BCA Indicators/Model

Dear Clients, Please note that next week's report will be a joint effort with our geopolitical team, focused on North Korea. The report will be sent to you two days later than usual, on Friday June 8. Best regards, Jonathan LaBerge, CFA, Vice President Special Reports Highlights Most episodes of negative relative Chinese equity performance this year have been driven by global stock market selloffs or related to the trade dispute with the U.S. Since Chinese ex-tech stocks have continued to outperform their global peers during this period, we recommend against downgrading China for now, barring hard evidence of a pernicious global slowdown or that severe protectionist action from the U.S. will indeed occur. Our list of charts to watch over the coming months highlights, among several other important points, that monetary conditions are not overly restrictive and that financial conditions are not tightening sharply. This is in spite of a recent clustering in corporate bond defaults that has concerned some investors. Besides broad-based stimulus in response to an impactful trade shock, a sustained pickup in housing construction remains the most plausible catalyst for an acceleration in domestic demand. For now tepid sales volume casts doubt on this scenario, but investors should continue to watch Chinese housing market dynamics closely. Feature There have been several developments affecting Chinese and global stock markets over the past two weeks. On the trade front, Secretary Mnuchin's statement on May 20 that the U.S. would be "putting the trade war" with China on hold was greeted by a material pushback from Congressional Republicans, particularly the administration's plan to ease previously announced sanctions on ZTE Group. The administration's trade rhetoric has since become more hawkish, as evidenced by yesterday's statement from the White House that referenced specific dates for the imposition of tariffs and the announcement of new restrictions on Chinese investment. This uptick in tough language sets the scene for Secretary Ross' Beijing visit this weekend to continue negotiations. More recently, a political crisis in Italy has caused euro area periphery bond yields to rise sharply, roiling global financial markets. The Italian President's rejection of Paolo Savona as proposed finance minister by the anti-establishment Five Star Movement (M5S) and Euroskeptic Lega has led to the installation of a caretaker government until the fall, when new elections are set to take place. The sharp tightening in financial conditions for Italy and Spain over the past week has exacerbated concerns about a potential growth slowdown in Europe, and has fed a relative selloff in emerging market equities that began in late-March. Despite the recent turmoil, our recommendation to investors is to avoid making any major changes to their allocation to Chinese ex-tech stocks within a global portfolio. Unless presented with hard evidence that the slowdown in the global economy is more than a simple deceleration from an above-trend pace, or that protectionist action from the U.S. will occur in a severe fashion, Table 1 suggests that investors should stay overweight Chinese ex-tech stocks (with a short leash). The table highlights that most episodes of negative relative Chinese ex-stock performance since the beginning of the year been driven by global stock market selloffs or related to the trade dispute with the U.S., despite the ongoing slowdown in China's industrial sector that we have repeatedly flagged. Since Chinese ex-tech stocks have continued to outperform their global peers during this period, our interpretation is that investors are well aware of the deceleration in China's economy, but do not yet regard it as a material threat to ex-tech equity prices. Table 1YTD Weakness In Chinese Stock Prices Has Been Driven By Global Events 11 Charts To Watch 11 Charts To Watch Clearly, however, this assessment on the part of global investors can change, underscoring that the situation in China merits continual re-assessment. With the goal of providing investors with a toolkit to continually monitor the state of the Chinese economy and the resulting implications for related financial asset prices, this week's report presents a list of 11 charts "to watch" across five categories of analysis. In our view these charts span key potential inflection points for the economic and profit outlook, and will serve as an important basis for us to update our view on China over the months ahead. Monetary & Fiscal Policy Chart 1: The Policy Rate Versus Borrowing Rates Chart 1Borrowing/Policy Rate Divergence Should Not Last,##br## But Is Worth Monitoring Borrowing/Policy Rate Divergence Should Not Last, But Is Worth Monitoring Borrowing/Policy Rate Divergence Should Not Last, But Is Worth Monitoring An interesting divergence has occurred lately between the 3-month interbank repo rate (currently the de-facto policy rate) and both corporate bond yields and the average lending rate. While the repo rate fell non-trivially after it became apparent in late-March that the PBOC would extend the deadline for the implementation of new regulatory standards for asset management products, corporate bond yields have recently risen sharply and China's weighted-average lending rate ticked higher in Q1. As we highlighted in last week's Special Report, the recent clustering of corporate bond defaults does not (for now) appear to be a source of systemic risk. First, by our estimation, the recent defaults cited above account for only 0.09% of outstanding corporate bonds. Second, the latest PBOC monetary report changed the tone from emphasizing "deleveraging" to "stabilizing leverage and restructuring", which shows that regulators are as concerned about the stability of the economy as they are about reducing excessive debts. But the possibility remains that the ongoing crackdown on China's shadow banking sector will cause some degree of persistence in the recent divergence between the interbank market and actual borrowing rates, implying that investors should continue to watch Chart 1 over the months for signs of materially tighter financial conditions. Chart 2: The Correlation Between Sovereign Risk And The Repo Rate We noted in a February Special Report that investors could use the rolling 1-year correlation between the 3-month interbank repo rate and the relative sovereign CDS spread between China and Germany as a gauge of whether Chinese monetary policy has become too restrictive for its economy.1 Despite the fact that actual sovereign credit risk in China is extremely low, Chart 2 shows that the relative CDS spread has acted as a good bellwether for growth conditions in the Chinese economy. It shows that the correlation between this spread and the 3-month interbank repo rate was initially positive in late-2016 (representing concern on the part of investors that monetary policy is restrictive), but has since come back down into negative territory. Interestingly, the correlation was consistently positive from mid-2011 to mid-2014, when average lending rates averaged 7% or higher and the benchmark lending rate exceeded the IMF's Taylor Rule estimate by about 1%.2 For now the correlation remains negative (as it was when we published our February report), meaning that it currently supports our earlier conclusion that monetary conditions are not overly restrictive and that financial conditions more generally are not tightening sharply (despite the recent rise in corporate bond yields). Chart 2No Sign Yet That Monetary Policy Is Overly Restrictive No Sign Yet That Monetary Policy Is Overly Restrictive No Sign Yet That Monetary Policy Is Overly Restrictive Chart 3Watch For Signs Of Fiscal Stimulus Watch For Signs Of Fiscal Stimulus Watch For Signs Of Fiscal Stimulus Chart 3: The Fiscal Spending Impulse Chart 3 presents the Chinese government's budgetary expenditure as an "impulse", calculated as expenditure over the past year as a percent of nominal GDP. Panel 2 shows the year-over-year change in the impulse. When compared with a similar measure for private sector credit, cyclical fluctuations in China's government spending impulse are relatively small. For this reason, BCA's China Investment Strategy service has not strongly emphasized fiscal spending as a major driver of China's business cycle. However, we also noted in a recent report that fiscal stimulus stands out as one of the "least bad" options available to policymakers to combat a negative export shock from U.S. protectionism, were one to occur.3 The potential for broader stimulus from Chinese authorities in response to an impactful trade shock raises the interesting possibility of another economic mini cycle in China, since the economy accelerated meaningfully in response to the last episode of material fiscal & monetary easing. As such, investors should closely watch over the coming months for signs that fiscal spending is accelerating, particularly if combined with potential signs of easing monetary policy. External Demand Chart 4: Global Demand And Chinese Export Growth Chart 4For Now, Resilient Exports ##br##Are Supporting China's Economy For Now, Resilient Exports Are Supporting China's Economy For Now, Resilient Exports Are Supporting China's Economy We have noted in several recent reports that a resilient export sector remains the most favorable pillar of Chinese growth. Besides the clear risk to Chinese trade from U.S. protectionism, two other factors have the potential to negatively impact the trend in export growth. The first (and most important) of these risks is a reduction in global demand, which some investors have recently been concerned about given the decline in global manufacturing PMIs. However, Chart 4 highlights that our global PMI diffusion indicator has done an excellent job of leading the global PMI over the past few years, and has barely registered a decline over the past few months. From our perspective, the odds are good that the recent deceleration in the PMI has been caused by sudden caution (even in developed countries) over the Trump administration's protectionist actions, and does not reflect a material or long-lasting slowdown in the global economy. But we will be closely watching the PMI releases over the coming months to rule out a more painful slowdown in global demand. Importantly, we have also highlighted that stronger exports may actually presage a further slowdown in China's industrial sector if it emboldens policymakers to intensify their reform efforts over the coming year. We argued in our May 2 Weekly Report that China's reform pain threshold is positively correlated with global growth momentum,4 meaning that the external sector of China's economy may have less potential to counter weakness in the industrial sector than many investors believe. In this regard, extreme export readings (to the up and downside) should be regarded by investors as a potentially problematic development. Chart 5: The Competitiveness Impact Of A Rising RMB Chart 5 highlights the second non-protectionist risk to Chinese export growth, namely the significant appreciation in the RMB that has occurred since mid-2017. The chart shows the percentile rank of three different trade-weighted RMB indexes since 2014, and highlights that all three are between their 70th & 80th percentiles (with our BCA Export-Weighted RMB index having risen the most). Importantly, the 2015-high shown in Chart 5 represents the strongest point for the currency in over two decades, suggesting that further currency strength may exacerbate the significant deceleration in export prices that has already occurred. Chart 5A Surging RMB Could Undercut Competitiveness A Surging RMB Could Undercut Competitiveness A Surging RMB Could Undercut Competitiveness Housing Chart 6: Housing Sales Versus Starts We have presented a variation of Chart 6 several times over the past few months, but it is important enough that it deserves to be continually monitored by investors over the coming year. Chart 6 tells the story of China's housing market from the perspective of an investor who is primarily interested in the sector because of its implications for growth. The chart highlights that residential floor space started, our best proxy for the real contribution to growth from residential investment, has fallen significantly relative to sales since 2012-2014. This appears to have occurred because of a significant build up in housing inventories, which has since reversed materially (even though the level remains elevated). To us, this suggests that the gap between housing sales and construction that has persisted for the past several years may finally be over, suggesting that the latter may pick up durably if sales trend higher. For now sales volume remains tepid, but this will be a key chart for investors to watch over the coming year given our view that housing is a core pillar of China's business cycle. The Industrial Sector Chart 7: The BCA Li Keqiang Leading Indicator And Its Components Chart 7 presents our leading indicator for the Li Keqiang index (LKI), which we developed in a November Special Report.5 There are six components of the indicator, all of which are related to changing monetary/financial conditions, and the growth in money and credit. Chart 6Housing Construction Could Accelerate##br## If Sales Pick Up Housing Construction Could Accelerate If Sales Pick Up Housing Construction Could Accelerate If Sales Pick Up Chart 7A Downtrend In Our LKI Leading Indicator, ##br##Within A Wide Component Range A Downtrend In Our LKI Leading Indicator, Within A Wide Component Range A Downtrend In Our LKI Leading Indicator, Within A Wide Component Range The indicator is at the core of our view, and we have been presenting monthly updates of the series in our regular reports since late last year. However, Chart 7 looks at the indicator from a different perspective, by showing it within a range that identifies the weakest and strongest components at any given point in time. Two points are noteworthy from the chart: While the overall LKI indicator has been trending down since early-2017, there is currently a wide range between the components. This gap is in stark contrast to the very narrow range that prevailed from 2014-2015, when the economy slowed considerably. This could mean that some of the components of the indicator are unduly weak, which in turn could imply that the severity of the slowdown in China's industrial sector will be less intense than the overall indicator would otherwise suggest. At least one component provided a lead on the subsequent direction of the overall indicator from late-2011 to late-2012, the last time that a significant gap existed between the components. This is in contrast to the situation today, in that all of the components are currently in a downtrend (albeit with differing paces as well as magnitudes). The key point for investors from Chart 7 is that all of the components of our indicator are moving in the same direction, which suggests with high conviction that China's economy is slowing. However, the wide range among the components suggests that indicator's message about the intensity of the slowdown is less uniform than it has been in the past, meaning that investors should be sensitive to a sustained pickup in the top end of the range. Equity Market Signals Chart 8: The Beta Of Our BCA China Sector Alpha Portfolio Chart 8 revisits a unique insight that we presented in our May 16 Weekly Report.6 The chart shows the rolling 1-year beta of our BCA China Investable Sector Alpha Portfolio versus the investable benchmark alongside China's performance versus global stocks, and suggests that the former may reliably lead the latter. While we noted in the report that drawing market-wide inferences from the beta characteristics of risk-adjusted performers is a not a conventional approach, finance theory is supportive of the idea. If investors are seeking to maximize their risk-adjusted returns and are engaging in tactical allocation across sectors, then it is entirely possible that beta-adjusted sector returns reflect the risk-on/risk-off expectations of market participants. For the purposes of China-related investment strategy over the coming year, our emphasis on Chart 8 will increase markedly if we see a sharp decline in the beta of our Sector Alpha Portfolio. As we noted in our May 16 report, the model is for now sending a curiously bullish signal, which we see as partial validation of our view that investors should have a high threshold to cut exposure to China within a global equity portfolio. Chart 8Watch For A Decline In The Beta Of ##br##Our Sector Alpha Portfolio Watch For A Decline In The Beta Of Our Sector Alpha Portfolio Watch For A Decline In The Beta Of Our Sector Alpha Portfolio Chart 9Decelerating Earnings Growth Could##br## Undermine Investor Sentiment Decelerating Earnings Growth Could Undermine Investor Sentiment Decelerating Earnings Growth Could Undermine Investor Sentiment Chart 9: Ex-Tech Earnings Versus The Li Keqiang Index We noted above that predicting the Li Keqiang index (LKI) is at the core of our view, and Chart 9 highlights why. The chart shows that a model based on the LKI closely fits the year-over-year growth rate of Chinese investable ex-tech earnings and, crucially, provides a lead. While the chart does not suggest that an outright contraction in ex-tech earnings is in the cards over the coming year, it does show that earnings growth is about to peak. This is potentially problematic, and warrants close attention, for two reasons: First, our leading indicator for the LKI suggests that it will decelerate further over the coming year, which could push our earnings growth estimate towards or below zero. Second, the peak in earnings growth could dampen investor sentiment towards Chinese ex-tech stocks, especially since bottom up analyst estimates for 12-months forward earnings growth have recently moved higher and are currently above what is predicted by our model. Chart 10: The Alpha Of Chinese Banks By now, the narrative surrounding Chinese banks is well known among global investors. The enormous leveraging of China's non-financial corporate sector is viewed by many as a clear sign of capital misallocation, meaning that a (potentially material) portion of the loan book of Chinese banks will have to be written off as bad debt. The ultimate scope of the bad debt problem in China is far from clear, but these longstanding concerns about loan quality suggest that Chinese bank stocks are likely to materially underperform their global peers if China's shadow banking crackdown begins to pose a significant threat to growth via restrictions on the provision of credit to the real economy. As such, we recommend that investors monitor Chart 10 over the coming year, which shows the rolling 1-year alpha significance for Chinese banks vs their global peers. While the rolling 1-year alpha of small banks has become less positive over the past few weeks, it remains in positive territory, similar to that of investable bank stocks. So, for now, this indicator supports our earlier conclusion that recent divergence between the interbank market and actual borrowing rates highlighted in Chart 1 is not heralding a material tightening in Chinese financial conditions. Chart 10Investors Should Monitor Chinese Bank Alpha ##br##For Significant Declines 11 Charts To Watch 11 Charts To Watch Chart 11No Technical Breakdown (Yet) In Ex-Tech Relative Performance No Technical Breakdown (Yet) In Ex-Tech Relative Performance No Technical Breakdown (Yet) In Ex-Tech Relative Performance Chart 11: The Technical Performance Of Ex-Tech Stocks BCA's approach to forecasting financial markets rests far more on top-down macroeconomic assessments than it does on technical analysis. However, technical indicators do contain important information, particularly when our top-down macro approach signals that a change in trend may be imminent. In this regard, technical indicators can provide valuable opportunities to enter or exit a position. To the extent that the technical profile of Chinese ex-tech stocks is informative in the current environment, Chart 11 shows that it is telling investors to stay invested despite the myriad risks to the economic outlook. This message is consistent with that of Table 1, namely that the negative performance of Chinese ex-tech stocks has been in response to global rather than idiosyncratic, China-specific risk. From our perspective, a technical breakdown in relative Chinese ex-tech stock performance in response to China-specific news would serve as a strong basis for a downgrade within a global equity portfolio, and we will be monitoring closely for such a development over the coming weeks and months. Stay tuned! Jonathan LaBerge, CFA, Vice President Special Reports jonathanl@bcaresearch.com 1 Please see China Investment Strategy Special Report, "Seven Questions About Chinese Monetary Policy", dated February 22, 2018, available at cis.bcaresearch.com. 2 Please see China Investment Strategy Weekly Report, "Monetary Tightening In China: How Much Is Too Much?" dated January 18, 2018, available at cis.bcaresearch.com. 3 Please see China Investment Strategy Weekly Report, "The Question That Won't Go Away", dated April 18, 2018, available at cis.bcaresearch.com. 4 Please see China Investment Strategy Weekly Report, "China: A Low-Conviction Overweight", dated May 2, 2018, available at cis.bcaresearch.com. 5 Please see China Investment Strategy Special Report, "The Data Lab: Testing The Predictability Of China's Business Cycle", dated November 30, 2017, available at cis.bcaresearch.com. 6 Please see China Investment Strategy Weekly Report, "The Three Pillars Of China's Economy", dated May 16, 2018, available at cis.bcaresearch.com. Cyclical Investment Stance Equity Sector Recommendations

"Amongst all unimportant subjects, football is by far the most important." - Pope John Paul II From June 14 to July 15, billions of people will tune into the 2018 FIFA World Cup, with untold loss of productivity, working hours, and quality family time as a consequence. At BCA Research, we decided to stop pretending that we are indifferent to the quadrennial sporting pilgrimage - or to our staff hogging bandwidth by live-streaming games during business hours - and instead harness the best young minds of the company to deliver this most eminently unimportant forecast. To our clients who may worry that their hard-earned dollars are subsidizing fleeting research pursuits, we want to assure you that the research herein was produced "off the clock." In our firm's 70-year history, the principal question driving all analysis has been "so what?" Each and every piece of analysis produced at BCA Research is intended to conclude with an investment angle that makes sense of the noise produced by the cacophony of data. We do not intend to evolve out of that genetic material. At the same time, we are passionate about research. Even though this report does not have obvious investment implications, it is an excellent example of our research fundamentals.1 We develop a macro framework through qualitative analysis, enrich it with data manipulated in novel ways, and articulate it through empirical testing. Just as we produced this report "off the clock," we hope that our clients consume it in their down time. The following pages are a respite from President Trump's tweets, elevated equity valuations, prospects of tepid returns, renewed confrontation in the Middle East, and trade wars. It is also the first in a series of in-depth, multi-disciplinary reports that we intend to produce. Stay tuned. This BCA Special Report presents a unique two-step empirical approach to predict the outcome of each World Cup match. Our approach includes micro (player level) and macro (team level) factors to forecast game matches. We present this model in Section I. Section II then introduces several interesting qualitative narratives, focusing on specific teams attending this year's competition. Section III builds on the existing academic literature and our own unique framework to establish whether the World Cup has any market and economic implications. I. Two-Step Model: Forecasting The 2018 FIFA World Cup In this section we introduce a two-step simulation of the upcoming World Cup. Building on academic literature that has shown that teams respond differently to the various stages of the tournament, we develop two separate models for the group and knockout stages of the competition (Box 1).2 Box 1: The World Cup: A Quick Overview The quadrennial World Cup finals will take place in Russia over a period of one month. The tournament is referred to as the "finals" because the actual competition to select the 32 teams began in 2015. The qualification tournament whittled down the 209 FIFA member states (minus the already qualified host, Russia) over the course of three years via continental qualification tournaments. The 32 teams competing in the finals are separated into eight groups. Each team will play one game against each of their group opponents. The top two performers advance to the knockout stage of the competition. In this stage, the World Cup resembles the NCAA March Madness tournament in that teams face immediate elimination. The 2018 FIFA World Cup is the 21st edition of the competition that goes back to 1930. In the history of the competition, only eight nations have won the cup. Brazil has won five times, Germany and Italy four, Argentina and Uruguay twice, and France, England, and Spain once each. We are not the first to attempt to predict the World Cup. In a recent innovative approach, Clemente et al. (2015) use parameters from network analysis to analyze the tournament's matches. More traditional models rely on the FIFA World Ranking, which ranks teams based on past performance.3 Other conventional models draw on economic fundamentals to explain the success and failure of national teams. A literature overview from 2004 of both quantitative and qualitative models found that a simulator running a commercially produced computer game offered the most successful prediction of the World Cup.4 This should not come as a surprise. The computer gaming industry has overtaken Hollywood in terms of total revenue, while production costs of some computer games are running higher than those of blockbuster movies.5 Given that realism is an important feature of sport simulation, it should follow that computer games have a better forecasting track record than humans. In this report, we combine macro and micro variables to develop a unique two-step model. Our micro factors are based on an extensive database consisting of individual player statistics. Inspired by the 2004 study above, we rely on the computer gaming industry for player ranking, modifying it with our own collective soccertise. The list of potential explanatory variables that we considered including in our model is available in Table 1. Table 1Variables Considered And Used For The Models The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup By relying on individual player attributes we believe that we have avoided the pitfalls of more "macro"-oriented models, such as those relying on overall team rankings. Overall team rankings are established on the basis of team performance over a multi-year period of time. We find that this approach overstates team performance over team quality in present time. These models also fail to incorporate tactical analysis into the model. We introduce these macro and qualitative factors - such as reputation and long-term historical performance - in the qualitative part of our evaluation, while letting objective, player- and team-specific data, dominate our model. Data To determine which variables from Table 1 had the highest predictive power we relied on the 192 matches that occurred during the 2006, 2010 and 2014 FIFA World Cups. There are two reasons we focused only on the last three tournaments. The first is data availability. The second is that football has undergone dramatic evolution in strategy in the twenty-first century. Among the 192 matches, 48 took place in the knockout stage of the competition, while the remaining 144 games occurred during the group stages. We rely on the database of player statistics used in Electronic Arts (EA) Sports FIFA computer simulation. The database includes player statistics from August 2004 to present. At the time of the publication of this report, not all 32 teams had made official their final list of 23 players (the lists will be published on June 4th by the FIFA). Therefore, we relied on the most recent World Cup qualifiers, as well as all the friendly matches played year-to-date to come up with the most likely line-ups for these teams. Step One: The Group Stage Model To simulate the 2018 group stage matches, we developed an Ordered Probit (OP) model estimated on past World Cups group stage games. Ordered Probit models are powerful when modeling an ordinal outcome (i.e. the response value has a strictly increasing ordering known prior to the estimation). Football matches can end in either a loss, a draw, or a win, with a well-defined order from loss to win. The Ordered Probit model therefore allows us to use this information, increasing the predictive ability of the model.6 The Ordered Probit model selected is represented using a continuous latent variable yi* that is linearly determined by a set of explanatory variables χι: The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Where φ is the cumulative normal distribution function and ϒ are arbitrary thresholds selected via log-likelihood maximization.7 Based on our sample of group-stage matches from the past three World Cups, we found that the best explanatory variables are: Team Average Player Rating Average Age - Forwards Average Number of Caps - Defenders Speed Positions Average Rating Group Stage Explanatory Variables Team Average Rating "Talent wins games, but teamwork and intelligence wins championships." Thus said six-time NBA champion Michael Jordan. Our model confirms that the insight from the basketball legend applies to football as well, as average team talent - based on the mean of individual player overall attributes taken from the EA Sports database - is the most significant predictor of game success in both stages of the competition. However, this variable becomes less important in the knockout stages, as the ability to play as a team ("teamwork"), as well as particular tactical matchups ("intelligence"), become paramount in winning the tournament. This is because, in the late stages of the competition, the talent differential between teams narrows substantively. Average Age - Forwards Teams with younger forwards tend to perform better than teams with older forwards due to the highly physical demands of the position. Forwards perform the highest amount of high-intensity efforts (sprinting and sudden changes of direction) and experience the most amount of physical contact among all positions in football.8 Considering that these abilities tend to peak in the early to mid-20's for most professional athletes, it is no surprise that our model gives a premium to youth in this position.9 Moreover, the annals of World Cup history are chock-full of tales of youthful forwards making a name for themselves when it matters.10 Figure 1Position Definitions The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Average Number Of Caps - Defenders Teams with more experienced defenders will tend to do better than teams with less experienced defenders. The physical demands on defenders tend to be less strenuous relative to other positions, given that they dish out the pain. This means that defenders tend to peak later than any other field position in football and are able to maintain peak performance sometimes well into their 30's. What defenders lack in athleticism they must make up in game IQ, as anticipation, composure, and tactical awareness are all acquired skills that improve with experience. Speed Positions Average Rating Teams with superior talent in their speed positions (full backs and wingers) will perform better in group stages (Figure 1). Academic research has shown that counterattacking football is the most effective tactic against unbalanced defenses.11 Wingers and full backs are crucial for both developing and defending against a counterattack as they cover the least congested part of the pitch, where there is generally the most space to run. Interestingly, this variable becomes less important in the knockout stages. This is most likely due to the fact that teams with skilled and fast wingers can easily exploit the space conceded by the tactically disorganized defenses of easier opponents that populate the early stages of the competition. In the later stages, teams are generally more tactically disciplined. Modeling Group Stage Games For each game, our model uses the spread between Team 1 and Team 2 of each explanatory variable to determine the probability of Team 1 winning the game. Table 2 lists all 32 teams and their descriptive statistics on the four explanatory variables. Table 3 presents how the 32 teams are ranked based on their probability of passing to the next stage inferred by these explanatory variables. Table 2Descriptive Statistics: Group Stage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Table 3Group Stage Ranking The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Our predictive variables were selected based on the results of the 2006, 2010 and 2014 World Cups. The estimated coefficients are then used to produce 1,000 simulations of each game in order to obtain a distribution of the outcome. Table 4Marginal Effect Of Selected##BR##Variables: Group Stage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The team average ratings are the core variables in our group stage model. However, due to the high level of multicollinearity in the disaggregated player rating variables, we could not capture the entire set of information from these individual variables. As a result, our final probabilities are derived from the weighted average of two separate estimations, Model 1 and Model 2. This allows our model to capture the importance of the speed positions average rating in predicting the outcome of the group stage matches, which displays the highest marginal impact on the winning probability (Table 4). Therefore, our final model for the probability of winning a game is: The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Where: Model 1 (M1)=ƒ (Team Average Player Rating, Forward Average Age, Defenders Average Caps) and, Model 2 (M2)= ƒ(Speed Positions Average Rating, Forward Average Age, Defenders Average Caps), and α and (1- α) are the weights given to each models.12 The ultimate product of our modeling is a coefficient we have termed E(points). As a reminder, teams are awarded three points for a win, one point for a draw, and zero points for a loss. The maximum amount of points a team can gain is nine, given that there are three matches (Box 2). BOX 2: E(points) Calculation The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup E(points) allows us to determine which teams have the highest probability of passing to the knockout stages. We caution readers from reading too much into E(points) in terms of which teams move on to the knockout rounds as the competition in each group greatly determines the value. For example, Denmark has a higher E(points) coefficient than Serbia, but that is a function of its relatively easy group (it faces weaker competition). Group Stage: Results Group A Group A was one of the most challenging to forecast (Table 5). First, Uruguay is probably one of the weakest of the group favorites in the competition.13 Second, host Russia and Egypt are separated by few points in terms of overall quality. Table 5Group A Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup To make our lives easier, we decided to attribute a significant home advantage bonus to Russia.14 History tells us that hosting the World Cup provides a clear advantage to almost any team (Table 6). One out of every three hosts has won the competition, going back to the first tournament in 1930. A whopping 65% of all hosts have made it all the way to the semi-finals. It pains us to see the "Egyptian King," Liverpool F.C. superstar Mohamed Salah, exit at the group stage. However, history is arrayed against his squad. Russia would have to be only the second host team ever - aside from South Africa in 2010 - to fail to capitalize on its home court in the group stage. And, as history teaches us, you just don't beat Russians on Russian soil. Group B Group B was, by far, the easiest to forecast (Table 7). The two Iberian giants will make it through, a high-conviction view. Portugal will try to become the fourth team to win the European Championship and the World Cup consecutively, joining such great teams as West Germany (Euro 1972, World Cup 1974), France (World Cup 1998, Euro 2000), and Spain (Euro 2008, World Cup 2010, Euro 2012). Our knockout stage model likes Portugal, giving it the fifth-best probability of winning the entire competition. Its performance in the group stage will go far in validating our confidence in the team. Table 6Home Advantage Is Real The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Table 7Group B Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group C A pre-tournament favorite, France will look to avenge its stunning loss to Portugal sans Cristiano Ronaldo in the finals of the 2016 Euro. France starts its competition off in one of the weakest groups (Table 8). Group C lacks real competition, with Denmark having only a 25% probability of beating France. This is the lowest probability of success for any second-ranked team in the group stage of the competition. Anything less than 9 points for France would be a concern for the fans of Les Bleus. Table 8Group C Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group D Argentina and Croatia are the favorites to go through in Group D, but our model likes Nigeria's and Iceland's chances (Table 9). Iceland keeps turning forecasters into fools. As we discuss in Section II, its success in international football is empirical evidence of the divine. Its run in the 2016 Euro Cup, and win against England, may be the greatest upset in any major sporting competition. However, the element of surprise is gone. Table 9Group D Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Nigeria, on the other hand, could be a dark horse. Our model likes their chances, with 31% probability that they advance to the next stage. They have also won only one of their last 12 World Cup matches, which suggests that they may be overlooked coming into the tournament. Group E Brazil is likely to dominate Group E, but the fight for second place, and thus qualification into the knockout round, will be vicious. Switzerland and Serbia represent a real challenge to our model. Their values in the four explanatory factors are tantalizingly close (Chart 1). Serbia gets the narrowest of pushes on three out of the four, but our model assigns identical probabilities of winning (39%) to each team when they face each other head-to-head (Table 10). The model gives the slightest of chances to Switzerland, mainly due to its higher probability of stealing points in its game against Brazil (48% compared to 46% for Serbia). It is a stunning revelation backed by history. Serbia has no luck against non-European competition in the tournament. Other than its impressive win against the eventual finalist Germany in 2010, it has lost to Argentina (by 6-0!), Ivory Coast, Ghana, and Australia. Furthermore, momentum is not on the side of the Orlovi, given that they backed their way into the World Cup with a 3-2 loss to Austria and a nail-biter against Georgia. These results prompted the Serbian federation to replace the head coach months ahead of the World Cup, not a reassuring sign. Chart 1Serbia Vs. Switzerland The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Table 10Group E Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group F Germany, the 2014 World Cup winner, will find very little opposition in Group F and should easily avoid the fate of the past two reigning champions who were eliminated early (Table 11). Die Mannschaft was the only team in Europe that won all its qualification games, ending the qualifying tournament with an otherworldly +39 goal difference. Our model likes Mexico over Sweden. This makes sense given that El Tri have progressed past the group stage in the previous six World Cups (only to be promptly eliminated in the first game of the knockout stage every time). Table 11Group F Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group G Our model finds Group G exceedingly easy to predict. Belgium and England will go through (Table 12). Both teams have a lot in common, starting with quality squads that have failed to make a mark in recent international competitions. Belgium comes to the 2018 World Cup with top quality players in many positions on the field (Table 13). The two squads will find no real competition in this group, but the pressure will be immense on both young squads. How they handle lowly Tunisia and Panama will determine their mental readiness for the knockout stage of the competition. Table 12Group G Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Table 13Top Players In Every Position On The Pitch The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group H Group H is by far the most difficult to forecast (Table 14), with all four teams clustered around a similar E(points) value (Chart 2). Poland has the best player - Bayern Münich superstar Robert Lewandowski - and Colombia the best team. However, both Senegal and Japan could surprise. The forecasting error in Group H should have little bearing on further results given the relatively low ranking of all four teams (for example, both Switzerland and Serbia from Group E are ranked as superior teams). However, the top two teams from this group will "cross" against Belgium and England, the two chronic underachievers. As such, winning the group carries considerable upside as the winner would face the young, untested, and skittish England. This means that Colombia, Poland, or perhaps even Senegal and Japan, could end up stunning the world and finishing at least in the top eight. Table 14Group H Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Chart 2Probability Of Proceeding To The Knockout Stage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Group Stage Results Table 15 illustrates group stage results. We assign points based on the highest probability outcomes (win, draw, loss) of each game. Table 15Group Stage Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup How confident are we of the results? Other than groups E and H, we are highly confident. Box 3 goes over our three most likely "Dark Horses" to go through to the second round. BOX 3: BCA's "Dark Horses" Serbia The Serbian national football team has faced many challenges over the past two decades. Despite these hurdles, Serbia has produced top-class footballers that are integral parts of several European super teams. An experienced defense and aggressive midfield - headlined by one of the world's premier defensive midfielder Nemanja Matic of Manchester United - makes Serbia a hard team to play against. There is little of the Balkan flair on this team that defined Yugoslav football in the 1980s. But perhaps that is a good thing, given that World Cup games are tight, nervous, and often devolve into a defensive slog. Senegal It has been 16 years since Les Lions de la Teranga last appeared in a World Cup - and what a memorable appearance that was. In 2002, Senegal reached the quarter finals, the second African team in history to make it that far in the competition. The 2018 squad shows similar potential. The Lions' forwards pose the kind of threat to opponents' defenses that many of their World Cup rivals would envy. Keep in mind some of these names: Sadio Mané (Liverpool F.C.), Ismaila Sarr (Rennes), and Keita Balde (A.S. Monaco). Peru Peru is our choice for a long shot at this year's tournament. The Peruvian national football team is unbeaten since November 2016, winning eight matches and drawing four, with many coming against elite opposition. Although we do not take overall team rankings into consideration when making forecasts, it is notable that Peru is ranked 11th in the world, according to the ELO index. Peru's path to Russia took it through the South American qualifying tournament, known for being the most competitive in the world. Teams have to play 18 games, many at high altitude or in unforgiving climates, against some of the world's most talented squads. While it is true that they were the last team to qualify, having to beat New Zealand in the intercontinental playoffs, Peru came ahead of Copa America champions Chile and just one and two points behind Colombia and Argentina, respectively. While our model does not believe in the relatively young and unproven Peruvian team, placing it last in the group, the team's excellent coaching and fearless play make it a potential candidate to come out of Group C along with France, especially given that our model may be overstating Denmark's potential. Step Two: The Knockout Stage Model The knockout stage is somewhat easier to model given that the set of possible outcomes is reduced to only {loss; win}. This difference with the group stage is not only relevant for the math behind our model. It is also relevant for the strategy teams employ during the games. Therefore, we simulated this part of our analysis using a probit model estimated on a sample of only knockout stage games from the 2006, 2010 and 2014 World Cups. The binary choice probability of observing a specific outcome becomes: The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup In this stage of the competition, we found the following factors to be the most important: Team Average Player Rating Club Level Synergy Player GINI Coefficient Average Rating - Midfielders Knockout Stage Explanatory Variables Team Average Rating As with the group stage, the overall rating of the team - based on the average of individual rankings from the EA Sports database - is the most powerful explanatory variable. Despite the higher marginal effect of the rating variables in the knockout stage sample, the standard deviation and average of these variables are significantly smaller than in the group stage.15 In other words, the gap in player quality between teams in the group stage is often vast. However, the knockout stage culls the minnows, narrowing the gap in overall player quality between teams. At this stage of the competition, our model has to be supplemented with variables that test for teamwork and synergy. Club Level Synergy Teams with more players playing in the same club tend to perform better in the knockout stages. This is evident from all the World Cup winners in our sample.16 Given the limited practice time that national teams have ahead of the tournament, the year-round experience of playing with teammates in club competition can provide a huge advantage. Especially for football teams from countries with major leagues - such as Germany, Spain and Italy. Their players are more likely to cluster on the major clubs in those leagues, whereas players from smaller footballing nations have to ply their trade in dispersed leagues and teams across the globe. Great Man Theory (Player GINI coefficient) Teams win games, but heroes win World Cups. To test whether superstars are relevant to winning games, we designed a player quality GINI measure. We find that teams with a higher GINI coefficient outperform those with a lower measure. It seems that having a superstar, or two or three, surrounded with role players is a superior strategy to having balanced talent across all positions. This variable only becomes significant in the knockout stages, where the overall talent level between teams narrows. While more skilled teams tend to be more balanced (Chart 3), once we normalize for skill, a higher GINI becomes a predictor of success. Chart 3The Great Man Theory The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Intuitively we agree with this finding. When the stress and tension of knockout stages peak and the weight of one's entire nation begins to crush a lesser player's shoulders, the Great Man shines through. Exceptional players have the skill, stamina, and otherworldly confidence to unlock the highly disciplined and tactically sound defensive schemes found in the latter stages of the competition. This should be good news for Belgium, Portugal, and Argentina, but really bad news for England. Average Rating - Midfielders Once we decompose the different positions in the field, we find that the midfield is more important to success than other positions in the knockout matches. Research has shown that midfielders, particularly those forming the "spine" of the team, are the most involved in a team's passing play, regardless of the tactics or strategy used.17 Precise and creative passing is key in knockout matches where tactically disciplined defenses are difficult to unlock. Defensive prowess in the midfield is also paramount to prevent the opposition from developing their attack.18 As with the group stage model, the final probabilities for the knockout stage games were derived from the average of two models in order to maximize the information contained in the average midfield player rating variable (Table 16). Table 16Marginal Effect Of Selected Variables: Knockout Stage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Therefore, our final probability is: The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Where: Model 1 (M1) = ƒ(Team Average Player Rating, Club Level Synergy, Player GINI Coefficient) and, Model 2 (M2) = ƒ(Midfielders Average Rating, Club Level Synergy, Player GINI Coefficient), and a and (1 - α) are the weights given to each models.19 Table 17 summarizes the descriptive statistics of each team according to the four variables used to model their performance. Table 17Descriptive Statistics: Knockout Stage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Knockout Stage: Results The Round of 16 There were no surprises in the round of 16 (Table 18). The two matches worth paying attention to are England vs. Colombia and Portugal vs. Uruguay. Our model gives the two European teams a 60% chance of winning their respective games. We see the "Great Man Theory" carrying Portugal forward to the quarter-finals. We are essentially willing to bet that Ronaldo is at least worth a quarter-finals berth at the World Cup. Table 18Round Of 16 Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The England-Colombia matchup is essentially a coin toss. The young and untested Three Lions will face a tough challenge in Colombia. However, it is safer to carry England over to the next round as their "conditional probability" of progressing to the quarter-finals is significantly higher than that of Colombia (53% to 17%) (Chart 4). Why the difference? Conditional probability takes into account the original probability of passing the group stage. Given England's easy group, and Colombia's challenging competition, it is mathematically safer to bet on England. Chart 4Probability Of Advancing To The Quarter-Finals The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Quarter-Finals Making it into the quarter-finals of the World Cup is an extraordinary success reserved for only eight footballing nations. At this point, teams have played four intense and decisive games over three weeks. Fatigue sets in, especially given that the superstars are playing at the end of a grueling club season in what is normally their off-season. Our model bets strongly on the previous two World Cup winners (Table 19), while Brazil and France struggle against their opponents. Belgium and Portugal are left to wonder what might have been, although for Belgium the pain will be greater. Not only will they yet again fail to meet expectations, but also they will waste the highest conditional probability of advancing to the next stage of the four teams that do not advance (Chart 5). Table 19Quarter-Finals Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Chart 5Probability Of Advancing To The Semis The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup France Vs. Portugal: Setting The Record Straight The loss to Portugal in the final of the 2016 Euro, on home soil no less, still stings for France. The loss was particularly painful given that Portugal's superstar, Real Madrid's Cristiano Ronaldo, had to leave the game due to an injury and spent the majority of the game hopping on one leg, yelling instructions to his teammates from the sidelines. Our model gives France a 66% probability of winning the game. The French team is superior in every facet of the game, other than in the speed positions (Diagram 1). France also sports two of the game's best defensive midfielders, Chelsea's N'Golo Kanté and Juventus's Blaise Matuidi, to whom it will fall on to neutralize Portugal's Great Man. French coach Didier Deschamps has also benefited from the overall improvement in the French Ligue 1, calling upon players that play together in the local league. Diagram 1Les Bleus Vs. A Selecao das Quinas The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Diagram 2A Selecao Vs. The Red Devils The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup We expect a tight match, with intense man-on-man coverage of Ronaldo. France will dominate the ball, slowly building chances against Portugal's defense. In 2016, a young and inexperienced French team wasted a plethora of chances against Portugal's version of the Maginot Line. We do not see history repeating itself. Brazil Vs. Belgium: Red Devils Don't Dance Samba The second tight matchup predicted by our model pits Brazil against Belgium. As with the France-Portugal matchup, the underdog has a solid one-in-three chance of winning the game. What makes Belgium so dangerous is their overall midfielder rating, which we identified as one of the four most relevant factors for winning the game. The problem for Belgium is that it trails Brazil by a lot in other facets of the game (Diagram 2). Its defense is particularly suspect. Belgium has no players ranked in the top five of their defensive position, whereas Brazil sports a fifth of all the best defensive players in the entire tournament. Spain Vs. Argentina: Don't Cry For Me Lionel Our model has no time for La Albiceleste, giving Argentina a paltry 10% chance of defeating Spain. As a reminder, our model is completely ignorant of head-to-head matchups between teams, but it has essentially predicted the 6-1 drubbing that La Roja gave Argentina in a friendly in late March. Granted, Argentina's superstar, and F.C. Barcelona icon, Lionel Messi watched the spanking from the bench. But unless Messi can play defense at the same high level at which he finishes attacks, Argentina is in trouble. Argentina's 2018 squad is essentially the stereotype of every team we have watched from the nation in the last half-century. Its forwards are world class, perhaps the best in the tournament (Diagram 3). Four Argentines are amongst the 15 best forwards in our sample, an extraordinary number (see Table 13). Nevertheless, the rest of the team is subpar relative to other favorites, and particularly against Spain. Germany Vs. England: Hard Brexit Defending champions rarely meet expectations at the World Cup. The tournament is simply too grueling, the pressure too great, and the chasm of time between tournaments too wide to bridge within a single generation. However, Germany faces a young, untested English team in our bracket. Getting to the quarter finals would be seen as a success for England, one upon which to build a winning generation for the 2022 tournament in Qatar. Germany tops England in all facets of the game and across the pitch (Diagram 4). Our model gives England little chance against Die Mannschaft. England has greater chances of extracting a soft Brexit from Berlin than of penetrating Germany's clinical press. Diagram 3La Furia Roja Vs. A Albiceleste The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Diagram 4Die Mannschaft Vs. The Three Lions The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Semi-Finals Our model gives Belgium and Portugal decent odds for advancing to the semis, leaving a door open for surprises in the quarter-final round. In the semi-finals, however, the model tightens the odds, snuffing out the chances for an aging Germany and an inexperienced Brazil (Table 20). Brazil's and Germany's chances fall to just 15% and 11%, respectively (Chart 6). Table 20Semi-Finals Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Chart 6Conditional Probability Of##BR##Advancing To The Finals The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Spain's conditional probability of reaching the finals hovers at an extraordinarily high 54%. France remains at a respectable, and yet uncertain, 41%. France Vs. Brazil: Painful Memories We are not surprised that our model assigns Brazil just a 27% probability against France. France practically coasted to the finals of the Euro 2016, defeating Germany. In terms of quality, the two teams are even on forwards. Brazil takes a big advantage in speed positions and defenders, but France wins where it matters: midfield (Diagram 5). For many years now, Brazil has been known to produce some of the best forwards and defenders, but apart from Ballon d'Or winner Kaka, who retired a couple of years ago, it is difficult to name an outstanding Brazilian midfielder in recent years. Diagram 5Les Bleus Vs. A Selecao The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The French squad also displays better results on the club synergy variable, the second-best reading for the teams in the knockout stage behind Spain. This is the result of improvements to its domestic league. Meanwhile, Brazil's top players continue to be dispersed across a number of different top clubs. Brazil will avenge its disastrous result from 2014 with a solid showing in Russia. Its pride will be reestablished and memories of the 7-1 drubbing softened. However, it will fail yet again against a European power. Spain Vs. Germany: The Battle Of Champions The other semi-final game will feature the 2010 champion (La Furia Roja) against the 2014 champion (Die Mannschaft). You have to go back all the way to the 1990 World Cup to see the two previous World Cup winners face each other in the semi-finals. Our model is indifferent to Germany's more recent success, giving Spain an overwhelming 85% probability of winning. In 2018, Spain has superior quality across the pitch (Diagram 6). Diagram 6 La Furia Roja Vs. Die Mannschaft The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup In 2010, the year of Spain's last title, the team relied heavily on what made F.C. Barcelona unbeatable in Europe: heavy possession of the ball and gameplay built on crisp passes. This still holds true today, as tiki-taka has become part of Spain's footballing DNA. However, after underperforming in the 2014 World Cup and at the 2016 Euro, new coach Julen Lopetegui has slowly improved overall play. This has particularly involved raising the quality of the Spanish attack. As Spain discovered in the last two major competitions, it is not enough to have possession for 80% of the game if one cannot do anything with the ball. The Finals: Spain Vs. France BCA's Two-Step World Cup model predicts that, on July 15, 2018, the world will see Spain dispatch France in the finals of the world's sporting pilgrimage (Diagram 7). The two teams have the highest conditional probability of traveling down the grueling camino a la gloria eterna, from the group stages to the final (Chart 7). Intriguingly, of the teams knocked out in the quarter-finals, Portugal shows up with a slight conditional probability of winning the tournament. The England-Colombia matchup is essentially a coin toss. Diagram 7Road Map Of The World's Sporting Pilgrimage The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Chart 7Conditional Probability Of Entering Elysium The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Our model gives Spain a 59% probability of beating France (Chart 8), large enough to give us confidence, but not an overwhelming figure. It is undeniable that Spain has superior quality across the pitch (Diagram 8). Chart 8Final Matchup Probabilities The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Diagram 8La Furia Roja Vs. Les Bleus The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Moreover, Spain clearly excels over France in its defense - it conceded just three goals in its ten qualifier matches - and in its club synergy value, where it scores the highest mark. Five players sampled in our data play for F.C. Barcelona and five play for Real Madrid. What of the game for third place? Surprisingly, it may be the game of the tournament.20 Sure, third place means little. However, this game will mean a lot. If our model is correct, Brazil will face nemesis Germany in a revenge game. The young Brazilian team, completely revamped after the 7-1 Blitzkrieg on home soil, will be playing for the future, for revenge, and for national pride. Our model gives Brazil only a slight edge over Germany (Table 21). This is unsurprising, given that Brazil is, by far, superior in defense, forward, and speed positions, even though it is considerably inferior where it matters: the midfield (Diagram 9). Table 21Third Place Match Summary Results The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup How does our forecast compare with current betting odds? Chart 9 shows that our model gives Spain an extraordinarily high probability of winning the tournament. Spain's 32% odds are double what the bookies' favorites, Brazil and Germany, command. Overall, we think that the betting market is underestimating the odds of a Mediterranean champion, while overstating the odds of both Germany and Brazil. Diagram 9A Selecao Vs. Die Mannschaft The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Chart 9We Do Not Condone Betting!##BR##(But If You Were Wondering...) The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup II. Teams And Narratives To Watch In Russia: A Qualitative Assessment The World Cup is not just about winning. It is the most watched event - sporting or non-sporting - in the world because it weaves so many narratives and themes together. In this section, we explore some of these narratives and themes. Some are investment relevant, some are geopolitically relevant. Our qualitative assessments are ordered from the lowest-ranked teams to the highest-ranked. South Korea - A New Era Dawns, With The Same Football Results Coming from a country that has imprisoned every former president, including the one it just impeached, it should come as no surprise that South Korea's national team lacks stable leadership. The Koreans had a remarkable coach - the German sweeper Uli Stielike - who had dedicated three years to preparing them for this tournament. They fired him last year, however, due to a handful of losses. He has been replaced with a native son - a known advantage in World Cup football - Shin Tae-yong. Shin Tae-yong threw a special winter training camp for the team in order to establish himself as coach and get to know the squad given such a short time remaining before the tournament. Activities consisted of drinking tiger's blood and hiking up to the Unicorn Lair of Kiringul where, rumor has it, former North Korean dictator Kim Jong Il was conceived. South Korea made it to fourth place when it co-hosted the World Cup in 2002. While it had a lot to do with the Flying Dutchman coach, Guus Hiddink, he was not playing on the field. Rather, it was the insatiable Korean thirst to perform better than their co-host and erstwhile colonial overlord, Japan.21 In 2018, South Korea has a good defense: they seldom have let the ball in the net in recent tournaments. The only problem is that they do not know how to score goals. Young star player Son Heung-min, of Tottenham Hotspurs fame, has shown that he can score multiple goals late in the game - but only when playing against Uzbekistan. Of course, South Korea does have a tremendously compelling national story this year: the war with North Korea is finally over! President Moon Jae-in and his North Korean counterpart, dictator Kim Jong Un, held the third Inter-Korean summit on April 27, and declared for the first time since 1952 that they would stop trying to kill each other, at least formally. Peace is the word of the day. Can the optimism of an era free of war, with potential Korean reunification on the horizon, translate to a morale booster that helps Korean athletes? Probably not. Our model assigns Korea only a 3% chance of making it out of the group stage. The 2018 Winter Olympics were held in South Korea - they even entered the opening ceremony jointly with their communist neighbors. But while the South Koreans did well, winning several golden medals, they did not set a new record. Unlike the South Korean women's hockey team, the South Korean men's football team won't receive an infusion of North Koreans. Nor will they host a North Korean art troupe. Although we doubt that either would help their odds in Russia. Egypt, Morocco, Tunisia, And Saudi Arabia - Renaissance Or Regression? It is now eight long and turbulent years since the Arab Spring ignited with protests in Tunisia. Since then, stagnant regimes have been toppled, hundreds of thousands of people killed in civil wars, and an Islamic State has been created and crushed. Tunisia and Morocco stand out as clear outliers in terms of geopolitics. Since the 2014 parliamentary elections, Tunisia has begun paving the way towards becoming the first country in the Arab world with a functional democracy. Morocco is far from a democracy, but has remained stable over the past decade in stark contrast to the region. King Mohammed VI has enacted a number of reforms that have forestalled, for now, expression of grievances from the public. Saudi leadership has decided to follow the Moroccan example and preempt social unrest by enacting wide ranging reforms from the top. The 33-year old Crown Prince Mohammad Bin Salman has not only pledged to reform the economy, but also to fundamentally change conservative Saudi society. Women have been given the right to drive, the religious police has been regulated into irrelevance, and there is even talk of eliminating public gender segregation. BCA believes that these reforms are genuine and that they will be executed with vigor.22 Saudi leadership is reacting to the reality that the majority of the population is under 30 years of age (Chart 10). Mohammad Bin Salman therefore did not "come out of nowhere," as many commentators claim. He is the regime's response to the socio-economic realities of Saudi Arabia. Chart 10Saudis Are Catering To The Youth The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup In contrast with these three countries, Egypt has stagnated in terms of governance and political reforms. In many ways, the country has regressed back to the military-backed rule that preceded the Arab Spring protests. However, President Abdel Fattah el-Sisi has launched and followed through on some private-sector and macro reforms which should result in marginal improvements for the economy.23 If football is a reflection of socio-economic conditions, then the participation of these four states at the World Cup is a harbinger of a brighter future. In Russia, Arab countries will have the highest number of representatives ever in the World Cup. Given that three of the four countries qualified in the highly competitive African continental qualification tournament, the result is impressive. Our quantitative model gives Morocco, Saudi Arabia, and Tunisia little chance to reach the knockout rounds. Egypt, on the other hand, is only eliminated because we have decided to give Russia a host premium. Since 1930, as a host, only South Africa failed to qualify for the round of 16 in 2010. Egypt has a real chance to be one of the dark horses in this year's tournament. It has the requisite footballing tradition to be able to withstand the pressure of a major tournament. Egypt has won seven African Cup of Nations, the most among its continental peers, and nearly half of its likely squad play in major international leagues. The most important among these is Mohamed Salah, "The Egyptian King," of Liverpool F.C. Salah may be the best forward in the world at the moment and will certainly wreak havoc in the group stage against Russia and Saudi Arabia. These four Arab states have already accomplished a lot by qualifying. But if one of them progresses to the knockout stage, it would be notable. BCA Emerging Markets Strategy research has shown that genuine structural reforms that improve the quality of governance are required for long-term productivity growth, and thus, asset performance.24 But a change in attitude comes first. Confidence is an important part of that change, and seeing Mo Salah and Egypt take on the world's best could light the spark for a Middle East Renaissance. Iceland - A Footballing Black Swan Iceland's football team success is a great reminder to investors that the probability of unexpected events is often greater than the consensus expectation. Iceland has a population of 330,000, just slightly larger than Swansea, Wales' second-largest city and home of "the Swans," the perennial bottom-feeders of the English Premiership League. If one eliminates Iceland's female population, the infirm, and the too old and too young, the pool of available humans for the country's football team is about 40,000. Iceland's qualification for the World Cup is extraordinary. Its success at the Euro 2016, where the country advanced to the last eight teams after defeating England, is nothing short of evidence of the divine. It may be the greatest upset in sports. Ever. Consider that in order to even participate at the Euro 2016, Iceland had to survive a qualification group that included the Netherlands (which failed to qualify!), Czech Republic, Turkey, Kazakhstan, and Latvia.25 Iceland finished second in the group and thus qualified directly for the event. Then, to get to the final eight at the Euro, Iceland had to survive a group made up of Hungary, Portugal, and Austria (where it finished ahead of the eventual Euro champion Portugal!).26 It defeated England in the round of 16 by a score of 2 to 1, only to finally be vanquished by the host nation France.27 What makes Iceland's success so astonishing is statistics, specifically the concept of conditional probability. Because qualification for a major event, such as the Euro, requires successive results, the conditional probability of Iceland getting through to the quarter-finals eventually is close to zero.28 Especially when one considers that there is no element of surprise for a country like Iceland once it shocks the world by qualifying. This is not college basketball, where the knockout nature of the March Madness tournament - and lack of scouting of small universities - creates the potential for upsets. This is international football, played by grown men getting paid millions of dollars to do their job (i.e., not lose to Iceland). What is the secret behind the success of Strákarnir okkar (Our Boys)? Several theories have been advanced: investment in coaching, heated football pitches, "Nordic mentality," "The Breath of Odin," etc.29 To us, all of this smells of the hindsight rationalization that Nassim Taleb identified as the hallmark of Black Swan events.30 Icelandic footballing success is a shock, a rift in the space-time continuum, and unlikely to be replicated. Our model gives Iceland little chance against Argentina, Croatia, and Nigeria. Of course, the whole point of a Black Swan event is that it is impossible to model. So, sit back, relax, and enjoy the Viking War chant!31 Japan - Bad Timing For New Leadership Japan is a nation in ascendancy, having emerged from its "Lost Decades" in recent years to reclaim a leading position among the nations of the world. What stirred the giant from its slumber? A global financial crisis, earthquakes, tsunamis, nuclear meltdowns, debt-deflation, and the revival of its ancient Chinese foe. A new era is dawning. That is fortunate, as it means that Japan's failures in the World Cup will fall under the final year of the outgoing emperor's reign. The problem, once again, is the lack of stable leadership. Just as factions in the ruling Liberal Democratic Party are busy trying to remove the hugely successful Prime Minister Shinzo Abe before a critical party leadership vote, so factions within the Japan Football Association have removed the head coach, Vahid Halilhodzic, just two months ahead of the tournament. Unlike his Korean counterpart, Akira Nishino has not held a winter camp to familiarize himself with the team. Instead, he will rely on the rock-solid supposition that the Japanese would think it ignoble to be led into battle by a foreigner. There is a basis for believing that teams do better under the leadership of a coach who is a national. But there is only so far that ethnic solidarity can go. After all, this is a team that wrestled Haiti and Mali to a draw. To throw its coach under the bus at the last minute is to ensure that Japan's squad remains, in the words of William Durant, "inclined to picturesque suicide". Croatia & Serbia - Strength In Numbers Extrapolating into the future from their vantage point in 1990, leaders of Yugoslav federal republics Slovenia and Croatia made an understandable forecast. First, communism was dead, and it was time for economic and political reforms. Second, "economies of scale" no longer mattered. The future was in global trade and open borders. Therefore, membership in a still-communist federal Yugoslavia dominated by semi-authoritarian Serbia was suboptimal. Looking back at the decision today, it is hard to argue with the results. Slovenia is in the Euro Area with a GDP per capita of $21,304 (2016), while Croatia is close behind, in the EU with a GDP per capita of $12,090 (2016). Meanwhile, their neighbor and aggressive "big brother" Serbia, which fought a war against both to keep them in Yugoslavia, is on the outside looking in. Its GDP per capita ($5,348 in 2016) has only recently recovered to the 1989 Yugoslav levels! There is no argument that Croatia and Slovenia made the right choice, at the time, by choosing secession. But the question is whether they would have still done it knowing everything we know about the present. For one thing, we cannot extrapolate globalization into the future. We are, in fact, at its apex.32 Strength in numbers will matter in a de-globalized, multipolar world of the twenty-first century. And while Croatia and Slovenia are part of a bigger whole - the EU - their size relative to the EU population and economy is laughable relative to their importance in Yugoslavia (Chart 11). Chart 11Small Parts Of A Bigger Whole The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup And then, of course, there is sport. Both Serbia and Croatia are present in Russia. Our model ranks them 14th and 15th respectively. If we combine their rosters into 23 top-ranked players, their ranking jumps to eight, high enough that a semi-final berth becomes a possibility given good luck. This is not the best team that former Yugoslavia could have featured. In 1998, Croatia came third in France with an over-the-hill squad. But in 1994, Serbia and Montenegro (still called Yugoslavia at the time) were banned from competing while that same Croatia was unprepared to compete in the qualifications, which began just after the country's independence. That 1994 team would have featured in-prime Croatian stars33 - who came within two Lilian Thuram goals of the 1998 finals - and a team of Serbian,34 Montenegrin,35 and Macedonian36 players who took Red Star Belgrade to its Champions League title in 1991. Given the relatively weak competition in 1994 (Bulgaria and Sweden made the semi-finals), Yugoslavia could have been crowned World Cup champion in 1994. Of course, if pigs had wings, they would fly. On the other hand, the fact is that Slovenia and Croatia are among the two most Euroskeptic countries in the EU. Clearly, there is some buyers' remorse in both. Perhaps as part of a reformed, democratic, and federal Yugoslavia, both would have enjoyed greater clout in Brussels and thus, greater say in what kind of policies the EU imposed on them. And if that did not work, at least looking at the shiny 1994 FIFA World Cup Trophy would have helped put everything else into perspective. Russia - The World Cup As The Ultimate Lagging Indicator Expectations are low for the Russian national team at the 2018 World Cup. Despite its size, population, relative wealth, and strong tradition of government support for sports, Russia is just not a footballing nation. Its best result at the World Cup was way back in 1966 (fourth place), although it did win a European title in 1960. That was more than half a century ago! As a host, we would expect Russia to get a bump out of its woefully easy group. But from there, the focus will shift away from the Sbornaya to the organization of the tournament. This is the second time Russia is hosting a major international competition, following the 2014 Sochi Winter Olympics, and all eyes will be on the updated infrastructure and tight security measures. The World Cup presents a much greater challenge than the Sochi Olympics because there are 11 venues spread out over a geography the size of Western Europe. Three of the venues - Volgograd, Rostov-on-Don, and Sochi - are also close to the Caucasus region. Russian security forces essentially encircled Sochi in 2014, severely limiting movement into and out of the Black Sea resort town.37 This is unfeasible to do for the country's 11 largest cities. Chart 12Russia - The World Cup As##BR##The Ultimate Lagging Indicator The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup There is an additional problem of limiting hooligan violence. We assume that Russian law enforcement will be much better at limiting the influx of troublemakers than its European neighbors. However, Russia has plenty of domestic hooligans to create violence, which was on clear display at the Euro 2016 clash between English and Russian "fans." All of this brings up the question: Why is Russia organizing the World Cup in the first place? Especially given its semi-pariah status following a spate of controversial geopolitical events thus far in 2018? The answer is that major sporting events are the ultimate lagging indicator of a country's economy, its geopolitics, and market performance. The Sochi selection was made in 2007, amidst an epic commodity bull market and at the peak of the "BRICS are taking over" narrative. The World Cup selection was made in 2010, well before Crimea was annexed, Ukraine was destabilized, and Russia decided to play "peacemaker" in Syria (Chart 12). Just like Brazil in 2014, the World Cup therefore arrives to Russia at an odd time, with the market, economy, and the country's relations with the West hitting rock bottom. Of course, the 2014 World Cup was the bottom for Brazil, with a spate of seismic economic and political changes following. Brazilian equities are up nearly 50% from the closing ceremony of the World Cup to today. Can the same happen in Russia? Mexico - Will Anything Change? Mexican voters will go to the polls on July 1 as its national football team competes in Russia. The election has been dubbed the "biggest election in Mexican history" by the country's National Electoral Institute. We agree. For the first time in Mexico's history, a left-wing, anti-establishment movement has a chance to win (Chart 13). Chart 13The "Biggest Election In Mexican History"? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Will Mexico's football team also break with tradition? For the past six World Cups, Mexico has emerged from the group stage only to be promptly eliminated in the knockout stage every time. In 2018, El Tri again have a great chance to emerge from the group. While Germany is a formidable foe, the other European nation in the group, Sweden, barely qualified and is playing without its only top-class footballer, Zlatan Ibrahimovic. As such, Mexico should accomplish the usual feat relatively easily. The question is whether it can do more. Unfortunately, second-place finish in its group will pair it, baring a major surprise, with Brazil. This is where its road will most certainly end. Will politics mirror football? Is Andres Manuel Lopez Obrador (AMLO) just more of the same? We are hopeful that Mexico's left-wing answer to President Donald Trump could actually make some changes for the better. First, the Mexican Congress is unlikely to give AMLO free rein in governing the country. Second, AMLO has moderated his campaign, purposefully hiring centrists and technocrats to signal moderation. Third, AMLO has the charisma and the gravitas to sit down with President Trump and negotiate face-to-face, unlike his predecessor, Enrique Peña Nieto, who never quite got used to Trump's rhetoric and aggressive style. In the end, if AMLO fails to make serious inroads with economic reforms, ongoing drug-related crime, and negotiations with President Trump, at least Mexicans will have one major success to take from 2018. Mexico qualified for the 2018 World Cup, whereas its eternal football rival the U.S. did not! Belgium & Switzerland - The Upside Of Immigration Let us state the facts. Without footballers of immigrant descent, we would not be writing an analysis on the Belgian and Swiss football teams today. Of the 23 call-ups for the international friendlies ahead of the World Cup, 13 were of immigrant background for the Swiss team and 12 for the Belgian. This includes the majority of the stars of both teams, such as Granit Xhaka (Albanian) and Ricardo Rodriguez (Spanish-Chilean) for Switzerland, and Vincent Kompany (Congolese-Belgian), Marouane Fellaini (Moroccan), and Romelu Lukaku (Congolese) for Belgium. In large part, this is a tired topic already well covered by the Belgian and Swiss media. It also draws on the main narrative following the 1998 French World Cup win. Half of the legendary Les Bleus team was made up of players with immigrant backgrounds, including essentially all of its stars. Of the 14 goals scored by the French team, nine came from its multicultural stars Zinedine Zidane, Youri Djorkaeff, Thierry Henry, Lilian Thuram, and David Trezeguet. What separates the 1998 French team from the Belgian and Swiss teams today, however, is the geopolitical context. The 2015 migration crisis in Europe - when the influx of illegal immigrants reached a peak of 220,000 per month in October of that year - still dominates politics on the continent (Chart 14). Although the crisis is now definitively over - the monthly influx figure is below 3,000 - both anti-establishment and establishment parties have swung hard against immigration. Chart 14What Migration Crisis? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup But wait, are these players really immigrants to begin with? The answer is an obvious and definitive no. Only six Swiss and none of the Belgian players were born outside the country. Even the 1998 French team had only two players who were actually born outside of France (and an additional two were from overseas French territories). This reflects one of the greatest misunderstandings in the political narratives regarding European immigration. There essentially is no longer any non-European immigration to Europe. At least not via legal channels. Players who are visible minorities on these teams are almost universally the children and grandchildren of the post-World War II immigrants from the former colonies (and three of the six foreign-born Swiss players are not visible minorities at all, given that they are from the former Yugoslavia). Several European countries undoubtedly have a problem integrating non-European immigrants. Our colleague Peter Berezin has made a connection between a generous social welfare state and lack of successful integration, with Belgium and Sweden as prominent examples of the connection.38 However, few commentators understand that there is no ongoing large-scale influx of immigrants to Europe, aside from occasional migration crises prompted by wars. The migrations of the 1950s and 1960s, from former colonies, was followed in the 1960s and 1970s by "guest worker" migration. Since then, Europe has progressively tightened its migration policies to all but internal migration from EU member states. What does this mean for Europe's national football teams? Nothing. Football is a sport played by the middle and lower classes almost universally.39 The high proportion of visible minorities on Europe's football rosters is simply a reflection of the fact that many immigrants and their kids grow up in precisely such communities. Eventually, they become French, Belgian, or Swiss. After all, nobody would today doubt Michel Platini's Frenchness. But he is as much a product of immigration as Zinedine Zidane or Zlatan Ibrahimovic. England - A Football Brexit It is well known that the English national team is cursed in World Cup football. If not for the home team advantage, they never would have won their single trophy in 1966. Many commentators in England lament the strength of the English Premier League. Best teams are collections of superstars from other countries who receive the highest wages in football, yet naturally return every four years to play for their mother country. In other words, the English fight wars with mercenaries hired from abroad whose loyalties lie elsewhere. Imagine that! The English national team has chronically underperformed despite having one of the most expensive, and productive, leagues in the world. Like many other U.K. industries, English football leverages the high productivity of extraordinary immigrants. Indeed, the percentage of foreign players in the English Premier League is the highest in the world (Chart 15).40 Chart 15The Most Globalized Labor Market? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The immigrant free-for-all, however, almost ended. Just as concerns over immigration prompted Brexit, concerns that English football was being dominated by the non-English almost led to changes to the domestic league Homegrown Player Rule. A proposal to put more stringent rules on foreign players in order to fix the talent pipeline in English football almost took hold under the previous FA chairman, Greg Dyke. Fortunately, the worst effects of the Brexit vote have already passed: The team suffered their worst upset in 66 years at the hands of Iceland in the 2016 Euro Cup, just days after the referendum was held. Remember, there are fewer people in Iceland than there are Sikhs in the United Kingdom. This devastating upset was the first material instance of "Bremorse" in the wake of the referendum. The team is unlikely to suffer such an upset again. Instead, it will perform just like it did in 1066: defeating the likes of the King of Norway and Iceland only to be invaded and enslaved by the King of France. But unfortunately for England, it does not get a rematch with Iceland in group play. Rather, its group of four affords an occasion for an even greater humiliation, perhaps even the crowning humiliation of the entire Brexit saga: a loss to Brussels. Such a loss is likely if only because the signal lesson of British history teaches that Europeans can never successfully invade the island unless it happens to be suffering from deep internal divisions. And with Jeremy Corbyn and a second Brexit referendum on the horizon, such "intestine" divisions are utterly assured. Given all of the above, does it seem likely that this is the year in which England will break a half-century-long curse? No. But our model gives the young team great odds of reaching the knockout stage. Argentina - Promise Vs. Performance Every investor knows the old adage that at the turn of the twentieth century, Argentina was "one of the ten richest countries in the world," ahead of European heavyweights France, Germany, and Italy. And that in 1950, Argentina's GDP-per-capita was six times greater than the likes of South Korea. After half a century of Argentine economic mismanagement and unfulfilled promises, South Korea has today doubled Argentina's national wealth (Chart 16). Chart 16Argentina - Promise Vs. Performance The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup At probably the height of the country's economic and political mismanagement - during the nearly decade-long military rule between 1976 and 1983 - Argentina's football team won two World Cups: 1978 and 1986. Since then, however, it has left behind a legacy of unfulfilled promises and lost opportunities that largely mirror its last thirty years of economic performance. Since the 1990 World Cup in Italy, Argentina has entered every World Cup as one of the favorites. Its fans consider the team one of the world's five greatest footballing nations. To its credit, it has made the finals twice, in 1990 and 2014, and the quarter-finals three other times in that span. But its status as a perennial superpower is under threat. We would argue that Argentina is, in fact, the greatest underperformer in the last three decades of international football. Despite being the fourth highest-ranked team over the past two decades - according to Elo scores, which we interpret to represent the consensus view - the country has failed to lift the World Cup. The greatest threat to Argentina's legacy as a footballing superpower is that, in another twenty years of unfulfilled promises, fans will look back at its past success the way we think of Uruguay's two World Cups, won in 1930 and 1950. One could argue, for example, that its two titles are a product of a single, golden generation, not of eternal status as a superpower. Since Diego Maradona stopped playing, Argentina has only advanced past the quarterfinals once. Coming into the 2018 World Cup, the expectations are once again high. Argentina still has one of the world's top-two players in Lionel Messi. And Elo again ranks it as the fourth-best team, with expectations of at least a semi-final appearance. Our model is more skeptical, dropping it to a sixth-best ranking and predicting yet another quarter-final exit. If Argentina wastes its "Messi generation," the chances that its unfulfilled football promises mirror that of its economy will grow. Portugal - The "Great Man" Theory The "Great Man Theory" posits that a large portion of human history can be explained by the impact of "Great Men" (and women) on global events. Popular from the Greek and Roman classics through the nineteenth century, the theory has been contradicted by post-modern philosophers, sociologists, and historians. BCA's Geopolitical Strategy largely rejects the idea in its methodology. We focus on the constraints to "Great People," rather than their preferences. After all, preferences are optional and subject to constraints, whereas constraints are neither optional nor subject to preferences. Enter Cristiano Ronaldo dos Santos Aveiro. Perhaps the greatest footballer ever, Ronaldo is indeed a Great Man. He has willed his teams (whether at club level or internationally) to extraordinary feats. Even in injury, his defeat on the battlefield inspired his teammates to play incredible football in the Euro 2016 final against a superior France. Ronaldo lacks mortal constraints on the football pitch. He has scored 308 goals in 289 appearances for Real Madrid (since 2009), an incredible per-game ratio (Chart 17). In the 2017-2018 season, at the age of 33, he has scored 41 goals in 38 appearances, maintaining a per-goal average of his entire career at the Spanish club. His game appearance total this season is lower than at any other time since 2009-2010, suggesting that he will come into the World Cup fresh and rested. Chart 17The Ronaldo Effect The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Football, however, is a game played on a wide pitch by 11 players. How can one man, even one as great as Ronaldo, make a difference? Gravity. Ronaldo is such a dangerous player with the ball that defensive midfielders and defenders have to constantly gravitate towards his presence on the pitch. This opens up lanes for his Portuguese teammates to counterattack. And Ronaldo's teammates are not as bad as the consensus thinks. Our quantitative model ranks Portugal seventh, just a few points below Brazil. They are also fortunate to be in a relatively easy group. While the matchup with Spain will be a tense affair, Portugal will easily dispatch Iran and Morocco. And if they come second in the group, they will face the easiest group winner of the next round, Group A's Uruguay. While we do not subscribe to the Great Man theory as a methodology, we respect it. Over the long term, and over a great number of iterations, focusing on great individuals is folly. But when faced with specific events, there is definitely value in thinking about how extraordinary human effort matters. Portugal's path to the quarter finals is easy. Once there, games will become tighter, the pressure unbearable, the stage infinitely larger. It is in these moments that legends play like Great Men, while lesser footballers play like mere mortals. Brazil - Redemption? As host of the last World Cup in 2014, Brazil didn't so much hit rock bottom as descend to the ninth circle of hell. Its merciless 7-1 obliteration by Germany in the semifinals - the sharpest rebuke to home-team self-confidence in history - would have scarred the national psyche even if the country had not proceeded into the worst economic recession in a hundred years. At that point the country's politics became a Shakespearean tragedy in which a handful of corrupt aristocrats slaughtered each other on stage, for all the world to see. Chart 18No Neymar Home The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup "Brazil is back," say the fans - just as the bulls in the financial markets. The argument goes like this: The crisis has proven that the country has resilient institutions. Inflation (read: hubris) has collapsed to the lowest point in years, while the country has tightened its belt (both financially and figuratively), in a long-overdue act of self-discipline. The only difference is that the football comeback, unlike the economic one, is built on firm foundations. The Brazilian squad has made an impeccable run since Dunga stepped down as coach in 2016. It was the first team to qualify for this year's World Cup, going undefeated for quite a stretch of time. Today, the team is coached by Tite, a former player. The least we can say is that he reinvented the way Brazil plays. Brazil's lamentable disappointment under coach Luiz Felipe Scolari ultimately comes down to the need for a new football identity and a new crop of players. That's what they have today. They are a very good team, and freshly humbled. We expect them to go far, though not all the way. It is not certain that all of Brazil's demons have been purged. By way of illustration, the leading presidential candidate is an advocate of military dictatorship, with no pro-market candidate in sight (Chart 18). France - Marchons, Marchons! In 2016, France lost to Portugal in the final of the Euro Cup on home court. This was not the first gut-wrenching loss in the final of a major tournament. In 2006, its aging 1998 generation lost a bitterly fought final against Italy. In Russia, France comes as one of the favorites. Its team is celebrating the 20-year anniversary of its epic 1998 run. Meanwhile, at home, France is marching in lockstep to President Emmanuel Macron's reforms. The economy is booming, cantankerous unions are either defeated or capitulating, and even equity market performance is leaving traditional competitors in the dust (Chart 19). Chart 19France - Marchons, Marchons! The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Our model sees France reaching the final of the World Cup. However, it also predicts that France will lose this final, a repeat of the 2016 and 2006 performances. Les Bleus will come close, but not light the victory cigar. Football in France is embedded in the cultural fabric; its team is portrayed by the media as a reflection of the contemporary socio-economic narratives. When France won in 1998, for example, the team's multicultural heritage was touted as a model of integration and as a symbol of open-minded, post-colonial France. The 2010 debacle in South Africa, on the other hand, where French players literally went on strike following an altercation with their coach, was presented as a symbol of declining French relevance and growing national impotence. Since 2014, however, positive momentum has been building both on the football pitch and in real life. A young team managed to qualify for the 2014 World Cup, losing to the eventual champion Germany in a closely contested quarter-final game. Two years later, that same team defeated World Champion Germany in a semi-final of the Euro tournament. On the political front, 39-year old Emmanuel Macron swept aside populists and left-wing firebrands, stunning the world by campaigning from a staunchly centrist perspective. The question for France in 2018 is whether it will fall short of victory, yet again. We ask the question both in terms of the performance on the football pitch and in politics. The youthful, energetic, president mirrors the youthful, energetic, squad. But a lot is at stake and second place may not be good enough for either. We remain optimistic that genuine reforms are coming to France.41 And despite our model's preference for Spain in the finals of the World Cup, France has a great chance of repeating the glory from twenty years ago. Spain - Divided It Stands, And Wins On October 1, 2017, 92% of Catalan voters decided that Catalonia should secede from Spain in a referendum deemed illegal by Madrid.42 Soon after, images of old ladies and students being beat up by riot police flooded the internet, as Mariano Rajoy's government signaled its refusal to compromise on the question of independence. Chart 20Catalan Separatism: Overstated Risk The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Intriguingly, Spain's most independent-minded regions have often produced some of its best footballers. Cesc Fàbregas, Gerard Piqué, Carles Puyol, Xavi Hernandez, and Sergio Busquets are all of Catalan descent and have dutifully defended the Spanish colors in several successful Euro and World Cup campaigns. Meanwhile, the bitter Barcelona-Real Madrid rivalry has been at the heart of Spanish football and deeply embedded in the country's political and socioeconomic history for generations. In other countries, such divisions and rivalries would let politics get in the way. In Yugoslavia, the civil war is famously said to have begun with fan rioting at a 1991 Dinamo Zagreb vs. Crvena Zvezda match. But Spain, somehow, channels the chaos into beautiful, and effective, football. Pro-independence sentiment has begun to waver in Catalonia, as BCA's Geopolitical Strategy predicted (Chart 20).43 This is bullish for the Spanish economy and assets, but also for its football team. Spain conquered the world of football at the height of its economic crisis, in 2010. Will it do so again, on the heels of its political crisis? Our model predicts yes. And we agree. It would seem that Spain's greatest attribute as a nation is to produce diamonds under pressure. III. Investment Implications: Does The World Cup Matter? The FIFA World Cup is the most widely watched sporting event in the world by far (Chart 21). Does all that passion and emotion translate into any economic or market implications? Chart 21World Cup Is The Most Important Unimportant Event In The World The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Academic literature on the topic is, at best, inconclusive. International football, as an industry, is largely inconsequential. For instance, the wealthiest football club in the world - Manchester United - generates less than 8% of the revenue of the 500th company in the Fortune 500 index. To assess whether the World Cup matters, we take two approaches. First, we explore the implications of hosting the event on the economy of the host nation. Second, we examine the implications of the World Cup on equity markets. Hosting The World Cup Despite the size and the reach of the World Cup, we find little evidence that hosting the event is beneficial to the host's economy. There appears to be little to no effect on either a short or long-term horizon. We found no "World Cup effect" on any of the economic variables one would assume may be impacted. Nonetheless, some patterns are worth highlighting (Table 22). Table 22Is It Worth It? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup We chose to look at the behavior of each economic variable one year before and after the World Cup, with the exception of core consumer price inflation and the unemployment rate, which we observed on a three-year horizon around the event. The rationale behind this is that the labor market tends to react slowly to changes in underlying economic activity, due to factors such as the rigidity of employment contracts and stickiness of wages. The study starts with the 1990 World Cup hosted by Italy. Interestingly, some patterns are apparent. Out of all the variables we analyzed, inward FDI tends to consistently increase in the host country following the World Cup. Our suspicion is that hosting the event demonstrates, on the margin, that the country meets suitable conditions (governance, return on capital, etc.) to attract foreign capital. On the other hand, we may be capturing a trend already underway, which itself was revealed by the fact that the country was selected to host the World Cup in the first place. As Table 23 illustrates, the selection process lags the actual event, as with any major sporting event (such as the Olympics). Intriguingly, the time-lag is, on average, six years, almost precisely the length of an economic cycle. As such, the decision to award the hosting of the event to a particular country may already take into account the bullish macroeconomic cycle. By the time the event is hosted, the cycle is in full swing and may actually be peaking. Table 23Soccernomic Cycles? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Consistently, capital expenditures also tend to increase following the Word Cup. As more foreign capital flows in the country, more investment takes place. Currencies, however, on a one-year horizon, do not react to these FDI flows since the short-term gyrations are dictated by more volatile short-term portfolio flows. As FDI flows in, the unemployment rate falls and tends to continue falling after a World Cup. However, some caution is advised in interpreting these conclusions. Macroeconomic business cycles, commodity prices, and other structural factors are still dominating factors in determining the trend and health of economies. As we indicated in our analysis of Russia in the qualitative section above, the World Cup itself can be considered a lagging indicator of the economy. This certainly appears to be the case with the three members of the BRICS who have hosted the event: South Africa, Brazil, and Russia. As such, hosting the World Cup may signal the top of the bullish cycle, as it certainly did for Brazil. This warrants a few key observations: Brazil did not see a rise in capital expenditure after hosting the World Cup in 2014. In fact, it suffered one of the deepest recessions in the last century. Consistently, unemployment did not fall as it did for other countries as commodities entered a bear market in 2015 and weighed heavily on the economy. Although South Africa did see its unemployment rate fall before the World Cup, the trend did not continue following the event. In the few years leading up to the global financial crisis, the world economy was in the midst of a period of stellar growth. Therefore, it makes sense that Germany was experiencing rising FDI, capital expenditure, and inflationary pressures prior to hosting the event in 2006. In addition, the implementation of the Hartz IV reforms in the early part of the decade put an end to the long-term structural uptrend in the unemployment rate, which has since declined by more than 7%. The year 2002 was an important one for Japan, as the turn of the millennium marked the end of the "lost decade". Subsequently, Japan experienced a falling unemployment rate and rising GDP growth. However, the end of the "lost decade" also saw the beginning of the self-feeding deflationary expectations which have anchored expected inflation levels near zero. As such, consumer prices fell before and after the event. France went through a period of intense reforms in the late 1990s. Prior and subsequent to 1998, growth rates were falling simply because 1998 recorded a higher growth rate of 3.3%, compared to 1.6% in 1997 and 3.1% in 1999. The lowering of the VAT rate, income, and corporate tax by the Chirac government led to real capital expenditure and real household consumption increases during the late 1990s. Furthermore, substantial labor market reforms helped to decrease the unemployment rate. The 1990s were the longest period of economic expansion in U.S. history. In fact, 1994 was the year where the number of jobs created was one of the largest on record, at 3.85 million. As such, real GDP growth and business investment expenditures were high, and the unemployment rate was falling. The prospects of high growth also brought in increased foreign direct investment in the same period in which the World Cup coincided. However, emerging markets were seeing an equally stellar period of growth, which was supported by high capital inflows and appreciating EM currencies, thus leading to a lower U.S. dollar. Although the Federal Reserve began hiking interest rates in 1994, the effect on the dollar was not registered until the beginning of 1995, which also marked the beginning of a dollar bull market. In the 1990s, the Italian government was heavily liberalizing the economy, leading to its eventual inclusion in the Euro Area at the end of the decade. Italian growth during the late '80s and early '90s was high enough for it to surpass Britain and France in terms of GDP growth, consumption and investment growth in 1990 were high. However, the 1990s oil price shock depressed consumer and business activity. Bottom Line: In line with the academic literature, the results of our assessment are at best inconclusive. However, the data does suggest that hosting of the World Cup coincides with some positive macroeconomic developments, particularly in terms of FDI inflows. This, however, may reflect the fact that the decision to host the event is usually made at the beginning of a bullish upcycle, which means that the actual event marks the peak, or top, of the cycle. Equity Market Implications The World Cup may be most relevant, not as a catalyst for market action, but rather as a distraction. The European Central Bank, for example, concluded in a 2012 study that trading volumes dropped considerably during match times.44 The World Cup in Russia could be particularly distracting. Because of Russia's time zone, most matches will be played between 8:00 am and 2:00 pm in New York, and 1:00 pm and 7:00 pm in London. As opposed to the economic impact, there seems to be a general consensus in the academic literature regarding the equity market implications of World Cup matches. For instance, a cross-sectional analysis of 39 countries looking at World Cup and relevant qualifying matches between 1973 and 2004 (that represents 1,162 matches!) concludes that losses have an economically and statistically significant negative effect on the losing country's stock market.45 Furthermore, the study finds that this effect is stronger in small caps, which are disproportionately held by local investors and therefore are more strongly affected by sentiment following a football match. Another study found that being short NYSE and long Treasuries during World Cups would produce a higher return from 1950 to 2007.46 This is due to the decisions taken by foreign investors in U.S. bond and equity markets. During the World Cup, these investors would withdraw capital to invest in their local markets, which would put downward pressures on NYSE prices. Table 24 presents the daily stock return deviation from what one would expect outside of World Cups for all the matches in the past seven World Cups, starting with the quarter-finals (for the period from 1990 to 2014). We report the local equity market movement of the teams playing up to three days after a match took place. Table 24Do World Cup Matches Impact The Local Stock Returns Of The Teams Playing? The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup Several observations can be made: We observe positive "abnormal" returns, i.e., deviations from trend, in the day following the matches, for any of the teams playing, and larger "abnormal" returns when it comes to the winning team. While we observe positive abnormal returns for the winners at t+1, it appears that more than 70% of the daily returns used to compute this average turn out to be negative - implying that, on average, the winners' local equity markets experience small negative returns and, in a few instances, large abnormal positive returns. In line with the literature, we observe negative abnormal returns on t+1 for the losers, regardless of how we define our benchmark. A high number of abnormal daily returns, up to three days after the match, appear to be negative (this is also the case regardless of whether we compare it to the past average, median, or non-World Cup daily returns). Bottom Line: We do not recommend that investors "play" the World Cup. Generally speaking, equity markets react, on average, positively to wins and negatively to losses. However, this is not always the case. The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup The Most Important Of All Unimportant Forecasts: 2018 FIFA World Cup 1 There are, however, some less obvious investment implications ... but we do not encourage gambling on sporting events! 2 C. Cotta, A. M. Mora, and J. J. Merelo, "A Network Analysis of the 2010 FIFA World Cup Champion Team Play," Journal of Systems Science 26:1 (2013), 21-42. 3 For instance, please see Gregory T. Papanikos, "Economic, population and political determinants of the 2014 World Cup match results," Soccer & Society 18:4 (2017), 516-532; and Fabian Wunderlich and Daniel Memmert, "Analysis of the predictive qualities of betting odds and FIFA World Ranking: evidence from the 2006, 2010 and 2014 Football World Cups," Journal of Sports Sciences 34:24 (2016), 20176-20184. 4 Please see O'Donoghue et al, "An evaluation of quantitative and qualitative methods of predicting the 2002 FIFA World Cup. Part II: Game activity and analysis," Journal of Sports Sciences 22:6 (2004), 513-514 5 Please see, The Economist, "Why video games are so expensive to develop," September 25, 2014, available at www.economist.com. 6 The OP model makes two critical assumptions: (1) it assumes that the (χ'ι, β, γ) function has the form of a continuous probability distribution function which is the standard normal distribution function of a linear combination of our explanatory variables; (2) it assumes proportional odds between each category in the dependent variable. Assumption (2) was confirmed using the proportional odds Brant test, confirming that the ordered probit is the best suited model for our purpose. Please see Alexander, Carol, Market Risk Analysis: Practical Financial Econometrics (John Wiley & Sons) 2008, 426 pages. 7 Please see EViews 8.1 User's Guide II, pp.259-284. 8 Please see J. Bloomfield, R. Polman, and P. O'Donoghue, "Physical Demands of Different Positions in FA Premier League Soccer," Journal of Sports Science & Medicine 6:1 (2007), 63-70. 9 Please see Seife Dendir, "When do soccer players peak? A note," Journal of Sports Analytics 2 (2016), 89-105. 10 Think 22-year old Mario Götze's stunning game-winner four years ago. 11 Please see H. Sarmento et al, "Match analysis in football: a systematic review," Journal of Sports Sciences 32:20 (2014), 1831-1843. 12 In order to favor our core model (M1), which uses the team average player rating variable, we assigned a weight α = 0.66. 13 We are sure that this is a pure coincidence. Being paired with the host Russia had nothing to do with this. Nothing. 14 A "Novichok bonus" perhaps? 15 Remember that the marginal effect represents the impact of a 1-unit change in the rating variable on the probability of winning. This is why we need to be careful when comparing both stages' marginal effect. A 1-unit change in the rating variable is less frequent, but extremely important in the knockout stage, hence the higher coefficient. 16 In 2006, Italy had 10 players from either Juventus or AC Milan; in 2010, Spain had 12 players from either F.C. Barcelona or Real Madrid; and in 2014, Germany had 11 players from either Bayern Munich or Borussia Dortmund. 17 Please see M. Clemente et al, "Midfielder as the prominent participant in the building attack: A network analysis of national teams in FIFA World Cup 2014," International Journal of Performance Analysis in Sport 15:2 (2015), 704-722. 18 For instance, please see F. M. Clemente et al, "Using Network Metrics in Soccer: A Macro-Analysis," Journal of Human Kinetics 45 (2015), 123-134. 19 In order to favor our core model (M1), which uses the team average player rating variable, we assigned a weight α = 0.66. 20 This would not be the first time that the third place game steals the spotlight. Going back to 1978, the third place game has outscored the final by a considerable margin. Teams usually enter the game with no pressure and therefore commit to a flowing, attacking style of play. Some of the most exciting games in World Cup history were played for third place: think Germany's 3-2 win over Uruguay in 2010, or Turkey's thrilling 3-2 win against South Korea. 21 As well as ... yes, atrocious refereeing. 22 Please see BCA Geopolitical Strategy Special Report, "The Middle East: Separating The Signal From The Noise," dated November 15, 2017, available from gps.bcaresearch.com. 23 Please see BCA Frontier Market Strategy Special Report, "Egypt: Giving Benefit Of Doubt," dated February 6, 2018, available at fms.bcaresearch.com. 24 Please see BCA Emerging Markets Strategy and Geopolitical Strategy Special Report, "Ranking EM Countries Based On Structural Variables," dated August 2, 2017, available at ems.bcaresearch.com. 25 The Netherlands has a population 50 times greater than Iceland and a GDP 35 times greater. 26 Portugal has a population 30 times greater than Iceland and a GDP 10 times greater. 27 England has a population 200 times greater than Iceland and a GDP 110 times greater. 28 Just in case any of our Icelandic clients take offense to our premise, we want to point out that the country has competed in a number of Games of the Small States of Europe. Participants at this biannual event includes such sporting powerhouses as Andorra, Cyprus, Liechtenstein, Luxembourg, Malta, Monaco, Montenegro, and San Marino. Iceland actually trails Cyprus in the total tally of medals collected since 1985. 29 Please see Frode Telseth and Vidar Halldorsson, "The success of Nordic football: the cases of the national men's teams of Norway in the 1990s and Iceland in the 2010s," Sport In Society, November 1, 2017. 30 Please see Taleb, Nassim, Fooled By Randomness (New York: Random House), 316 pages. 31 Unless you are a U.S. soccer fan. The U.S. has a population 1,000 times greater than Iceland and a GDP 800 times greater. Please light yourself on fire responsibly. 32 Please see BCA Geopolitical Strategy Special Report, "The Apex Of Globalization - All Downhill From Here," dated November 14, 2014, available at gps.bcaresearch.com. 33 Robert Prosinecki would have been 25-years old, and at the time one of the highest paid players in the world. In 1998, due to a number of injuries, he was playing for Croatia and largely an afterthought in the team. Davor Šuker and Zvonimir Boban would have also been 25 in 1994 and both at the peak of their careers. 34 Vladimir Jugovic and Sinisa Mihajlovic, were both in their prime in 1994, and both were stars in Italy. 35 The eventual Real Madrid legend Predrag Mijatovic would have been in-prime 25 years of age in 1994 (and unlikely to make the starting 11 of the combined Yugoslav team!), while, by then, the three-time Champions League winner Dejan Savicevic would have already been A.C. Milan's best player. 36 Shout out to Darko Pancev, the 1991 European Golden Boot award winner! 37 For example, for the duration of the 2014 Winter Olympics, only vehicles registered in Sochi, or those with special permission, were allowed through security checkpoints set a 100 kilometers outside the city. 38 Please see BCA Global Investment Strategy Weekly Report, "The End Of Europe's Welfare State," dated June 26, 2015, available at gis.bcaresearch.com. 39 Except in the U.S., which is maybe why it continues to underperform in global competitions. 40 Contrary to most leagues, the Premier League does not have a cap on the number of foreign players a team can have. It just requires eight players to be home grown (three years with an English team before the age of 21). 41 Please see BCA Geopolitical Strategy Special Report, "The French Revolution," dated February 3, 2017, available at gps.bcaresearch.com. 42 The turnout of the referendum was a woeful 43%, however. 43 Please see BCA Geopolitical Strategy Weekly Report, "Why So Serious?" dated October 11, 2017, available at gps.bcaresearch.com. 44 Ehrmann, M. and Jansen D., "The Pitch Rather Than The Pit: Investor Inattention During FIFA World Cup Matches," European Central Bank, Working Paper 1424, February 2012. 45 Please see Edmans, A., Garcia, D., & Norli, O. "Sports Sentiment and Stock Returns," The Journal of Finance 62:4 (2007), 1967-1998. 46 G. Kaplanski, and H. Levy, "Exploitable Predictable Irrationality: the FIFA World Cup effect on the U.S. Stock Market," Journal of Financial and Quantitative Analysis 45:2 (2010), 535-553. References Alexander, Carol, Market Risk Analysis: Practical Financial Econometrics (John Wiley & Sons) 2008, 426 pages. J. K. Ashton, B. Gerrard, and R. Hudson (2011), "Do national soccer results really impact on the stock market?" Applied Economics, 43:26, 3709-3717. M. H. Berument, and N. B. Ceylan (2012), "Effects of soccer on stock markets: the return-volatility relationship," The Social Science Journal 49, 368-374. J. Bloomfield, R. Polman, and P. O'Donoghue (2007), "Physical Demands of Different Positions in FA Premier League Soccer," Journal of Sports Science & Medicine, 6(1): 63-70. F. M. Clemente, M. S. Couceiro, F. Martins, and R. Mendes (2015), "Using Network Metrics in Soccer: A Macro-Analysis," Journal of Human Kinetics 45/2015, 123-134. M. Clemente et al. (2015), "Midfielder as the prominent participant in the building attack: A network analysis of national teams in FIFA World Cup 2014," International Journal of Performance Analysis in Sport 15(2): 704-722. C. Cotta, A. M. Mora, and J. J. Merelo (2013) "A Network Analysis of the 2010 FIFA World Cup Champion Team Play," Journal of Systems Science 26(1): 21-42. Seife Dendir (2016), "When do soccer players peak? A note," Journal of Sports Analytics, 2: 89-105. A. Edmans, D. Garcia, and O. Norli (2007), "Sports Sentiment and Stock Returns," The Journal of Finance, 62 (4), 1967-1998. J. Geyer-Klingeberg, M. Hang, M. Walter, and A. Rathgeber (2018), "Do stock markets react to soccer games? A meta-regression analysis," Applied Economics, 50:19, 2171-2189. G. Kaplanski, and H. Levy (2010), "Exploitable Predictable Irrationality: the FIFA World Cup effect on the U.S. Stock Market," Journal of Financial and Quantitative Analysis, 45(2): 535-553. Jorge Knijnik (2014), "Playing for freedom: Sócrates, futebol-arte and democratic struggle in Brazil," Soccer & Society, 15:5, 635-654. Christian Koller (2017), "Le football suisse: Des pionniers aux professionnels, " Soccer & Society, 18:4, 597-598 Lindsay Sarah Krasnoff (2017), "Devolution of Les Bleus as a symbol of a multicultural French future," Soccer & Society, 18:2-3, 311-319 S. Kuper, and S. Szymanski (2009), "Soccernomics: Why England Loses, Why Germany and Brazil Win, and Why the U.S., Japan, Australia, Turkey - and Even Iraq - Are Destined To Become the Kings of the World's Most Popular Sport". H. Liu, M.A. Gomez, C. Lago-Peñas, and J. Sampaio (2015), "Match statistics related to winning in the group stage of 2014 Brazil FIFA World Cup," Journal of Sports Sciences, 33:12, 1205-1213. R. Mackenzie, and C. Cushion (2013), "Performance Analysis in Football: A Critical Review and Implications for Future Research", Journal of Sports Sciences 31(6). J.A. Mangan, Hyun-Duck Kim, A. Cruz, and Gi-Heun Kang (2013), "Rivalries: China, Japan and South Korea - Memory, Modernity, Politics, Geopolitics - and Sport," The International Journal of the History of Sport, 30:10, 1130-1152. M. Martiniello, and G. W Boucher (2017), "The colours of Belgium: red devils and the representation of diversity," Visual Studies, 32:3, 224-235. Brank Milanovic (2010), "The World at Play: Soccer Takes on Globalization," YaleGlobal Online. Richard Mills (2009), "'It All Ended in an Unsporting Way': Serbian Football and the Disintegration of Yugoslavia, 1989-2006," The International Journal of the History of Sport, 26:9, 1187-1217. Shakya Mitra (2014), "Spanish football: from underachievers to world beaters," Soccer & Society, 15:5, 709-719. O'Donoghue et al. (2004), "An evaluation of quantitative and qualitative methods of predicting the 2002 FIFA World Cup. Part II: Game activity and analysis," Journal of Sports Sciences 22(6):513-514. Gregory T. Papanikos (2017), "Economic, population and political determinants of the 2014 World Cup match results," Soccer & Society, 18:4, 516-532. Guy Podoler (2008), "Nation, State and Football: The Korean Case," The International Journal of the History of Sport, 25:1, 1-17. Alejandro Quiroga (2017), "Spanish football and social change: sociological investigations," Soccer & Society, 18:1, 160-162. H. Sarmento et al. (2014), "Match analysis in football: a systematic review," Journal of Sports Sciences, 32:20, 1831-1843. B. Scholtens, and W. Peenstra (2009), "Scoring on the stock exchange? The effect of football matches on stock market returns: an event study," Applied Economics, 41:25, 3231-3237. Stefan Szymanski (2010), "The Economic Impact Of The World Cup," Football Economics and Policy 226-235. Taleb, Nassim, Fooled By Randomness (New York: Random House), 316 pages. F. Telseth, and V. Halldorsson (2017), "The success culture of Nordic football: the cases of the national men's teams of Norway in the 1990s and Iceland in the 2010s," Sport in Society. Dag Tuastad (2014), "From football riot to revolution. The political role of football in the Arab world," Soccer & Society, 15:3, 376-388. Scott Waalkes (2017), "Does soccer explain the world or does the world explain soccer?" Soccer and globalization, Soccer & Society, 18:2-3, 166-180. D. Wong, and S. Chadwick (2017), "Risk and (in)security of FIFA football World Cups - outlook for Russia 2018," Sport in Society, 20:5-6, 583-598. F. Wunderlich, and D. Memmert (2016), "Analysis of the predictive qualities of betting odds and FIFA World Ranking: evidence from the 2006, 2010 and 2014 Football World Cups," Journal of Sports Sciences, 34:24, 20176-20184. D. Zec, and M. Paunovic (2015), "Football's positive influence on integration in diverse societies: the case study of Yugoslavia," Soccer & Society, 16:2-3, 232-244.

Feature Chart I-1Recent Defaults Have Focused Attention ##br##On Corporate Health Recent Defaults Have Focused Attention On Corporate Health Recent Defaults Have Focused Attention On Corporate Health The recent spike in defaults on bonds and loans in China, including missed debt repayments by local government financing vehicles (LGFV) and some listed companies, has unsettled investors over the past few weeks.1 The yield spread between 5-year government bonds and 5-year corporate bonds AA minus in China's domestic bond market, has recently hit their widest level in nearly two years (Chart I-1). As a result, some investors are concerned about the possibility of widespread defaults as the Chinese government's deleveraging campaign continues to roll out, and sweeping new rules on shadow banking take effect. Given the report focus on corporate health, this week we are updating our China Industry Watch thematic chartpack to present a visual presentation of the changing situation in China's corporate sector, and its relevance to the broader stock market performance. Overall, the Chinese corporate sector has continued to deleverage and its financial situation has improved modestly. Our Corporate Health Monitor (CHM),2 which is an equally weighted average of net income margin, return on capital, EBIT-to-debt ratio, debt-to-asset ratio and interest coverage ratio, shows that the health of most sectors are improving. Specifically, for steel, construction materials, automobile, food& beverage and tech, our CHMs are in healthy territory. For oil & gas, coal, non-ferrous metals and machinery, CHMs are still below zero but are recovering. In terms of profit growth, it has remained robust for most of the sectors shown in the report. In particular, profit growth has accelerated substantially in the coal and steel sectors, as higher selling prices helped offset the impact of production constraints on revenue and aggressive cost cutting increased gross margins. Firms in the energy sector have also enjoyed higher profit growth as oil prices rebounded. In terms of the leverage picture, the liabilities-to-assets ratio has continued to decline broadly across sectors (Chart I-2). However, in regards of debt sustainability, the interest-to-sales ratio has increased substantially in coal, steel, and non-ferrous sectors, due to dramatic decline in sales resulting from production constraints. The interest coverage ratio in these sector is less problematic because of improving gross margins. For the tech sector, however, there has been a spike in the interest-to-sales ratio and a sharp decline in interest coverage. Looking beyond the fairly broad-based improvement in our overall non-financial CHM, we doubt that a broad-based default wave will occur in response to the crackdown on shadow banking. First, by our estimation, the recent defaults cited above account for only 0.09% of outstanding corporate bonds. Second, the latest PBOC monetary report changed the tone from emphasizing "deleveraging" to "stabilizing leverage and restructuring", which shows that regulators are as concerned about the stability of the economy as they are about reducing excessive debts. One problem that is worth monitoring is the negative trend in overall industrial enterprises sales, which had a negative growth rate in Q1 relative to the same quarter last year. Part of this negative growth rate is likely due to base effects, given that Q1 2017 itself was abnormally strong. Nevertheless, comparing first three month of the sales this year to that of previous years, it is clear that 2018's value did not reflect an uptrend in the data (Chart I-3). This weak top line performance is somewhat worrisome and we will continue to watch for signs of a further slowdown. Chart I-2A Continued Decline In Debt-To-Assets A Continued Decline In Debt-To-Assets A Continued Decline In Debt-To-Assets Chart I-3Tepid Topline Growth Is Worrisome Tepid Topline Growth Is Worrisome Tepid Topline Growth Is Worrisome Lin Xiang, Research Analyst linx@bcaresearch.com Jonathan LaBerge, CFA, Vice President Special Reports jonathanl@bcaresearch.com BCA China Industry Watch includes four categories of financial ratios to monitor a sector's leverage, profitability, growth and efficiency, respectively. Some of these ratios, as shown in Table 1, are slightly tweaked from conventional definitions due to data availability. The financial data in our exercise are from the official statistics on overall industrial firms, of which the listed companies are a subset, but most financial ratios based on the two sets of data are very similar, especially for the heavy industries that dominate the Chinese stock markets - both onshore and offshore. The financial ratios on leverage, growth and profitability are almost identical for some sectors, while some other sectors that are not well represented in the stock market, such as technology, healthcare and consumer sectors, show notable divergences. As the Chinese equity universe continues to expand, we expect that the two sets of data will increasingly converge. Table 1The China Industry Watch Messages From BCA's China Industry Watch Messages From BCA's China Industry Watch 1 More than 10 companies, several of them listed, from a variety of industries have defaulted on 17 bonds worth more than 16.5 billion yuan (US$2.6 billion), according to figures from Choice. 2 Please see China Investment Strategy Special Report, “Introducing The BCA China Industry Watch,” dated February 10, 2016, available at cis.bcaresearch.com. Appendix: China Industry Watch All Firms Chart II-1Non-Financial Firms: Stock Price & Valuation Indicators Non-Financial Firms: Stock Price & Valuation Indicators Non-Financial Firms: Stock Price & Valuation Indicators Chart II-2Non-Financial Firms: Relative Performance Of Valuation Indicators Non-Financial Firms: Relative Performance Of Valuation Indicators Non-Financial Firms: Relative Performance Of Valuation Indicators Chart II-3Non-Financial Firms: Leverage Indicators Non-Financial Firms: Leverage Indicators Non-Financial Firms: Leverage Indicators Chart II-4Non-Financial Firms: Growth Indicators Non-Financial Firms: Growth Indicators Non-Financial Firms: Growth Indicators Chart II-5Non-Financial Firms: Profitability Indicators Non-Financial Firms: Profitability Indicators Non-Financial Firms: Profitability Indicators Chart II-6Non-Financial Firms: Efficiency Indicators Non-Financial Firms: Efficiency Indicators Non-Financial Firms: Efficiency Indicators Oil & Gas Sector Chart II-7Oil&Gas Sector: Stock Price & Valuation Indicators Oil&Gas Sector: Stock Price & Valuation Indicators Oil&Gas Sector: Stock Price & Valuation Indicators Chart II-8Oil&Gas Sector: Relative Performance Of Valuation Indicators Oil&Gas Sector: Relative Performance Of Valuation Indicators Oil&Gas Sector: Relative Performance Of Valuation Indicators Chart II-9Oil&Gas Sector: Leverage Indicators Oil&Gas Sector: Leverage Indicators Oil&Gas Sector: Leverage Indicators Chart II-10Oil&Gas Sector: Growth Indicators Oil&Gas Sector: Growth Indicators Oil&Gas Sector: Growth Indicators Chart II-11Oil&Gas Sector: Profitability Indicators Oil&Gas Sector: Profitability Indicators Oil&Gas Sector: Profitability Indicators Chart II-12Oil&Gas Sector: Efficiency Indicators Oil&Gas Sector: Efficiency Indicators Oil&Gas Sector: Efficiency Indicators Coal Sector Chart II-13Coal Sector: Stock Price & Valuation Indicators Coal Sector: Stock Price & Valuation Indicators Coal Sector: Stock Price & Valuation Indicators Chart II-14Coal Sector: Relative Performance Of Valuation Indicators Coal Sector: Relative Performance Of Valuation Indicators Coal Sector: Relative Performance Of Valuation Indicators Chart II-15Coal Sector: Leverage Indicators Coal Sector: Leverage Indicators Coal Sector: Leverage Indicators Chart II-16Coal Sector: Growth Indicators Coal Sector: Growth Indicators Coal Sector: Growth Indicators Chart II-17Coal Sector: Profitability Indicators Coal Sector: Profitability Indicators Coal Sector: Profitability Indicators Chart II-18Coal Sector: Efficiency Indicators Coal Sector: Efficiency Indicators Coal Sector: Efficiency Indicators Steel Sector Chart II-19Steel Sector: Stock Price & Valuation Indicators Steel Sector: Stock Price & Valuation Indicators Steel Sector: Stock Price & Valuation Indicators Chart II-20Steel Sector: Relative Performance Of Valuation Indicators Steel Sector: Relative Performance Of Valuation Indicators Steel Sector: Relative Performance Of Valuation Indicators Chart II-21Steel Sector: Leverage Indicators Steel Sector: Leverage Indicators Steel Sector: Leverage Indicators Chart II-22Steel Sector: Growth Indicators Steel Sector: Growth Indicators Steel Sector: Growth Indicators Chart II-23Steel Sector: Profitability Indicators Steel Sector: Profitability Indicators Steel Sector: Profitability Indicators Chart II-24Steel Sector: Efficiency Indicators Steel Sector: Efficiency Indicators Steel Sector: Efficiency Indicators Non Ferrous Metals Sector Chart II-25Non Ferrous Metals Sector: Stock Price & Valuation Indicators Non Ferrous Metals Sector: Stock Price & Valuation Indicators Non Ferrous Metals Sector: Stock Price & Valuation Indicators Chart II-26Non Ferrous Metals Sector: Relative Performance Of Valuation Indicators Non Ferrous Metals Sector: Relative Performance Of Valuation Indicators Non Ferrous Metals Sector: Relative Performance Of Valuation Indicators Chart II-27Non Ferrous Metals Sector: Leverage Indicators Non Ferrous Metals Sector: Leverage Indicators Non Ferrous Metals Sector: Leverage Indicators Chart II-28Non Ferrous Metals Sector: Growth Indicators Non Ferrous Metals Sector: Growth Indicators Non Ferrous Metals Sector: Growth Indicators Chart II-29Non Ferrous Metals Sector: Profitability Indicators Non Ferrous Metals Sector: Profitability Indicators Non Ferrous Metals Sector: Profitability Indicators Chart II-30Non Ferrous Metals Sector: Efficiency Indicators Non Ferrous Metals Sector: Efficiency Indicators Non Ferrous Metals Sector: Efficiency Indicators Construction Material Sector Chart II-31Construction Material Sector: Stock Price & Valuation Indicators Construction Material Sector: Stock Price & Valuation Indicators Construction Material Sector: Stock Price & Valuation Indicators Chart II-32Construction Material Sector: Relative Performance Of Valuation Indicators Construction Material Sector: Relative Performance Of Valuation Indicators Construction Material Sector: Relative Performance Of Valuation Indicators Chart II-33Construction Material Sector: Leverage Indicators Construction Material Sector: Leverage Indicators Construction Material Sector: Leverage Indicators Chart II-34Construction Material Sector: Growth Indicators Construction Material Sector: Growth Indicators Construction Material Sector: Growth Indicators Chart II-35Construction Material Sector: Profitability Indicators Construction Material Sector: Profitability Indicators Construction Material Sector: Profitability Indicators Chart II-36Construction Material Sector: Efficiency Indicators Construction Material Sector: Efficiency Indicators Construction Material Sector: Efficiency Indicators Machinery Sector Chart III-37Machinery Sector: Stock Price & Valuation Indicators Machinery Sector: Stock Price & Valuation Indicators Machinery Sector: Stock Price & Valuation Indicators Chart III-38Machinery Sector: Relative Performance Of Valuation Indicators Machinery Sector: Relative Performance Of Valuation Indicators Machinery Sector: Relative Performance Of Valuation Indicators Chart III-39Machinery Sector: Leverage Indicators Machinery Sector: Leverage Indicators Machinery Sector: Leverage Indicators Chart III-40Machinery Sector: Growth Indicators Machinery Sector: Growth Indicators Machinery Sector: Growth Indicators Chart III-41Machinery Sector: Profitability Indicators Machinery Sector: Profitability Indicators Machinery Sector: Profitability Indicators Chart III-42Machinery Sector: Efficiency Indicators Machinery Sector: Efficiency Indicators Machinery Sector: Efficiency Indicators Automobile Sector Chart III-43Automobile Sector: Stock Price & Valuation Indicators Automobile Sector: Stock Price & Valuation Indicators Automobile Sector: Stock Price & Valuation Indicators Chart III-44Automobile Sector: Relative Performance Of Valuation Indicators Automobile Sector: Relative Performance Of Valuation Indicators Automobile Sector: Relative Performance Of Valuation Indicators Chart III-45Automobile Sector: Leverage Indicators Automobile Sector: Leverage Indicators Automobile Sector: Leverage Indicators Chart III-46Automobile Sector: Growth Indicators Automobile Sector: Growth Indicators Automobile Sector: Growth Indicators Chart III-47Automobile Sector: Profitability Indicators Automobile Sector: Profitability Indicators Automobile Sector: Profitability Indicators Chart III-48Automobile Sector: Efficiency Indicators Automobile Sector: Efficiency Indicators Automobile Sector: Efficiency Indicators Food & Beverage Sector Chart III-49Food&Beverage Sector: Stock Price & Valuation Indicators Food&Beverage Sector: Stock Price & Valuation Indicators Food&Beverage Sector: Stock Price & Valuation Indicators Chart III-50Food&Beverage Sector: Relative Performance Of Valuation Indicators Food&Beverage Sector: Relative Performance Of Valuation Indicators Food&Beverage Sector: Relative Performance Of Valuation Indicators Chart III-51Food&Beverage Sector: Leverage Indicators Food&Beverage Sector: Leverage Indicators Food&Beverage Sector: Leverage Indicators Chart III-52Food&Beverage Sector: Growth Indicators Food&Beverage Sector: Growth Indicators Food&Beverage Sector: Growth Indicators Chart III-53Food&Beverage Sector: Profitability Indicators Food&Beverage Sector: Profitability Indicators Food&Beverage Sector: Profitability Indicators Chart III-54Food&Beverage Sector: Efficiency Indicators Food&Beverage Sector: Efficiency Indicators Food&Beverage Sector: Efficiency Indicators Information Technology Sector Chart III-55Information Technology Sector: Stock Price & Valuation Indicators Information Technology Sector: Stock Price & Valuation Indicators Information Technology Sector: Stock Price & Valuation Indicators Chart III-56Information Technology Sector: Relative Performance Of Valuation Indicators Information Technology Sector: Relative Performance Of Valuation Indicators Information Technology Sector: Relative Performance Of Valuation Indicators Chart III-57Information Technology Sector: Leverage Indicators Information Technology Sector: Leverage Indicators Information Technology Sector: Leverage Indicators Chart III-58Information Technology Sector: Growth Indicators Information Technology Sector: Growth Indicators Information Technology Sector: Growth Indicators Chart III-59Information Technology Sector: Profitability Indicators Information Technology Sector: Profitability Indicators Information Technology Sector: Profitability Indicators Chart III-60Information Technology Sector: Efficiency Indicators Information Technology Sector: Efficiency Indicators Information Technology Sector: Efficiency Indicators Utilities Sector Chart III-61Utilities Sector: Stock Price & Valuation Indicators Utilities Sector: Stock Price & Valuation Indicators Utilities Sector: Stock Price & Valuation Indicators Chart III-62Utilities Sector: Relative Performance Of Valuation Indicators Utilities Sector: Relative Performance Of Valuation Indicators Utilities Sector: Relative Performance Of Valuation Indicators Chart III-63Utilities Sector: Leverage Indicators Utilities Sector: Leverage Indicators Utilities Sector: Leverage Indicators Chart III-64Utilities Sector: Growth Indicators Utilities Sector: Growth Indicators Utilities Sector: Growth Indicators Chart III-65Utilities Sector: Profitability Indicators Utilities Sector: Profitability Indicators Utilities Sector: Profitability Indicators Chart III-66Utilities Sector: Efficiency Indicators Utilities Sector: Efficiency Indicators Utilities Sector: Efficiency Indicators Cyclical Investment Stance Equity Sector Recommendations

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 Uncovered Interest Rate Parity still works for currencies. However, it needs to be based on a combination of short- and long-term real rates. Currencies are also affected by global risk appetite, as approximated by corporate spreads and commodity prices. For the next six months, the euro has additional downside, while the dollar's rebound could run further. The CAD also looks attractive. Feature In July 2016, in a Special Report titled, "In Search Of A Lost Timing Model," we introduced a set of intermediate-term models to complement our long-term fair value models for various currencies.1 These groups of models provide additional discipline - a sanity check if you will - to our regular analysis. Additionally, these models can help global equity investors manage their currency exposure, having increased the Sharpe ratio of global equity portfolios vis-à-vis other hedging strategies, and also for a host of base-currencies.2 In this report, we review the logic underpinning these intermediate-term models and provide commentary on their most recent readings for the G10 currencies vis-à-vis the USD. UIP, Revisited The Uncovered Interest Rate Parity (UIP) relationship is at the core of this modeling exercise. This theory suggests that an equilibrium exchange rate is the one that will make an investor indifferent between holding the bonds of Country A or Country B. This means that as interest rates rise in Country A relative to Country B, the currency of Country B will fall today in order to appreciate in the future. These higher expected returns are what will drive investors to hold the lower-yielding bonds of Country B. Chart 1Interest Rate Parity: ##br##Generally Helpful, But... Interest Rate Parity: Generally Helpful, But... Interest Rate Parity: Generally Helpful, But... There has long been debate as to whether investors should focus on short rates or long rates when looking at exchange rates through the prism of UIP. This debate has regained vigor in the past six months as the dollar has greatly lagged the levels implied by 2-year rate differentials (Chart 1). Research by the Federal Reserve and the IMF suggests incorporating longer-term rates to UIP models increase their accuracy.3 This informational advantage works whether policy rates are or aren't close to their lower bound.4 Incorporating long-term rates as an explanatory variable increases the performance of UIP models because exchange rate movements do not only reflect current interest rate conditions, but currency market investors also try to anticipate the path of interest rates over many periods. By definition, long-term bonds do just that, as they are based on the expected path of short rates over their maturity - as well as a term premium, which compensates for the uncertain nature of future interest rates. There is another reason why long-term rate differential changes improve the power of UIP models. Since UIP models are based on the concept of indifference of investors between assets in two countries, changes in the spreads between 10-year bonds in these two countries will create more volatility in the currency pair than changes in the spreads between 3-month rates. This is because an equivalent delta in the 10-year spread will have much greater impact on the relative prices of the bonds than on the short-term paper, courtesy of their much more elevated duration. To compensate for these greater changes in prices, the currency does have to overshoot its long-term PPP to a much greater extent to entice investors trading the long end of the curve. Bottom Line: The interest rate parity relationship still constitutes the bedrock of any shorter-term currency fair value model. However, to increase its accuracy, both long-term and short-term rates should be used. Real Rates Really Count Another perennial question regarding exchange rate determination is whether to use nominal or real rate differentials. At a theoretical level, real rates are what matter. Investors can look through the loss of purchasing power created by inflation. Therefore, exchange rates overshoot around real rate differentials, not nominal ones. On a practical level, there are additional reasons to believe that real rates should matter, especially when trying to explain currency moves beyond a few weeks. Indeed, various surveys and studies on models used by forecasters and traders show that FX professionals use purchasing power parity as well as productivity differential concepts when setting their forex forecasts.5 Indeed, as Chart 2 illustrates, real rate differentials have withstood the test of time as an explanatory variable for exchange rate dynamics, albeit with periods where rate differentials and the currency can deviate from one another. It is true that very often, nominal rate differentials can be used as a shorthand for real rate differentials, as both interest rate gaps tend to move together. However, regularly enough, they do not. In countries with very depressed inflation expectations (Japan immediately comes to mind), nominal and real rate differentials can in fact look very different (Chart 3). With the informational cost of incorporating market-based inflation expectations being very low, we find the shorthand unnecessary when building UIP-based models. Chart 2Real Rates Work Better Over The Long Run Real Rates Work Better Over The Long Run Real Rates Work Better Over The Long Run Chart 3Real And Nominal Rate Spreads Can Differ Real And Nominal Rate Spreads Can Differ Real And Nominal Rate Spreads Can Differ Finally, it is important to remark that in environments of high inflation, inflation differentials dominate any other factor when it comes to exchange rate determination. However, the currencies discussed in this report currently are not like Zimbabwe or Latin America in the early 1980s. Bottom Line: When considering an intermediate-term fair value model for exchange rates, investors should focus on real, not nominal, long-term rate differentials. Global Risk Aversion And Commodity Prices Chart 4The Dollar Benefits From Global Stresses The Dollar Benefits From Global Stresses The Dollar Benefits From Global Stresses Global risk appetite is also a key factor in trying to model exchange rates. Risk-aversion shocks tend to lead to an appreciation in the U.S. dollar, which benefits from its status as the global reserve currency.6 Literature has often focused on the use of the VIX as a gauge for global risk appetite. Our exercise shows stronger explanatory power with options-adjusted spreads on junk bonds (Chart 4). Commodity prices, too, play a key role. Historically, commodity prices have displayed a very strong negative correlation with the dollar.7 This correlation is obviously at its strongest for commodity-producing nations, as rising natural resource prices constitute a terms-of-trade shock for them. However, this relationship holds up for the euro as well, something already documented by the European Central Bank.8 The Models The models for each cross rate are built to reflect the insight gleaned above. Each cross is modeled on three variables, with the model computed on a weekly timeframe. Real rates differentials: We use the average of 2-year and 10-year real rates. The rates are deflated using inflation expectations. Global risk appetite, approximated by junk OAS. Commodity prices: We use the Bloomberg Continuous Commodity Index. For all countries, the variables are statistically highly significant and of the expected signs. These models help us understand in which direction the fundamentals are pushing the currency. We refer to these as Fundamental Intermediate-Term Models (FITM). We created a second set of models, based on the variables above, which also include a 52-week moving average for each cross. The real rates differentials, junk spreads and commodity prices remain statistically very significant and of the correct sign. They are therefore trend- and risk-appetite adjusted UIP-deviation models. These models are more useful as timing indicators on a three- to nine-month basis, as their error terms revert to zero much faster. We refer to these as Intermediate-Term Timing Models (ITTM). Mathieu Savary, Vice President Foreign Exchange Strategy mathieu@bcaresearch.com The U.S. Dollar Chart 5Dollar Back In Line With Fundamentals Dollar Back In Line With Fundamentals Dollar Back In Line With Fundamentals Chart 6More Upside For Now More Upside For Now More Upside For Now To model the dollar index (DXY), we used two approaches. In the first one, we took all the deviation from fair value for the pairs constituting the index, based on their weights in the DXY. In the second approach, we ran the model specifically for the DXY, using the three variables described above. U.S. real rates were compared to an average of euro area, Japanese, Canadian, British, Swiss and Swedish real rates, weighted by their contribution to the DXY. We then averaged both approaches, which gave us very similar results to begin with. After a short period when it traded below its FITM, the dollar's recent strength has pushed the greenback back to its equilibrium, suggesting the easy gains are behind us. However, the rising risks in EM along with the continued widening in rate differentials between the U.S. and the rest of the world could put upward pressure on the dollar for a few more months (Chart 5). When the trend in the dollar is included, the greenback also trades in line with the ITTM (Chart 6). This confirms the idea that the dollar could experience some more upside for the remainder of 2018, as periods of undervaluation to the ITTM tend to be followed by overshoots. The return of inflation, along with the injection of large amounts of fiscal stimulus in the U.S., could be the narratives that push the greenback up by another 5%. Despite a positive outlook for 2018, we remain concerned about the dollar's longer-term performance. Not only is it still trading at a 16% premium on a PPP basis, European rates have room to increase substantially once euro area economic slack is fully absorbed. We are not there yet, but continued robust growth in the euro area will let the ECB increase rates more aggressively than the Fed beyond 2020. The Euro Chart 7The Euro Is Not A Bargain Anymore... The Euro Is Not A Bargain Anymore... The Euro Is Not A Bargain Anymore... Chart 8...And Has More Downside Before Year End ...And Has More Downside Before Year End ...And Has More Downside Before Year End The FITM for EUR/USD continues to point south, dragged down by widening interest rate differentials in favor of the greenback. However, unlike in early 2017, the euro is no longer trading at a big discount to its fair value (Chart 7). As a result, unlike last year, the euro is not able to avoid the downward gravitational pull of a falling FITM. More worrisome for the euro's performance over the coming six months, EUR/USD is still trading at a premium to its ITTM, which adjusts our FITM by taking account of the euro's trend (Chart 8). Currently, the fair value for EUR/USD stands at 1.15, but the euro tends to undershoot its equilibrium after large overshoots such as when EUR/USD traded around 1.25. Moreover, if China's economic slowdown deepens, commodity prices will suffer, which will drag down both the FITM and the ITTM for the euro. We are not yet willing buyers of the euro at current levels. While we espouse a bearish short-term view on the euro, we will be looking to purchase it once it moves to the 1.15-1.10 range. On longer-term metrics, EUR/USD still trades at a significant discount to its fair value. Moreover, long-term rates could rise in Europe relative to the U.S. once investors begin to lift their expectations for future euro area policy rates more aggressively. As such, we continue to closely monitor the slowdown in both euro area and global growth. Once we see signs of stabilization, the euro should again catch a durable bid. The Yen Chart 9A Dovish BoJ Is Pushing Down ##br##The Yen's Fundamentals A Dovish BoJ Is Pushing Down The Yen's Fundamentals A Dovish BoJ Is Pushing Down The Yen's Fundamentals Chart 10Tactically, The Yen Is At Risk, But Softening Global ##br##Growth Will Limit Its Downside This Year Tactically, The Yen Is At Risk, But Softening Global Growth Will Limit Its Downside This Year Tactically, The Yen Is At Risk, But Softening Global Growth Will Limit Its Downside This Year The FITM for the yen is falling fast, and as a result the JPY cannot rally anymore against the dollar (Chart 9). The ITTM provides a very similar message: the yen still trades at a premium to its short-term equilibrium, and is vulnerable to the dollar's strength (Chart 10). Softness in the yen has materialized despite growing stresses in emerging markets and budding signs of a slowdown in global growth - two normally yen-bullish developments - making it clear that the breakdown between USD/JPY and interest rate differentials could not withstand a period of generalized strength in the dollar. While the yen could weaken against the dollar, it is likely to rally further against the euro. Weakness in global growth is likely to limit the yen's downside to the equilibrium implied by its ITTM. Meanwhile, EUR/USD is likely to undershoot this same equilibrium. This contrast points to further weakness in EUR/JPY. The British Pound Chart 11The Pound Is ##br## At Equilibrium The Pound Is At Equilibrium The Pound Is At Equilibrium Chart 12GBP/USD May Be Dragged Lower By A Falling ##br## EUR/USD, But Cable Is Less At Risk Than The Euro GBP/USD May Be Dragged Lower By A Falling EUR/USD, But Cable Is Less At Risk Than The Euro GBP/USD May Be Dragged Lower By A Falling EUR/USD, But Cable Is Less At Risk Than The Euro GBP/USD is in a very different position than EUR/USD. While the pound's FITM points south, driven by interest rate differentials, cable trades below its equilibrium level (Chart 11). For the FITM to move up from this point onward, the U.K. economy needs to stabilize. We do think this will happen as British inflation slows, which will support household real incomes, and thus consumer spending. This message is also confirmed by the fact that unlike EUR/USD, GBP/USD does not trade at a premium to the ITTM, which incorporates the trend in the pair (Chart 12). While investors bid up the pound against the dollar as the greenback weakened in 2017 and early 2018, they are still embedding a risk premium in the GBP, a consequence of the murky political outlook that has shrouded the U.K. ever since the Brexit referendum. The models confirm our analysis of two weeks ago: that the pound could experience some downside against the dollar if the euro were to weaken, but that nonetheless cable will suffer less than EUR/USD.9 As a result, EUR/GBP is likely to experience downside as the correction in EUR/USD unfolds. On a longer-term basis, traditional valuation metrics such as PPP suggest that the GBP remains cheap. However, for this judgment to be true, much will depend on the evolution of the negotiations between the U.K. and the rest of the EU. A British exit from the common market will invalidate the message from PPP models, as the economic relationship between the U.K. and its largest trading partner will change drastically, implying that the models are specified over a sample that is not relevant anymore. However, it remains far from clear what form Brexit will ultimately take. The Canadian Dollar Chart 13NAFTA Risk Premia Evident Here... NAFTA Risk Premia Evident Here... NAFTA Risk Premia Evident Here... Chart 14...And Here ...And Here ...And Here Not only is the loonie trading well below the levels implied by the FITM, but augmented interest rate differential models for the CAD still point north, suggesting its fundamental drivers are currently very supportive (Chart 13). The ITTM for the Canadian dollar confirms this message; even after adjusting for its trend the CAD still trades at a discount to equilibrium (Chart 14). Both formulations of the models highlight that a risk premium has been embedded into the Canadian dollar, reflecting still-possible hazards and setbacks surrounding NAFTA negotiations. However, BCA expects a benign outcome for Canada in the coming weeks, which should help the loonie down the road. Not only does the absence of a major overhaul to NAFTA imply that trade flows between the U.S. and Canada will avoid a major shock, it also means that the Bank of Canada can resume tightening monetary policy. The biggest risk for the Canadian dollar versus the greenback is global growth. So long as global growth has not stabilized, the CAD will find it hard to rally durably against the USD. As a result, we prefer to buy the CAD versus other currencies, the EUR and AUD in particular. The Swiss Franc Chart 15No Evident Deviation From ##br## Fundamentals In The Franc No Evident Deviation From Fundamentals In The Franc No Evident Deviation From Fundamentals In The Franc Chart 16Rising EM Stresses And Better Value Will ##br##Help The Swiss Franc Versus The Euro Rising EM Stresses And Better Value Will Help The Swiss Franc Versus The Euro Rising EM Stresses And Better Value Will Help The Swiss Franc Versus The Euro The FITM for the Swissie continues to move upward (Chart 15). In fact, the franc currently trades at a discount to its ITTM. This suggests that downside for the Swiss franc versus the dollar is limited for the remainder of the year (Chart 16). Since the Swiss franc already trades at a discount to the USD, but the euro does not, logically, the EUR/CHF is currently very pricey. Hence, it will be difficult for the euro to rally further against the franc this year. Moreover, the slowdown in global growth and the trouble facing EM assets and currencies are likely to further contribute to the current deceleration in European economic data. As a result, both short-term valuation metrics and economic considerations argue for selling EUR/CHF on a six-month basis. Longer term, the Swiss franc's strength in recent years has contributed to a sharp deterioration in Swiss competitiveness. Since the Swiss economy is very flexible, this has mostly been translated into strong deflationary pressures in the alpine state. As a result, the Swiss National Bank will continue to fight off any appreciation in the franc, maintaining very easy monetary conditions. Thus, long-term investors should not short EUR/CHF, but instead, they should use any weakness in this cross this year to accumulate larger bets on the long side. The Australian Dollar Chart 17AUD Fundamentals At Risk AUD Fundamentals At Risk AUD Fundamentals At Risk Chart 18AUD Not Cheap Enough To Flash A Buy Signal AUD Not Cheap Enough To Flash A Buy Signal AUD Not Cheap Enough To Flash A Buy Signal The FITM for the Aussie is currently in a holding pattern (Chart 17). Meanwhile, AUD/USD trades at a marginal discount to the trend-augmented version of the model, the ITTM (Chart 18). Do not get lulled into a sense of comfort by these observations. First, AUD/USD never stops a move at the ITTM; it tends to overshoot its equilibrium. In fact, undershoots tends to culminate at an 8% discount to the short-term fair value. Additionally, the global economic environment suggests that both the AUD's FITM and ITTM could experience downside in the coming months. Slowing global activity and budding EM stress weigh on commodity prices - key components of the models. They also weigh on Australian interest rate differentials vis-à-vis the U.S. - especially as the Australian economy is replete with slack - keeping wage pressures, inflationary pressures, and consequently the Reserves Bank of Australia at bay. This picture is in sharp contrast to Canada. Canadian labor market conditions are tight and the BoC is likely to resume its hiking campaign once uncertainty around NAFTA dissipates. Since the CAD trades at a much larger discount to both its FITM and ITTM, the relative economic juncture supports being short AUD/CAD. The New Zealand Dollar Chart 19NZD Weaker Than ##br##Fundamentals Imply NZD Weaker Than Fundamentals Imply NZD Weaker Than Fundamentals Imply Chart 20NZD Is Cheap Enough To Warrant ##br## A Buy Versus The AUD NZD Is Cheap Enough To Warrant A Buy Versus The AUD NZD Is Cheap Enough To Warrant A Buy Versus The AUD As was the case with the Aussie, the FITM for the kiwi has stabilized (Chart 19). However, unlike with the AUD, the NZD trades at a meaningful discount to the ITTM (Chart 20). The NZD has greatly suffered in response to a deceleration in New Zealand economic data and to investors' worries about the Adern government - a coalition of the left-leaning Labour and populist New Zealand First parties. Investors are especially concerned over limitation to immigration on long-term growth, as well as risks to the Reserve Bank of New Zealand's independence. These concerns are real, and warrant taking a cautious stance on the NZD. New Zealand growth has greatly benefited from decades of a large immigration influx and from a staunchly independent central bank. Moreover, slowing global growth and trade as well as rising EM stresses are also likely to exert downward pressure on the NZD's short-term fair-value estimates. We have been taking advantage of the NZD's discount to its FITM and ITTM by selling the Aussie/kiwi cross. AUD/NZD trades at a premium to its relative ITTM. Moreover, the deceleration in global growth and the stress in EM are likely to exact a greater toll on metals than agricultural prices. This represents a greater negative terms-of-trade shock for Australia than New Zealand. Since Australia displays greater labor market slack than New Zealand, this disinflationary shock will bit the larger of the two economies harder. Therefore, interest rate differentials should move against the AUD, pushing the relative ITTM and FITM down. The Norwegian Krone Chart 21NOK Still A Value Play Among ##br## Commodity Currencies... NOK Still A Value Play Among Commodity Currencies... NOK Still A Value Play Among Commodity Currencies... Chart 22...But It Could Experience Further Downside ##br##Against The Dollar This Year ...But It Could Experience Further Downside Against The Dollar This Year ...But It Could Experience Further Downside Against The Dollar This Year The fundamental model for the Norwegian krone remains in an uptrend, established since the beginning of 2016 (Chart 21). This reflects rallying oil prices, the key determinant of Norwegian terms-of-trade and growth. However, the NOK still trades slightly above its ITTM, its fundamentals adjusted for the trend in the currency pair (Chart 22). Over the next six months, the Norwegian krone could experience further downside versus the USD. Corrections in this pair tends to end when it trades 4% below its ITTM. Additionally, the rise in EM volatility and the great sensitivity of the Norwegian krone to USD fluctuations adds an economic impetus to this risk. Moreover, EUR/USD normally exerts a gravitational pull on the NOK/USD. Since we expects the euro to weaken further, this should drag the krone along for a ride. However, we continue to see downside in EUR/NOK as short-term valuations are not attractive, and as oil is likely to outperform the broad commodity complex. In the longer term, we are positive on the NOK. It is cheap based on long-term models that take into account Norway's stunning net international position of 220% of GDP. Moreover, the high inflation registered between 2015 and 2016 is now over as the pass-through from the weak trade-weighted krone between 2014 and 2015 is gone. This means that the NOK's PPP fair value has stopped deteriorating. The Swedish Krona Chart 23The SEK Has Been Clobbered ##br##Beyond Fundamentals... The SEK Has Been Clobbered Beyond Fundamentals... The SEK Has Been Clobbered Beyond Fundamentals... Chart 24...And Is Becoming Attractive,##br## But Beware The Riskbank ...And Is Becoming Attractive, But Beware The Riskbank ...And Is Becoming Attractive, But Beware The Riskbank The Swedish krona's short-term valuations are attractive. As was the case with the krona, the SEK's FITM remains in an uptrend (Chart 23), and the SEK trades at a sizeable discount to its ITTM (Chart 24). Despite this benign picture, we are reluctant to bet on the SEK. To begin with, the SEK displays the greatest sensitivity to the dollar of all the G-10 currencies; our dollar-bullish stance for the rest of the year thus bodes poorly for the krona, pointing to greater undervaluation ahead. Additionally, despite an economy running 2% above potential GDP, the Riksbank still runs an extremely accommodative monetary policy. In fact, recent communications by the Swedish central bank demonstrate a high degree of comfort with the SEK's weakness. It seems as though Riksbank Governor Stefan Ingves wants to competitively devalue the krona. With global growth softening, the Riksbank is likely to encourage further SEK depreciation as the Swedish business cycle is tightly linked to EM growth. We were long NOK/SEK until two weeks ago, when our target level was hit. While we look to re-open this position, the NOK/SEK currently trades at a small premium to its relative ITTM, and thus the corrective episode could run a few more months. Meanwhile, the relative short-term valuation picture suggests that the recent bout of weakness in EUR/SEK could run a bit further. However, weakening global growth and the Riksbank's dovish proclivities suggest that visibility on this cross remains exceptionally low. 1 Please see Foreign Exchange Strategy / Global Investment Strategy Special Report titled, "Assessing Fair Value In FX Markets", dated February 26, 2018, available at fes.bcaresearch.com and gis.bcaresearch.com. 2 Please see Foreign Exchange Strategy / Global Asset Allocation Special Reports titled, "Currency Hedging: Dynamic Or Static? - A Practical Guide For Global Equity Investors", dated September 29, 2017, and "Currency Hedging: Dynamic Or Static? - A Practical Guide For Global Equity Investors (Part II)", dated October 13, 2017, available at fes.bcaresearch.com and gaa.bcaresearch.com. 3 Ravi Balakrishnan, Stefan Laseen, and Andrea Pescatori, "U.S. Dollar Dynamics: How Important Are Policy Divergence And FX Risk Premiums?" IMF Working Paper No.16/125 (July 2016); and Michael T. Kiley, "Exchange Rates, Monetary Policy Statements, And Uncovered Interest Parity: Before And After The Zero Lower Bound", Finance and Economics Discussion Series 2013-17, Board of Governors of the Federal Reserve System (January 2013). 4 Michael T. Kiley (January 2013). 5 Please see Yin-Wong Cheung and Menzie David Chinn, "Currency Traders and Exchange Rate Dynamics: A Survey of the U.S. Market", CESifo Working Paper Series No. 251 (February 2000); and David Hauner, Jaewoo Lee, and Hajime Takizawa, "In which exchange rate models do forecasters trust?" IMF Working Paper No.11/116 (May 2010) for revealed preference approach based on published forecasts from Consensus Economics. 6 Ravi Balakrishnan, Stefan Laseen, and Andrea Pescatori (July 2016) 7 Ravi Balakrishnan, Stefan Laseen, and Andrea Pescatori (July 2016) 8 Francisco Maeso-Fernandez, Chiara Osbat, and Bernd Schnatz, "Determinants Of The Euro Real Effective Exchange Rate: A BEER/PEER Approach", Working Paper No.85, European Central Bank (November 2001). 9 Please see Foreign Exchange Strategy Special Report titled, "A Long, Strange Cycle", dated May 4, 2018, available at fes.bcaresearch.com. Trades & Forecasts Forecast Summary Core Portfolio Tactical Trades Closed Trades

Highlights An examination of the three pillars of China's economy provides an unambiguous signal that a slowdown is underway. This would normally warrant, at most, a neutral allocation to Chinese stocks, but several factors argue against cutting exposure for now. Stay overweight, but with a short leash. Recent changes in the BCA China Investable Sector Alpha Portfolio's recommended allocation have validated two of our recent investment recommendations. In addition, the model is providing a curiously bullish signal about the relative performance of Chinese vs global stocks that heightens our reluctance to reduce Chinese equity exposure. Our China Reform Monitor signals that investors do not view the current pace of structural reforms as being overly burdensome for the economy. In addition, while Chinese policymakers have made some significant gains in improving China's air quality over the past 18 months, these changes have mostly occurred from a near-hazardous starting point (suggesting that more progress will be needed). As such, we recommend that investors stick with our long ESG leaders / short investable benchmark trade over the coming year. Feature Global investor sentiment improved modestly on Monday, in response to statements from President Trump indicating a possible détente between the U.S. and China on the issue of trade. In particular, Mr. Trump signaled a willingness to assist ZTE, a Chinese telecommunications equipment maker, whose operations would have been enormously impacted by the U.S. Commerce Department's decision last month to ban American companies from selling to the firm. In the view of our Geopolitical Strategy Service, announcements like these should be viewed as marginally positive developments within the context of a serious downtrend in U.S./China relations. Investors appear to be eager to respond to positive news about waning U.S. protectionism, but the reality is that several important decisions related to the U.S.' section 301 probe have yet to be announced.1 As we noted in last week's Special Report,2 this underscores that the near-term risks to China from the external sector are clearly to the downside. Abstracting from the day-to-day assessment of the trade picture, we have emphasized that other core elements of the China outlook have deteriorated. As we present below, an aggregate view of the three pillars of China's economy continues to argue for a (contained) slowdown, with protectionism acting as a downside risk to an already sober economic outlook. Extremely cheap valuation and the high-beta nature of Chinese ex-tech stocks continue to justify an overweight stance versus global equities, but we recommend that investors keep Chinese stocks on downgrade watch for the remainder of Q2 as the risks to the Chinese economy warrant an ongoing assessment of what is currently a finely balanced equity allocation decision. Assessing The Three Pillars Chart 1 presents our stylized framework for analyzing China's economy. It highlights that China's business cycle is largely driven by three "pillars": industrial activity, the housing market, and trade. While the services sector, the Chinese consumer, and/or the technology sector are of interesting secular relevance, generally-speaking China's business cycle continues to be subject to its "old" growth model centered on investment and exports. Chart 1The Three Pillars Of China's Business Cycle The Three Pillars Of China's Economy The Three Pillars Of China's Economy Industrial Activity: We took an empirical approach to predicting China's industrial sector activity in our November 30 Special Report,3 and tested the ability of 40 different macro data series to lead the Li Keqiang index (LKI). While the LKI is closely followed and somewhat cliché, we have focused on it because of its strong correlation with ex-tech earnings and import growth. The results of our November report pointed to the success of monetary condition indexes, money supply, and credit measures to reliably predict the LKI since China's real GDP growth peaked in 2010. We constructed our BCA Li Keqiang Leading Indicator based on these measures, and we have frequently highlighted over the past few months that the indicator is pointing to a continued deceleration in China's industrial activity (Chart 2). Housing: We noted in our November report that housing market data also correlates with the LKI, albeit less well than the components of our Leading Indicator. One important observation about China's housing market that we highlighted in our February 8 Weekly Report is that residential floor space sold appears to have reliably led floor space started (a proxy for real residential investment) since 2010 (Chart 3). Over the past 6-8 months, however, floor space started appears to have diverged from the trend in floor space sold, which may have been caused by a non-trivial reduction in housing inventories over the past few years.4 Nonetheless, we also noted that the level of inventories remains quite elevated, suggesting that the uptrend in floor space started is unlikely to continue without a renewed uptrend in sales volume. In our view, this conclusion implies that the housing outlook over the coming 6-12 months is neutral, at best. Chart 2China's Industrial Sector ##br##Will Continue To Slow China's Industrial Sector Will Continue To Slow China's Industrial Sector Will Continue To Slow Chart 3Resi Sales Volume Does Not Point To ##br##A Sustained Pickup In Construction Resi Sales Volume Does Not Point To A Sustained Pickup In Construction Resi Sales Volume Does Not Point To A Sustained Pickup In Construction Trade: The third pillar of China's economy is the external sector, which remains important even though net exports have fallen quite significantly in terms of contribution to China's growth. We noted in our April 18 Weekly Report that there is a strongly positive relationship between the annual change in contribution to growth from China's net exports and subsequent gross capital formation, highlighting that external demand provides an important multiplier effect for Chinese activity. For now, nominal export growth (in CNY terms) remains at the high end of its 5-year range, reflecting the strength of the global economy. But three significant risks remain to the export outlook: 1) the clear and present danger of U.S. import tariffs, 2) the possibility that Chinese policymakers may accelerate their reform efforts to take advantage of the "window of opportunity" provided by robust global demand,4 and 3) the very substantial rise in the export-weighted RMB (Chart 4), which is fast approaching its 2015 high. As a final point on trade, Chart 5 highlights that the recent divergence between the LKI and nominal import growth is resolved when examining the latter in CNY terms. The chart suggests that while export growth has been buoyed by a strong global economy, China's contribution to the global growth impulse is diminishing. The very tight link demonstrated in Chart 5 also suggests that industrial activity is the most important pillar to watch among the three noted above, which means that Chart 2 argues for a negative export outlook for China's major trading partners. Chart 4A Non-Trivial Deterioration ##br##In Competitiveness A Non-Trivial Deterioration In Competitiveness A Non-Trivial Deterioration In Competitiveness Chart 5The Rise In CNYUSD Is Flattering ##br##Imports Measured In Dollars The Rise In CNYUSD Is Flattering Imports Measured In Dollars The Rise In CNYUSD Is Flattering Imports Measured In Dollars Our assessment of the three pillars of China's economy points to a conclusion that we have highlighted frequently in our recent reports: China's industrial sector is slowing, and there are downside risks to the export outlook. The character of the slowdown does not suggest that a major shock to the global economy is likely to emanate from China over the coming 6-12 months, but the outlook is more consistent with a reduction than an expansion in China's contribution to global growth. Under normal circumstances, at best this would warrant a neutral asset allocation outlook to China-related financial assets. Chart 6The Uptrend In Relative Chinese ##br##Ex-Tech Performance Is Intact The Uptrend In Relative Chinese Ex-Tech Performance Is Intact The Uptrend In Relative Chinese Ex-Tech Performance Is Intact However, we have also argued that the relatively attractive valuation and the technical profile of Chinese equities suggests that investors should have a high threshold for reducing their exposure to China within a global equity portfolio. Chart 6 highlights that Chinese ex-tech share prices continue to demonstrate resilient performance versus their global peers, despite the ongoing slowdown in China's economy. In addition, as we will note below, our BCA China Investable Sector Alpha Portfolio is providing a curiously bullish signal about the relative performance of Chinese stocks, which heightens our reluctance to cut exposure. Bottom Line: An examination of the three pillars of China's economy provides an unambiguous signal that a slowdown is underway. This would normally warrant, at most, a neutral allocation to Chinese stocks, but several factors argue against cutting exposure for now. Stay overweight, but with a short leash. Reading The Tea Leaves From Our Sector Alpha Portfolio We introduced our BCA China Investable Sector Alpha Portfolio in a January Special Report, in part to demonstrate that the concept of alpha persistence (i.e. alpha that is persistently positive or negative) has material implications for portfolio returns. In particular, we noted that the portfolio's strategy of allocating to China's investable equity sectors based on the significance of alpha has resulted in over 200bps of long-term outperformance versus the investable benchmark, without taking on any additional risk (Table 1). Table 1An Alpha-Based Sector Model Has Historically Outperformed China's Investable Stock Market The Three Pillars Of China's Economy The Three Pillars Of China's Economy Table 2 presents the portfolio's current allocation, relative to the current benchmark weights for each sector as well as the portfolio's sectoral allocation when we published our January report. Two observations are noteworthy: The model recommends an overweight allocation to resources; consumer staples; health care; utilities; and real estate, at the expense of industrials; consumer discretionary; financials; technology; and telecom services. These positions are largely in-line with the model's recommendations in January, except for a non-trivial increase in exposure to energy and financials, and a significant reduction in technology and consumer discretionary. The portfolio's reduced exposure to technology and consumer discretionary stocks validate two recent investment recommendations from BCA's China Investment Strategy team: we recommended a long consumer staples / short consumer discretionary trade on November 16,5 and we recommend that investors retain cyclical exposure to investable Chinese stocks while neutralizing exposure to the tech sector on February 15.6 Table 2Our Sector Alpha Portfolio Has Validated Two Of Our Recent Recommendations The Three Pillars Of China's Economy The Three Pillars Of China's Economy Chart 7 highlights another interesting insight from the model, by presenting the beta of the portfolio relative to the investable benchmark alongside the benchmark's performance versus global stocks. First, the chart underscores the limited systemic risk of the portfolio, as the portfolio's beta rarely deviates materially from 1. But more importantly, it appears that the portfolio's beta versus the investable benchmark is somewhat correlated with (and leads) China's performance versus global stocks: Chart 7A Curiously Bullish Signal From ##br##Our Sector Alpha Portfolio A Curiously Bullish Signal From Our Sector Alpha Portfolio A Curiously Bullish Signal From Our Sector Alpha Portfolio Prior to the global financial crisis, the portfolio's beta was above 1 and rising, until early-2007 (preceding the peak in relative performance by about a year). Following the crisis, the portfolio beta steadily declined until late-2014/early-2015, interrupted only by a brief rise back above 1 from 2009-2010. Chinese stock prices steadily underperformed global equities during this period. The portfolio beta rose back to 1 in mid-2015, and stayed flat until early last year. Chinese stocks technically underperformed global stocks during this period, but by a much more modest amount than what occurred on average from 2009 to 2014. In this case, the rise in the portfolio beta in 2015 appeared to correctly signal that a sharply underweight stance towards Chinese stocks was no longer warranted. Finally, the portfolio beta surged rapidly higher last year, in line with a material rise in the relative performance of Chinese stocks. It has fallen modestly since January, but remains at one of the highest levels seen over the past 15 years. Drawing pro-cyclical inferences from the beta characteristics of risk-adjusted performers is a novel approach for BCA's China Investment Strategy service, and for now we regard the results of Chart 7 as a curious signal that warrants further examination. Still, this bullish sign is consistent with the general resilience of Chinese stocks that we have observed over the past several months, which continues to argue in favor of a high threshold to cut exposure to China within a global equity portfolio. Bottom Line: Recent changes in the BCA China Investable Sector Alpha Portfolio's recommended allocation have validated two of our recent investment recommendations. In addition, the model is providing a curiously bullish signal about the relative performance of Chinese vs global stocks that heightens our reluctance to reduce Chinese equity exposure. An Update On The "Reform Trade" We noted in the aftermath of last November's Communist Party Congress that China was likely to step up its reform efforts in 2018, and make meaningful efforts to: Pare back heavy-polluting industry Hasten the transition of China's economy to "consumer-led" growth7 Halt leveraging in the corporate/financial sector Eliminate corruption and graft As a result of this outlook, we highlighted that the pace of renewed structural reforms would be a key theme to watch this year, in order to ensure that the pursuit of these policies would not unintentionally cause a repeat of the significant slowdown in the economy that occurred in 2014/2015. We presented our framework for monitoring this risk in our November 16 Weekly Report, which was to track an index that we called the BCA China Reform Monitor. The monitor is calculated as an equally-weighted average of four "winner" sectors that outperformed the investable benchmark in the month following the Party Congress relative to an equally-weighted average of the remaining seven sectors. We argued that significant underperformance of "loser" sectors could be a sign that reform intensity has become too burdensome for the economy (and thus a material headwind ex-tech equity performance), and highlighted that we would be watching for signs that our monitor was rising largely due to outright declines in the denominator. Using this framework, Chart 8 suggests that structural reform efforts are ongoing but that investors do not view the current pace of these reforms as overly burdensome for the economy. In particular, panel 2 highlights that recent movements in our Reform Monitor have been driven by fairly steady outperformance of the "winner" sectors, with "loser" sectors simply trending sideways. While it is possible that Chinese policymakers will intensify their efforts to reform the economy over the coming 6-12 months,4 for now our China Reform Monitor continues to support an overweight stance towards Chinese ex-tech stocks vs their global peers. However, given the message of our Reform Monitor, it is somewhat surprising that another of our reform-themed trades has fared so poorly over the past three months. Chart 9 presents the performance of our long investable environmental, social and governance (ESG) leaders / short investable benchmark trade, which was up approximately 4% since inception in late-January but is now down 1.4%. The basis of this trade was to overweight stocks that are best positioned to deliver "sustainable" growth, which we argued would fare well in a reform environment. Does the underperformance of this trade suggest that the reform theme is unlikely to be investment-relevant over the coming year? Chart 8Structural Reforms Not Viewed As ##br##Economically Restrictive By Investors Structural Reforms Not Viewed As Economically Restrictive By Investors Structural Reforms Not Viewed As Economically Restrictive By Investors Chart 9ESG Leaders Should Fare Quite ##br##Well In A Reform Environment ESG Leaders Should Fare Quite Well In A Reform Environment ESG Leaders Should Fare Quite Well In A Reform Environment In our view, the answer is no. First, while the MSCI ESG leaders index maintains roughly similar sector weights as the investable benchmark (which limits the beta risk of the trade), Table 3 highlights that differences do exist. These modest differences in sector allocation do appear to be impacting performance (Chart 10), in particular the underweight allocation to energy stocks (which are outperforming) and the overweight allocation to technology (which has sold off since mid-March). Table 3Sector Allocation Has Impacted The Recent Performance Of China's ESG Leaders The Three Pillars Of China's Economy The Three Pillars Of China's Economy Chart 10Sector Allocation Impacting Recent ##br##Performance Of ESG Leaders Sector Allocation Impacting Recent Performance Of ESG Leaders Sector Allocation Impacting Recent Performance Of ESG Leaders Second, while China made significant gains last year in improving air quality in several major population centers (such as Beijing and Shanghai), these improvements have mostly occurred from a near-hazardous starting point and have simply rendered China's air to be less unhealthy. Even in Beijing, Chart 11 highlights that PM2.5 readings have started to increase again, from a level that only briefly reached "good" quality. In addition, Chart 12 highlights that some of the improvement in air quality last year occurred, at least in part, because China shifted polluting activity from one province to another. This implies that Chinese policymakers will continue to wrestle with improving the country's air quality for some time to come, which in our view continues to favor ESG leaders over the coming year and beyond. Chart 11Some Significant Recent Gains In Air ##br##Quality, But Part Of An Ongoing Battle Some Significant Recent Gains In Air Quality, But Part Of An Ongoing Battle Some Significant Recent Gains In Air Quality, But Part Of An Ongoing Battle Chart 12Air Quality Gains In Some Provinces, At The Expense Of Others The Three Pillars Of China's Economy The Three Pillars Of China's Economy Bottom Line: Our China Reform Monitor signals that investors do not view the current pace of structural reforms as being overly burdensome for the economy. In addition, while Chinese policymakers have made some significant gains in improving China's air quality over the past 18 months, these changes have mostly occurred from a near-hazardous starting point (suggesting that more progress will be needed). As such, we recommend that investors stick with our long ESG leaders / short investable benchmark trade over the coming year. Jonathan LaBerge, CFA, Vice President Special Reports jonathanl@bcaresearch.com 1 Please see Geopolitical Strategy Weekly Report "Inside The Beltway," dated May 2, 2018, available on gps.bcaresearch.com 2 Please see Geopolitical Strategy and China Investment Strategy Special Report "China's "Red Line" In The Trade Talks," dated May 9, 2018, available on cis.bcaresearch.com 3 Please see China Investment Strategy Special Report "The Data Lab: Testing The Predictability Of China's Business Cycle," dated November 30, 2017, available on cis.bcaresearch.com 4 Please see China Investment Strategy Weekly Report "China: A Low-Conviction Overweight," dated May 2, 2018, available on cis.bcaresearch.com 5 Please see China Investment Strategy Weekly Report "Messages From The Market, Post-Party Congress," dated November 16, 2017, available on cis.bcaresearch.com 6 Please see China Investment Strategy Weekly Report "After The Selloff: A View From China," dated February 15, 2018, available on cis.bcaresearch.com 7 Investors should note that BCA's China Investment Strategy service has long been skeptical of calls to shift China's economy to a consumption-driven growth model, because it significantly raises the odds that the country will not be able to escape the middle income trap. For example, please see Please see China Investment Strategy Special Report, "On A Higher Note", dated October 5, 2017, available at cis.bcaresearch.com Cyclical Investment Stance Equity Sector Recommendations

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 At just under 3-in-10 odds, the probability Brent crude oil prices will exceed $80/bbl by year-end is now more than double what it was at the beginning of the year, following President Trump's announcement he would withdraw the U.S. from the 2015 Joint Comprehensive Plan of Action (JCPOA), and re-impose all economic sanctions against Iran (Chart of the Week). Chart of the WeekProbability Brent Exceeds $90/bbl Is Understated By Markets Feedback Loop: Spec Positioning & Oil Price Volatility Feedback Loop: Spec Positioning & Oil Price Volatility We believe these odds are too low. Indeed, we think the odds of Brent prices ending above $90/bbl this year are higher than the 1-in-8 chance being priced in the markets presently, even though this is up from just under 4% at the beginning of the year. We also expect sharper down moves going forward, as news flows become noisier. Speculators have loaded the boat on the long side, and they will be exquisitely sensitive to any unexpected softening in fundamentals - e.g., a supply increase or the whiff of lower demand - given their positioning (Chart 2). Chart 2Specs Have Loaded the Boat##BR##Getting Long Brent and WTI Exposure Specs Have Loaded the Boat Getting Long Brent and WTI Exposure Specs Have Loaded the Boat Getting Long Brent and WTI Exposure Our research indicates that spec positioning in the underlying futures can, under some circumstances, dominate the evolution of oil options' implied volatility, the markets' key gauge of risk and the essential component of option pricing. As new risk factors arising from Trump's decision emerge, we expect option implied volatility to increase, as the frequency of spec re-positioning increases. Energy: Overweight. We are getting long Feb/19 $80/bbl Brent calls expiring in Dec/18 vs. short Feb/19 $85/bbl calls, given our assessment that the odds of ending the year above $90/bbl are higher than the market's expectation. We also recommend getting long Aug/19 $75 Brent calls vs. short Aug/19 $80/bbl calls. We already are long Dec/18 $65/bbl Brent calls vs. short $70/bbl calls expiring at the end of Oct/18, which are up 74.2% since they were recommended in Feb/18. Rising vol favors long options positions. The new positions will put on at tonight's close. Base Metals: Neutral. Refined copper imports in China grew 47% y/y in March. For the first four months of 2018 they are up 15% y/y. Imports of copper ores and concentrates were up 9.7% y/y in the January - April period. Precious Metals: Neutral. We remain strategically long gold and tactically long spot silver. A stronger USD continues to weigh on both. Ags/Softs: Underweight. The USDA's weekly Crop Progress report indicates farmers in the U.S. are catching up in their spring planting, converging toward averages for this time of year. Nevertheless, the condition of winter wheat remains a concern. Feature The wild swings in crude oil prices following President Trump's decision not to waive nuclear-related sanctions against Iran - down ~ 2% after Trump's announcement Tuesday, then up more than 2.5% the following morning - resolved one of the more important "known unknowns" ahead of schedule - to wit, would the U.S. re-impose nuclear-related sanctions against Iran, or continue to waive them.1 Ahead of Trump's announcement this week, speculators clearly were building long positions in Brent and WTI, as seen in Chart 2. Among other things, stout fundamentals, which we have been highlighting, and a possible tightening of supply on the back of the re-imposition of U.S. sanctions were obvious catalysts for building the bullish positions. We find specs do not Granger-cause oil prices, and typically these traders are reacting to fundamental news.2 This is consistent with other research into this topic.3 In other words, we find specs essentially follow the fundamentals, they don't lead them, and, as a result, the level of oil prices largely is explained by supply, demand and inventories. Based on the Granger-causality tests and our fundamental modeling, we believe oil markets are, to a very large extent, efficient in the sense that prices reflect most publicly available information.4 This is not to say, however, that the role of speculation can be dismissed as trivial to price formation. Spec Positioning Matters For Implied Volatility In Oil Our most recent research, building on earlier work on speculation in oil markets, finds that the concentration of speculators on the long side or the short side of the market actually does play a significant role in how volatility evolves (Chart 3, bottom panel).5 Other factors are important to the evolution of volatility, as well - i.e., U.S. financial conditions, particularly the stress in the system as measured by the St. Louis Fed's Financial Stress Index; EM equity volatility; and y/y percent changes in WTI oil prices themselves (Chart 3). But spec positioning clearly dominates: In periods of rising or elevated volatility, it explains most of the change in WTI option implied volatilities (Chart 4). This can push volatility higher when it occurs. However, on the downside, this does not hold - Working's T Index is not material to the evolution of implied volatility when uncertainty about future oil prices is low or decreasing. Chart 3Key Variables##BR##Explaining Volatility Key Variables Explaining Volatility Key Variables Explaining Volatility Chart 4Spec Positioning Dominates##BR##Evolution of WTI Implied Volatility Spec Positioning Dominates Evolution of WTI Implied Volatility Spec Positioning Dominates Evolution of WTI Implied Volatility Working's T Index and implied volatility are independent of price direction - they are directionless, therefore they cannot be used to forecast prices.6 These variables tend to increase when the quality of information available to the market deteriorates - i.e., when it becomes more difficult to form expectations about future oil prices. This is, we believe, an attractive time for informed speculators to enter the market and use their information to make profits. We find two-way Granger-causality between WTI implied volatility and Working's T, when the annual change in excess speculation is one-standard deviation above or below its mean. This means the more specs are concentrated on one side of the market in the underlying futures - long or short - the more influence their positioning has on volatility, and that the higher volatility is the more specs are drawn to the market. Given that specs' beliefs are different, this means there is a rising number of long or short spec contracts relative not only to specs on the other side of the market, but also to long and short hedgers. Why Speculation Is Important Prices do not suddenly manifest themselves in markets fully aligned with fundamentals. They are made efficient by hedgers off-loading risk based on their marginal costs, and speculators uncovering information that is material to the level at which prices clear markets. The goal of speculation is to buy low and sell high. Hedging and speculation are both done in the presence of noise, or pseudo-information that has no real connection with where markets clear.7 Information is to noise as substance is to a void. Noise can look like information, as Black (1986) notes, and people can trade on it, but they will lose money and eventually go out of business. Information, on the other hand, is costly, as Grossman and Stiglitz (1980) point out. To incentivize someone (a speculator) to gather it and feed it into prices via the market clearing - i.e., buying and selling based on information - they have to be able to make a profit. Speculators supply the liquidity necessary for trading - and, most importantly, hedging - to occur. Successful speculators make profits. Therefore, the information on which they trade is more often germane to the market-clearing process than not. To be successful they have to be willing to buy when prices are low, expecting them to go higher, and to sell when prices are high, expecting them to go lower. As Paul Samuelson wryly observed, "Is there any other kind of price than 'speculative' price? Uncertainty pervades real life and future prices are never knowable with precision. An investor is a speculator who has been successful; a speculator is merely an investor who last lost his money."8 Known Unknowns Will Keep Vol Elevated Chart 5BCA's Oil Price Forecast Unchanged,##BR##Following Trump's Iran Announcement BCA's Oil Price Forecast Unchanged, Following Trump's Iran Announcement BCA's Oil Price Forecast Unchanged, Following Trump's Iran Announcement In the wake of Trump's announcement, the fundamental and geopolitical landscape has been re-cast, creating additional "known unknowns", particularly re how the U.S. will implement the renewed sanctions and the timing of these moves. Among the new known unknowns, which can only be resolved with the passage of time, are: The precise timing and extent of the re-imposed sanctions on the part of the U.S., which will evolve over the next 90 to 180 days. Demand-side implications of higher prices, particularly in EM economies where policymakers used the low prices following OPEC's 2014 - 16 market-share war to eliminate fuel subsidies, which prevented high prices from being experienced by their citizens. The supply-side implications of higher prices on U.S. shale production - does production and investment, including pipeline take-away capacity, take another leg higher? The Kingdom of Saudi Arabia's (KSA) ability to raise output, given the Kingdom said it would be raising output in the event Iranian volumes are lost to export markets. The fate of the Saudi Aramco IPO, and how the re-imposition of sanctions by the U.S. on Iran affects the royal family's decision on whether to float 5% of the company publicly. Will production in distressed states in- and outside of OPEC be negatively affected by increasing geopolitical risk?9 Among the "known unknowns," Iran's next moves rank high, as do responses to such moves by the U.S. and its allies. The U.S. and its Gulf allies clearly view Iran as a threat and, with the re-imposition of sanctions against Iran, are confronting it. Iran has a similar view vis-à-vis the U.S. and its Gulf allies. Left to be determined: Does Iran increase its level of direct action against KSA, upping the ante, so to speak, in its ongoing proxy wars with the Kingdom? Is Gulf production threatened? Are U.S. - European relations threatened by Trump's action? Thus far, European leaders have indicated they remain committed to the sanctions deal Trump walked away from. What would it take for OPEC 2.0 to restore actual production cuts we estimate at 1.1 to 1.2mm b/d to the market? What would it take to trigger a release of the U.S. Strategic Petroleum Reserve (SPR), estimated at just under 664-million-barrel, which could be released to the market at a rate of 500k to 1mm b/d? These known unknowns are not causing us to change our price forecast for this year - $74/bbl for Brent and $70/bbl for WTI, based on our fundamental modeling (Chart 5). However, we do think price risk is to the upside in both markets, given the elevated geopolitical tensions in the market. We continue to expect more frequent prices excursions to and through $80/bbl for the balance of the year, particularly for Brent. Robert P. Ryan, Senior Vice President Commodity & Energy Strategy rryan@bcaresearch.com Hugo Bélanger, Senior Analyst Commodity & Energy Strategy HugoB@bcaresearch.com 1 We lay out some of these "known unknowns" in BCA Research's Commodity & Energy Strategy Weekly Report "Tighter Balances Make Oil Price Excursions To $80/bbl Likely," published April 19, 2018. In addition to the Iran issues, which have been resolved, Venezuela looms large. Oil production declined by 900k b/d between December 2015 and March 2018, with half of that occurring in the past six months. We are carrying Venezuela's current production at ~ 1.5mm b/d, although other estimates have it lower. With the country moving closer to collapsing as a functioning state, the risk to its oil output and exports is high. 2 Granger-causality refers to an econometric test developed by Clive Granger, the 2003 Nobel laureate in economics. It determines whether past values of one variable can be said to predict, or cause, the present value of another variable. 3 Please see BCA Research's Commodity & Energy Strategy Weekly Report, "Specs Back Up The Truck For Oil," published April 26, 2018, available at ces.bcaresearch.com. See also the International Energy Agency's "Oil: Medium-Term Market Report 2012;" and "The Role of Speculation in Oil Markets: What Have We Learned So Far?" by Bassam Fattouh, Lutz Kilian and Lavan Mahadeva, published by The Oxford Institute For Energy Studies. Also, see "Speculation, Fundamentals, and The Price of Crude Oil," by Kenneth B. Medlock III, published by the James A. Baker III Institute for Public Policy at Rice University, August 2013. 4 This is the semi-strong form of market efficiency. For a discussion of how markets impound information in prices, please see Eugene Fama's Noble lecture, "Two Pillars of Asset Pricing," which was reprinted in the June 2014 issue of The American Economic Review (p. 1467). 5 Please see BCA Research's Commodity & Energy Strategy Weekly Report, "Specs Back Up The Truck For Oil," published April 26, 2018, in which we introduce Holbrook Working's "T Index," a measure of speculative concentration in futures and options markets. It is available at ces.bcaresearch.com. Briefly, Working's T Index shows how much speculative positioning exceeds the net demand for hedging from commercial participants in the market. Excessive speculation - spec positioning in excess of hedging demand by commercial interests - could be read into index values above 1.0. However, the U.S. CFTC notes values of Working's T at or below 1.15 do not provide sufficient liquidity to support hedging, even though "there is an excess of speculation, technically speaking." Formally, Working's T Index looks like this: Feedback Loop: Spec Positioning & Oil Price Volatility Feedback Loop: Spec Positioning & Oil Price Volatility 6 Please see Irwin, S. H. and D. R. Sanders (2010), "The Impact of Index and Swap Funds on Commodity Futures Markets: Preliminary Results", OECD Food, Agriculture and Fisheries Working Papers, No. 27. 7 Please see Black, Fischer (1986), "Noise," in the Journal of Finance, 41:3; and Grossman, Sanford J., and Stiglitz, Joseph E. (1980), "On the Impossibility of Informationally Efficient Markets," in the June issue of the American Economic Review. 8 Please see Samuelson, Paul A. (1973), "Mathematics Of Speculative Price," in the January 1973 SIAM Review, 15:1. 9 Please see "Geopolitical Certainty: OPEC Production Risks Are Playing To Shale Producers' Advantage," published by BCA's Energy Sector Strategy on May 9, 2018, which discusses these production risks in depth. Investment Views and Themes Recommendations Strategic Recommendations Tactical Trades Commodity Prices and Plays Reference Table Feedback Loop: Spec Positioning & Oil Price Volatility Feedback Loop: Spec Positioning & Oil Price Volatility Trades Closed in 2018 Summary of Trades Closed in 2017 Feedback Loop: Spec Positioning & Oil Price Volatility Feedback Loop: Spec Positioning & Oil Price Volatility

There is scant evidence that the character of the equity market advance is changing and the fact that weak balance sheet stocks are no longer outperforming strong balance sheet stocks is giving us pause (Chart 1). Chart 1Time To Pause And Reflect Time To Pause And Reflect Time To Pause And Reflect Using the Goldman Sachs equity baskets - that utilize the 'Altman Z-score' framework to select stocks - via Bloomberg, we find that the weak balance sheet over strong balance sheet share price ratio leads the broad market at both peaks and is coincident at troughs. The most recent peak occurred in early 2017 and it is rather surprising that a proxy for this ratio using the fixed income market, i.e. the total return high yield bond index versus the total return investment grade bond index, is moving in the opposite direction and not confirming the equity market's message (Chart 2). This begs the question: Which market signal is right, stocks or fixed income, and what are the equity sector investment implications? But before trying to answer these questions, we first zoom out and look at the broad U.S. debt picture. How Will It All End? In our travels and conference calls one common question keeps coming up: What will end all this? The short answer is that rising interest rates will eventually deal a blow to the debt overhang and the expansion will give way to a fresh deleveraging cycle. In other words, a whiff of inflation will entice the Fed to keep on raising the fed funds rate to the point where the business cycle turns down. As demand falters, a decreasing cash flow backdrop will not be able to service the debt overload, as both coupon payments and principal repayments will become a big burden. This will ignite a jump in the default rate, a message the yield curve is already sending (Chart 3). Chart 2Which Market Is Right? Which Market Is Right? Which Market Is Right? Chart 3Has The Junk Default Rate Troughed? Has The Junk Default Rate Troughed? Has The Junk Default Rate Troughed? Peering back to the onset of the GFC, a U.S. financial sector debt crisis engulfed the world. Subsequently, this morphed into a government sector debt problem in the Eurozone and more recently into a non-financial corporate sector debt overhang mostly in the commodity complex and the emerging markets. Debt Supercycle Lives On The investment world is obsessed with China's excess debt uptake and that is a valid concern. However, investors should also be aware that U.S. debt has not been fully purged. Rather, it has moved around between different domestic sectors. The debt supercycle lives on.1 The implication is that an interest rate-induced debt bubble pricking would be deflationary, and thus identifying the U.S. domestic sector most exposed to such risk is important. Chart 4 breaks down U.S. total debt into the four largest sectors using flow of funds data. While households and the financial sector have significantly de-levered, the government and the non-financial business sector have been picking up the slack and aggressively re-levering. While the Trump Administration has embarked on a two-year fiscal policy easing period that will add to the government debt profile, the nonfinancial corporate debt overhang is more vulnerable and thus troublesome in our view (fed funds rate shown inverted, Chart 5). Worrisomely, since the GFC, nonfinancial corporates have been issuing debt and partially using this debt to retire equity and pay handsome dividends. According to the flow of funds data, the cumulative nonfinancial net equity retirement figure stands near $4tn over the past decade (middle panel, Chart 6). Undoubtedly, this has been a large contributor to equity market returns (top panel, Chart 6), and will likely gain further momentum this year on the back of the tax repatriation holiday. Some sell side equity retirement estimates for the S&P 500 hover around $800bn for calendar 2018 or roughly twice the past decade's annual average. AAPL's recent announcement of a $100 billion share repurchase program confirms that the buyback bonanza is persevering and will continue to boost equities. Clearly, such breakneck equity retirement pace is unsustainable and will converge down to a lower trend rate in 2019 and beyond, especially given the drying liquidity as the Fed continues to pursue a tighter monetary policy. Chart 4Debt Is Moving Around Debt Is Moving Around Debt Is Moving Around Chart 5Tight Monetary Policy Pricks Bubbles, And... Tight Monetary Policy Pricks Bubbles, And… Tight Monetary Policy Pricks Bubbles, And… Chart 6...Threatens To End The Equity Retirement Binge …Threatens To End The Equity Retirement Binge …Threatens To End The Equity Retirement Binge Introducing BCA's Sector Insolvency Risk Monitor (IRM) The purpose of this Special Report is to identify debt soft spots and outliers in the U.S. GICS1 equity sectors. What follows is a financials statement-heavy analysis of sector indebtedness. We introduce the 'Altman Z-score' sector analysis that gauges sector credit strength, with a rising score indicating improving health and a declining Z-score signifying deteriorating health.2 In absolute terms, a score below 1.8 warns of a possible credit event, whereas any reading above 3 signals that bankruptcy risk is very low (see appendix below). Our analysis includes our flagship Bank Credit Analyst's Corporate Health Monitor framework that breaks down corporate health in the different sectors3 (see appendix below). We also sift through a number of different stock market reported ratios/data to gauge each sector's health, with net debt-to-EBITDA and interest coverage at the forefront of our analysis, and try to identify outliers (see appendix below). Finally, with the invaluable help of BCA's Chief Quantitative Strategist, David Boucher, we created our new insolvency risk monitor (IRM) per U.S. equity sector incorporating the respective 'Altman Z-scores', BCA's corporate health monitor readings and net debt-to-EBITDA ratios. In more detail, we ranked each sector (ex-financials and real estate) on a monthly basis on each of these three measures. Then we used a simple average of the ranked measures per sector to come up with the final sector ranking. We also selected the median sector ranking per measure and used the average of the three metrics as a proxy for the broad market.4 This way we were able to compare each sector IRM to the overall market. Note that the IRMs are designed so that a higher IRM ranking means better solvency. Charts 7 & 8 summarize the results and showcase this new all-inclusive relative ranking alongside relative share price performance. Chart 7Unsustainable... Unsustainable… Unsustainable… Chart 8...Divergences ...Divergences ...Divergences Sector Outliers Consumer discretionary stocks are the clearest outliers and the message from the relative IRM is to expect a significant underperformance phase in the coming quarters (top panel, Chart 7). AMZN's juggernaut is blurring the discretionary landscape given its 20% index weight, and artificially boosting relative share prices. Ex-AMZN, this early cyclical sector is behaving similar to previous episodes when the Fed embarked on a tightening interest rate cycle. We reiterate our recent downgrade to a below benchmark allocation.5 Consumer staples equities are steeply deviating from their increasing relative IRM score, underscoring that investors are unduly punishing staples stocks (second panel, Chart 8). We maintain our overweight stance and treat this sector as a small portfolio hedge to our otherwise general dislike of defensives (as a reminder we are underweight both the S&P health care and the S&P telecom services sectors). Chart 9Cyclicals Have The Upper Hand Cyclicals Have The Upper Hand Cyclicals Have The Upper Hand The utilities share price ratio is also deviating from the IRM relative reading (fourth panel, Chart 8). The implication is that extreme bearishness toward the sector is overdone and we reiterate our mid-February upgrade to a neutral stance.6 Energy stocks have fallen behind the energy IRM rebound reading (top panel, Chart 8). We expect a catch up phase on the back of the global capex upcycle, still improving debt profile, favorable underlying commodity supply/demand dynamics and firming oil prices. The S&P energy sector remains a high-conviction overweight. The niche materials sector is also trailing the sector's slingshot IRM recovery. Keep in mind that, as expected, the materials IRM is one of the most volatile series (second panel, Chart 8). Materials manufacturers are capital intensive and high operating leverage businesses and despite the debt dynamic betterment since the recent global manufacturing recession, this sector is still saddled with a large amount of debt that makes it extremely sensitive to the ebbs and flows of global economic growth. We continue to recommend a benchmark allocation. The remaining sectors' (tech, health care, telecom services and industrials) relative share prices are moving in tandem with their respective IRM readings (Charts 7 & 8). In addition, we have complied all the cyclical and defensive IRMs in two distinct series and the relative IRM ratio is giving the all-clear sign to continue to prefer cyclicals over defensives on a 9-12 month time horizon (Chart 9). So What? In sum, the IRM is one new additional metric we are using to gauge the validity of our sector positioning and should not be used in isolation. To answer our original question, while the weak balance sheet versus strong balance sheet stock underperformance is alarming and we will continue to closely monitor this stock price ratio, it is premature to change our constructive overall equity market view on a 9-12 month horizon. We therefore continue to recommend a cyclical over defensive portfolio bent. Finally, for completion purposes, the appendix below shows a number of debt-related indicators we track, including the absolute 'Altman Z-score' and corporate health monitor readings, in two charts per sector along with the cyclicals over defensives compilation and the overall market (ex-financials). Anastasios Avgeriou, Vice President U.S. Equity Strategy anastasios@bcaresearch.com 1 For a primer on the debt super cycle please refer to Box 1 in the BCA Special Year End Issue: "Outlook 2013: Fewer Storms, More Sunny Breaks," dated December 19, 2012, available at bca.bcaresearch.com. 2 Altman Z-Score = 1.2A + 1.4B + 3.3C + 0.6D + 1.0E. Where: A = working capital / total assets, B = retained earnings / total assets, C = earnings before interest and tax / total assets, D = market value of equity / total liabilities and E = sales / total assets. Source: https://www.investopedia.com/terms/a/altman.asp 3 Please see BCA The Bank Credit Analyst Report, "U.S. Corporate Health Gets A Failing Grade," dated January 28, 2016, available at bca.bcaresearch.com. 4 We refrained from using the top down computed S&P 500 'Altman Z-Score' and net debt-to-EBITDA as the financials sector really skewed the results and therefore opted to use the median sector 'Altman Z-score' and net debt-to-EBITDA as a proxy for the broad market because using the mean also skewed the results largely because of the tech sector. Staying consistent in our analysis, we also used the median sector BCA corporate health monitor to proxy the broad market. 5 Please see BCA U.S. Equity Strategy Weekly Report, "Reflective Or Restrictive?" dated March 12, 2018, available at uses.bcaresearch.com. 6 Please see BCA U.S. Equity Strategy Weekly Report, "Manic-Depressive?" dated February 12, 2018, available at uses.bcaresearch.com. Appendix U.S. Non-Financial Broad Market I U.S. Non-Financial Broad Market I U.S. Non-Financial Broad Market I U.S. Non-Financial Broad Market II U.S. Non-Financial Broad Market II U.S. Non-Financial Broad Market II U.S. S&P Industrials I U.S. S&P Industrials I U.S. S&P Industrials I U.S. S&P Industrials II U.S. S&P Industrials II U.S. S&P Industrials II U.S. S&P Energy I U.S. S&P Energy I U.S. S&P Energy I U.S. S&P Energy II U.S. S&P Energy II U.S. S&P Energy II U.S. S&P Consumer Staples I U.S. S&P Consumer Staples I U.S. S&P Consumer Staples I U.S. S&P Consumer Staples II U.S. S&P Consumer Staples II U.S. S&P Consumer Staples II U.S. S&P Tech I U.S. S&P Tech I U.S. S&P Tech I U.S. S&P Tech I U.S. S&P Tech I U.S. S&P Tech I U.S. S&P Utilities I U.S. S&P Utilities II U.S. S&P Utilities II U.S. S&P Utilities II U.S. S&P Utilities II U.S. S&P Utilities II U.S. S&P Materials I U.S. S&P Materials I U.S. S&P Materials I U.S. S&P Materials II U.S. S&P Materials II U.S. S&P Materials II U.S. S&P Consumer Discretionary I U.S. S&P Consumer Discretionary I U.S. S&P Consumer Discretionary I U.S. S&P Consumer Discretionary II U.S. S&P Consumer Discretionary II U.S. S&P Consumer Discretionary II U.S. S&P Telecom Services I U.S. S&P Telecom Services I U.S. S&P Telecom Services I U.S. S&P Telecom Services II U.S. S&P Telecom Services II U.S. S&P Telecom Services II U.S. S&P Health Care I U.S. S&P Health Care I U.S. S&P Health Care I U.S. S&P Health Care II U.S. S&P Health Care II U.S. S&P Health Care II U.S. S&P Cyclicals Vs. Defensives I U.S. S&P Cyclicals Vs. Defensives I U.S. S&P Cyclicals Vs. Defensives I U.S. S&P Cyclicals Vs. Defensives II U.S. S&P Cyclicals Vs. Defensives II U.S. S&P Cyclicals Vs. Defensives II

GAA DM Equity Country Allocation Model Update The GAA DM Equity Country Allocation model is updated as of April 30, 2018. There are no significant changes in the model's allocation this month, as shown in Table 1. Table 1Model Allocation Vs. Benchmark Weights GAA Quant Model Updates GAA Quant Model Updates Table 2Performance (Total Returns In USD, %) GAA Quant Model Updates GAA Quant Model Updates As shown in Table 2 and Charts 1, 2 and 3, the overall model outperformed its benchmark by 20 bps in April, largely driven by the Level 2 model which outperformed by 44 bps while the Level 1 model outperformed only by 2 bps. Since going live, the overall model outperformed the MSCI World by 156 bps, due to the 493 bps of outperformance from the Level 2 model which allocates funds among 11 non-U.S. countries. The Level 1 model (which allocates funds between U.S. and the non-U.S.) is on par with the MSCI world benchmark.Please see also the website http://gaa.bcaresearch.com/trades/allocation_performance. Chart 1GAA DM Model Vs. MSCI World GAA DM Model Vs. MSCI World GAA DM Model Vs. MSCI World Chart 2GAA U.S. Vs. Non U.S. Model (Level 1) GAA U.S. Vs. Non U.S. Model (Level1) GAA U.S. Vs. Non U.S. Model (Level1) Chart 3GAA Non U.S. Model (Level 2) GAA Non U.S. Model (Level 2) GAA Non U.S. Model (Level 2) Text below For more details on the models, please see Special Report, "Global Equity Allocation: Introducing The Developed Markets Country Allocation Model," dated January 29, 2016, available at https://gaa.bcaresearch.com. Please note that the overall country and sector recommendations published in our Monthly Portfolio Update and Quarterly Portfolio Outlook use the results of these quantitative models as one input, but do not stick slavishly to them. We believe that models are a useful check, but structural changes and unquantifiable factors need to be considered too in making overall recommendations. GAA Equity Sector Selection Model The GAA Equity Sector Selection Model (Chart 4) is updated as of April 30, 2018. For the third consecutive month, the model maintains a defensive positioning generating an alpha of 60 bps for the month of April. Following the end of trade threats (for now), the growth component of the model has stabilized. But, overall the model maintains the same weights from last month with an aggregate tilt of 1.3% towards defensive sectors. Energy remains the only cyclical sector with an overweight on the back of favorable valuations and improving momentum. Among defensive sectors, utilities maintains a large overweight of 5% on the back of better momentum. Chart 4Overall Model Performance Overall Model Performance Overall Model Performance Table 3Allocations GAA Quant Model Updates GAA Quant Model Updates Table 4Performance Since Going Live GAA Quant Model Updates GAA Quant Model Updates For more details on the model, please see the Special Report "Introducing The GAA Equity Sector Selection Model," dated July 27, 2016, available at https://gaa.bcaresearch.com. Xiaoli Tang, Associate Vice President xiaoliT@bcaresearch.com Aditya Kurian, Senior Analyst adityak@bcaresearch.com