mentions in Q2, when there was skepticism that a tax plan would pass this year. In the Q4 2016 reporting season following the November election, tax and reform were cited 16 times. BCA's Geopolitical Strategy service has consistently expected a tax package to pass by the end of Q1 2018.1 We are encouraged by the upward trajectory of EPS estimates for 2017 and 2018 (Chart 7). It is odd that the recent downtick in 2017 EPS is mirrored by an uptick in the 2018 figure. That said, the divergence can be explained by the impact of the hurricanes on the financial sector's earnings in 2017 and probable snapback in early 2018. Analysts expect 2019 EPS growth to slow from 2018's clip, which matches BCA's view. However, unlike estimates for 2017 and 2018, we anticipate that EPS estimates for 2019 will move lower throughout 2018 and 2019, ahead of a recession in late 2019.2 Bottom Line: The BCA earnings model shows that S&P 500 EPS growth is peaking and should decelerate through 2018 toward a level commensurate with 3 ½-4% nominal GDP growth (Chart 8). Accordingly, BCA believes that the earnings backdrop will remain a tailwind for the equity market, albeit a smaller tailwind. This forecast excludes any positive effect on growth from tax cuts, which would be positive for EPS and the S&P 500 price index in the short term, although this would also bring forward Fed rate hikes. The entire Treasury curve has readjusted to reflect this view. Chart 7Stability In '17 & '18 EPS Estimates,##BR##But '19 Likely To Move Lower Chart 8Strong EPS Growth Ahead,##BR##Will Start To Slow Soon 10-Year Treasury Update BCA's view is that the 10-year Treasury yield will head higher in the coming months. However, is the move from 2.03% in early September to 2.43% last week sustainable? BCA's fair value model for the 10-year Treasury yield (based on Global PMI and dollar sentiment) places fair value at 2.65% (Chart 9, panel 1). Moreover, BCA's three-factor version of the model (that includes the Global Economic Policy Uncertainty Index), puts fair value slightly higher at 2.63% (Chart 9, panel 3). Investors should continue to position for a steeper curve by favoring the 5-year bullet versus a duration-matched 2/10 barbell. Chart 9Treasury Fair Value Models BCA's U.S. Bond Strategy service will publish updated fair models after the November 1 release of October's global PMI data. The latest readings on Citi's Economic Surprise index also support BCA's stance on rates. How Long Can The Economic Surprise Index Stay Positive? The Citi Economic Surprise Index crossed into positive territory on October 2nd, remaining above zero for 20 business days, and risk assets are responding (Chart 10). Since 2010, once the Index turns positive, it continues to rise for 46 days. The implication for investors is that the economic data will continue to be remarkable for another two months. Table 2 shows that risk assets outperform as the economic surprise index rises from zero toward its zenith. Risk assets have also outperformed since the June bottom in economic surprises, matching the historical performance.3 Oil (+17%), small caps and investment grade corporates are all standouts and the gains may not be over. The track record of risk assets as the Economic Surprise Index climbs suggests that additional increases are in prospect for risk assets. On average, equities (relative to treasuries) and oil are the best performers during these intervals. Chart 10May Still Be Room To Run On Economic Surprise Table 2Risk Assets Perform Well As Economic Surprise Rises Post-Hurricane Macro Backdrop The strength of the Citi Economic Surprise Index following the hurricanes duplicates the historical trend and supports the rise in risk assets. The Index moves higher for the first month post-storm, and then remains above zero for an additional three weeks (Chart 11, panel 4). This bolsters BCA's stance that the direction of the Index will continue to lift risk assets in the next few months. Financial assets are also adhering to the post-Hurricane playbook,4 with a few notable exceptions (Chart 12). The stock-to-bond ratio moved higher and the VIX has declined since Hurricane Harvey, matching the typical post-storm performance. However, the 10-year Treasury yield, the S&P 500 and the Fed funds rate, all have bucked historical trends. The S&P 500 rose by 5.6% since late August; stocks typically drift lower in the first few months after a major storm. In addition, the 10-year Treasury yield climbed but it usually moves down in the two months following a hurricane. Post- storm, the Fed typically continues to do whatever it was doing prior to the storm. Accordingly, we expect the Fed to hike rates at its December meeting. Chart 11Major Hurricane Impact##BR##On Activity Data Chart 12Major Hurricane Impact On##BR##Financial Markets And The Fed The economic, inflation and sentiment data are also mixed. Housing data frequently lags in the wake of a storm, but both new and existing home sales moved up in the month after Harvey and Irma; housing starts declined in recent months which is counter to the historical pattern (Chart 13). Both IP and employment plunged after the storms, however, these indicators tend to rise after major weather. Initial claims for unemployment insurance were typically volatile in the six weeks since Harvey hit Texas, but have resumed their downtrend. Average hourly earnings in inflation climbed after Harvey and Irma, while consumer confidence dipped, matching history. However, the bump in gasoline prices since late August runs counter to historical precedent. Gasoline prices tend to decline after major storms (Chart 14). Chart 13Major Hurricane Impact##BR##On Housing Data Chart 14Major Hurricane Impact On##BR##Sentiment And Inflation Data Investment Conclusions: The macro backdrop remains bullish for risk assets, especially since synchronized growth has reduced fears of secular stagnation. Bond yields will rise, but won't be a headwind for stocks yet.5 Rising bond yields because of growth, without rising inflation, are bullish for risk assets, but this will change as inflation reaches 2% and inflation expectations start to rise. At that point, the Fed will be behind the curve. This will lead to faster Fed rate hikes, historically a headwind for equities. John Canally, CFA, Senior Vice President U.S. Investment Strategy johnc@bcaresearch.com 1 Please see BCA's Geopolitical Strategy Weekly Report, "Xi Jinping: Chairman Of Everything," October 25, 2017. Available at gps.bcaresearch.com. 2 Please see BCA's Global Investment Strategy Weekly Report, "Strategy Outlook Fourth Quarter 2017: Goldilocks And The Recession Bear," October 4, 2017. Available at gis.bcaresearch.com. 3 Please see BCA's U.S. Investment Strategy Weekly Report, "Global Monetary Policy Recalibration," April 17, 2017. Available at usis.bcaresearch.com. 4 Please see BCA's U.S. Investment Strategy Weekly Report, "Shelter From The Storm," September 5, 2017. Available at usis.bcaresearch.com. 5 Please see BCA's U.S. Investment Strategy Weekly Report, "Still In The Sweet Spot" June 19, 2017. Available at usis.bcaresearch.com.