Yield Curve

Highlights Editor's Note: I am pleased to return to U.S. Investment Strategy (USIS). I worked with the service when I joined BCA in 2010, and previously led it from August 2013 through September 2014. Sara Porrello, who has been with the team for over 20 years, and I look forward to re-aligning USIS with its original mandate. We hope you will find it consistently insightful. Best regards, Doug Peta U.S. Investment Strategy is getting back to basics: Today's report, plainly stating our position on the near-term direction of interest rates, is the first in an ongoing series meant to stake out our views on the macro issues that are most important to investors. Rates are headed higher, consistent with a booming economy that may well overheat, ... : Assuming trade tensions don't short-circuit the expansion, the U.S. economy is poised to grow above trend well into 2019. ...thanks to a tightening labor market and dubious fiscal spending, ... : Employers will be forced to bid up wages as the pool of idled and under-utilized workers dries up, and the fiscal stimulus package is all but certain to goose inflation pressures. ... and neither tweets nor testy interviews nor other expressions of presidential pique are likely to stay the Fed from its appointed rounds: The Federal Reserve cherishes its independence, and it is extremely unlikely to bow to presidential pressure. Feature U.S. Investment Strategy is meant to provide analyses of the U.S. economy and its future direction for the purpose of helping our clients make asset-allocation decisions. Starting with this report, we are going back to the basics of meeting that mandate. Over the rest of the summer, we intend to outline our positions on the key macro drivers of financial markets: rates, credit, the business cycle, and the state of monetary policy. Laying out our big-picture views, and the rationale underpinning them, will establish a framework for evaluating incoming data. The goal is to allow our clients to think along with us as new information is disseminated, and to distinguish signals from noise. We also want to make it easier for clients to anticipate the evolution of our views. To that end, will make frequent use of checklists highlighting the specific elements that might lead us to change our take on the evolution of the key cycles. The ultimate goal is to stay on top of cyclical inflection points, and to use them to inform asset-allocation decisions. The Fed Gets Its Way On Rates Monetary policy is a blunt instrument that works with indeterminate lags, and its effect has been roundly questioned. At the ends of the armchair-quarterback continuum, the Fed is mocked as a clueless bumbler, turning dials at random like a fumbling Mr. Magoo, or bemoaned as an omnipotent manipulator of financial markets and real-world activity. Strictly speaking, it controls nothing more than short rates. As its post-crisis communications strategy has shown, however, its reach extends well beyond its official policy-rate dominion. Talk of last decade's "conundrum" aside, changes in the fed funds rate reverberate along the entire yield curve. As the Chart Of The Week demonstrates, the aggregate yield on all outstanding Treasury issues is joined at the hip, directionally, with the fed funds rate. Aggregate weighted-average Treasury duration sits squarely in the belly of the curve, and it is a not-quite-perfect proxy for the long end, where the Fed's gravitational pull wanes (Table 1). Its pull is still powerful, though; the 90% correlation between the fed funds rate and the 30-year bond testifies eloquently to the Fed's significant influence at all points of the curve (Chart 2). Chart of the WeekThe Fed Gets Its Way The Fed Gets Its Way The Fed Gets Its Way The investment takeaway is that the Fed gets what it wants across the full spectrum of rates the vast majority of the time. Given the FOMC's repeatedly expressed intention to continue on its normalization course, the path of least resistance for rates at all maturities is higher. Despite the money markets' resistance to extrapolate the 25-bps-a-quarter "gradual pace" across the rest of this year and next (Chart 3), six more quarters of that pace is our baseline expectation provided an economic shock does not occur. Investors should be prepared for a higher peak in the fed funds rate than the consensus expects. Table 1Correlation With The Fed Funds##BR##Rate By Bond Maturity The Rates Outlook The Rates Outlook Chart 2The Long Arm##BR##Of The Fed The Rates Outlook The Rates Outlook Chart 3Rates Have Room To##BR##Surprise To The Upside Rates Have Room To Surprise To The Upside Rates Have Room To Surprise To The Upside Bottom Line: The Treasury curve faithfully reflects changes in the fed funds rate. In the absence of a shock that would cause the FOMC's repeatedly expressed plans to change, monetary policy is a catalyst for higher rates. But What About An Inverted Yield Curve? The yield curve typically inverts in the latter stages of a rate-hiking campaign, so it is more correct to say a higher fed funds rate implies higher Treasury yields until the yield curve inverts. An inverted yield curve is a classic recession indicator, albeit often a very early one (Table 2), and it should not be taken as a signal to immediately de-risk portfolios. The yield curve may be prone to invert even earlier than it otherwise would this time around, given that QE1, QE2, and QE3 may well have depressed the term premium on long-term bonds,1 as The Bank Credit Analyst noted in its August edition. The question of how much the Fed's asset purchases have affected the term premium, if at all, is far from settled within either the Fed or BCA, but its potential to impact the signal from the yield curve reinforces our conviction to look to other indicators to confirm its recession message before declaring the end of the bull markets in equities and spread product. Table 2The Yield Curve Is Early The Rates Outlook The Rates Outlook The Inflation Outlook As the tepid post-crisis expansion has stretched on and on, investors have grown accustomed to sleepy inflation readings and begun to regard the prospects for a pickup in inflation with skepticism, if not outright disdain. Even within BCA, there has been spirited debate about the relevance of the Phillips Curve - the formalization of the idea that there is an inverse relationship between wage growth and the unemployment rate. Despite the stagflation of the 1970s and the lengthy post-crisis dry spell that have undermined the Phillips Curve's credibility with the rigorously empirically-minded, we do not find it controversial. The relationship between unemployment and compensation may not be perfectly linear, but the Phillips Curve is nothing more than an extension of the laws of supply and demand to wage negotiations. We can accept that the Phillips Curve is kinked - that compensation growth is utterly indifferent to changes in the unemployment rate when labor supply is glutted (as can be seen in Chart 4 when covering all of the observations below 7%), but rather sensitive to its moves when it is in the neighborhood of full employment (as can be seen when covering all of the observations above 5%). We believe the U.S. labor market has reached the point at which employers will have to compete fiercely to attract new talent. After nine years, the economy has finally worked down nearly all of the hidden slack that had padded the broader U-6 unemployment rate.2 The pool of discouraged workers - those who are not counted as officially unemployed because they're not actively looking for a job, but would start tomorrow if offered one - has shrunk below its 2000 and 2007 levels (Chart 5, top panel). Similarly, the share of the labor force that is working part time but would prefer to be working full time is approaching its pre-crisis bottom (Chart 5, bottom panel). The prospects for inflation gained another boost last December upon the passage of the spending package on the coattails of the tax-cut bill. The U.S. economy is poised to receive a substantial dose of fiscal stimulus this year and next (Chart 6). Mainstream macroeconomic thought holds that stimulus injected into an economy that is already operating at full capacity is prone to kindle inflation.3 Chart 4The Phillips Curve Can't Handle Copious Slack ... The Rates Outlook The Rates Outlook Chart 5... But Almost All Of It Has Been Worked Off ... But Almost All Of It Has Been Worked Off ... But Almost All Of It Has Been Worked Off Chart 6Goosing Inflation Along With Output Goosing Inflation Along With Output Goosing Inflation Along With Output Bottom Line: The U.S. labor market has tightened considerably and competition between employers to attract scarce talent should soon translate to a pickup in wage growth. Unneeded fiscal stimulus is also likely to push prices higher. There are plenty more inflation green shoots behind the ones that have already begun to emerge. White House-Fed Tension Is Nothing New It is not beyond the realm of possibility that presidential pressure could deter the Fed from following through on its intentions and present a risk to our above-consensus terminal rate estimate. The bond market immediately discounted the potential of a less independent Fed by selling off at the long end after the president stated he was "not thrilled" with ongoing rate hikes in an interview with CNBC. There would seem to be little doubt that a captive Fed would be more reluctant to remove the punch bowl than a Fed which was free to pursue its inflation mandate without outside interference. After all, elected officials would be happy to trade long-term pain for near-term gain (at least through the next campaign). The president may have upended convention by publicly airing his displeasure, but there is a natural tension between the White House and the Fed. There have been dust-ups in the past, and there will be dust-ups in the future for as long as elected officials shudder at the thought of an economic downturn. Alan Greenspan wrote frankly in his memoir about friction with the first Bush administration, which included public criticism from the sitting president. "I do not want to see us move so strongly against inflation that we impede growth," President Bush told the press at the beginning of his term, in response to hawkish congressional testimony from Greenspan.4 By all accounts, however, the conflict between Bush père and Greenspan was of a lower-pressure variety than the conflicts between LBJ and William McChesney Martin, and Nixon and Arthur Burns. The legendarily intimidating LBJ summoned Martin to his ranch following an unwelcome rate hike. According to several accounts (and consistent with his longstanding negotiating practices in the Senate), LBJ backed the smaller Martin up against a wall before giving full voice to his complaints. Martin did not budge, pointing out that the Fed had acted in accordance with the legislation governing its actions.5 If Martin represents the heroic Fed chief, standing his ground in the face of heavy pressure from a larger-than-life figure, Arthur Burns is the poster child for folding like a cheap lawn chair. The Nixon tapes capture Nixon and his proxies repeatedly pressuring Burns to prime the pump ahead of the 1972 election, which Burns ultimately did.6 Our view is that Fed Chair Powell is more likely to follow Martin than Burns. The Fed is more transparent today, and its independence is more firmly established than it was in the 1970s. Even if Powell were amenable to doing the president's bidding, he would be held back by the realization that it would ultimately be self-defeating: any hint of political manipulation in the rate-setting process would risk a bond market riot that would blast rates far beyond the levels where a 3.5% fed funds rate would take them. Bottom Line: We are not concerned that the FOMC will yield to pressure from the White House to back away from their rate hike plans. Attempted influence of the Fed is nothing new, and investors need not worry about it now. Investment Implications If we are correct in our view that rates have not yet peaked, the bond market is likely to face continued headwinds. Long-dated Treasuries will come under more pressure than shorter-maturity issues. Thanks to positive carry, spread product will be less vulnerable to higher rates, but our bond strategists are lukewarm on the risk-reward offered by investment-grade and high-yield bonds given the late stage of the cycle and historically tight spreads. We acknowledge the potential seriousness of the current spate of geopolitical risks, headlined by trade tensions, and advocate temporarily de-risking portfolios in line with the BCA house view (equal weight equities, underweight bonds, overweight cash). We are more constructive than the BCA consensus, however, because we remain constructive on the business cycle, the monetary policy cycle, and the credit cycle. If the key cycles aren't over, the equity bull market probably isn't over, and neither spread widening nor a pickup in defaults is likely to wipe out spread product's excess returns. We will express all of our calls in a basket of ETF recommendations once we have completed our review of the most impactful macro questions, but for now we recommend maintaining below-benchmark positioning in Treasury portfolios while overweighting TIPS at the expense of nominal Treasuries. Doug Peta, Senior Vice President U.S. Investment Strategy dougp@bcaresearch.com 1 Long-term bond yields can be decomposed into the expected path of short-term rates and a term premium, which compensates an investor for the uncertainties that can arise over the extended time period that s/he is locking up his/her money by buying a longer-maturity instrument. 2 In the monthly employment report, the headline unemployment rate, which includes only jobless workers who are actively seeking work, is labeled U-3 unemployment. The U-6 series broadens the definition of unemployment to include the jobless who aren't actively searching and those who are working part time only because they cannot find a full-time position. 3 Please see the November 7, 2016 U.S. Investment Strategy Weekly Report, "Policy, Polls, Probability," available at usis.bcaresearch.com, for a discussion of fiscal multipliers under a range of scenarios. 4 Greenspan, Alan. The Age of Turbulence: Adventures in a New World, Penguin (New York): 2007, p.113. To this day, several members of the G.H.W. Bush administration continue to pin a large measure of blame for its 1992 electoral loss on overly conservative monetary policy. The ex-president himself, in a 1998 television interview, said, "I reappointed him [Greenspan], and he disappointed me." 5 Granville, Kevin. "A President at War With His Fed Chief, 5 Decades Before Trump," New York Times, June 15, 2017, page B3 (updated July 19, 2018). https://www.nytimes.com/2017/06/13/business/economy/a-president-at-war-with-his-fed-chief-5-decades-before-trump.html 6 "How Richard Nixon Pressured Arthur Burns: Evidence from the Nixon Tapes, Vol. 20, No. 4," Journal of Economic Perspectives (Fall 2006). https://fraser.stlouisfed.org/title/1167/item/2388, accessed on July 24, 2018.

Highlights President Trump has taken the next step in the trade war by charging some of America's major trading partners with outright currency manipulation. However, we are not headed for Plaza Accord 2.0, because neither the ECB nor the PBOC will re-orient policy until their own economic and inflation dynamics warrant it. Moreover, we doubt the FOMC will be bullied into keeping rates lower than policymakers deem appropriate. With the labor market showing signs of overheating, the Fed will stick with its current game plan and ignore President Trump's tweets. The worsening trade dispute is the key risk that investors face and there are growing signs that uncertainty regarding the future of the world trade order is dampening animal spirits and global capital spending. Risk tolerance should be no more than benchmark. Based on previous late cycle periods, the fact that S&P 500 profit margins are still rising suggests that most risk assets will outperform bonds and other defensive sectors in the near term. Nonetheless, timing is always difficult and we have decided to focus on capital preservation given extended valuations and a raft of risks that could cause a premature end to the bull market. The flattening U.S. yield curve is also worrying. We would not ignore the signal if the curve inverts, although there are reasons to believe that it is not as good a recession signal as it has been in the past. We wish to see corroborating evidence from our other favorite indicators before trimming risk asset exposure to underweight. A peak in the S&P 500 operating margin would be a strong sign that the end of the cycle is drawing close. Even if trade tensions soon die down and global growth holds up, the extended nature of the U.S. economic and profit cycle make asset allocation particularly tricky. Attractive late-cycle assets to hold include structured product, Timberland and Farmland. High-quality bonds will of course outperform in the next recession, but yields are likely to rise in the meantime. We believe that U.S. Agency MBS are unattractively valued, but should remain insulated from negative shocks such as a trade war or higher Treasury yields. We also like Agency CMBS. Oil and related plays are not a reliable late-cycle play, but we are bullish because of the favorable supply-demand outlook. However, this does not carry over to base metals, where we are more cautious. Feature We warned in last month's Overview that investors had not yet seen "peak pessimism" on the global trade front. Right on cue, President Trump raised the stakes again in July by threatening to impose tariffs on virtually all imports of Chinese goods. Congress is pushing the President to be tough on China because American voters have soured on trade. China will not easily back down with the authorities responding in kind to the U.S. President's trade threats. They have also allowed the RMB to depreciate to cushion the trade blow (Chart I-1). It is not clear whether the authorities purposely depressed the RMB or whether they simply failed to lean against market pressures. Either way, it is a dangerous approach because it has clearly raised the U.S. President's ire. Chart I-1RMB Is Much Weaker Across The Board RMB Is Much Weaker Across The Board RMB Is Much Weaker Across The Board President Trump has taken the next step in the broader trade war by charging some major trading partners with outright currency manipulation. The script appears to be following previous times that the U.S. sought trade adjustment via tariffs and currency re-alignment: the early 1970s and the 1985 Plaza Accord. Adjusting currencies on a sustained basis requires much more than simply "talking down" the dollar. There must be major changes in relative monetary and/or fiscal policies vis-à-vis U.S. trading partners. On the fiscal front, expansionary U.S. policy is working at cross purposes with the desire to have a weaker dollar and a smaller trade gap. We do not foresee the U.S. President having any success in changing the broad thrust of monetary policy either. Europe and Japan enjoyed booming economies in the early 1970s and mid-1980s, and thus had the luxury of placating the U.S. by adjusting monetary policy and thereby appreciating their currencies. Today, it is difficult to see how either Europe or China can afford significant monetary policy tightening that generates major bull markets in their currencies. Neither the ECB nor the People's Bank of China (PBOC) will re-orient policy until their own economic and inflation dynamics warrant it.1 It is also unlikely that the Bank of Japan will raise the 10-year yield target to either strengthen the yen or to help bank profits. This is not Plaza Accord 2.0. Powell Isn't Arthur Burns As for the Fed, we doubt the FOMC will be bullied into keeping rates lower than policymakers deem appropriate. The Fed is more open and independent today than in the 1970s and 1980s. Even if Fed Chair Powell were amenable, any hint that he is being politically manipulated to change course would result in a bond market riot that would rattle investors to their core. More likely, the Fed will stick with its current game plan and ignore President Trump's tweets. Powell could not be any clearer in his July Congressional Testimony: "With a strong job market, inflation close to our objective, and the risks to the outlook roughly balanced, the FOMC believes that-for now-the best way forward is to keep gradually raising the federal funds rate." Investors should not be fooled by the uptick in the U.S. unemployment rate in June. The rise reflected a pop in the labor force participation rate. However, the labor force figures are volatile and there is no upward trend evident in the participation rate. The real story is that the labor market continues to tighten. The number of people outside the labor force who want a job, as a percentage of the total working-age population, is back to pre-recession lows. The Employment Cost Index for private-sector workers shows that wage growth is accelerating. Moreover, the New York Fed's Underlying Inflation Gauge, which leads core CPI inflation by 18 months, has already jumped to almost 3 ½% (Chart I-2). Small businesses are increasingly able to pass on cost increases to consumers (Chart I-3). Chart I-2U.S. Inflation Is Percolating U.S. Inflation Is Percolating U.S. Inflation Is Percolating Chart I-3U.S. Pricing Power On The Rise U.S. Pricing Power On The Rise U.S. Pricing Power On The Rise The Minutes from the mid-June FOMC meeting included a lengthy discussion of the growing signs of inflation pressure and labor shortage. Firms are responding to the lack of qualified labor by offering training, automating, and boosting wages. Anecdotal evidence suggests that bottlenecks and other cost pressures are boiling over in the transportation sector. Despite an acute shortage of truck drivers, the average hourly earnings data do not show any acceleration in their wages (Chart I-4, second panel). However, these data do not include bonuses, which have been on the rise. The PPI for truck transportation services was up 7.7% year-over-year in June, while the Cass Freight Index that tracks full-truckload prices rose 15.9% year-over-year. The latter does not even include fuel costs. These pipeline cost pressures have implications not only for the Fed, but for corporate profit margins as well (see below). Chart I-4U.S. Transportation Is Boiling Over U.S. Transportation Is Boiling Over U.S. Transportation Is Boiling Over The U.S. Yield Curve: A Red Flag? The FOMC expects that the fed funds rate will continue to rise and will temporarily exceed its 2.9% estimate of the neutral rate. If the true neutral rate is higher than the Fed's estimate, then the FOMC could find itself hiking too slowly and the economy could severely overheat. And vice versa if the true neutral rate is below 2.9%. We are keeping a close eye on the yield curve as an indication of policy tightness. If the curve inverts with a few more Fed rate hikes, it would signal that the market believes that policy is turning restrictive. It is possible that the yield curve is not as good a recession signal as it has been in the past. First, there is a lot of uncertainty regarding the neutral fed funds rate in the post-GFC world. The collective market wisdom on this could be wrong. Indeed, BCA's Chief Global Strategist, Peter Berezin, makes the case that the neutral rate is rising faster than most investors believe.2 Structural factors have depressed the neutral rate, including population aging and low productivity growth. However, these structural tailwinds for bond prices are now slowly turning into headwinds. Moreover, as Peter argues, cyclical pressures are acting to lift the neutral rate. Private credit growth is rising faster than nominal GDP growth again. The same is true for housing and equity wealth, at a time when the personal saving rate is falling. All this implies strong desired spending which, in turn, suggests a higher neutral rate of interest. It will be important to watch the housing market; if it remains healthy in the face of rate hikes, it means that the neutral rate is still north of the actual fed funds rate. Chart I-5 presents today's market expectation for the real fed funds rate, based on the forward OIS curve and the forward CPI swaps curve. Technical issues may be distorting forward rates in 2019, but we are more interested in expectations further into the future. The real fed funds rate is expected to hover in the 55-75 basis point range until 2024. It then rises to about 1%, but not until almost the end of the next decade. This appears overly complacent to us, suggesting that the risks are to the upside for market expectations of the terminal, or neutral, short-term interest rate. If the neutral rate is indeed higher than the market is currently discounting, then an inverted curve may be premature in signaling that policy is too tight and that an economic slowdown is on the horizon. Moreover, the term premium on long-term bonds may still be depressed by asset purchases by the Fed and the other major central banks, again suggesting that the curve will more easily invert than in the past. There is much disagreement on this issue, even among FOMC members and among BCA strategists. This publication is sympathetic to the work done by the Fed Staff which suggests that the term premium has been substantially depressed by quantitative easing. Chart I-6 shows the annual change in the size of G4 central bank balance sheets (inverted), along with an estimate of the term premium in the 10-year government bonds of the major countries. The chart is far from conclusive, but it is consistent with the view that QE has depressed term premia worldwide. Moreover, forward guidance and the low level of inflation since the GFC have undoubtedly dampened interest-rate volatility, which theory suggests is a key driver of the term premium. Chart I-5Policy Rate Expectations Policy Rate Expectations Policy Rate Expectations Chart I-6Depressed Term Premiums ##br##Distort Yield Curves Depressed Term Premiums Distort Yield Curves Depressed Term Premiums Distort Yield Curves The factors that have depressed the term premium are beginning to reverse, including G4 central bank balance sheets. Still, the premium will trend higher from a low starting point, suggesting that an inverted curve today may not necessarily signal a recession. That said, it would be wrong to completely dismiss a U.S. curve inversion, given its excellent track record. Historically, the 3-month/10-year Treasury slope has worked better than the 2/10 yield slope in terms of calling recessions. An inversion of the 3-month/10-year curve has successfully heralded all seven recessions in the past 50 years with one false positive signal. Nonetheless, the curve tends to be very early, inverting an average of almost 12 months before the recession. And, given the possible distortion to the term premium, we would want to see corroborating evidence before jumping to the conclusion that an inverted curve is sending a correct recession signal. For example, the U.S. and/or global Leading Economic Indicator would need to turn negative. The bottom line is that a curve inversion would not be enough on its own to further trim risk asset exposure to underweight. Nonetheless, we are not dismissing the message from the yield curve either, especially in the context of a trade war that could prematurely end the expansion. Trade War Hitting Economy? Estimates based on macro models suggest that the damage to global GDP growth from higher tariffs would be quite small. Nonetheless, these models do not incorporate the indirect, or second-round, effects of rising tariff walls. Business leaders abhor uncertainty, and will no doubt hold off on major capital expenditure plans until the trade dust settles. The uncertainty can then ripple through the economy to industries that are not directly affected by the trade action. The extensive use of global supply chains reinforces this ripple effect. Labor is not free to move between countries or between industries to facilitate shifts in production that are required by changing tariffs. Capital is more mobile, but it is still expensive to shift machinery. Some of the world's capital stock could become "stranded", raising the cost of the tariffs to the world economy. Finally, important economies-of-scale are lost when firms no longer have access to a single large global market. This month's Special Report, beginning on page 18, sorts out the U.S. equity sector winners and the losers from a trade war with China. Spoiler alert: there are not many winners! The bottom line is that the trade threat for the global economy and risk assets is far from trivial. The negative trade headlines have not had a meaningful economic impact so far, but there are some worrying signs. A number of indicators suggest that global growth continues to slow, including the BCA Global Leading Economic Indicator diffusion index, the Global ZEW sentiment index and the BCA Global Credit Impulse index (Chart I-7). The softness in these indicators predates the latest flaring of trade tensions. Nonetheless, business confidence outside the U.S. has dipped (fourth panel). Growth in capital goods imports for an aggregate of 20 countries continues to decelerate, along with industrial production for capital goods and machinery & electrical equipment in the major advanced economies (production related to energy, consumer products and IT remain strong; Chart I-8). Chart I-7Global Growth Is Still Moderating... Global Growth Is Still Moderating... Global Growth Is Still Moderating... Chart I-8...In Part Due To Capital Spending ...In Part Due To Capital Spending ...In Part Due To Capital Spending None of these data are flagging a disaster, but they all support the view that uncertainty regarding the future of the world trade order is dampening animal spirits and global capital spending. Even if trade tensions soon die down, the extended nature of the U.S. economic and profit cycle make asset allocation particularly tricky. Late Cycle Investing Some of our economic and policy analysis over the past year has focused on previous late-cycle periods. Specifically, we analyzed the growth, inflation and policy dynamics after the point when the economy reached full employment (i.e. when the unemployment rate fell below the CBO estimate of full employment). This month we look at asset class returns during late cycle periods. We wanted to use as broad a range of asset classes as possible, although data limitations mean that we can only analyze the late-cycle periods at the end of the 1990s and the mid-2000s (Chart I-9). To refine the analysis, we split the late-cycle periods into two parts: before and after S&P 500 profit margins peak. One could use other signposts to split the period, such as a peak in the ISM manufacturing index. However, using the S&P operating profit margin proved to be a more useful break point across the cycles in terms of timing trend changes in risk assets. Table I-1 presents total returns for the following periods: (1) the full late-cycle period - i.e. from the point at which full employment is reached until the following recession; (2) from the point of full employment to the peak in the S&P margin; (3) from the peak in the margin to the recession; and (4) during the subsequent recession. All returns are annualized for comparison purposes, and the data shown are the average of the late 1990s and mid-2000 late-cycle periods. Chart I-9Margin Peak Signals Very Late Cycle Margin Peak Signals Very Late Cycle Margin Peak Signals Very Late Cycle Table I-1Late-Cycle Asset Returns August 2018 August 2018 We must be careful in interpreting the results because no two cycles are exactly the same, and we only have two cycles in our sample of data. Nonetheless, we make the following observations: Treasury bond returns are positive across the board, which seems odd at first glance. However, in both cases the selloff occurred before the late-cycle period began. Yields then fluctuated in a range, and then began to fall after margins peaked. Global factors also contributed to Greenspan's "conundrum" of stable bond yields in the years before the Great Recession. We do not expect a replay this time around given the low starting point for real yields and the fact that the Fed is encouraging an overshoot of the inflation target. Bonds are unlikely to provide positive returns on a six month horizon. Similar to Treasurys, investment-grade (IG) corporate bond returns were positive across the board for the same reason. However, IG underperformed Treasurys after margins peaked and into the recession. High-yield bonds followed a similar pattern, but suffered negative absolute returns after margins peaked. U.S. stocks began to sniff out the next recession after margins peaked. Small caps outperformed large caps in the recessions, but relative performance was mixed after margins peaked. We are avoiding small caps at the moment based on poor fundamentals and valuations. Growth stocks had a mixed performance versus value stocks before and after margins peaked, but tended to outperform in the recessions. Dividend Aristocrats performed well relative to the overall equity market after margins peaked and into the recessions on average, but the performance was not consistent across the two late cycles. EM stocks performed well before margins peak, and poorly during the recessions. However, the performance is mixed in the period between the margin peak and the recession. We recommend an underweight allocation because of poor macro fundamentals and tightening financial conditions. In theory, Hedge Funds are supposed to be able to perform well in any environment, but returns were a mixed bag after margins peaked. The return performance of Private Equity, Venture Capital and Distressed Debt were similar to the S&P 500, albeit with more volatility. Avoid them after margins peak. Structured Product is one of the few categories that performed well across all periods and cycles. The index we used includes MBS, CMBS and ABS. Farmland and Timberland returns were attractive across all periods and cycles, except for Timberland during one of the recessions. Oil and non-oil commodities tended to perform poorly during recession, but returns were inconsistent in the other phases shown in the table. Gold was also a mixed bag. The historical return analysis underscores that it is dangerous to remain aggressively positioned late in an economic cycle because risk assets can begin to underperform well before evidence accumulates that the economy has fallen into recession. Using the peak in the S&P 500 operating profit margin as a signal to lighten up appears prudent. Based on this approach, investors should generally remain overweight risk assets generally, including stocks, corporate bonds, hedge funds, private equity and real estate, as long as margins are still rising. Investors should scale back in most of these areas as soon as margins peak. For fixed income, investors should be looking to raise exposure but move up in quality after margins peak. Oil and related plays are not a reliable late-cycle play, but we are bullish because of the favorable supply-demand outlook. However, this does not carry over to base metals, where we are more cautious. There are some assets other than government bonds that generated a positive average return late in the cycle and during the recession periods, suggesting that they are good late-cycle assets to hold. However, this is misleading because in some cases they experienced a significant correction either during or slightly before the recession (see the maximum drawdown columns in Table I-1; blank cells indicate that the asset did not experience a correction). These include IG credit, CMBS, ABS, Gold and Dividend Aristocrats. The only assets in our list that provided both a positive return across all the phases in Table I-1 and avoided a correction during the recessions, were mortgage-backed securities, Timberland and Farmland. A Special Report from BCA's Global Asset Allocation service found that Timberland is a superior inflation hedge to Farmland, but the latter is a superior hedge against recessions and equity bear markets.3 We believe that Agency MBS are unattractively valued, but should remain insulated from negative shocks such as a trade war or higher Treasury yields (as long as the Treasury selloff is not extreme). Our fixed income team also likes Agency CMBS.4 When Will U.S. Margins Peak? It is impressive that S&P 500 after-tax operating margins are extremely elevated and still rising. The trend has been aided by tax cuts, but corporate pricing power has improved and wage growth has not yet accelerated enough to damage margins. Chart I-10 presents some indicators to monitor as we await the cyclical peak in profit margins. These are generally not leading indicators, but they do provide some warning when they roll over late in the cycle. The first is the BCA Margin Proxy, which is the ratio of selling prices for the non-financial corporate sector to unit labor costs. Margins have tended to fall historically when the growth rate of this ratio is below zero. The same is true for nominal GDP growth minus aggregate wages. The aggregate wage bill incorporates both changes in wages/hour and in total hours worked. We are also watching a diffusion index of the changes in margins for the industrial components of the S&P 500, as well as BCA's Corporate Pricing Power indicator. The latter takes into consideration price changes at the detailed industry level. Chart I-10U.S. Profit Margin Indicators To Watch U.S. Profit Margin Indicators To Watch U.S. Profit Margin Indicators To Watch None of these indicators are signaling an imminent top in margins, but all appear to have peaked except the Corporate Pricing Power indicator. An equally-weighted average of these four indicators, labelled the U.S. Composite Margin Indicator in Chart I-10, is falling but is still above the zero line. We would not be surprised to see S&P 500 margins peak for the cycle late early in 2019. Conclusions: The S&P 500 has so far been largely immune to shocking trade headlines with the help of a solid start to the U.S. Q2 earning season. Based on previous late cycle periods, the fact that S&P 500 profit margins are still rising suggests that investors should remain fully-exposed to most risk assets. Nonetheless, timing is always difficult and we have decided to focus on capital preservation given extended valuations and a raft of risks that could cause a premature end to the bull market. These risks include a possible hard economic landing in China, crises in one or more EM countries, and an escalation in the trade war among others. Some investors appear to believe that the U.S. can "win" the trade war, but there are no winners when tariff walls are rising. We are not yet ready to go underweight on risk assets, but risk tolerance should be no more than benchmark. This includes equities, corporate bonds, EM assets and other risky sectors. An inversion of the yield curve could trigger a shift to underweight, although this signal would have to be corroborated by our other favorite U.S. and global indicators. Attractive late-cycle assets to hold include structured product, Timberland and Farmland. The first statements by Jay Powell as FOMC Chair underscored that it is too early to hide in Treasurys. Market expectations for real short-term interest rates are overly benign out to the middle of the next decade. Moreover, the Fed is not in a position to be proactive in leaning against the negative impact of rising tariffs because inflation is near target and the labor market is showing signs of overheating. This means that bond yields are headed higher until economic pain is clearly evident. Keep duration short of benchmark. Long-term rate expectations for the Eurozone appear even more complacent than they do for the U.S. The real ECB policy rate is expected to remain in negative territory until 2028 (Chart I-5)! At some point there will be a convergence of real rate expectations with the U.S., which will boost the value of the euro. Nonetheless, we believe that it is too early to position for rate convergence. Core inflation is still well below target and Eurozone economic growth has softened recently, suggesting that the ECB will be in no hurry to lift rates once asset purchases have ended. ECB policymakers will be disinclined to cater to President's Trump's desire for tighter monetary policy in Europe, which means that the U.S. dollar has more upside versus the euro and in broad trade-weighted terms. An escalation in the trade war would augment upward pressure on the greenback. As the dollar's behavior during the Global Financial Crisis illustrates, even major shocks that originate from the U.S. tend to attract capital inflows into the safe-haven Treasury market. Emerging market assets are particularly vulnerable to another upleg in the dollar because of the high level of U.S. dollar-denominated debt. Favor DM to EM equity markets and currencies. Mark McClellan Senior Vice President The Bank Credit Analyst July 26, 2018 Next Report: August 30, 2018 1 For more information on why a replay of the 1985 Plaza Accord is unlikely, please see BCA Geopolitical Strategy Weekly Report "The Dollar May Be Our Currency, But It Is Your Problem," dated July 25, 2018, available on gps.bcaresearch.com 2 Please see BCA Global Investment Strategy Weekly Report "U.S. Housing Will Drive the Global Business Cycle...Again," dated July 6, 2018, available on gis.bcaresearch.com 3 Please see BCA Global Asset Allocation Service Special Report "U.S. Farmland & Timberland: An Investment Primer," dated October 24, 2017, available on gaa.bcaresearch.com 4 Please see BCA's U.S. Bond Strategy Weekly Report, "The Fed's Balance Sheet Problem," dated July 17, 2018, available on usbs.bcaresearch.com II. U.S. Equity Sectors: Trade War Winners And Losers In this Special Report, we shed light on the implications of the U.S./Sino trade war for U.S. equity sectors. The threat that trade action poses to the U.S. equity market is greater than in past confrontations. Perhaps most importantly, supply chains are much more extensive, globally and between China and the U.S. Automobile Components, Electrical Equipment, Materials, Capital Goods and Consumer Durables have the most extensive supply chain networks. The USTR claims that it is being strategic in the Chinese goods it is targeting, focusing on companies that will benefit from the "Made In China 2025" initiative. The list of Chinese goods targeted in both the first and second rounds covers virtually all of the broad import categories. The only major items left for the U.S. to hit are apparel, footwear, toys and cellphones. Beijing is clearly targeting U.S. products based on politics in order to exert as much pressure on the President's party as possible. Based on a list of products that comprise the top-10 most exported goods of Red and Swing States, China will likely lift tariffs in the next rounds on civilian aircraft, computer electronics, healthcare equipment, car engines, chemicals, wood pulp, telecommunication and integrated circuits. Supply chains within and between industries and firms mean that the impact of tariffs is much broader than the direct impact on exporters and importers. We measure the relative exposure of 24 GICs equity sectors to the trade war based on their proportion of foreign-sourced revenues and the proportion of each industry's total inputs that are affected by U.S. tariffs. The Semiconductors & Semiconductor Equipment sector stands out, but the Technology & Hardware Equipment, Capital Goods, Materials, Consumer Durables & Apparel and Motor Vehicle sectors are also highly exposed to anti-trade policy action. Energy, Software, Banks and all other service sectors are much less exposed. China may also attempt to disrupt supply chains via non-tariff barriers, placing even more pressure on U.S. firms that have invested heavily in China. Wholesale Trade, Chemicals, Transportation Equipment, Computers & Electronic Parts and Finance & Insurance are most exposed. U.S. technology companies are particularly vulnerable to an escalating trade war. Virtually all U.S. manufacturing industries will be negatively affected by an ongoing trade war, even defensive sectors such as Consumer Staples. The one exception is defense manufacturers, where we recommend overweight positions. Our analysis highlights that the best shelter from a trade war can be found in services, particularly services that are insulated from trade. Financial Services appears a logical choice, and the S&P Financial Exchanges & Data subsector is one of our favorites. The trade skirmish is transitioning to a full-on trade war. The U.S. has imposed a 25% tariff on $50 billion worth of Chinese goods, and has proposed a 10% levy on an additional $200 billion of imports by August 31. China retaliated with tariffs on $50 billion of imports from the U.S., but Trump has threatened tariffs on another $300 billion if China refuses to back down. That would add up to over $500 billion in Chinese goods and services that could be subject to tariffs, only slightly less than the total amount that China exported to the U.S. last year. BCA's Geopolitical Strategy has emphasized that President Trump is unconstrained on trade policy, giving him leeway to be tougher than the market expects.1 This is especially the case with respect to China. There will be strong pushback from Congress and the U.S. business lobby if the Administration tries to cancel NAFTA. In contrast, Congress is also demanding that the Administration be tough on China because it plays well with voters. Trump is a prisoner of his own tough pre-election campaign rhetoric against China. The U.S. primary economic goal is not to equalize tariffs but to open market access.2 The strategic goal is much larger. The U.S. wants to see China's rate of technological development slow down. Washington will expect robust guarantees to protect intellectual property and proprietary technology before it dials down the pressure on Beijing. The threat that the trade war poses to the U.S. equity market is greater than in past confrontations, such as that between Japan and the U.S. in the late 1980s. First, stocks are more expensive today. Second, interest rates are much lower, limiting how much central banks can react to adverse shocks. Third, and perhaps most importantly, supply chains are much more extensive, globally and between China and the U.S. Nearly every major S&P 500 multinational corporation is in some way exposed to these supply chains. Chart II-1 shows that Automobile Components, Electrical Equipment, Materials, Capital Goods and Consumer Durables have the most extensive supply chain networks. The Global Value Chain Participation rate, constructed by the OECD, is a measure of cross-border value-added linkages.3 In this Special Report, we shed light on the implications of the trade war for U.S. equity sectors. Complex industrial interactions make it difficult to be precise in identifying the winners and losers of a trade war. Nonetheless, we can identify the industries most and least exposed to a further rise in tariff walls or non-tariff barriers to trade. We focus on the U.S./Sino trade dispute in this Special Report, leaving the implications of a potential trade war with Europe and the possible failure of NAFTA negotiations for future research. Chart II-1Measuring Global Supply Chains August 2018 August 2018 Trade Channels There are at least five channels through which rising tariffs can affect U.S. industry: The Direct Effect: This can be positive or negative. The impact is positive for those industries that do not export much but are provided relief from stiff import competition via higher import tariffs. The impact is negative for those firms facing higher tariffs on their exports, as well as for those firms facing higher costs for imported inputs to their production process. These firms would be forced to absorb some of the tariff via lower profit margins. Some industries will fall into both positive and negative camps. U.S. washing machines are a good example. Whirlpool's stock price jumped after President Trump announced an import tariff on washing machines, but it subsequently fell back when the Administration imposed an import tariff on steel and aluminum (that are used in the production of washing machines); Indirect Effect: The higher costs for imported goods are passed along the supply chain within an industry and to other industries that are not directly affected by rising tariffs. This will undermine profit margins in these indirectly-affected industries to the extent that they cannot fully pass along the higher input costs; Foreign Direct Investment: Some Chinese exports emanate from U.S. multinationals' subsidiaries in China, or by Chinese or foreign OEM suppliers for U.S. firms. Even though it would undermine China's economy to some extent, the Chinese authorities could make life more difficult for these firms in retaliation for U.S. tariffs on Chinese goods. Macro Effect: A trade war would take a toll on global trade and reduce GDP growth globally. Besides the negative effect of uncertainty on business confidence and, thus, capital spending, rising prices for both consumer and capital goods will reduce the volume of spending in both cases. Moreover, corporate profits have a high beta with respect to economic activity. We would not rule out a U.S. recession in a worst-case scenario. Obviously, a recession or economic slowdown would inflict the most pain on the cyclical parts of the S&P 500 relative to the non-cyclicals, in typical fashion. Currency Effect: To the extent that a trade war pushes up the dollar relative to the other currencies, it would undermine export-oriented industries and benefit those that import. However, while we are bullish the dollar due to diverging monetary policy, the dollar may not benefit much from trade friction given retaliatory tariff increases by other countries. Some of the direct and indirect impact can be mitigated to the extent that importers facing higher prices for Chinese goods shift to similarly-priced foreign producers outside of China. Nonetheless, this adjustment will not be costless as there may be insufficient supply capacity outside of China, leading to upward pressure on prices globally. Targeted Sectors: (I) U.S. Tariffs On Chinese Goods As noted above, the U.S. has already imposed tariffs on $50 billion of Chinese imports and has published a list of another $200 billion of goods that are being considered for a 10% tariff in the second round of the trade war. The first round focused on intermediate and capital goods, while the second round includes consumer final demand categories such as furniture, air conditioners and refrigerators. The latter will show up as higher prices at retailers such as Wal Mart, having a direct and visible impact on U.S. households. Appendix Table II-A1 lists the goods that are on the first and second round lists, grouped according to the U.S. equity sectors in the S&P 500. The U.S. Trade Representative (USTR) claims that the Chinese items are being targeted strategically. It is focusing on companies that will benefit from China's structural policies, such as the "Made In China 2025" initiative that is designed to make the country a world leader in high-tech areas (see below). Table II-1 reveals the relative size of the broad categories of U.S. imports from China, based on trade categories. The top of the table is dominated by Motor Vehicles, Machinery, Telecommunication Equipment, Computers, Apparel & Footwear and other manufactured goods. The list of Chinese goods targeted in both the first and second rounds covers virtually all of the broad categories in Table II-1. The only major items left for the U.S. to hit are Apparel and Footwear, as well as two subcategories; Toys and Cellphones. These are all consumer demand categories. Table II-1U.S. Imports From China (January-May 2018) August 2018 August 2018 (II) Chinese Tariffs On U.S. Goods Total U.S. exports to China were less than $53 billion in the first five months of 2018, limiting the amount of direct retaliation that China can undertake (Table II-2). The list of individual U.S. products that China has targeted so far is long, but we have condensed it into the broad categories shown in Table II-3. The U.S. equity sectors that the new tariffs affect so far include Food, Beverage & Tobacco, Automobiles & Components, Materials and Energy. China has concentrated mainly on final goods in a politically strategic manner, such as Trump-supported rural areas and Harley Davidson bikes whose operations are based in Paul Ryan's home district in Wisconsin. Table II-2U.S. Exports To China (January-May 2018) August 2018 August 2018 Table II-3China Tariffs On U.S. Goods August 2018 August 2018 What will China target next? Chart II-2 shows exports to China as percent of total state exports, and Chart II-3 presents the value of products already tariffed by China as a percent of state exports. Other than Washington, the four states most targeted by Beijing are conservative: Alaska, Alabama, Louisiana and South Carolina. Chart II-2U.S. Exports To China By State August 2018 August 2018 Chart II-3Value Of U.S. Products Tariffed By China (By State) August 2018 August 2018 Beijing is clearly targeting products based on politics in order to exert as much pressure on the President's party as possible. To identify the next items to be targeted, we constructed a list of products that comprise the top-10 most exported goods of Red States (solidly conservative) and Swing States (competitive states that can go either to Republican or Democratic politicians). Appendix Tables II-A2 and II-A3 show this list of products, with those that have already been flagged by China for tariffs crossed out. Table II-4 shows the top-10 list of products that are not yet tariffed by China, but are distributed in a large proportion of Red and Swing states. What strikes us immediately is how important aircraft exports are to a large number of Swing and Red States. In total, 27 U.S. states export civilian aircraft, engines and parts to China. This is an obvious target of Beijing's retaliation. In addition, we believe that computer electronics, healthcare equipment, car engines, chemicals, wood pulp, telecommunication and integrated circuits are next. Table II-4Number Of U.S. States Exporting To China By Category August 2018 August 2018 Market Reaction Chart II-4 highlights how U.S. equity sectors performed during seven separate days when the S&P 500 suffered notable losses due to heightened fears of protectionism. Cyclical sectors such as Industrials and Materials fared worse during days of rising protectionist angst. Financials also generally underperformed, largely because such days saw a flattening of the yield curve. Tech, Health Care, Energy and Telecom performed broadly in line with the S&P 500. Consumer Staples outperformed the market, but still declined in absolute terms. Utilities and Real Estate were the only two sectors that saw absolute price gains. The market reaction seems sensible based on the industries caught in the cross-hairs of the trade action so far. At least some of the potential damage is already discounted in equity prices. Nonetheless, it is useful to take a closer look at the underlying factors that should determine the ultimate winners and losers from additional salvos in the trade war. Chart II-4S&P 500: Impact Of Trade-Related Events August 2018 August 2018 Determining The Winners And Losers The U.S. sectors that garner the largest proportion of total revenues from outside the U.S. are obviously the most exposed to a trade war. For the 24 level 2 GICS sectors in the S&P 500, Table II-5 presents the proportion of total revenues that is generated from operations outside the U.S. for the top five companies in the sector by market cap. Company reporting makes it difficult in some cases to identify the exact revenue amount coming from outside the U.S., as some companies regard "domestic" earnings as anything generated in North America. Nonetheless, we believe the data in Table II-5 provide a reasonably accurate picture. Table II-5Foreign Revenue Exposure (2017) August 2018 August 2018 Semiconductors, Tech Equipment, Materials, Food & Beverage, Software and Capital Goods are at the top of the list in terms of foreign-sourced revenues. Not surprisingly, service industries like Real Estate, Banking, Utilities and Telecommunications Services are at the bottom of the exposure list. U.S. companies are also exposed to U.S. tariffs that lift the price of imported inputs to the production process. This can occur directly when firm A imports a good from abroad, and indirectly, when firm A sells its intermediate good to firm B at a higher price, and then on to firm C. In order to capture the entire process, we used the information contained in the Bureau of Economic Analysis' Input/Output tables. We estimated the proportion of each industry's total inputs that are affected by already-implemented U.S. tariffs and those that are on the list for the next round of tariffs. These estimates, shown in Appendix Table II-A4 at a detailed industrial level, include both the direct and indirect effects of higher import costs. At the top of the list is Motor Vehicles and Parts, where Trump tariffs could affect more than 70% of the cost of all material inputs to the production process. Electrical Equipment, Machinery and other materials industries are also high on the list, together with Furniture, Computers & Electronic Parts and Construction. Unsurprisingly, service industries and Utilities are in the bottom half of the table.4 We then allocated all the industries in Appendix Table II-A4 to the 24 GICs level 2 sectors in the S&P 500, in order to obtain an import exposure ranking in S&P sector space (Table II-6). Table II-6U.S. Import Tariff Exposure August 2018 August 2018 Chart II-5 presents a scatter diagram that compares import tariff exposure (horizontal axis) with foreign revenue exposure (vertical axis). The industries clustered in the top-right of the diagram are the most exposed to a trade war. Chart II-5U.S. Industrial Exposure To A Trade War With China August 2018 August 2018 The Semiconductors & Semiconductor Equipment sector stands out by this metric, but the Technology & Hardware Equipment, Capital Goods, Materials, Consumer Durables & Apparel and Motor Vehicle sectors are also highly exposed to anti-trade policy action. Energy, Software, Banks and all other service sectors are much less exposed. Food, Beverage & Tobacco lies between the two extremes. Joint Ventures And FDI Table II-7Stock Of U.S. Direct ##br##Investment In China (2017) August 2018 August 2018 As mentioned above, most U.S. production taking place in China involves a joint venture. The Chinese authorities could attempt to disrupt the supply chain of a U.S. company by hindering production at companies that have ties to U.S. firms. Data on U.S. foreign direct investment (FDI) in China will be indicative of the industries that are most exposed to this form of retaliation. The stock of U.S. FDI in China totaled more than $107 billion last year (Table II-7). At the top of the table are Wholesale Trade, Chemicals, Transportation Equipment, Computers & Electronic Parts and Finance & Insurance. Apple is a good example of a U.S. company that is exposed to non-tariff retaliation, as the iPhone is assembled in China by Foxconn for shipment globally with mostly foreign sourced parts. Our Technology sector strategists argue that U.S. technology companies are particularly vulnerable to an escalating trade war (See Box II-1).5 BOX II-1 The Tech Sector The U.S. has applied tariffs on the raw materials of technology products rather than finished goods so far. At a minimum, this will penalize smaller U.S. tech firms which manufacture in the U.S. and provide an incentive to move production elsewhere. Worst case, the U.S. tariffs might lead to component shortages which could have a disproportionately negative impact, especially on smaller firms. Although it has not been proposed, U.S. tariffs on finished goods would be devastating to large tech companies such as Apple, which outsources its manufacturing to China. China appears determined to have a vibrant high technology sector. The "Made In China 2025" program, for example, combines ambitious goals in supercomputers, robotics, medical devices and smart cars, while setting domestic localization targets that would favor Chinese companies over foreigners. The ZTE sanctions and the potential for enhanced export controls have had a traumatic impact on China's understanding of its relatively weak position with respect to technology. As a result, because most high-tech products are available from non-U.S. sources, Chinese engineers will likely be encouraged to design with non-U.S. components; for example, selecting a Samsung instead of a Qualcomm processor for a smartphone. Similarly, China is a major buyer of semiconductor capital equipment as it follows through with plans to scale up its semiconductor industry. Most such equipment is also available from non-U.S. vendors, and it would be understandable if these suppliers are selected given the risk which would now be associated with selecting a U.S. supplier. The U.S. is targeting Chinese made resistors, capacitors, crystals, batteries, Light Emitting Diodes (LEDs) and semiconductors with a 25% tariff. For the most part these are simple, low cost devices, which are used by the billions in high-tech devices. Nonetheless, China could limit the export of these products to deliver maximum pain, leading to a potential shortage of qualified parts. A component shortage can have a devastating impact on production since the manufacturer may not have the ability to substitute a new part or qualify a new vendor. Since the product typically won't work unless all the right parts are installed, want of a dollar's worth of capacitors may delay shipping a $1,000 product. Thus, the economic and profit impact of a parts shortage in the U.S. could be quite severe. Conclusions: When it comes to absolute winners in case of a trade war, we believe there are three conditions that need to be met: Relatively high domestic input costs. Relatively high domestic consumption/sales; the true beneficiaries of a tariff are those industries who are allowed to either raise prices or displace foreign competitors, with the consumer typically bearing the cost. Relatively low direct exposure to global trade - international trade flows will certainly slow in a trade war. There are very few manufacturing industries that meet all of these criteria. Within manufacturing, one would typically expect the Consumer Staples and Discretionary sectors to be the best performers. However, roughly a third of the weight of Staples is in three stocks (PG, KO and PEP) that are massively dependent on foreign sales. Moreover, a similar weight of Discretionary is in two retailers (AMZN and HD) that are dependent on imports. As such, consumer indexes do not appear a safe harbor in a trade war. Nevertheless, if the trade war morphs into a recession then consumer staples (and other defensive safe-havens) will outperform, although they will still decrease in absolute terms. Transports are an industry that has relatively high domestic labor costs and an output that is consumed virtually entirely within domestic borders. However, their reliance on global trade flows - intermodal shipping is now more than half of all rail traffic - means they almost certainly lose from a prolonged trade dispute. There is one manufacturing industry that could be at least a relative winner and perhaps an absolute winner: defense. Defense manufacturers certainly satisfy the first two criteria above, though they do have reasonably heavy foreign exposure. However, we believe high switching costs and the lack of true global competitors mean that U.S. defense company foreign sales will be resilient. After all, a NATO nation does not simply switch out of F-35 jets for the Russian or Chinese equivalent. Further, if trade friction leads to rising military tension, defense stocks should outperform. Finally, the ongoing global arms race, space race and growing cybersecurity requirements all signal that these stocks are a secular growth story, as BCA has argued in the recent past.6 Still, as highlighted by the data presented above, the best shelter from a trade war can be found in services, particularly services that are insulated from trade. Financial Services appears a logical choice, especially the S&P Financial Exchanges & Data subsector (BLBG: S5FEXD - CME, SPGI, ICE, MCO, MSCI, CBOE, NDAQ). Another appealing - and defensive - sector is Health Care Services. With effectively no foreign exposure and a low beta, these stocks would outperform in the worst-case trade war-induced recession. Mark McClellan Senior Vice President The Bank Credit Analyst Marko Papic Senior Vice President Geopolitical Strategy Chris Bowes Associate Editor U.S. Equity Strategy 1 Please see BCA Geopolitical Strategy Special Report, "Constraints & Preferences Of The Trump Presidency," dated November 30, 2016, available at gps.bcaresearch.com. 2 Please see BCA Geopolitical Strategy Weekly Report, "Trump's Demands On China," dated April 4, 2018, available at gps.bcaresearch.com. 3 For more information, please see: "Global Value Chains (GVSs): United States." May 2013. OECD website. 4 Please see BCA U.S. Equity Strategy Special Report, "Brothers In Arms," dated October 31, 2016, available at uses.bcaresearch.com. 5 Please see BCA Technology Sector Strategy Special Report "Trade Wars And Technology," dated July 10, 2018, available at tech.bcaresearch.com 6 Please see BCA U.S. Equity Strategy Special Report, "Brothers In Arms," dated October 31, 2016, available at uses.bcaresearch.com. Appendix Table II-1 Allocating U.S. Import Tariffs To U.S. GICS Sectors August 2018 August 2018 Appendix Table II-2 Exports By U.S. Red States August 2018 August 2018 Appendix Table II-3 Exports By U.S. Swing States August 2018 August 2018 Appendix Table II-4 Exposure Of U.S. Industries To U.S. Import Tariffs August 2018 August 2018 III. Indicators And Reference Charts Our equity-related indicators flashed caution again in July, despite robust U.S. corporate earnings indicators. Forward earnings estimates continued to surge in July. The net revisions ratio and the earnings surprises index remained well above average, suggesting that forward earnings still have upside potential in the coming months. However, several of our indicators suggest that it is getting late in the bull market. Our Monetary Indicator is approaching very low levels by historical standards. Equities are still close to our threshold of overvaluation, at a time when our Composite Technical Indicator appears poised to break down. An overvalued reading is not bearish on its own, but valuation does provide information on the downside risks when the correction finally occurs. Equity sentiment is close to neutral according to our composite indicator, but the low level of implied volatility suggests that investors are somewhat complacent. Our U.S. Willingness-to-Pay (WTP) indicator has fallen significantly this year, and the Japanese WTP appears to be rolling over. The WTP indicators track flows, and thus provide information on what investors are actually doing, as opposed to sentiment indexes that track how investors are feeling. Flows into the U.S. stock market are waning, and those into the Japanese market are wavering. Flows into European stocks have flattened off. Finally, our Revealed Preference Indicator (RPI) for stocks remained on a ‘sell’ signal in July. The RPI combines the idea of market momentum with valuation and policy measures. It provides a powerful bullish signal if positive market momentum lines up with constructive signals from the policy and valuation measures. Conversely, if constructive market momentum is not supported by valuation and policy, investors should lean against the market trend. These indicators are not aligned at the moment, further supporting the view that caution is warranted. The U.S. 10-year Treasury is slightly on the inexpensive side and our Composite Technical Indicator suggests that the bond has still not worked off oversold conditions. This suggests that the consolidation period has further to run, although we still expect yields to move higher over the remainder of the year. This month’s Overview section discusses the upside potential for the term premium in the yield curve and for market expectations of the terminal fed funds rate. This year’s dollar rally has taken it to very expensive levels according to our purchasing power parity estimate. The long-term trend in the dollar is down, but we still believe it has some upside while market expectations for the terminal fed funds rate adjust upward. EQUITIES: Chart III-1U.S. Equity Indicators U.S. Equity Indicators U.S. Equity Indicators Chart III-2Willingness To Pay For Risk Willingness To Pay For Risk Willingness To Pay For Risk Chart III-3U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators U.S. Equity Sentiment Indicators Chart III-4Revealed Preference Indicator Revealed Preference Indicator Revealed Preference Indicator Chart III-5U.S. Stock Market Valuation U.S. Stock Market Valuation U.S. Stock Market Valuation Chart III-6U.S. Earnings U.S. Earnings U.S. Earnings Chart III-7Global Stock Market And Earnings: ##br##Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance Chart III-8Global Stock Market And Earnings: ##br##Relative Performance Global Stock Market And Earnings: Relative Performance Global Stock Market And Earnings: Relative Performance FIXED INCOME: Chart III-9U.S. Treasurys And Valuations U.S. Treasurys and Valuations U.S. Treasurys and Valuations Chart III-10U.S. Treasury Indicators U.S. Treasury Indicators U.S. Treasury Indicators Chart III-11Selected U.S. Bond Yields Selected U.S. Bond Yields Selected U.S. Bond Yields Chart III-1210-Year Treasury Yield Components 10-Year Treasury Yield Components 10-Year Treasury Yield Components Chart III-13U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor U.S. Corporate Bonds And Health Monitor Chart III-14Global Bonds: Developed Markets Global Bonds: Developed Markets Global Bonds: Developed Markets Chart III-15Global Bonds: Emerging Markets Global Bonds: Emerging Markets Global Bonds: Emerging Markets CURRENCIES: Chart III-16U.S. Dollar And PPP U.S. Dollar And PPP U.S. Dollar And PPP Chart III-17U.S. Dollar And Indicator U.S. Dollar And Indicator U.S. Dollar And Indicator Chart III-18U.S. Dollar Fundamentals U.S. Dollar Fundamentals U.S. Dollar Fundamentals Chart III-19Japanese Yen Technicals Japanese Yen Technicals Japanese Yen Technicals Chart III-20Euro Technicals Euro Technicals Euro Technicals Chart III-21Euro/Yen Technicals Euro/Yen Technicals Euro/Yen Technicals Chart III-22Euro/Pound Technicals Euro/Pound Technicals Euro/Pound Technicals COMMODITIES: Chart III-23Broad Commodity Indicators Broad Commodity Indicators Broad Commodity Indicators Chart III-24Commodity Prices Commodity Prices Commodity Prices Chart III-25Commodity Prices Commodity Prices Commodity Prices Chart III-26Commodity Sentiment Commodity Sentiment Commodity Sentiment Chart III-27Speculative Positioning Speculative Positioning Speculative Positioning ECONOMY: Chart III-28U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop U.S. And Global Macro Backdrop Chart III-29U.S. Macro Snapshot U.S. Macro Snapshot U.S. Macro Snapshot Chart III-30U.S. Growth Outlook U.S. Growth Outlook U.S. Growth Outlook Chart III-31U.S. Cyclical Spending U.S. Cyclical Spending U.S. Cyclical Spending Chart III-32U.S. Labor Market U.S. Labor Market U.S. Labor Market Chart III-33U.S. Consumption U.S. Consumption U.S. Consumption Chart III-34U.S. Housing U.S. Housing U.S. Housing Chart III-35U.S. Debt And Deleveraging U.S. Debt And Deleveraging U.S. Debt And Deleveraging Chart III-36U.S. Financial Conditions U.S. Financial Conditions U.S. Financial Conditions Chart III-37Global Economic Snapshot: Europe Global Economic Snapshot: Europe Global Economic Snapshot: Europe Chart III-38Global Economic Snapshot: China Global Economic Snapshot: China Global Economic Snapshot: China Mark McClellan Senior Vice President The Bank Credit Analyst

Highlights Global Yields: Flattening government yield curves in the developed world have raised concerns about a potential future growth slowdown. Yet real policy rates will need to move into positive territory before monetary policy becomes truly restrictive and curves invert. This means global bond yields have not yet peaked for this cycle. UST-Bund Spread: The U.S. Treasury-German Bund spread has overshot our fair value estimates, and relative positive data surprises are turning more in favor of Europe. We are taking profits on our tactical UST-Bund spread widening trade, after a gain of 7% (hedged into U.S. dollars). UST Technicals: Some of the oversold technical conditions in the U.S. Treasury market have turned more neutral, but sentiment remains bearish. With both U.S. growth and inflation accelerating, we recommend sticking with a strategic below-benchmark U.S. duration stance rather than playing for a tactical short-covering Treasury rally. Feature In most years, investment professionals can look forward to taking some well-deserved time off in July to hit the beach and read a good book. This year, those same investors are forced to keep an eye on their Bloombergs while responding to the public musings of Donald Trump. The president made comments late last week that threatened the independence of the Federal Reserve, while also accusing China and Europe of currency manipulation. While those headlines can briefly move markets on a sunny summer day, they represent more Trump-ian bluster than any potential change in the conduct of U.S. monetary or currency policy. Chart of the WeekCan Policy Be Truly "Tight"##BR##With Negative Real Rates? Can Policy Be Truly 'Tight' With Negative Real Rates? Can Policy Be Truly 'Tight' With Negative Real Rates? The underlying dynamic remains one of mixed global growth (strong in the U.S., slowing almost everywhere else) but with low unemployment and rising inflation in most major economies. That means that independent, inflation-fighting central bankers must focus on their inflation mandates. In the U.S., that means more Fed rate hikes and a firm U.S. dollar, regardless of the desires of President Trump - the author of the large fiscal stimulus, at full employment, which is forcing the Fed to continue hiking rates. In other countries, however, the economic backdrop is leading to varying degrees of central banker hawkishness. That ranges from actual rate hikes (Canada) to tapering of bond buying (Europe, Japan) to merely talking up the potential for rate increases (U.K., Sweden, Australia). The aggregate monetary policy stance of the major developed market central banks is now tilted more hawkishly. So it is no surprise that global government bond yield curves have been flattening and returns on risk assets have been underwhelming (Chart of the Week). Yet the reality is that all major global curves still have a positive slope, even in the U.S. and Canada where central banks have been most actively tightening, while real policy interest rates remain below zero. It would be highly unusual for yield curves to invert before real rates turned positive, especially if central bankers must move to an outright restrictive stance given tight labor markets and rising realized inflation. This implies that there is more scope for global bond yields to rise over the next 6-12 months. We continue to recommend that investors maintain a defensive overall duration stance ... and to focus more on that good book on the beach and less on Trump's Twitter feed. Where To Next For The Treasury-Bund Spread? Chart 2A Pause In The Rising Yield Trend,##BR##Not A Reversal A Pause In The Rising Yield Trend, Not A Reversal A Pause In The Rising Yield Trend, Not A Reversal The rise in bond yields in both the U.S. and euro area seen in the first quarter of 2018 has been partly reversed since then. One of the culprits has been a stalling of the rally in oil markets, which has prompted a pause in the rise of inflation expectations on both sides of the Atlantic (Chart 2). Yet another factor has been the larger decline in real bond yields, which have fallen around 20bps in the both the U.S. and euro area since the peak in mid-May (bottom two panels). A potential driver of those lower real yields is the growing concern over the potential hit to global growth from rising trade tensions between the U.S. and China (and Europe, Canada, Mexico, etc). This comes at a time when China's economic growth was already slowing and acting as a drag on global trade activity and commodity prices. There has been significant weakness in China's currency and equity market of late, which raises the specter of another broader global selloff as occurred during the Chinese turbulence of 2015/16. Yet the declines in industrial metals prices and emerging market corporate debt have been far more modest so far in 2018 (Chart 3). A big reason for that has been the more subdued performance of the U.S. dollar this year, unlike the massive surge in 2015/16 that crushed risk assets worldwide (Chart 4). A more likely driver of the recent drop in real yields in the U.S. and core Europe was the slump in euro area economic data earlier in 2018. That move not only drove yields lower, but also pushed out the market-implied timing of the first ECB rate hike (Chart 5) and drove the spread between U.S. Treasuries and German Bunds to new wides. In our last Weekly Report, we updated our list of indicators in the U.S. and euro area that we have been monitoring to assess if our below-benchmark duration stance was still appropriate.1 The conclusion was that the underlying trends in growth and inflation on both sides of the Atlantic still supported higher bond yields on a cyclical basis, although the pressures were greater in the U.S. Yet at the same time, the gap between U.S. and euro area government bond yields has remained historically wide, with the 10-year Treasury-German Bund spread now sitting at 255bps - the highest level since the late 1980s. Chart 3Slowing Growth##BR##In China... Slowing Growth In China... Slowing Growth In China... Chart 4...But Not Yet Enough To Threaten##BR##Global Financial Stability ...But Not Yet Enough To Threaten Global Financial Stability ...But Not Yet Enough To Threaten Global Financial Stability Monetary policy differences have historically been the biggest driver of that spread. Today, the Fed is well into an interest rate hiking cycle that began nearly three years ago, and is now in the process of unwinding its balance sheet. Meanwhile, the ECB has been keeping policy rates at or below 0% while engaging in large-scale bond buying (Chart 6). Chart 5A Turn In European Yields##BR##On The Horizon? A Turn In European Yields On The Horizon? A Turn In European Yields On The Horizon? Chart 6Wide UST-Bund Spread Reflects##BR##Monetary Policy Divergences Wide UST-Bund Spread Reflects Monetary Policy Divergences Wide UST-Bund Spread Reflects Monetary Policy Divergences When looking at more typical fundamental drivers of the Treasury-Bund spread, many of the cross-regional differences are already "in the price". The spread appears to have overshot relative to the three main factors that go into our Treasury-Bund spread valuation model (Chart 7): The gap between Fed and ECB policy rate The ratio of the U.S. unemployment rate to the euro area equivalent The gap between headline inflation in the U.S. and euro area The Fed's rate hikes have now widened the policy rate differential versus the ECB equivalent (the short-term repo rate) to 200bps. At the same time, the rapidly improving situation in the euro area labor market now means that the unemployment ratio has been constant over the past couple of years, while euro area inflation has also caught up a bit toward U.S. levels in recent months. Adding it all up together in our Treasury-Bund valuation model - which also includes the sizes of the Fed and ECB balance sheets to quantify the impact on yields of bond-buying programs - and the conclusion is that the current spread level of 255bps is 50bps above "fair value" (Chart 8). Chart 7UST-Bund Spread Overshooting Fundamentals UST-Bund Spread Overshooting Fundamentals UST-Bund Spread Overshooting Fundamentals Chart 8UST-Bund Spread Looks Wide On Our Model UST-Bund Spread Looks Wide On Our Model UST-Bund Spread Looks Wide On Our Model Importantly, fair value is still rising, primarily because of the widening policy rate differential. We have consistently argued that the true cyclical peak in the Treasury-Bund spread will occur when the Fed is done with its rate hike cycle. Yet there are opportunities to play that spread more tactically, based on shorter-term indicators. For example, the gap between the data surprise indices for the U.S. and euro area has been a correlated to the momentum of the Treasury-Bund spread, measured as the 13-week change of the level of the spread (Chart 9). Data surprises are now bottoming out in the euro area while they continue to drift lower in the U.S. As a result, the Treasury-Bund spread momentum has begun to fade, right in line with the narrowing of the data surprise differential. Also from a more technical perspective, the deviation of the Treasury-Bund spread from its 200-day moving average is at one of the more stretched levels of the past decade. Combined with the extended spread momentum, this suggests that the Treasury-Bund spread should expect to see a period of consolidation in the next few months (Chart 10). Chart 9Relative Data Surprises No Longer##BR##Support A Wider UST-Bund Spread Relative Data Surprises No Longer Support A Wider UST-Bund Spread Relative Data Surprises No Longer Support A Wider UST-Bund Spread Chart 10UST-Bund Spread Momentum##BR##Got To Stretched Extremes UST-Bund Spread Momentum Got To Stretched Extremes UST-Bund Spread Momentum Got To Stretched Extremes We have been recommending both a structural short U.S./long core Europe position in our model bond portfolio for over a year now. We also entered into a trade that directly played for a wider 10-year Treasury-Bund spread in our Tactical Trade portfolio. We initiated that recommendation on August 8th, 2017 when the spread was at 162bps. With the spread now at 255bps, we are now closing out that recommendation this week, taking a profit of 7% (inclusive of the gains from hedging the Bund exposure into U.S. dollars).2 At the same time, we feel that it is too early to position for a narrowing of the Treasury-Bund spread. The large U.S. fiscal stimulus will continue to put upward pressure on U.S. bond yields over the next year, both through higher U.S. inflation and the associated need for tighter Fed policy. Already, the Treasury-Bund spread reflects both the relatively larger dearth of spare capacity in the U.S. economy (Chart 11) and the expected widening of the U.S. federal budget deficit compared to reduced deficits in the euro area (Chart 12). Much like the rise in the fair value of the Treasury-Bund spread, this suggests that there is limited downside for the spread on a more medium-term basis. Chart 11UST-Bund Spread Narrowing Will Be##BR##Limited By Faster U.S. Growth... UST-Bund Spread Narrowing Will Be Limited By Faster U.S. Growth... UST-Bund Spread Narrowing Will Be Limited By Faster U.S. Growth... Chart 12...The Result Of Looser##BR##U.S. Fiscal Policy ...The Result Of Looser U.S. Fiscal Policy ...The Result Of Looser U.S. Fiscal Policy We are taking profits on our tactical spread based on our read of all of our relevant indicators. There is a good chance, however, that we could consider re-entering a spread widening trade on any meaningful narrowing of the spread or adjustment in our indicators. Bottom Line: The fundamental drivers of the 10-year U.S. Treasury-German Bund spread continue to point to the spread staying wide over the next 6-12 months. Yet the spread has overshot our fair value estimates, and relative positive data surprises are turning more in favor of Europe. We are taking profits on our tactical UST-Bund spread widening trade, after a gain of 7% (hedged into U.S. dollars). A Quick Update On U.S. Treasury Market Technicals One of the overriding aspects of the U.S. Treasury market over the past few months has been the stretched technical backdrop. The combination of oversold price momentum, bearish sentiment and aggressive short positioning have helped keep yields in check, even as U.S. growth and inflation accelerate and the Fed continues to signal more future rate hikes. Back in March, we presented a study of previous episodes of an oversold U.S. Treasury market since the year 2000.3 Our goal was to determine how long it typically took for a resolution of oversold Treasury market conditions. Unsurprisingly, we concluded that the longest episodes of oversold Treasuries occurred when U.S. economic growth and core inflation were both accelerating, and vice versa. At the time of that report, all of the technical indicators that we looked at were signaling that Treasury bearishness was deeply entrenched (Chart 13). Now, four months later, there has been some change in those indicators: Chart 13UST Technical Indicators##BR##Are More Mixed Now UST Technical Indicators Are More Mixed Now UST Technical Indicators Are More Mixed Now The 10-year Treasury yield relative to its 200-day moving average: then, +43bps; now, +18bps The trailing 26-week total return of the Bloomberg Barclays U.S. Treasury index: then, -4.3%; now, -0.6% The J.P. Morgan client survey of bond managers and traders: then, very large underweight duration positioning; now, positioning is neutral The Market Vane index of bullish sentiment for Treasuries: then, near the bottom of the range since 2000; now, still near that same level The CFTC data on speculator positioning in 10-year U.S. Treasury futures: then, a large net short of -8% (scaled by open interest); now, still a large net short of -11%. Therefore, the message from the technical indicators is more mixed now than in March. Price momentum and duration positioning is now neutral, while sentiment and speculative positions remain stretched. The former suggests that there is scope for Treasury yields to begin climbing again, while the latter implies that there may still be room for some counter-trend short-covering Treasury rallies in the near term. In our March study, we defined the duration of each episode of an oversold Treasury market by the following conditions: The start date was when the 10-year Treasury yield was trading at least 30bps above its 200-day moving average and the Market Vane Treasury bullish sentiment index dipped below 50; The end-date was when the yield declined below its 200-day moving average. The details of each of those episodes can be found in Table 1. This is the same table that we presented back in March, but we have now added the current episode. At 150 days in length, this is already the fourth longest period of an oversold Treasury market since 2000. Yet perhaps most surprising is the fact that Treasury yields are essentially unchanged since the start date of the current episode (March 20th, 2018). There is no other period in our study that where yields did not decline while the oversold market resolved itself. Table 1A Look At Prior Episodes Of An Oversold U.S. Treasury Market The Bond Bear Market Is Not Over The Bond Bear Market Is Not Over Perhaps this can be interpreted as a sign that there is still scope for a final short-covering Treasury rally before this current oversold episode can truly end. Yet as we concluded in our March study, it took an average of 156 days for an oversold market to be fully corrected if U.S. growth was accelerating (i.e. the ISM manufacturing index was rising) and core PCE inflation were both rising at the same time - as is currently the case (Chart 14). Chart 14U.S. Growth/Inflation Backdrop Points To Yields Consolidating, Not Reversing U.S. Growth/Inflation Backdrop Points To Yields Consolidating, Not Reversing U.S. Growth/Inflation Backdrop Points To Yields Consolidating, Not Reversing The longest such episode in 2003/04 lasted for 203 days before the 10-year yield fell below its 200-day moving average. Yet the second longest episode (196 days) occurred in 2013/14, and Treasury yields ended up climbing to a new cyclical high before eventually peaking. Given the underlying positive momentum in both U.S. economic growth and inflation, but with a mixed message from the technical indicators, we suspect that this current oversold episode may have further to run. Yet as we concluded back in March, and still believe today, it will prove difficult to earn meaningful returns betting on a counter-trend decline in yields this time, as any such move will likely be modest in size and lengthy in duration. Bottom Line: Some of the oversold technical conditions in the U.S. Treasury market have turned more neutral, but sentiment remains very bearish and there are large speculative short positions. With both U.S. growth and inflation accelerating, we recommend sticking with a strategic below-benchmark U.S. duration stance rather than playing for a tactical short-covering Treasury rally. Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 Please see BCA Global Fixed Income Strategy Weekly Report, "The Trendless, Friendless Bond Market", dated July 17th 2018, available at gfis.bcaresearch.com. 2 The return on this trade is calculated using the Bloomberg Barclays 7-10-year government bond indices for the U.S. and Germany, adjusted for duration differences between the indices. The German return is hedged into U.S. dollars, as this trade was done on a currency-hedged basis. 3 Please see BCA Global Fixed Income Strategy Weekly Report, "Bond Markets Are Suffering From Withdrawal Symptoms", dated March 20th 2018, available at gfis.bcaresearch.com. Recommendations The GFIS Recommended Portfolio Vs. The Custom Benchmark Index The Bond Bear Market Is Not Over The Bond Bear Market Is Not Over Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Highlights Chart 1Corporate Health: Improving Everywhere, ##br##Down In The U.S. Corporate Health: Improving Everywhere, Down In The U.S. Corporate Health: Improving Everywhere, Down In The U.S. Dollar bull markets are often accompanied by positive returns for the S&P 500. While a strong dollar hurts the earnings outlook for the S&P 500, it supports an expansion of multiples by putting downward pressure on rates and elongating the U.S. business cycle. The dollar and stocks are most positively correlated when the U.S. yield curve slope is between zero and 50-basis points, and flattening. Today's environment fits this bill. BCA is neutral on U.S. in a balanced portfolio. While the USD's strength should be associated with rising U.S. equity prices, the quality of U.S. stock returns is deteriorating. This warrants a certain degree of de-risking relative to our former overweight stance. Feature For the past two weeks, we have warned investors that the dollar rally was over-extended, and that a correction was likely to ensue. However, we also argued that this correction was likely to prove a countertrend move, and that the dollar was likely to end the year at higher levels. BCA has a neutral stance on equities on both a cyclical and tactical horizon. BCA is also neutral on U.S. equities within a global equity portfolio. For investors, it becomes important to understand whether a stronger dollar constitutes an additional downside risk for stocks. This is especially relevant in the U.S., where equity valuations are comparatively elevated, and where corporate health is deteriorating relative to the rest of the world (Chart 1). In this report, we built on the research of our colleague Anastasios Avgeriou, who spearheads BCA's U.S. Equity Sector Strategy service, who has shown that the dollar and the S&P often do rise in unison.1 Ultimately, while the dollar can have an impact on the relative performance of the U.S., it is generally not a strong determinant of the trend in the S&P 500. Strong Dollar And The S&P: Good Friends Indeed A picture is worth a thousand words. As Chart 2 illustrates, a strong dollar has never really been enough to slay a bull run in the S&P 500. Between late 1978 and early 1985, the real trade-weighted dollar rallied by 45%, yet the S&P 500 was able to advance by 102%. Between 1995 and 2002, the real trade-weighted dollar increased by 33% but rallied by nearly 92%. If one were to confine their observations to 1995 to August 2000 window, the dollar would have been up 16.5% and the S&P an outstanding 223%. Finally, from its most recent cyclical bottom in 2011 to the end of 2016, the trade-weighted dollar rallied by 22%, but the S&P 500 managed to rise by another impressive 68%. It is true that the magnitude of the strength of U.S. equities in the face of a strong dollar has decreased over time. This essentially reflects the fact that in the early 1980s, 20% of S&P 500 revenues were garnered outside the U.S. versus roughly 40% today, which in turn has increased the drag on earnings created by a stronger dollar. This problem is illustrated by the negative relationship present between the dollar and U.S. earnings revisions (Chart 3). Chart 2Strong Dollar, No Problem Strong Dollar, No Problem Strong Dollar, No Problem Chart 3Dollar Is Dangerous For The Earnings Outlook Dollar Is Dangerous For The Earnings Outlook Dollar Is Dangerous For The Earnings Outlook Yet, despite this negative link between earnings revisions and the dollar, the S&P can still rise when the dollar increases. What explains this seeming paradox? The answer is almost tautological: It is multiples. A strong dollar tends to be associated with a rising P/E ratio. This is because a strong dollar has a dampening impact on inflation. As a result, when the dollar rises, the Federal Reserve can keep interest rates lower than would otherwise be the case, fomenting periods of declining bond yields (Chart 4). Thanks to lower bond yields, not only do multiples get a boost, but additionally the domestically driven U.S. economic cycle also gets elongated. This further helps stocks in the process. Another more international dimension helps explain the positive correlation between stocks and the dollar. The dollar tends to experience its strongest rallies when U.S. growth is superior to that of the rest of the G-10. As Chart 5 illustrates, the bulk of the early 1980s dollar rally, of the late 1990s rally, and of the 2011 to early 2017 rally materialized when U.S. economic activity was outperforming. In all these instances, the relative strength of the U.S. economy attracted funds from abroad. This also meant that foreign funds flowing into the U.S. economy bolstered liquidity in the U.S. economy. Not only did this liquidity support economic activity, thereby counterbalancing the drag created by a stronger dollar, these funds also found their way into asset markets, generating higher multiples in the U.S. in the process. Chart 4Strong Dollar Hurts Yields Strong Dollar Hurts Yields Strong Dollar Hurts Yields Chart 5Growth Differentials Matter For The Dollar Growth Differentials Matter For The Dollar Growth Differentials Matter For The Dollar Bottom Line: A strong dollar in and of itself has never been enough to derail a bull market in the S&P 500. While a strong dollar creates a hurdle for foreign earnings accruing to U.S. firms, higher multiples often compensate for this negative. Essentially, a higher dollar causes downside to bond yields, warranting lower hurdle rates and higher valuations. Moreover, a stronger dollar diminishes inflationary pressures in the U.S., warranting easier Fed policy than would otherwise be the case. Since the U.S. economy is domestically driven, this elongates the business cycle, helping stocks in the process. Correlation And The Yield Curve Slope While a strong dollar does not seem to be a death threat for the equity market, are there environments when the dollar and the S&P 500 are more correlated than others? Table 1Dollar Versus S&P 500 Correlation: ##br##A Function Of The Yield Curve The S&P Doesn't Abhor A Strong Dollar The S&P Doesn't Abhor A Strong Dollar The answer to this question is yes. As Table 1 illustrates, the correlation between the dollar and the S&P 500 fluctuates significantly based on both the slope of the yield curve and whether the yield curve is flattening or not. Interestingly, when the yield curve is steep (defined as greater than a 50-basis-point spread between 10-year and 2-year Treasury yields), the dollar and U.S. stock prices tend to move in opposite directions. However, when the yield curve is flatter but before it has yet to invert (a yield curve slope of between zero and 50 basis points), the correlation between the dollar and the S&P 500 changes: it becomes positive. In fact, the time at which the correlation between stocks and the dollar is the highest is when the yield curve slope is in that zone and is also flattening. This is surprising, but at the same time it makes sense. We know that when the yield curve is flat but not inverted, the stock market tends to still rally (Chart 6). However, this flattening yield curve indicates that monetary conditions are not as accommodative as they once were. Interestingly, while the dollar performs poorly in the early innings of a monetary tightening campaign, it performs much better when monetary conditions are not so easy anymore that they juice up global growth, but they are not yet tight enough to cause an imminent recession in the U.S.2 This corresponds to a an environment with a flatter yield curve that has yet to invert, such as the one in place today. In light of these observations, the close correlation between the S&P 500 and the dollar in this environment should not be very surprising. Chart 6Flat And Flattening: No Problem For Stocks Flat And Flattening: No Problem For Stocks Flat And Flattening: No Problem For Stocks Bottom Line: The dollar and the stock market are not always positively correlated. However, when the U.S. yield curve slope stands between zero and 50 basis points and is flattening, the positive correlation between the S&P 500 and the dollar is at its strongest. This defines today's environment. Investment Implications BCA thinks the U.S. dollar has ample downside on a long-term basis. After all, the U.S. dollar trades at a significant premium to its PPP fair value, and this kind of overvaluation historically indicates significant downside for the greenback on a multi-year time horizon (Chart 7). Moreover, the Trump administration's fiscal policy is likely to result in a widening of both the fiscal and current account deficits. While a twin deficit rarely impacts the dollar negatively, so long as U.S. real rates rise relative to the rest of the world, it nonetheless often ends up being a harbinger of long-term weakness in the greenback.3 It is hard to make any inference for the S&P 500 based on a bearish long-term dollar view as historically, during a structural dollar bear market, the relationship between the greenback and the S&P has been rather ambiguous. However, BCA also thinks the 2018 dollar rally is not over. As Chart 8 shows, when U.S. rates are in the top of the distribution of interest rates among G-10 economies, the dollar tends to perform well. The U.S.'s status as the global high-yielder is currently unchallenged. This suggests the dollar has a natural advantage over other currencies through the remainder of the year. Chart 7Long-Term Downside For The Dollar... Long-Term Downside For The Dollar... Long-Term Downside For The Dollar... Chart 8...But 2018 Rally Is Not Over ...But 2018 Rally Is Not Over ...But 2018 Rally Is Not Over Moreover, as the U.S. economy is less exposed to the global industrial cycle than the rest of the world is, the U.S. dollar will benefit from the softening global economic environment. This is even truer, given that the U.S. economy was already set to outperform other G-10 economies even before the soft patch in global trade began. As a result, long-term flows into the U.S. are strong, which is generating a basic balance-of-payments surplus (Chart 9). American investors are not blind to this reality; the higher expected rate of returns on U.S. projects along with U.S. corporations bringing earnings back home to take advantage of the Trump tax cuts is generating outsized repatriation flows into the country, historically a good correlate of a strong dollar (Chart 10). This phenomenon is likely to remain alive through the remainder of the year. Chart 9Money Is Making Its Way Into The U.S. Money Is Making Its Way Into The U.S. Money Is Making Its Way Into The U.S. Chart 10Americans Like Their Dollar Americans Like Their Dollar Americans Like Their Dollar Since the U.S. yield curve slope currently stands between zero and 50 basis points while it is flattening in response to the Fed's interest rate hikes, we are in the part of the cycle where the dollar and stocks are positively correlated, and where they in fact often rise together. This suggests the S&P 500 has more upside ahead for the rest of the year as well. It is important to note that the tech sector is now the most at risk from the dollar strength as it has the largest percentage of foreign sales (Chart 11). However, BCA is neutral on stocks on a cyclical horizon. This is not because stocks will not be able to eke out some positive returns; it is because we are acutely aware that we stand close to the end of the bull market. Moreover, the end of an equity bull market is often marked by a pick-up in volatility. Accordingly, risk-adjusted returns for U.S. equities are declining. Hence, while an underweight stance on stocks is not yet warranted, a neutral stance is appropriate as we believe that it is better to be early and leave some money on the table than to be late.4 There remains a big risk that could cause the dollar to rally and stocks to fall, despite where we stand in the cycle: trade disputes. As Chart 12 illustrates, since May, tariff announcements and protectionist pronouncements have buoyed the dollar. However, the same announcements ultimately represent a real risk to profits as they create a real danger for global supply chains and imply higher cost of goods sold by U.S. corporations. Investors should monitor these risks closely. Chart 11S&P 500: Aggregate Sector International Revenue Exposure (%) The S&P Doesn't Abhor A Strong Dollar The S&P Doesn't Abhor A Strong Dollar Chart 12While Tariffs Can Help The Dollar, ##br##They Will Not Help Stocks While Tariffs Can Help The Dollar, They Will Not Help Stocks While Tariffs Can Help The Dollar, They Will Not Help Stocks Bottom Line: BCA anticipates the dollar to be able to rise over the course of the next six to nine months, as U.S. rates are in favor of the greenback and domestic growth outperformance will continue to favor inflows into the U.S. This bullish view on the U.S. dollar currently does not constitute a reason to downgrade stocks to underweight. In fact, at this stage of the cycle, U.S. stocks and the dollar tend to rise in unison. However, since the quality of the equity gains is likely to deteriorate as equity volatility is on an uptrend, BCA prefers to maintain a neutral cyclical stance on equities within a balanced portfolio rather than an overweight stance. Mathieu Savary, Vice President Foreign Exchange Strategy mathieu@bcaresearch.com 1 Please see U.S. Equity Sector Strategy Insight Report, titled "Can the S&P 500 Continue Rising Alongside the U.S. Dollar?", dated October 13, 2016, available at uses.bcaresearch.com 2 Please see Foreign Exchange Strategy Weekly Report, titled "This Time Is NOT Different," dated May 25 2018, available at fes.bcaresearch.com 3 Please see Foreign Exchange Strategy Weekly Report, titled "Twin Deficits: Bearish Or Not, The Fed Holds The Trump Card," dated February 23 2018, available at fes.bcaresearch.com 4 Please see The Bank Credit Analyst Special Report, titled "U.S. Fiscal Policy: An Unprecedented Macro Experiment," dated June 28, 2018 available at bcaresearch.com. Trades & Forecasts Forecast Summary Core Portfolio Tactical Trades Closed Trades

Dear Client, Geopolitical analysis is a fundamental part of the investment process. My colleague, and BCA's Chief Geopolitical Strategist, Marko Papic will introduce a one-day specialized course - Geopolitics & Investing - to our current BCA Academy offerings. This special inaugural session will take place on September 26 in Toronto and is available, complimentary, only to those who sign up to BCA's 2018 Investment Conference. The course is aimed at investors and asset managers and will emphasize the key principles of our geopolitical methodology. Marko launched BCA's Geopolitical Strategy (GPS) in 2012. It is the financial industry's only dedicated geopolitical research product and focuses on the geopolitical and macroeconomic realities which constrain policymakers' options. The Geopolitics & Investing course will introduce: The constraints-based methodology that underpins BCA's Geopolitical Strategy; Best-practices for reading the news and avoiding media biases; Game theory and its application to markets; Generating "geopolitical alpha;" Manipulating data in the context of political analysis. The course will conclude with two topical and market-relevant "war games," which will tie together the methods and best-practices introduced in the course. We hope to see you there. Click here to join us! Space is limited. Ryan Swift, Vice President U.S. Bond Strategy Highlights Chart 1Inflations Expectations Hard To Shake Inflations Expectations Hard To Shake Inflations Expectations Hard To Shake Low inflation expectations are proving difficult to shake. Year-over-year core PCE inflation moved to within 5 bps of the Fed's 2% target in May, but long-maturity TIPS breakeven inflation rates barely budged (Chart 1). Instead, breakevens are taking cues from commodity prices which are being held down by flagging global growth (bottom panel). The minutes from the June FOMC meeting revealed that "one participant" advocated postponing rate hikes in an attempt to re-anchor inflation expectations, but we do not expect the Fed to pursue this course. Instead, the Fed will continue to lift rates at a pace of 25 bps per quarter until a risk-off episode in financial markets prompts a delay, hoping that the incoming inflation data are strong enough to send TIPS breakevens higher in the meantime. Ultimately we think that strategy will be successful, but Fed hawkishness in the face of weakening global growth threatens the near-term performance of corporate credit. We recommend only a neutral allocation to spread product versus Treasuries, while maintaining a below-benchmark duration bias. Feature Investment Grade: Neutral Chart 2Investment Grade Market Overview Investment Grade Market Overview Investment Grade Market Overview Investment grade corporate bonds underperformed the duration-equivalent Treasury index by 60 basis points in June, dragging year-to-date excess returns down to -181 bps. Value is no longer stretched in the investment grade corporate bond market, though it is not attractive enough to compensate for being in the late stages of the credit cycle or for the looming collision between a hawkish Fed and decelerating global growth. These factors led us to reduce exposure to corporate bonds two weeks ago.1 With inflation running close to the Fed's 2% target and the 2/10 Treasury slope between 0 bps and 50 bps, our research shows that small positive excess returns are the best case scenario for corporate bonds. The likelihood that leverage will rise in the second half of this year is also a concern (Chart 2). Profit growth is only just keeping pace with debt growth and will soon have to contend with rising wage costs and the drag from recent dollar strength. The Fed's staunch hawkishness in the face of decelerating global growth is reminiscent of 2015. Then, the end result was a period of spread widening that culminated in the Fed pausing its rate hike cycle. In recent weeks we also explored how to position within the investment grade corporate bond sector, considering both the maturity spectrum and the different credit tiers.2 We concluded that in the current environment investors should favor long maturities and maintain a balanced or slightly up-in-quality bias (Table 3). Table 3ACorporate Sector Relative Valuation And Recommended Allocation* The Deflationary Mindset The Deflationary Mindset Table 3BCorporate Sector Risk Vs. Reward* The Deflationary Mindset The Deflationary Mindset High-Yield: Neutral Chart 3High-Yield Market Overview High-Yield Market Overview High-Yield Market Overview High-Yield outperformed the duration-equivalent Treasury index by 40 basis points in June, bringing year-to-date excess returns up to +76 bps. The average index option-adjusted spread widened 1 bp on the month, and currently sits at 365 bps. Our measure of the excess spread available in the High-Yield index after accounting for expected default losses has widened to 262 bps, just above its long-run mean (Chart 3). This tells us that if default losses during the next 12 months are in line with our expectations, we should expect excess high-yield returns of 262 bps over duration-matched Treasuries, assuming also that there are no capital gains/losses from spread tightening/widening. However, we showed in last week's report that the default loss expectations embedded in our calculation are extremely low relative to history (panel 4).3 Our assumption, derived from the Moody's baseline default rate forecast and our own forecast of the recovery rate, calls for default losses of 1.03% during the next 12 months. The only historical period to show significantly lower default losses was 2007, a time when corporate balance sheets were in much better shape than they are today. While most indicators suggest that default losses will in fact remain low for the next 12 months, historical context clearly demonstrates that the risks to that forecast are to the upside. It will be critically important to track real-time indicators of the default rate such as job cut announcements, which remain low relative to history but have perked up in recent months (bottom panel), for signals about whether current default forecasts are overly optimistic. MBS: Neutral Chart 4MBS Market Overview MBS Market Overview MBS Market Overview Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 3 basis points in June, bringing year-to-date excess returns up to -24 bps. The conventional 30-year zero-volatility MBS spread widened 1 bp on the month, driven entirely by a 1 bp widening of the option-adjusted spread (OAS). The compensation for prepayment risk (option cost) held flat. The MBS option-adjusted spread has widened since the beginning of the year (Chart 4), though by much less than the investment grade corporate bond spread (panel 3). The year-to-date OAS widening has been offset by a contraction in the option cost component of spreads, and this has kept the overall nominal MBS spread flat at very tight levels (bottom panel). Going forward, rising interest rates will limit mortgage refinancing activity and this will ensure that MBS spreads remain low. In other words, while MBS valuation is not attractive, the downside is limited. Our Bond Maps show an unfavorable risk/reward trade-off in the MBS sector. This analysis, based on volatility-adjusted breakeven spreads, shows that only 7 days of average spread widening are required for the MBS sector to lose 100 bps versus duration-matched Treasuries. While this speaks to the low spread buffer built into current MBS valuations, the message from the Bond Map must be weighed against the macro outlook which suggests that the odds of significant spread widening are quite low. Government-Related: Underweight Chart 5Government-Related Market Overview Government-Related Market Overview Government-Related Market Overview The Government-Related index outperformed the duration-equivalent Treasury index by 5 basis points in June, bringing year-to-date excess returns up to -35 bps. Sovereign debt outperformed the Treasury benchmark by 33 bps on the month, bringing year-to-date excess returns up to -210 bps. Foreign Agencies outperformed by 10 bps on the month, bringing year-to-date excess returns up to -46 bps. Local Authorities underperformed by 9 bps on the month, dragging year-to-date excess returns down to +28 bps. Supranationals outperformed by 5 bps on the month, bringing year-to-date excess returns up to +7 bps. Domestic Agency bonds underperformed by 7 bps, dragging year-to-date excess returns down to zero. The escalating tit-for-tat trade war and increasing divergence between U.S. and non-U.S. growth is a clear negative for USD-denominated Sovereign debt. Relative valuation also shows that U.S. corporate bonds are more attractive than similarly rated Sovereigns (Chart 5). Maintain an underweight allocation to Sovereign debt. Within the universe of Emerging Market Sovereign debt, we showed in a recent report that only Russian debt offers an attractive spread relative to the U.S. corporate sector.4 In contrast, the Foreign Agency and Local Authority sectors continue to offer a favorable risk/reward trade-off compared to other fixed income sectors (please see the Bond Maps). Maintain overweight allocations to these two sectors. The Bond Maps also show that the Supranational and Domestic Agency sectors are very low risk, but offer feeble return potential compared to other sectors. The Supranational and Domestic Agency sectors should be avoided. Municipal Bonds: Overweight Chart 6Municipal Market Overview Municipal Market Overview Municipal Market Overview Municipal bonds outperformed the duration-equivalent Treasury index by 10 basis points in June, bringing year-to-date excess returns up to +120 bps (before adjusting for the tax advantage). The average Aaa-rated Municipal/Treasury yield ratio fell 1% in June to reach 85%, close to one standard deviation below its post-crisis mean. It is also only slightly higher than the average 81% level that was observed in the late stages of the previous cycle, between mid-2006 and mid-2007. The technical picture remains favorable for Muni / Treasury yield ratios. Fund inflows increased in recent weeks, and visible supply has contracted substantially compared to this time last year (Chart 6). State & local government credit fundamentals are also fairly robust. Net borrowing is on the decline and this should ensure that municipal ratings upgrades continue to outpace downgrades (bottom panel). Despite relatively tight valuation compared to history, the Total Return Bond Map on page 16 shows that municipal bonds offer a fairly attractive risk/reward trade-off compared to other U.S. fixed income sectors, particularly for investors exposed to the top marginal tax rate. Given the favorable reading from our Bond Map and the steadily improving credit fundamentals, we recommend an overweight allocation to Municipal bonds. Treasury Curve: Favor 7-Year Bullet Over 1/20 Barbell Chart 7Treasury Yield Curve Overview Treasury Yield Curve Overview Treasury Yield Curve Overview The Treasury curve bear-flattened in June. The 2/10 Treasury slope flattened 10 bps and the 5/30 slope flattened 7 bps. At present, the 2/10 slope sits at 29 bps, its flattest level of the cycle. The yield curve has flattened relentlessly in recent months as the impact of Fed rate hikes at the short-end of the curve have not been offset by higher inflation expectations at the long end. As explained in a recent report, we think it is unlikely that curve flattening can maintain this rapid pace.5 At 2.34%, the 1-year Treasury yield is already priced for 100 bps of Fed rate hikes during the next 12 months, assuming no term premium. Meanwhile, long-maturity TIPS breakeven inflation rates remain below levels that are consistent with the Fed's 2% inflation target. While curve flattening will proceed as the Fed lifts rates, higher breakeven inflation rates at the long-end of the curve will offset some flattening pressure during the next few months. With that in mind, we continue to recommend a position long the 7-year bullet and short the duration-matched 1/20 barbell. According to our models, this butterfly spread currently discounts 41 bps of 1/20 curve flattening during the next six months (Chart 7). This is considerably more than what is likely to occur. Table 4 of this report shows the output from our valuation models for each butterfly combination across the entire yield curve, as explained in a recent Special Report.6 Table 4Butterfly Strategy Valuation (As Of July 6, 2018) The Deflationary Mindset The Deflationary Mindset TIPS: Overweight Chart 8Inflation Compensation Inflation Compensation Inflation Compensation TIPS outperformed the duration-equivalent nominal Treasury index by 35 basis points in June, bringing year-to-date excess returns up to +129 bps. The 10-year TIPS breakeven inflation rate increased 4 bps on the month and currently sits at 2.12%. The 5-year/5-year forward TIPS breakeven inflation rate increased 5 bps and currently sits at 2.16% (Chart 8). Both the 10-year and 5-year/5-year TIPS breakeven inflation rates remain below the range of 2.3% to 2.5% that has historically been consistent with inflation expectations that are well-anchored around the Fed's 2% target. We expect breakevens will return to that target range as investors become increasingly convinced that the risk of deflation has faded. Consistent inflation prints at or above the Fed's 2% target will be the deciding factor that eventually leads to this upward re-rating of inflation expectations. In that regard, the current outlook is promising. Core PCE inflation has printed above the 0.17% month-over-month level that is consistent with 2% annual inflation in four of the past five months (panel 4). Year-over-year trimmed mean PCE inflation is at 1.84% and should continue to rise based on the 2.03% reading from the 6-month trimmed mean PCE (bottom panel). Finally, our Pipeline Inflation Indicator continues to point toward mounting inflationary pressures in the economy (panel 3). Maintain an overweight allocation to TIPS relative to nominal Treasury securities. We will reduce exposure to TIPS once long-maturity TIPS breakeven inflation rates return to our 2.3% to 2.5% target range. ABS: Neutral Chart 9ABS Market Overview ABS Market Overview ABS Market Overview Asset-Backed Securities outperformed the duration-equivalent Treasury index by 1 basis point in June, bringing year-to-date excess returns up to -2 bps. The index option-adjusted spread for Aaa-rated ABS widened 2 bps on the month and now stands at 43 bps, 16 bps above its pre-crisis low. The Bond Maps show that consumer ABS continue to offer relatively attractive return potential compared to other low-risk spread products. However, we maintain only a neutral allocation to this space because credit quality trends are moving against the sector. The household debt service ratio on consumer credit ticked down slightly in the first quarter, but its multi-year uptrend remains intact (Chart 9). Consumer credit delinquency rates follow the household debt service ratio with a lag. Meanwhile, banks are noticing the decline in credit quality and have begun tightening lending standards (bottom panel). Tighter lending standards tend to coincide with upward pressure on delinquencies and spreads. Non-Agency CMBS: Underweight Chart 10CMBS Market Overview CMBS Market Overview CMBS Market Overview Non-Agency Commercial Mortgage-Backed Securities underperformed the duration-equivalent Treasury index by 11 basis points in June, dragging year-to-date excess returns down to +61 bps. The index option-adjusted spread for non-agency Aaa-rated CMBS widened 4 bps on the month and currently sits at 74 bps (Chart 10). The gap between decelerating commercial real estate prices and tight CMBS spreads continues to send a worrying signal for CMBS (panel 3). However, delinquencies continue to decline and banks recently started to ease lending standards on nonfarm nonresidential loans (bottom panel). Agency CMBS: Overweight Agency CMBS underperformed the duration-equivalent Treasury index by 6 basis points in June, dragging year-to-date excess returns down to +6 bps. The index option-adjusted spread widened 2 bps on the month and currently sits at 51 bps. The Bond Maps show that Agency CMBS offer high potential return compared to other low risk spread products. An overweight allocation to this defensive sector continues to make sense. The BCA Bond Maps The following page presents excess return and total return Bond Maps that we use to assess the relative risk/reward trade-off between different sectors of the U.S. fixed income market. The Maps employ volatility-adjusted breakeven spread/yield analysis to show how likely it is that a given sector will earn/lose money during the subsequent 12 months. The Maps do not impose any macroeconomic view. The Excess Return Bond Map The horizontal axis of the excess return Bond Map shows the number of days of average spread widening required for each sector to lose 100 bps versus a position in duration-matched Treasuries. Sectors plotting further to the left require more days of average spread widening and are therefore less likely to see losses. The vertical axis shows the number of days of average spread tightening required for each sector to earn 100 bps in excess of duration-matched Treasuries. Sectors plotting further toward the top require fewer days of spread tightening and are therefore more likely to earn 100 bps in excess of Treasuries. The Total Return Bond Map The horizontal axis of the total return Bond Map shows the number of days of average yield increase required for each sector to lose 5% in total return terms. Sectors plotting further to the left require more days of yield increases and are therefore less likely to lose 5%. The vertical axis shows the number of days of average yield decline required for each sector to earn 5% in total return terms. Sectors plotting further toward the top require fewer days of yield decline and are therefore more likely to earn 5%. Chart 11Excess Return Bond Map (As Of July 6, 2018) The Deflationary Mindset The Deflationary Mindset Chart 12Total Return Bond Map (As Of July 6, 2018) The Deflationary Mindset The Deflationary Mindset Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Jeremie Peloso, Research Analyst jeremiep@bcaresearch.com 1 Please see U.S. Bond Strategy Special Report, "Go To Neutral On Spread Product", dated June 26, 2018, available at usbs.bcaresearch.com 2 Please see U.S. Bond Strategy Weekly Report, "Rigidly Defined Areas Of Doubt And Uncertainty", dated June 19, 2018, for further details on positioning across different credit tiers. Please see U.S. Bond Strategy Weekly Report, "Out Of Sync", dated July 3, 2018, for further details on positioning across the maturity spectrum. Both reports available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, "Out Of Sync", dated July 3, 2018, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, "Threats & Opportunities In Emerging Markets", dated June 12, 2018, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, "Rigidly Defined Areas Of Doubt And Uncertainty", dated June 19, 2018, available at usbs.bcaresearch.com 6 Please see U.S. Bond Strategy Special Report, "More Bullets, Barbells And Butterflies", dated May 15, 2018, 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 Monetary Policy: Position for rate hikes of 25 bps per quarter for the next 6-12 months and watch nominal GDP growth, cyclical spending and the price of gold for signals about the position of the fed funds rate relative to its equilibrium level. Yield Curve: Curve flattening will proceed as the Fed lifts rates, but some flattening pressure will be mitigated by the re-anchoring of long-dated inflation expectations. Against this back-drop, and given currently attractive valuations, a position long the 7-year bullet and short the duration-matched 1/20 barbell makes the most sense. IG Credit: Moving down-in-quality has a greater positive impact on the risk-adjusted performance of a credit portfolio when excess return volatility and index duration-times-spread are low. At present, down-in-quality allocations within investment grade credit are only marginally attractive. Feature "You just let the machines get on with the adding up," warned Majikthise, "and we'll take care of the eternal verities, thank you very much. [...] "That's right," shouted Vroomfondel, "we demand rigidly defined areas of doubt and uncertainty!" - The Hitchhiker's Guide To The Galaxy, By Douglas Adams Jerome Powell put his stamp on Fed communications at last week's FOMC meeting. He trimmed 100 words from the policy statement and began his post-meeting press conference with a concise "plain-English" summary of how the economy is doing. In short: "the economy is doing very well". But while he expressed confidence in the Fed's assessment of the economy, he was also keen to point out areas where the outlook is cloudier. His central theme seemed to be that we must delineate between those questions that can be addressed by the Fed's reading of the economic data and those that are better left to the philosophers in Douglas Adams' novel. The Chairman stressed the uncertainty surrounding two concepts in particular: the non-accelerating inflation rate of unemployment (NAIRU) and the neutral (or equilibrium) interest rate, even advising that "we can't be too attached to these unobservable variables." But what can we say about these traditionally important policy guideposts? And more importantly, how should we think about them when formulating an investment strategy? The Importance Of NAIRU Chart 1The Fed's Projections The Fed's Projections The Fed's Projections One issue that came up repeatedly in the Chairman's press conference was the seeming disconnect between the Fed's labor market projections and its inflation projections. The Fed expects the unemployment rate to fall far below NAIRU during the next two years, and yet it anticipates only a mild overshoot of its inflation target (Chart 1).1 Ultimately this disconnect will be resolved in one of two ways. Either the Fed is underestimating the inflation pressures that will result from running the unemployment rate so far below NAIRU and will be forced to hike rates more quickly than anticipated, or it will eventually revise its estimate of NAIRU downward. From an investment perspective, this disconnect will only matter if inflation starts to rise more quickly than anticipated and the Fed is forced to ramp up the pace of rate hikes. We discussed this possibility in a recent report and concluded that, on a 6-12 month horizon, the odds of the Fed hiking more quickly than its current 25 bps per quarter pace are low.2 This is principally because the Fed will likely tolerate a fairly substantial overshoot of its inflation target before it feels the need to tighten more quickly. The Importance Of The Neutral Rate For bond investors the theoretical concept of the neutral (or equilibrium) interest rate is much more important. This interest rate represents the threshold between accommodative and restrictive monetary policy. When the fed funds rate is above neutral we should expect the pace of economic growth to slow and inflation pressures to dissipate. At present, the majority of FOMC participants estimate that the neutral fed funds rate is between 2.75% and 3%. At the Fed's current 25 bps per quarter pace, the funds rate will reach neutral by the middle of next year (Chart 2). Chart 2The Federal Funds Rate Will Hit Neutral Next Year The Federal Funds Rate Will Hit Neutral Next Year The Federal Funds Rate Will Hit Neutral Next Year The important question for investors is whether the Fed will start to slow its rate hike pace at that time, or whether it will revise its estimate of the neutral rate based on trends in the economy. Chairman Powell's emphasis on uncertainty makes us lean toward the latter. In a recent report we outlined three factors to monitor that will help us determine whether monetary policy is accommodative (fed funds rate below neutral) or restrictive (fed funds rate above neutral).3 The first factor is the year-over-year growth rate in nominal GDP relative to the fed funds rate (Chart 3). Historically, the year-over-year growth rate in nominal GDP falling below the fed funds rate is a reliable (though often lagging) signal that monetary policy has turned restrictive. A more leading signal of restrictive monetary policy is the proportion of nominal GDP that comes from the most cyclical (or interest rate sensitive) sectors of the economy. Those sectors being consumer spending on durable goods, residential investment and investment on equipment & software. When cyclical spending declines as a proportion of overall growth it is often a sign that the fed funds rate is above its neutral level (Chart 3, panel 2). Finally, we also recommend monitoring the price of gold for clues about the neutral rate of interest. Gold tends to appreciate when the stance of monetary policy becomes more accommodative and depreciate when it becomes more restrictive. The steep decline in the gold price between 2013 and 2016 even preceded downward revisions to the Fed's estimate of the neutral rate (Chart 4). Going forward, an upside breakout in the price of gold would be a signal that we should revise our estimate of the neutral fed funds rate higher. Conversely, a large decline would suggest that monetary policy is turning restrictive and we should think about calling the cyclical peak in bond yields. Chart 3Tracking The Neutral Rate I Tracking The Neutral Rate I Tracking The Neutral Rate I Chart 4Tracking The Neutral Rate II Tracking The Neutral Rate II Tracking The Neutral Rate II Bottom Line: Rather than rely on current estimates of unobservable variables like NAIRU and the neutral rate of interest, investors should monitor developments in the economy and consider how those estimates might evolve over time. For now, investors should expect a rate hike pace of 25 bps per quarter and watch nominal GDP growth, cyclical spending and the price of gold for signals about the position of the fed funds rate relative to its equilibrium level. Gradualism And The Slope Of The Curve The Fed's fairly explicit guidance that rates will rise by 25 bps per quarter is quite helpful when formulating expectations about the slope of the yield curve. For example, we know that the current 1-year par coupon Treasury yield of 2.35% is priced for exactly 100 bps of rate hikes during the next 12 months with no term premium. In other words, investors today should be indifferent between an investment in cash and an investment in a 1-year Treasury note if they are 100% certain that the Fed will stick to its 25 bps per quarter hike pace for the next 12 months. We can also forecast where the 1-year Treasury yield will be six months from now under a few different scenarios (Table 1). The forward curve is consistent with a 1-year Treasury yield of 2.69% six months from now, and we calculate that it will be 2.83% if the market moves to fully discount a rate hike pace of 25 bps per quarter until the end of 2019. If the market only prices in the Fed's median funds rate projection, which calls for three hikes in 2019, then the 1-year Treasury yield will be between 2.62% and 2.81% six months from now, depending on which meetings in 2019 those three rate hikes are delivered. Table 1Forecasting The 1-Year Treasury Yield Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty The main takeaway from these observations is that even in the most hawkish scenario the 1-year Treasury yield will only rise to 2.83%. This is 48 bps above its current level and a mere 14 bps more than what is already priced into the forward curve. Now let's consider the long-end of the curve. The 10-year and 20-year TIPS breakeven inflation rates currently sit at 2.12% and 2.10%, respectively. If inflation expectations become re-anchored around the Fed's 2% target during the next six months, which we expect they will, then both of these rates will reach a range between 2.3% and 2.5% (Chart 5). This alone will apply between 20 bps and 40 bps of upward pressure to the 20-year Treasury yield. The nominal 20-year Treasury yield is currently 2.98% and the forward curve is priced for it to rise to 3.01% in six months. In the most hawkish scenario where the Fed lifts rates 25 bps per quarter and long-maturity yields remain constant, the 1/20 Treasury slope will flatten by 48 bps during the next six months. In the more likely scenario where Fed rate hikes coincide with the re-anchoring of long-dated inflation expectations, the 1/20 slope will flatten by 28 bps or less. Meanwhile, our model of the 1/7/20 butterfly spread shows that it is priced for 55 bps of 1/20 flattening during the next six months (Chart 6). Or put differently, there is so much extra yield pick-up in the 7-year bullet relative to the duration-matched 1/20 barbell that being long the bullet and short the barbell will be profitable unless the 1/20 slope flattens by more than 55 bps. Chart 5Inflation Expectations Are Still Too Low Inflation Expectations Are Still Too Low Inflation Expectations Are Still Too Low Chart 6Butterfly Spread Fair Value Model Butterfly Spread Fair Value Model Butterfly Spread Fair Value Model Bottom Line: Curve flattening will proceed as the Fed lifts rates, but some flattening pressure will be offset by the re-anchoring of long-dated inflation expectations. Against this back-drop, and given currently attractive valuations, a position long the 7-year bullet and short the duration-matched 1/20 barbell makes the most sense. Risk Update On May 22 we initiated a tactical long duration position premised on extended net short positioning in the bond market and the high likelihood of negative near-term data surprises.4 We have seen considerable movement in our indicators during the past two weeks - positioning is now much closer to neutral (Chart 7) and our model no longer expects data surprises to turn negative (Chart 8). Therefore, this week we remove our tactical long duration recommendation. The biggest current risk to our below-benchmark duration stance is the large divergence that has opened up between U.S. growth and the rest of the world (Chart 9). This divergence is putting upward pressure on the U.S. dollar and, much like in 2015, is starting to hurt growth in emerging markets, as we discussed last week. Chart 7Bond Market Positioning Bond Market Positioning Bond Market Positioning Chart 8Data Surprises Should Remain Positive Data Surprises Should Remain Positive Data Surprises Should Remain Positive Chart 9Foreign Growth Is The Greatest Risk Foreign Growth Is The Greatest Risk Foreign Growth Is The Greatest Risk But dollar strength and emerging market weakness is not an imminent threat to higher U.S. yields. Using the 2015 experience as a template, we see in Chart 9 that U.S. yields did not fall until after emerging market financial conditions and global growth had already troughed. In fact, it was not until dollar strength and weak global growth culminated in a dramatic tightening of U.S. financial conditions that the Fed finally signaled a slower pace of rate hikes and Treasury yields declined (Chart 9, bottom panel). Similarly, we don't think the Fed will react to a strong dollar and weak foreign growth until the impact is felt by U.S. risk assets. With U.S. growth still elevated and the dollar having appreciated only modestly so far, we think Treasury yields will avoid this risk during the next few months. Nonetheless, the divergence between U.S. and foreign growth is a risk that bears close monitoring. We will not hesitate to alter our duration stance if the dollar continues to appreciate and the divergence appears close to a breaking point. The Best Time To Move Down In Quality In last week's report we reviewed our assessment of where we stand in the credit cycle. That assessment determines whether we should be overweight or underweight investment grade corporate bonds relative to a duration-equivalent position in Treasuries. This week we zero-in on our allocation to investment grade corporate bonds and consider how we should allocate between the different credit tiers (Aaa, Aa, A and Baa). In next week's report we will look at positioning across the different maturity buckets and industries. We begin our analysis with the four Bond Maps presented in Charts 10-13. These Bond Maps show risk-adjusted return potential on the y-axis. Specifically, the number of months of average spread tightening necessary to achieve the excess return threshold listed in each map's title. The risk-adjusted potential for losses is shown on the x-axis. In this case, it shows the number of months of average spread widening required to underperform Treasuries by the amount listed in the title. Chart 10Investment Grade Corporate Excess Return Bond Map:##br## +/- 50 BPs Threshold Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty Chart 11Investment Grade Corporate Excess Return Bond Map: ##br##+/- 100 BPs Threshold Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty Chart 12Investment Grade Corporate Excess Return Bond Map: ##br##+/- 200 BPs Threshold Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty Chart 13Investment Grade Corporate Excess Return Bond Map:##br## +/- 300 BPs Threshold Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty Credit tiers plotting closer to the bottom-left of the Bond Maps have less potential for return and less risk. Credit tiers plotting closer to the upper-right have greater potential for return and more risk. What we find particularly interesting is that when we set a low return threshold, such as +/- 50 bps, the credit tiers plot almost right on top of each other. In other words, an allocation to Baa-rated corporate bonds gives you a much greater chance of earning 50 bps with about the same risk of losing 50 bps as the other credit tiers. But as we increase the excess return threshold the risk/reward trade-off between the different credit tiers becomes more linear. In Chart 13 we see that Baa-rated bonds have a greater chance of earning 300 bps than the other credit tiers, but also carry a significantly greater risk of losing 300 bps. Chart 14Down-In-Quality Works ##br##Best When Vol Is Low Down-In-Quality Works Best When Vol Is Low Down-In-Quality Works Best When Vol Is Low This leads to an interesting conclusion. A macro environment where we would expect low excess return volatility is also one where moving down in quality within investment grade corporate bonds is most beneficial from a risk/reward perspective. Conversely, moving down in quality will improve the risk-adjusted performance of your portfolio by less (and might even hurt the risk-adjusted performance of your portfolio) in a highly volatile return environment. To test this theory, we first recognize that the excess return volatility of the investment grade corporate bond index is tightly linked with its duration-times-spread (DTS). Low DTS environments have lower excess return volatility, and also less of a spread differential between the lower and higher credit tiers (Chart 14). With this in mind we split the historical time series of monthly corporate bond excess returns into four quartiles based on the index DTS (Table 2). We also exclude recessions from our sample, meaning this analysis is only valid during periods of economic recovery. Not surprisingly, the results show that the standard deviation of monthly excess returns increases alongside index DTS. But we also see that the average return advantage in the Baa-rated credit tier is lower when the index DTS is higher. Table 2Investment Grade Corporate Bond Excess Returns By Credit Tier (1989-Present)* Rigidly Defined Areas Of Doubt And Uncertainty Rigidly Defined Areas Of Doubt And Uncertainty When the index DTS is between 3 and 4.5, the reward/risk ratio in the Baa-rated credit tier exceeds the average of the other three credit tiers by 0.13. This advantage falls to 0.07 when the DTS is between 4.5 and 6.7; and falls further to 0.04 when the DTS is between 6.7 and 9.7. In the highest DTS quartile, the Baa-rated credit tier provides a lower reward/risk ratio than the average of the other three credit tiers. At present the index DTS is 8.4. This puts us in the second highest quartile relative to history, and is consistent with a 12-month standard deviation of monthly excess returns of roughly 77 bps for the corporate bond index. In this environment we should expect down-in-quality allocations to positively impact the risk-adjusted performance of a credit portfolio, but not by as much as in lower DTS environments. Bottom Line: Moving down-in-quality has a greater positive impact on the risk-adjusted performance of a credit portfolio when excess return volatility and index duration-times-spread are low. At present, down-in-quality allocations within investment grade credit are only marginally attractive. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com 1 In order to display a longer history, Chart 1 shows the Congressional Budget Office's estimate of NAIRU rather than the Fed's. At present both estimates are very close. The CBO estimates NAIRU to be 4.65% and the Fed's median projection calls for a NAIRU of 4.5%. 2 Please see U.S. Bond Strategy Weekly Report, "Breaking Points", dated May 29, 2018, available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, "A Signal From Gold?", dated May 1, 2018, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, "Pulling Back And Looking Ahead", dated May 22, 2018, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Highlights After having written about the role of the U.S. yield curve in forecasting recessions, we are devoting this Special Report to addressing the widely asked question on the effectiveness of the yield curve in determining asset allocation. A naïve, rules-based approach is applied to the yield curve in each of seven countries/regions to produce a dynamic allocation signal between equities and bonds in each country/region. Despite its simplicity, we find that the dynamic portfolio systematically outperforms the 60/40 equity/bond benchmark portfolio in the U.S., Canada, euro area, Switzerland, U.K. and Australia from a long-term perspective (four years), with Japan being the outlier. Despite the dominance of the U.S. in the global economy and also in global asset markets, the equity/bond performance cycle outside the U.S. does not necessarily follow the U.S. Instead, the yield curve in each country provides a consistently better signal than just following U.S. decisions alone. Currently, signals from yield curves still favor equities over bonds. Feature U.S. yield curve inversion has been a good leading indicator for recessions in the U.S. Since the mid-1950s, every U.S. recession has been preceded with curve inversion, as shown in Chart 1. The lead time, however, varies from one month to 18 months. In addition, even though it is true that stocks underperform bonds in a recession, stocks can begin to underperform bonds long before a recession starts and can also continue to underperform long after a recession ends. For example, U.S. stocks/bonds performance ratio peaked in December 1999 and then troughed in September 2002 with a more than 50% drawdown, yet only about 6% occurred between March 2001 and November 2001 - the NBER official dates for the 2001 recession. So could information from the U.S. yield curve itself systematically add value to a stock-bond allocation decision in the U.S.? Even if it could in the U.S., could the same apply elsewhere, given that yield curves in different countries do not move in a synchronized fashion? (Chart 2) Chart 1U.S. Yield Curve Vs. Recession U.S. Yield Curve Vs. Recession U.S. Yield Curve Vs. Recession Chart 2Global Yield Curve Cycle Global Yield Curve Cycle Global Yield Curve Cycle In this Special Report, we use a simplified naïve, rules-based approach to attempt to demonstrate if information from yield curves in seven countries - the U.S., Japan, the U.K, Euro Area, Canada, Australia and Switzerland - can systematically add value in asset allocation decisions. Yield Curves Are An Effective Indicator For Long-Term Asset Allocation The test results are quite encouraging, despite the simplicity and need for further refinement. Except in Japan, yield curves in all six other countries provide value-add information for stock-bond allocation decisions. The solid lines in Chart 3 are the relative total return performance of the active stock/bond portfolio versus the benchmark for each country. The active portfolio is simply constructed based on a naïve rule such that a 10% underweight is given to equities and a 10% overweight is given to bonds when the yield curve reaches the lower band from above. Once the yield curve reaches the upper band from below, the allocation is reversed. The upper and lower bands are explained in our methodology section on page 5, we omit Japan from these charts because, as explained on page 9, its stock/bond ratio has not had a consistent relationship with the yield curve. The dash lines in Chart 3 are the monthly four-year rolling return differentials between the active portfolios and the benchmarks. It is encouraging to see that the four-year rolling performance in each country has suffered only very limited downside. Chart 4 is the same as Chart 3 except that the active bet is maxed out to 40% over- or underweight relative to the 60/40 equity/bond benchmark - i.e. when the signal is bullish for stocks, 100% is in stocks, and when it is bullish for bonds, the weights are 80% bonds and 20% stocks. This is a more extreme version of risk-taking, though the upside/downside trade-off is still quite impressive. This simple approach illustrates that in the long run, the yield curve is a useful indicator for equity/bond allocations. However, it does not do very well on a shorter-term time horizon. As shown in Chart 5, the one-year performance differentials are less appealing. Chart 3Backtest Base Case Backtest Base Case Backtest Base Case Chart 4Backtest Aggressive Case Backtest Aggressive Case Backtest Aggressive Case Chart 5Short-Term Risk Reward Less Appealing Short-Term Risk Reward Less Appealing Short-Term Risk Reward Less Appealing So how are the back tests conducted? The Methodology The Passive Benchmark: A 60/40 fixed-weight equity/bond benchmark is constructed for each country using the MSCI equity total return index and Bloomberg/Barclays Treasury Total Return Index, all in local currencies. The Active Allocation Rule: For each country, a range is set for its yield curve with an upper band and a lower band. The bands are set based on yield curve cycles and also their correlation with stock/bond performance cycles. When the curve reaches the upper band from below, an overweight is assigned to equities until the yield curve reaches the lower band from above, at which point the overweight then shifts to bonds. To determine how the size of the over- and underweight positions impacts the efficacy of the signal, we tested four different bet sizes - from 10% to 40% - in 10% increments, since no short selling is allowed. Objective: The active portfolio in each country is aimed to outperform its passive benchmark with a minimal four-year rolling drawdown. The same approach is applied to all seven countries. In terms of yield curve, the 3M/10 curve works better than the 2/10 curve for the U.S. because the former has better cyclicality. For all other countries, 2/10 yield curves are used. Despite the simplicity of our approach, some interesting observations are worth highlighting: U.S. And Canada: Reduce Risk When Yield Curve Inverts As shown in Chart 6, yield curve inversion in these two countries has historically been a good indication to reduce risk in equities. Bonds in general start to outperform equities after the curve is inverted and continue to do so as the yield curve steepens. However, when the curves steepens near to its cyclical high, then it's time to add risk in equities. Historically, the upper threshold for the U.S. 3M/10 is 3.4%, while for the Canadian 2/10 it is 1.8%. Currently, this indicator alone still favors equities in these two countries. Chart 6AU.S. & Canada: Curve Inversion ##br##Triggers Risk Reduction (I) U.S. & Canada: Curve Inversion Triggers Risk Reduction (I) U.S. & Canada: Curve Inversion Triggers Risk Reduction (I) Chart 6BU.S. & Canada: Curve Inversion ##br##Triggers Risk Reduction (II) U.S. & Canada: Curve Inversion Triggers Risk Reduction (II) U.S. & Canada: Curve Inversion Triggers Risk Reduction (II) Euro Area And Switzerland: Reduce Risk Before Yield Curve Approaches Inversion As shown in Chart 7, the yield curve of the euro area does not invert often, while the Swiss curve has never gone into inversion during the short period for which we have historical data. However, both curves have good cyclicality, which makes the 0.2%-1.8% range works very well for both. Chart 7AEuro Area & Swiss: Reduce Risk##br## Before Curve Inverts (I) Euro Area & Swiss: Reduce Risk Before Curve Inverts (I) Euro Area & Swiss: Reduce Risk Before Curve Inverts (I) Chart 7BEuro Area & Swiss: Reduce Risk ##br##Before Curve Inverts (II) Euro Area & Swiss: Reduce Risk Before Curve Inverts (II) Euro Area & Swiss: Reduce Risk Before Curve Inverts (II) U.K And Australia: Reduce Risk After Yield Curve Has Inverted 2/10 yield curves in both the U.K. and Australia invert more often than in other countries. However, unlike other countries, equities can continue to outperform bonds even after the curve is inverted. The turning point is about minus 50 basis points, as shown in Chart 8. The upper band for Australia is 1.25% and 0.9% for the U.K. Chart 8AU.K. & Australia: Reduce Risk ##br##After Yield Curve Has Inverted (I) U.K. & Australia: Reduce Risk After Yield Curve Has Inverted (I) U.K. & Australia: Reduce Risk After Yield Curve Has Inverted (I) Chart 8BU.K. & Australia: Reduce Risk ##br##After Yield Curve Has Inverted (II) U.K. & Australia: Reduce Risk After Yield Curve Has Inverted (II) U.K. & Australia: Reduce Risk After Yield Curve Has Inverted (II) Japan: Yield Curve Does Not Provide Consistent Information The Japanese stock/bond ratio does not have a consistent relationship with the 2/10 yield curve, as shown in Chart 9. This makes it very difficult to apply the simple approach employed here. Country Divergence U.S. economic cycles have been widely studied. But as shown in Chart 1, correctly identifying recessions in the U.S. does not systematically capture equity/bond relative performance cycles because even U.S. equities can underperform bonds before a recession starts and after a recession ends. Using the yield curve, on the other hand, does a much better job in capturing the equity/bond performance cycle in each country. Chart 10 shows that investors in different countries should pay more attention to local yield curve cycles other than just following a U.S.-centric analysis, even though the U.S. does play a dominant role in the global economy and in global equity and bond indices. Chart 9Japan Is The Outlier Japan Is The Outlier Japan Is The Outlier Chart 10Country Divergences Country Divergences Country Divergences Bottom Line: The yield curve is an effective indicator for equity/bond allocation in most developed countries from a long-run perspective. Currently, yield curve-based signals from the U.S., Canada, Euro Area, Switzerland, the U.K. and Australia all still favor equities over bonds. Xiaoli Tang, Associate Vice President xiaoliT@bcaresearch.com

Highlights 0 To 3 Months: Extended net short positioning and the recent moderation in economic data suggest that Treasury yields are ripe for a near-term pullback. Investors who are able should consider tactically buying bonds on a 0-3 month horizon, but with a tight stop loss. 6 to 12 Months: We recommend that investors maintain below-benchmark portfolio duration on a 6-12 month horizon, consistent with our Two Stage Bond Bear Market framework. While the credit cycle is in its late stages, it is still too soon to reduce exposure to corporate bonds. We will pare exposure to corporate bonds once our TIPS breakeven inflation targets are met. Total Return Forecasts: Our simple framework for estimating total bond returns reveals that risk/reward arguments clearly favor below-benchmark portfolio duration on a 12-month horizon. Feature Chart 1Two Milestones Two Milestones Two Milestones The U.S. bond market reached one noteworthy milestone last week and is quickly closing in on another. The first milestone is that the 10-year Treasury yield decisively broke through the 3% level that had defined its most recent peak (Chart 1). The second milestone is that the market is now close to fully pricing-in the likely near-term path for Fed rate hikes. We noted in a recent report that the Fed's "gradual" rate hike path is quite clearly defined as one 25 basis point rate hike per quarter.1 This equates to 100 bps on our 12-month Fed Funds Discounter, which currently sits at 91 bps, just below this key level (Chart 1, bottom panel). We continue to see upside in Treasury yields on a cyclical horizon. Though tactically, the likelihood of a near-term pullback in yields has increased greatly during the past few days. In this week's report we outline the case for a near-term (0-3 month) pullback in Treasury yields, but also look ahead by introducing a simple framework investors can use to make total return forecasts for all different U.S. bond sectors. The Case For A Near-Term Pullback In addition to the fact that the market is closer to fully discounting the likely near-term path of rate hikes than it has been for some time, there are two other reasons to expect a near-term, temporary pullback in yields. The first is that the below-benchmark duration trade has become the consensus position in the market (Chart 2). Net speculative short positions in 10-year Treasury futures have rarely been greater, and since the financial crisis large net short positions have correlated quite strongly with a decline in the 10-year yield during the subsequent three months. Similarly, positions reported in the JP Morgan Duration Survey are firmly in "net short" territory for both the "all clients" and "active clients" surveys. The Marketvane survey of bond sentiment has also turned bearish for only the fourth time since 2010. Each of the other three times has coincided with a near-term drop in yields. Chart 2Bond Market Looks Oversold Bond Market Looks Oversold Bond Market Looks Oversold But positioning alone would not be enough to convince us that yields might decline in the near-term. Investors also need a catalyst. An excuse to take profits on large net short positions that have been working well. That catalyst is typically a period of worse-than-expected economic data. To judge the trend in economic data relative to expectations we turn to the Economic Surprise Index. Chart 3Economic Surprise Index Economic Surprise Index Economic Surprise Index In a report from last year we demonstrated that if the Economic Surprise Index ends a month below (above) the zero line, it is very likely that Treasury yields fell (rose) during that month.2 Also, we know that the surprise index is mean reverting by its very nature. A long period of positive (negative) data surprises will certainly be followed an upward (downward) revision to investors' economic expectations. Eventually expectations become so elevated (depressed) that they become impossible to surpass (disappoint). The index will then start to mean revert. In that same report from last year we also introduced a simple auto-regressive model of the surprise index, designed to capture its average speed of mean reversion. Based on that model, which is purely a function of the index's own lags, we would expect the surprise index to dip slightly into negative territory in one month's time (Chart 3). Though given the large amount of uncertainty in the model, a fairer assessment would be that it is no longer a given that the surprise index will remain above the zero line in the near-term. Bottom Line: Extended net short positioning and the recent moderation in economic data suggest that Treasury yields are ripe for a near-term pullback. Investors who are able should consider tactically buying bonds on a 0-3 month horizon, but with a tight stop loss. Less nimble investors are better off riding out any potential near-term volatility and maintaining below-benchmark portfolio duration on a 6-12 month horizon. The Cyclical Picture Is Unchanged On a 6-12 month investment horizon, we are sticking with the playbook of our Two-Stage Bond Bear Market.3 The first stage is characterized by the re-anchoring of inflation expectations, and here, long-maturity TIPS breakeven inflation rates are still slightly below our target range of 2.3% to 2.5% (Chart 4). We also think bond investors should maintain an overweight allocation to spread product, though the time to trim exposure is approaching. Because the Fed's support for credit markets will weaken as inflation pressures mount, we will start reducing exposure to spread product once both the 10-year and 5-year/5-year forward TIPS breakeven inflation rates are within our target 2.3% to 2.5% band. The intuition that the credit cycle is long in the tooth is further supported by the fact that the 2/10 Treasury curve is close to 50 bps (Chart 4, bottom panel). In a recent report we showed that while corporate bond excess returns relative to Treasuries usually remain positive until the yield curve inverts, they decline dramatically once the slope dips below 50 bps.4 Valuation also remains tight in the corporate bond market. While investment grade corporate bond spreads have widened in recent months, the junk spread is still close to its post-crisis low, as is the differential between the junk and investment grade spread (Chart 5). Chart 4Inflation Compensation Inflation Compensation Inflation Compensation Chart 5Flirting With The Lows Flirting With The Lows Flirting With The Lows The recent widening of investment grade corporate spreads appears to simply reflect a reversion to more reasonable valuation levels, after they had been extremely expensive at the start of the year. Chart 6 shows the 12-month breakeven spread for each investment grade credit tier. We look at the breakeven spread - defined as the spread widening required to lose money versus Treasuries on a 12-month horizon - in order to adjust for the changing duration of the index over time. Chart 6 also shows the breakeven spread as a percentile rank relative to history. In other words, it shows the percentage of time that the breakeven spread has been lower in the past. Notice that earlier in the year investment grade corporate spreads had been approaching all-time expensive levels. They are now closer to the 25th percentile, much more in line with similar spreads for the High-Yield credit tiers (Chart 7). Chart 6Investment Grade Breakeven Spreads Investment Grade Breakeven Spreads Investment Grade Breakeven Spreads Chart 7High-Yield Breakeven Spreads High-Yield Breakeven Spreads High-Yield Breakeven Spreads There is no longer a risk-adjusted opportunity in high-yield corporate bonds relative to investment grade. Bottom Line: We recommend that investors maintain below-benchmark portfolio duration on a 6-12 month horizon, consistent with our Two Stage Bond Bear Market framework. While the credit cycle is in its late stages, it is still too soon to reduce exposure to corporate bonds. We will pare exposure to corporate bonds once our TIPS breakeven inflation targets are met. A Simple Framework For Forecasting Total Returns In a recent report we observed that, using a 12-month investment horizon, the difference between market expectations for the change in the federal funds rate and the actual change in the federal funds rate closely tracks the price return from the Bloomberg Barclays Treasury index.5 With that in mind, this week we extend that analysis to develop a simple framework for forecasting bond total returns. The framework relies on the fact that the "12-month rate hike surprise" described above is correlated with the 12-month change in Treasury yields. The Appendix to this report shows the historical correlation between the 12-month rate hike surprise and the 12-month change in several different par-coupon Treasury yields. Unsurprisingly, the correlation is very strong for short maturity yields, and gradually weakens as we move further out the curve. This is important because it means that the total return forecasts we generate from this exercise will be more accurate for bond sectors with low duration than for those with high duration. Table 1 shows the total return forecasts we generated for the Bloomberg Barclays Treasury Master Index and for several of its maturity buckets. The results are presented in such a way that readers can impose their own forecasts for the number of Fed rate hikes that will occur during the next 12 months, and then map that forecast to a reasonable expectation for Treasury total returns. Table 1Treasury Index Total Return Forecasts Pulling Back And Looking Ahead Pulling Back And Looking Ahead For example, in a scenario where the Fed lifts rates four times (100 bps) during the next year, given current market pricing the rate hike surprise will be modestly negative.6 Using the historical correlations shown in the Appendix, we map that rate hike surprise to changes in the par-coupon Treasury curve and then use the duration and convexity attributes of each individual index to determine how that shift in the Treasury curve will impact index returns. In the scenario described above we would expect the Treasury Master Index to return +2.13% during the next year. While this is a slightly positive number, it is close enough to zero that it does not provide much insulation from changes in long-dated yields that are unrelated to the near-term path for rate hikes. Further, in the four rate hike scenario, investors moving from the Treasury Master Index to the 1-3 year index need only sacrifice 12 bps of expected return to reduce their duration risk by a factor of three. Such a risk/reward trade-off clearly favors a below-benchmark duration stance on a 12-month investment horizon. Table 2 repeats the same exercise but for the major spread sectors of the U.S. bond market. To estimate spread sector total returns we need to forecast both the shift in the Treasury curve and whether spreads will widen, tighten or remain constant. Specifically, we assume that spreads either widen or tighten by the standard deviation of annual spread changes for each index, calculated using a post-crisis interval. Table 2Spread Product Total Return Forecasts Pulling Back And Looking Ahead Pulling Back And Looking Ahead The results show that, in a four rate hike scenario, we should expect 12-month investment grade corporate bond total returns of approximately 3.4%, assuming also that spreads stay flat. In a scenario where the average index spread widens by 42 bps, we should expect total returns of only 1%. Bottom Line: Our simple framework for estimating total bond returns reveals that risk/reward arguments clearly favor below-benchmark portfolio duration on a 12-month horizon. Spread product returns should continue to beat Treasuries for the time being, but the window for outperformance is starting to close. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Appendix Chart 8Change In 1-Year Yield Vs. 12-Month ##br## Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 9Change In 2-Year Yield Vs. 12-Month ##br## Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 10Change In 3-Year Yield Vs. 12-Month ##br##Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 11Change In 5-Year Yield Vs.12-Month ##br##Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 12Change In 7-Year Yield Vs. 12-Month ##br##Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 13Change In 10-Year Yield Vs. 12-Month ##br##Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead Chart 14Change In 30-Year Yield Vs. 12-Month ##br## Fed Funds Rate Surprise Pulling Back And Looking Ahead Pulling Back And Looking Ahead 1 Please see U.S. Bond Strategy Portfolio Allocation Summary, "Coming To Grips With Gradualism", dated May 8, 2018, available at usbs.bcaresearch.com 2 Please see U.S. Bond Strategy Weekly Report, "How Much Higher For Yields?", dated October 31, 2017, available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, "A Signal From Gold?", dated May 1, 2018, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, "Back To Basics", dated April 17, 2018, available at usbs.bcaresearch.com 6 The 12-month rate hike surprise is defined as the 12-month Fed Funds Discounter less the actual change in the fed funds rate during the following 12 months. Fixed Income Sector Performance Recommended Portfolio Specification

Highlights Butterfly Trades: Duration-neutral butterfly trades are the best way to gain pure exposure to changes in the slope of the yield curve while remaining insulated from parallel shocks. Yield Curve Models: In this report we present models for each different butterfly spread combination across the entire Treasury curve. The models allow us to pinpoint the most attractively valued parts of the yield curve at any given point in time. We also demonstrate how trading rules based on our valuation models have delivered excellent investment results. Current Curve Valuation: Our models show that the most attractively valued butterfly spread at the moment is a position long the 7-year bullet and short the 1/20 barbell. We recommend closing our current position long the 5-year bullet and short the 2/10 barbell, and shifting into the 7-year over 1/20. Feature Last summer we published a Special Report that explained why duration-neutral butterfly trades are the best way to gain exposure to changes in the slope of the yield curve.1 That report focused on the 2/5/10 butterfly spread, which is defined as the spread between the 5-year Treasury note and a barbell consisting of the 2-year and 10-year notes. For this method to work the 2-year and 10-year notes must be weighted so that the dollar duration (DV01) of the 2/10 barbell matches the DV01 of the 5-year bullet.2 Chart 1Butterfly Strategy Valuation More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies The report demonstrated how, when using the above weighting scheme, a long position in the 5-year bullet versus a short position in the 2/10 barbell allows investors to profit from a steepening of the 2/10 Treasury slope while remaining insulated from small parallel yield curve shocks. Similarly, we showed that investors who want to gain exposure to 2/10 curve flattening should go long the 2/10 barbell and short the 5-year bullet. The report also presented a fair value model for the 2/5/10 butterfly spread based on the 2/10 slope. The model allows us to incorporate initial valuation into our yield curve trading framework. For example, while the 5-year bullet will tend to outperform the 2/10 barbell when the 2/10 slope is steepening, it will require very little 2/10 steepening for it to outperform when the 5-year appears cheap on our model. More 2/10 steepening is required when the 5-year is initially expensive. In this follow-up Special Report we extend the above modeling framework to all different segments of the yield curve. The results of our analysis, shown in Chart 1, allow us to quickly scan the entire Treasury curve and identify which butterfly combinations are most attractively valued. We can then consider the message from our valuation models alongside our macro view of how the slope of the yield curve will evolve. These two factors together will suggest appropriate butterfly trades to implement. This Special Report proceeds in three sections. The first section provides a quick re-cap of the theory of butterfly trades with a focus on the importance of valuing butterfly spreads relative to the slope. The second section explains the process we followed to extend our 2/5/10 butterfly model to the rest of the yield curve. The final section presents the results of two trading rules based on the read-out from our yield curve models. Butterfly Theory Revisited: The Importance Of Valuation In our report from last year we showed that, because both the bullet and barbell have the same DV01, a position long one and short the other is immune from small parallel yield curve shifts. However, because the longest maturity bond contributes more DV01 to the barbell than the short maturity bond, the barbell will underperform (outperform) the bullet when the curve steepens (flattens). This dynamic also means that the butterfly spread - defined as the bullet yield over the barbell yield - is positively correlated with the slope of the curve (Chart 2). The logic of this relationship depends on the fact that the yield curve tends to mean revert over time. A steep yield curve implies that it is more likely to flatten in the future. This means that when the curve is steep investors will demand greater compensation to enter trades that profit from further steepening. The bullet yield will therefore be bid up relative to the barbell. This is the relationship we exploit to create our yield curve models. Chart 2The Butterfly Spread And Slope Are Positively Correlated The Butterfly Spread And Slope Are Positively Correlated The Butterfly Spread And Slope Are Positively Correlated Trade Performance When The Butterfly Spread Is At Fair Value For example, let's consider the 2/5/10 butterfly spread once more. Our analysis shows that the butterfly spread is fairly valued when it is 0.14 times the slope of the 2/10 curve. The "first scenario" in Table 1 shows hypothetical returns to a position that is long the 5-year bullet and short the 2/10 barbell in four different yield curve scenarios. All four scenarios assume that the 2/5/10 butterfly spread is always fairly valued relative to the 2/10 slope (i.e. it is equal to 0.14 multiplied by the 2/10 slope). Table 1Hypothetical Butterfly Trade Performance More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies Notice that the bullet outperforms the barbell in both scenarios where the 2/10 slope steepens and underperforms in both scenarios where the 2/10 slope flattens. It does not matter whether yields move higher or lower, only changes to the slope of the curve impact returns. Trade Performance When The Butterfly Spread Deviates From Fair Value Next, let's consider the "second scenario" shown in Table 1. Here we assume that the butterfly spread is initially different from its model-implied fair value and then reverts to fair value by the end of the investment horizon. Now, in the bear-steepening scenario the 5-year bullet actually underperforms the 2/10 barbell even though the yield curve steepens. This is because the 5-year bullet is initially expensive relative to the barbell. Notice that the 2/5/10 butterfly spread is initially only 4 bps. A fairly valued butterfly spread would have been 7 bps (0.14 * 50 bps). The point of this analysis is to demonstrate the importance of initial valuation. When the butterfly spread is initially below fair value, more curve steepening is necessary for the bullet to outperform the barbell. Similarly, the bottom half of Table 1 shows that when the butterfly spread is initially above fair value, more curve flattening is required for the barbell to outperform. Modeling The Entire Curve With that in mind, we decided to extend our simple modeling framework to every segment of the yield curve. Using par-coupon bond yields from the Federal Reserve we considered all possible butterfly combinations consisting of 1-year, 2-year, 3-year, 5-year, 7-year, 10-year, 20-year and 30-year Treasury securities. We then estimated models of each possible butterfly spread (bullet over barbell) versus the slope between the two maturities used in the barbell. Chart 3 shows that the effectiveness of these models varies considerably between the different butterfly combinations. Chart 31-Factor Model Adjusted R2 More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies To understand why some butterfly combinations are more easily modeled than others we need to rely on an alternative theory for the positive correlation between the butterfly spread and the slope. This theory relates to the fact that implied interest rate volatility is also highly correlated with the slope of the yield curve (Chart 4). The reasoning is fairly straightforward. Investors demand more compensation to bear duration risk when the economic outlook is more uncertain and interest rate volatility is higher. Greater volatility therefore causes investors to bid up the term premium embedded in long-maturity Treasury securities, leading to a steeper curve. The strong relationship between implied volatility and the slope of the yield curve is important because another property of DV01-matched butterfly trades is that the barbell always has greater convexity than the bullet. Elevated convexity is a desirable property when interest rate volatility is high, meaning that the side of the trade with lower convexity (the bullet) will need to offer a higher yield to entice investors when rate volatility is elevated and the yield curve is steep. The key point is that while the barbell has greater convexity than the bullet in every butterfly combination, some butterfly combinations have a greater difference in convexity between the bullet and barbell than others. Chart 5 shows that those butterfly combinations with a larger convexity difference between the bullet and barbell are more sensitive to changes in the slope of the curve, and are thus easier to model using our framework. Chart 4The Yield Curve ##br##And Volatility The Yield Curve And Volatility The Yield Curve And Volatility Chart 5Models Work Better When The ##br## Convexity Mismatch Is Large More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies Finally, because there are strong theoretical arguments for why the butterfly spread should be positively correlated with both the slope of the yield curve and interest rate volatility, we tried adding the MOVE index of implied rate volatility as a second independent variable in each of our yield curve models. We found that this second variable only materially improved the accuracy of the models for a handful of butterfly combinations: the 5/7/10, 5/7/30, 1/20/30, 2/20/30, 3/20/30, 5/20/30, 7/20/30 and 10/20/30. We will rely on two-factor models (using both the curve slope and the MOVE index) for those combinations, while using one-factor models (with the slope only) for the others. One advantage of using a model based only on the slope is that we can reverse the model to ask the question: What change in the slope is necessary in order for the butterfly spread to be considered "fairly valued" at its current level? By framing the valuation question in this context it is easier to link the message from our valuation models to our macro view on the yield curve. For example, our 2/5/10 butterfly spread model shows that the 5-year bullet is currently 6 bps cheap. Alternatively, we can also state that the 2/5/10 butterfly spread is priced for 32 bps of 2/10 flattening during the next six months (Chart 6).3 If we expect the 2/10 slope to flatten by more than what is discounted we should enter the barbell over the bullet. Conversely, if we think the slope will flatten by less than what is discounted we should favor the bullet. Chart 62/5/10 Butterfly Spread Fair Value Model 2/5/10 Butterfly Spread Fair Value Model 2/5/10 Butterfly Spread Fair Value Model Chart 7 shows the current valuation for every butterfly combination in this manner. Rather than showing whether the bullet is cheap/expensive relative to the barbell (as in shown in Chart 1), it shows what change in the slope between the two components of the barbell is currently being discounted by the butterfly spread. We omit the butterfly combinations that are modeled using both the slope and volatility from this exercise. Chart 7Discounted Slope Change During Next Six Months (BPs) More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies Performance Tests We performed two tests to see whether our suite of yield curve models adds value to the investment process. Test #1 First, we considered each butterfly combination individually and tested the following trading rule: When the bullet is more than 0.5 standard deviations cheap on our model, we go long the bullet and short the barbell. When the barbell is more than 0.5 standard deviations cheap on our model, we go long the barbell and short the bullet. If nether the bullet nor the barbell is more than 0.5 standard deviations cheap we take no position. The trades are re-balanced daily and tested on a horizon from 1988 to the present. The results of this first test are shown in Chart 8. Here we see the annualized excess returns earned from each butterfly combination over the course of the testing horizon. In Chart 9 we also show the average number of times per year that the above trading rule would have recommended switching between the bullet, barbell and taking no position. Chart 10 shows the average annualized excess return divided by the average number of annual position changes. Chart 8Trading Rule Annualized Excess Returns Since April 1988 (BPs) More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies Chart 9Average Number Of Trades Per Year More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies Chart 10Excess Return Per Trade (BPs) More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies While the test results are encouraging insofar as every combination delivers positive excess returns, we note that due to limits in the amount of historical data at our disposal, most of the back-test is performed in sample. Although our robustness checks suggest that the regression coefficients are fairly stable through time, so we expect the results to be replicable going forward. Chart 11Excess Returns Versus Model Fit More Bullets, Barbells And Butterflies More Bullets, Barbells And Butterflies We also observe that the performance is not equally distributed amongst the different curve models. In fact, we notice that the models with the best fit - and hence largest convexity mismatches between the bullet and barbell - deliver better results than models with worse fit (Chart 11). This is not very surprising, but it does reinforce that we should put more weight on the message from the valuation models with greater convexity mismatches than on those with smaller mismatches. Test #2 In practice, we would not recommend trying to implement every butterfly trade that appears cheap according to our models. Rather, the real power of our modeling framework is that we can choose the most attractive segment of the yield curve and implement that trade only - assuming it synchs up with our macro view of the yield curve. In our second performance test we did just that. Each month we chose the most attractively valued yield curve trade based on our models and implemented only that trade. Chart 12 shows that not only does that method deliver excellent excess returns over time, it also outperforms a benchmark where we take the average of all yield curve trades recommended by our models. Chart 12Test #2 Results Test #2 Results Test #2 Results At present, the most attractive butterfly trade according to our models is the 7-year bullet over the 1/20 barbell. This trade is directionally similar to our currently recommended position long the 5-year bullet over the 2/10 barbell, in that both will benefit from curve steepening (or less curve flattening than is currently priced). Given the more attractive value in the 7-year over 1/20 combination, we recommended investors shift their yield curve allocation away from the 2/5/10 butterfly to favor the 7-year bullet over the 1/20 barbell. Alex Wang, CFA, Senior Analyst alexw@bcaresearch.com Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com 1 Please see U.S. Bond Strategy Special Report, "Bullets, Barbells And Butterflies", dated July 25, 2017, available at usbs.bcaresearch.com 2 The dollar duration (DV01) is the dollar value of a basis point. It measures the dollar change in the price of a given bond assuming a one basis point change in yield. It is calculated as the bond's duration times its price, divided by 104. 3 We assume an investment horizon of 6 months, a length of time that approximates the average length of time it takes for the butterfly spread to revert to our model's fair value.

Highlights Chart 1Interest Rate Expectations Interest Rate Expectations Interest Rate Expectations Last week the Federal Reserve made some necessary tweaks to the language in its statement. Namely, with the year-over-year core PCE deflator now up to 1.88%, the Fed was forced to upgrade its assessment of inflation and note that it has "moved close" to the 2 percent target. To assuage concern that such a change might lead to a quicker pace of rate hikes, the statement also emphasized that the inflation target is "symmetric" and noted that its policy of "gradual increases in the federal funds rate" will continue. While the recent increase in inflation is not sufficient to nudge the Fed away from "gradualism", the more important observation is that yields are still not high enough to discount the Fed's gradual approach (Chart 1). The Fed has tightened policy once per quarter since December 2016, tapering asset purchases in place of a rate hike in September 2017. It should be obvious that, absent an economic shock, one rate hike per quarter is the Fed's definition of "gradual". And yet, the market is still priced for barely more than two hikes for the balance of 2018, and not even two rate hikes for all of 2019! Maintain a below-benchmark duration stance until the market comes to grips with the Fed's gradualism. Feature Investment Grade: Overweight Chart 2Investment Grade Market Overview Investment Grade Market Overview Investment Grade Market Overview Investment grade corporate bonds outperformed the duration-equivalent Treasury index by 4 basis points in April, bringing year-to-date excess returns up to -77 bps. The Corporate index option-adjusted spread tightened somewhat in the first half of April, but widened anew during the past couple of weeks and recently made a new high for the year. Despite this sell-off, valuation remains expensive for investment grade corporates. The 12-month breakeven spread for an A-rated bond has only been tighter 27% of the time since 1989 (Chart 2). The same measure for a Baa-rated bond has only been tighter 28% of the time. We are preparing to cyclically scale back our corporate bond exposure, and will start the process once TIPS breakeven inflation rates reach our target range, signaling that monetary conditions are sufficiently restrictive. Our target range is 2.3% to 2.5% for both the 10-year and 5-year/5-year forward TIPS breakeven inflation rates. Those rates currently sit at 2.16% and 2.23%, respectively. In a recent report we noted that corporate bond excess returns fall sharply once the 2/10 Treasury yield curve flattens to below 50 bps, though they typically remain positive until the curve actually inverts.1 The 2/10 Treasury slope currently sits at 45 bps. That same report also notes that while the outlook for corporate revenue growth is strong, rising employee compensation costs will likely soon put a dent in profit margins and cause gross leverage to resume its uptrend (panel 4). This will apply further widening pressure to spreads later in the year. Table 3ACorporate Sector Relative Valuation And Recommended Allocation* Coming To Grips With Gradualism Coming To Grips With Gradualism Table 3BCorporate Sector Risk Vs. Reward* Coming To Grips With Gradualism Coming To Grips With Gradualism High-Yield: Overweight Chart 3High-Yield Market Overview High-Yield Market Overview High-Yield Market Overview High-Yield outperformed the duration-equivalent Treasury index by 121 basis points in April, bringing year-to-date excess returns up to 102 bps. The average index option-adjusted spread tightened 16 bps on the month, and currently sits at 343 bps. The 12-month trailing speculative grade default rate moved higher for the second consecutive month, hitting 3.92% in March. Moody's baseline forecast still calls for it to fall to 1.7% by March of next year. Based on Moody's default rate projection and our estimate of the recovery rate, we forecast High-Yield default losses of 0.85% for the next 12 months. This translates to a 12-month excess return of 257 bps for the High-Yield index versus Treasuries, assuming an unchanged junk spread (Chart 3). One hundred basis points of spread widening would lead to an excess return of -140 bps during this time horizon, and 100 bps of spread tightening would lead to an excess return of +654 bps. However, such a large spread tightening is almost certainly over-optimistic. As inflation continues to rise and the Fed applies the brakes, a floor will likely remain under the VIX index of implied equity volatility and this will prevent junk spreads from recovering their cyclical lows (top panel). This would be consistent with behavior typically seen late in the cycle, once the 2/10 Treasury slope flattens to below 50 bps.2 MBS: Neutral Chart 4MBS Market Overview MBS Market Overview MBS Market Overview Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 18 basis points in April, bringing year-to-date excess returns up to -22 bps. The conventional 30-year zero-volatility MBS spread tightened 4 bps on the month, split between a 1 bp tightening of the option-adjusted spread (OAS) and a 3 bps decline in the compensation for prepayment risk (option cost). While mortgages are no longer excessively cheap compared to corporate credit (Chart 4), we still see limited potential for spread widening during the next 6-12 months. Rising interest rates should serve to limit mortgage refinancing, and muted refis are closely linked to tight MBS spreads (bottom panel). We also view extension risk as relatively limited for conventional 30-year MBS. Using a model of excess MBS returns that we introduced in February, we estimate that despite the 25 bps increase in duration-matched Treasury yields that occurred in April, extension risk trimmed only 2 bps off monthly excess returns.3 Our excess return Bond Map also shows that conventional 30-year MBS require far fewer days of average spread tightening to earn 100 bps of excess return than most other Aaa-rated structured products (Non-Agency Aaa-rated CMBS being the exception), although they are also more likely to deliver losses. But given the benign refinancing back-drop, we remain reasonably positive on the sector.4 Government-Related: Underweight Chart 5Government-Related Market Overview Government-Related Market Overview Government-Related Market Overview The Government-Related index underperformed the duration-equivalent Treasury index by 9 basis points in April, dragging year-to-date excess returns down to -7 bps. Sovereign debt underperformed the Treasury benchmark by 37 bps on the month, while Foreign Agencies underperformed by 15 bps and Domestic Agencies underperformed by 14 bps. Local Authorities delivered 14 bps of outperformance and Supranationals bested duration-equivalent Treasuries by 5 bps. Dollar strength hurt the performance of Sovereign debt last month, and relative valuation continues to show that Sovereigns are expensive relative to similarly-rated U.S. corporate bonds (Chart 5). We remain underweight USD-denominated Sovereign debt. Conversely, Foreign Agencies and Local Authorities continue to offer very attractive spreads, especially considering the duration and spread volatility characteristics of those sectors. Our excess return Bond Map shows that both sectors offer a superior risk/reward trade-off than the Barclays Aggregate and almost all of its components.5 The large presence of state-owned energy companies in the Foreign Agency sector means it should also benefit from higher oil prices in the coming months. Municipal Bonds: Underweight Chart 6Municipal Market Overview Municipal Market Overview Municipal Market Overview Municipal bonds outperformed the duration-equivalent Treasury index by 65 basis points in April, bringing year-to-date excess returns up to 94 bps (before adjusting for the tax advantage). The average Aaa-rated Municipal/Treasury yield ratio declined 2% in April as fund inflows returned to the sector (Chart 6). Persistently low visible supply is also contributing to the strong technical environment for yield ratios. The tax-adjusted yield for a 10-year municipal bond is now about 46 bps below the yield offered by an equivalent-duration corporate bond. As we have shown in prior research, investors typically get an opportunity to shift out of corporates and into munis at a positive spread differential before the end of the cycle.6 We will await this more attractive entry point before aggressively shifting our allocation in favor of munis. In a recent report we noted that state and local governments are still working to repair their budgets.7 More states enacted tax increases than decreases in fiscal year 2018 and the projected nominal budget increase across all states is a paltry 2.3%. Fortunately, our Municipal Health Monitor indicates that the hard work is paying off, and suggests that ratings upgrades should continue to outpace downgrades for the time being (bottom panel). Treasury Curve: Favor 5-Year Bullet Over 2/10 Barbell Chart 7Treasury Yield Curve Overview Treasury Yield Curve Overview Treasury Yield Curve Overview The Treasury curve rose considerably in April, steepening a touch out to the 5-year maturity point and flattening thereafter. The 2/10 Treasury slope flattened 1 basis point in April, and currently sits at 45 bps. The 5/30 slope flattened 9 bps on the month and currently sits at 34 bps. The trade-off between the pace of Fed rate hikes on the one hand, and the re-anchoring of long-dated TIPS breakeven inflation rates on the other will dictate the slope of the yield curve during the next six months. With the 10-year TIPS breakeven inflation rate at 2.16%, it remains slightly below the range of 2.3% to 2.5% that is consistent with well-anchored inflation expectations. It will be difficult for the yield curve to flatten aggressively until that target is met. After that, curve flattening becomes much more likely. We continue to recommend a position in the 5-year bullet versus the duration-matched 2/10 barbell, primarily due to extremely attractive starting valuation. Our model suggests that the 2/5/10 butterfly spread is priced for 17 bps of 2/10 curve flattening during the next six months (Chart 7). With long-maturity TIPS breakevens still below target, we think that is too high a bar. TIPS: Overweight Chart 8TIPS Market Overview TIPS Market Overview TIPS Market Overview TIPS outperformed the duration-equivalent nominal Treasury index by 93 basis points in April, bringing year-to-date excess returns up to 161 bps. The 10-year TIPS breakeven inflation rate rose 12 bps on the month and currently sits at 2.16%. The 5-year/5-year forward TIPS breakeven inflation rate increased 6 bps and currently sits at 2.23%. As we explained in a recent report, we view the first stage of the bond bear market as being driven by the re-anchoring of inflation expectations.8 We will consider inflation expectations well anchored when both the 10-year and 5-year/5-year forward TIPS breakeven inflation rates are in a range between 2.3% and 2.5%, where they were the last time that inflation was well anchored around the Fed's target. If the recent trend in inflation continues, then this re-anchoring will occur relatively soon. The annualized 6-month rate of change in the trimmed mean PCE deflator has already returned to the Fed's target, and the annual rate of change jumped from 1.71% to 1.77% in March (Chart 8). Pipeline measures of inflation pressure also continue to strengthen. Our Pipeline Inflation Indicator is in a strong uptrend and the prices paid component of the ISM manufacturing survey is closing in on 80, a level last seen in 2011 (panel 4). ABS: Neutral Chart 9ABS Market Overview ABS Market Overview ABS Market Overview Asset-Backed Securities outperformed the duration-equivalent Treasury index by 13 basis points in April, bringing year-to-date excess returns up to -6 bps. The index option-adjusted spread for Aaa-rated ABS narrowed 4 bps on the month and now stands at 40 bps, 7 bps above its pre-crisis low. Our recently introduced excess return Bond Map shows that both Aaa-rated credit card and Aaa-rated auto loan ABS exhibit lower risk and less potential for gains than the Barclays Aggregate index.9 It also confirms that credit card ABS are somewhat more attractive than auto loan ABS, offering approximately the same potential for excess return with less risk. Compared to other fixed income sectors, Aaa-rated ABS offer greater potential return and higher risk than Agency CMBS, Domestic Agencies and Supranationals. But the ABS sector also has a less attractive risk/reward profile than the Foreign Agency, Local Authority and Investment grade corporate sectors. Fundamentally, while consumer delinquencies remain low, they are heading higher alongside a rising household debt service coverage ratio (Chart 9). The persistent (though mild) deterioration in credit quality causes us to maintain a neutral allocation to the sector, despite reasonably attractive valuations. Non-Agency CMBS: Underweight Chart 10CMBS Market Overview CMBS Market Overview CMBS Market Overview Non-Agency Commercial Mortgage-Backed Securities outperformed the duration-equivalent Treasury index by 60 basis points in April, bringing year-to-date excess returns up to 71 bps. The index option-adjusted spread for non-agency Aaa-rated CMBS tightened 4 bps on the month and currently sits at 69 bps, close to one standard deviation below its pre-crisis mean. Our excess return Bond Map shows that Aaa-rated non-Agency CMBS offer greater potential reward, but also greater risk, than the majority of other high-rated spread products. The exception is conventional 30-year Agency MBS, which offer a less attractive risk/reward trade-off.10 That being said, the fundamental picture for commercial real estate is less appealing than on the residential side. CMBS spreads continue to diverge from commercial property prices (Chart 10). Agency CMBS: Overweight Agency CMBS outperformed the duration-equivalent Treasury index by 26 basis points in April, bringing year-to-date excess returns up to 12 bps. The index option-adjusted spread was flat on the month and currently sits at 47 bps. According to our Bond Map, Agency CMBS offer greater potential excess return and less risk than both the Supranational and Domestic Agency sectors. We continue to view the Agency CMBS space as an attractive low-risk spread sector. Treasury Valuation Chart 11Treasury Fair Value Models Treasury Fair Value Models Treasury 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 2.70%. The drop in the model's fair value stems from a decline in the global PMI to 53.5 from a recent peak of 54.5. While global growth has undoubtedly lost momentum in recent months, we also suspect that our 2-factor model is finally breaking down. The 2-factor model does not contain a variable to capture the degree of resource utilization in the economy. Logically, as slack dissipates in the economy and inflationary pressures mount, then the same level of global growth should be associated with a higher Treasury yield, all else equal. This means that at some point, as we approach the end of the cycle, the model will break down and consistently produce fair value readings that are too low. We suspect that we may be reaching this point. When we augment our model with an additional variable to measure the degree of resource utilization, in this case the employment-to-population ratio, we find that the new model projects a fair value of 3.28% for the 10-year Treasury yield (Chart 11). This 3-factor model would not have worked as well as our 2-factor model during the zero-lower bound period, as can be seen by looking at how rolling regression betas from each of the three variables moved sharply following the recession (bottom three panels). However, as we move further away from the zero-lower bound we expect the regression coefficients to return to pre-crisis levels, meaning that it will be important to monitor both trends in global growth and the amount of resource slack in the economy. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Jeremie Peloso, Research Analyst jeremiep@bcaresearch.com 1 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 2 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 3 Please see U.S. Bond Strategy Weekly Report, "On The MOVE", dated February 13, 2018, available at usbs.bcaresearch.com 4 For details on the Bond Map please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 6 Please see U.S. Bond Strategy Weekly Report, "Monetary Restraints", dated February 27, 2018, available at usbs.bcaresearch.com 7 Please see U.S. Bond Strategy Weekly Report, "Profiting From A Higher LIBOR", dated March 20, 2018, available at usbs.bcaresearch.com 8 Please see U.S. Bond Strategy Weekly Report, "A Signal From Gold?", dated May 1, 2018, available at usbs.bcaresearch.com 9 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, available at usbs.bcaresearch.com 10 Please see U.S. Bond Strategy Weekly Report, "As Good As It Gets For Corporate Debt", dated April 24, 2018, 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)