Commodities & Energy Sector

Special Report

Highlights In this report, we build and present models designed to predict the odds of Chinese investable equity sector outperformance, based on a set of macroeconomic and equity market factors. BCA Research's China Investment Strategy service will aim to use our newly developed sector outperformance probability models to help investors to better understand the drivers of performance at any given moment, and to make more active equity sector recommendations in the future. Among the top six factors explaining historical periods of sector performance, three were macroeconomic in orientation, and two were directly related to the broad Chinese equity market. We see this as strongly supportive of the potential returns to be earned from active top-down sector rotation within China’s investable market. Cyclical stocks are very depressed relative to defensives, and we would favor them versus defensives over the coming year if China strikes a trade deal with the US and the Chinese economy incrementally improves, as we expect. Feature In our June 19 Special Report, we reviewed the predictability and cyclicality of equity sector earnings in China's investable & domestic markets, and examined the relevance of earnings in predicting relative sector performance over the past decade. We noted that a few sectors scored highly in terms of earnings predictability and the relevance of those earnings in predicting relative performance. But we also highlighted that most of China's equity sectors, in both the investable and domestic markets, either demonstrated earnings trends that were difficult to predict based on the trend in overall market earnings or exhibited relative performance that was difficult to explain based on the relative earnings profile. Our models are designed to predict equity sector relative performance using a series of macroeconomic and equity market factors. In short, our June report underscored that China’s equity sectors warranted a closer examination, with a particular emphasis on understanding the specific macroeconomic or equity market factors that have historically predicted relative sector performance. Today’s report examines this question in depth, focused on China’s investable equity market. We hope to extend our research to the A-share market in the near future. Our approach focuses on constructing and presenting models that quantify a checklist-based approach to determining the odds of equity sector performance. The aim is to use these models to better understand the drivers of performance at any given moment, and to make more active equity sector recommendations in the future. These recommendations will not mechanically follow the models; rather, we plan to use them as a stand in for what typically would be expected given the macro and financial market environment, and as a basis to investigate “abnormal” relative performance. We conclude by highlighting the substantial underperformance of cyclical vs defensives sectors over the past two years, and argue that it is highly unlikely that cyclicals will underperform defensives over the coming 12 months if China strikes a trade deal with the US and the economy incrementally improves, as we expect. We also explain the importance of monitoring the relative performance of health care & utilities stocks over the coming few months, and present a unique sector-based barometer for gauging China’s reflationary stance. The latter two relative performance trends are likely to assist investors in positioning for the big call: the outperformance of Chinese investable stocks vs the global benchmark. Detailing Our Approach In our effort to better understand historical periods of sector outperformance, we have chosen to model the probability of outperformance of each level 1 GICS sector (plus banks) based on a set of macro and equity market variables. Specifically, we use an analytical tool called a logistic regression, which forecasts the probability of a discrete event rather than forecasting the value of a dependent variable. We utilized this approach when building our earnings recession model for China (first presented in our January 16 Special Report1), and investors will often see it (in its conceptually different but practically similar probit form) employed when analyzing the likelihood of an economic recession. The New York Fed’s US recession model is a notable example of the latter,2 which has received much attention by market participants over the past year following the inversion of the US yield curve. The “events” that we modeled are historical periods of individual Chinese investable sector outperformance from 2010 to 2018, relative to the MSCI China index (the “broad market”). Charts I-1A and I-1B illustrate these periods with shading in each panel. We then attempt to explain these episodes of outperformance with the following macro predictors: Chart I-1AThis Report Builds Models Aimed At... Chart I-1B...Predicting The Shaded Regions Of These Charts Periods of accelerating economic activity, represented by our BCA's China Activity Index Periods of rising leading indicators of economic activity, represented by our BCA Li Keqiang Leading Indicator Episodes of tight monetary policy, defined as periods where China’s 3-month interbank repo rate is rising Periods of accelerating inflation, measured both by headline and core inflation We also include several equity market variables: uptrends in relative sector earnings, periods of rising broad market stock prices, uptrends in broad market earnings, and episodes of extreme technical conditions and relative over/undervaluation for the sector in question. In the case of energy stocks, we also include oil prices as a predictor. Charts I-2A and I-2B illustrate these periods as well as the macro & market variables that we have included as predictors. Chart I-2AWe Use These Macroeconomic And Equity Market Factors... Chart I-2B...To Predict Periods Of Equity Sector Outperformance Our approach also accounts for the existence of any leading or lagging relationships between the macro and market variables we have used as predictors and sector relative performance. In most cases the predictors lead relative sector performance, but in some cases it is the opposite. In the case of the latter, we have limited the lead of any variable in our models to 3 months in order to reduce the need to forecast. The link between tight monetary policy and industrial sector performance is one exception to this rule that we detail below. Finally, our approach also limits the extent to which we consider a leading relationship between our predictors and relative sector performance, in order to avoid picking up overlapping economic cycles. This issue, and the evidence supporting the existence of a 3½-year credit cycle in China, are detailed in Box 1. Box 1 Accounting For China’s 3½-Year Credit Cycle Over the course of the analysis detailed in this report, judgments concerning how much of a lead or lag to allow when accounting for any leading or lagging relationships between sector relative performance and either macroeconomic & stock market predictors were necessary. In cases where sector relative performance led any of our predictors, we capped the lead at 3-months to reduce the need to forecast the predictors when using the models. As explained below, the 8-month lead between industrial sector relative performance and tight monetary policy was the only exception to this rule. We also did not include any leading relationship between relative sector stock performance and the trend in relative sector EPS, and allowed at most a co-incident relationship. Limits were also required in the cases where our predictors led relative sector performance. While more lead time is usually better from the perspective of investment strategy, Chart I-B1 presents strong evidence of a 3½ -year credit cycle in China. Chart I-B2 illustrates the problem with including significant lags between predictors and relative sector performance when economic cycles are short. The chart shows the lead/lag correlation profile of the stylized cycle shown in Chart I-B1, and highlights that lags greater than 12-14 months risk picking up the impact of the previous economic cycle. Given this, we have limited the extent to which our predictors can lead relative sector performance in our models, and in practice lead times are generally less than one year. Chart I-B1Over The Past Decade, China Has Experienced A 3½-Year Credit Cycle Chart I-B2With Short Cycles, Excessive Lags Risk Picking Up The Previous Cycle The Key Drivers Of Chinese Investable Equity Sectors Pages 12-23 present the results of each sector’s outperformance probability model, along with a list of factors that were found to be useful predictors and a summary of the results. The importance of the factors included in the models is shown in each of the tables at the top right of pages 12-23 by a score of 1-3 stars, (loosely representing key levels of statistical significance) as well as each factor’s optimal lead or lag. A minus sign shows that the predictor leads sector relative performance, whereas a plus sign shows that it lags. Rising core inflation in China is the most important signal of sector performance that emerged from our analysis. Chart I-3China’s Sectors Linked Strongly To Core Inflation, Monetary Policy, And Growth Chart I-3 summarizes the significance of the factors in predicting sector performance in general, by summing up each predictor’s number of stars across all of the models. The chart shows that rising core inflation in China is the most important signal of sector performance that emerged from our analysis, followed by tight monetary policy, rising economic activity, rising broad market stock prices, oversold technical conditions, and rising broad market earnings. Chart I-3 highlights two important points: If regarded through the lens of causality alone, the strong relationship between rising core inflation and sector performance is somewhat surprising: normally, pricing power is subordinate to revenue/sales/demand as the primary factor driving fundamental performance. However, given that inflation is a lagging economic variable, we suspect that the significance of inflation in our models actually reflects the middle phase of the economic cycle in which sectors tend to best exhibit meaningful out/underperformance. It is also a stronger predictor of periods of tight monetary policy in China than headline inflation.3 This is an encouraging result for investors, as it suggests good odds that future episodes of meaningful sector outperformance can be identified given a particular macro view. Among the top six factors explaining historical periods of sector performance, three were macroeconomic in orientation, and two were directly related to the broad Chinese equity market. While Chinese equity sector performance can sometimes be idiosyncratic, we see this as strongly supportive of the idea that investors can earn positive excess returns by actively shifting between China’s equity sectors using a top-down approach. Turning to the specific results of our sector models, we present the following big-picture findings of our research: Defining China’s Cyclical & Defensive Sectors From a top-down perspective, the most important element of sector rotation typically involves shifting from defensive to cyclical stocks when economic activity is set to improve (and vice versa). In China, it is clear from the results of our models that the investable energy, materials, industrials, consumer discretionary, and information technology sectors are cyclical sectors. The relative performance of these sectors exhibits a positive relationship to pro-cyclical macro variables, or broad market trends. Following last year’s GICS changes, we also include the media & entertainment industry group (within the new communication services sector) in this list. Correspondingly, investable consumer staples, health care, financials, telecom services, utilities, and real estate are defensive sectors in China. Chart I-4Cyclical Stocks Are Bombed Out Versus Defensives Chart I-4 illustrates how these sectors have performed over the past decade by grouping them into equally-weighted cyclical and defensive stock price indexes, as well as the relative performance of cyclicals versus defensives. The chart makes it clear that cyclical stock performance is essentially as weak as it has ever been relative to defensives over the past decade, with the exception of a brief period in 2013. Panel 2 highlights that all of the underperformance of cyclicals over the past two years has been due to de-rating, rather than due to underperforming earnings. The Atypical Case Of Financials & Real Estate The fact that financial and real estate stocks are defensive in China is somewhat curious. In the case of financials, the abnormality is straightforward: most global equity portfolio managers would consider financials to be cyclical, and our work suggests that this is not true for the investable market. Our explanation for this apparent discrepancy is also straightforward: while small and medium banks in China have obviously grown in prominence over the past decade, large state-owned or state-affiliated commercial banks are still dominant in the provision of credit to China's old economy. In most cases China’s large banks lend to state-owned enterprises with implicit government guarantees, meaning that the earnings risk for Chinese banks has typically been lower than for the investable market in the aggregate. It remains to be seen whether this will remain true in a world where Chinese policymakers are keen to slow the pace at which China’s macro leverage ratio rises and to render the existing stock of debt more sustainable for the non-financial sector. Indeed, over a multi-year time horizon, the risk are not trivial that banks will be forced to recapitalize as a result of forced changes to loan terms (eg: significant increases in the amortization period of existing loans) or the recognition of sizeable loan losses, which would clearly increase the cyclicality of the Chinese investable financial sector. Chart I-5A Seeming Contradiction: Real Estate Is High-Beta, But Defensive On the real estate front, the anomaly is not that real estate stocks respond defensively to macroeconomic and stock market variables, it is that real estate stock prices are considerably more volatile than this defensive characterization would suggest. Globally (and especially in the US), real estate stocks are often viewed as bond proxies and thus are typically low-beta, but Chart I-5 shows that this is not the case in China. In our view, this issue is reconciled by the fact that Chinese investable real estate stocks are also highly positively linked to Chinese house price appreciation, with relative performance typically leading a pickup in house prices by up to 1 year. This strongly leading relationship has meant that real estate stocks have often outperformed the broad market as economic activity is slowing, in anticipation that policy easing will lead to an eventual recovery in house prices. Chart I-6Still Following The Defensive Playbook This Year In effect, investable real estate stocks are a high-beta sector that have acted counter-cyclically due to the historical interplay between economic activity, monetary policy, and the housing market. Real estate performance this year has not deviated from this playbook (Chart I-6), and so for now we are content to include real estate stocks in our defensive index. But similar to the case of financials, we can conceive of scenarios in which ongoing Chinese financial sector reform may change this relationship in the future. The Unique Monetary Policy Sensitivity Of Industrials And Consumer Staples Pages 14 and 16 highlight that industrials and consumer staples stocks have typically been sensitive to periods of tight monetary policy. In the case of industrials the relationship is negative, whereas consumer staples relative performance has been positively linked to these periods. In both cases, relative performance has led periods of tight monetary policy, significantly so in the case of industrials (by an average of 8 months). While the relative performance of banks, tech, and real estate stocks have also been linked to periods of tight monetary policy, industrials and consumer staples are the only sectors that have tended to lead these periods. Chart I-7Diverging Corporate Health Explains Industrials/Staples Monetary Policy Sensitivity This is a revelatory finding, and in our view it is explained by divergences in corporate health and leverage for the two sectors. We reviewed Chinese corporate health in our August 28 Special Report,4 and noted that the food & beverage sub-industry was a clear (positive) outlier based on our corporate health monitors. In particular, Chart I-7 highlights that food & beverage corporate health is markedly better than that for machinery companies or for industrial firms in general, supporting the notion that high (low) leverage is impacting the relative performance of industrials (consumer staples). The Leading Nature Of Health Care & Utilities Health care and utilities exhibit similar key drivers of relative performance: in both cases, periods of rising economic activity, rising core inflation, and rising broad market stock prices are all negatively associated with performance. Health care and utilities relative performance also happens to lead all three of those predictors, by 1-3 months on average depending on the variable in question. Our modeling work highlights that these are the only sectors whose relative performance has led multiple factors, suggesting that health care & utilities stocks are particularly interesting market bellwethers to monitor. Core Inflation Matters More Than Headline, Except For Energy & Real Estate As highlighted in Chart I-3, rising core inflation has been a much more important signal about relative sector performance than headline inflation. Chart I-8In China, Food Prices (Not Energy) Account For Headline/Core Differences The two exceptions to this rule relate to the energy and real estate sectors, with the former positively linked to headline inflation and the latter negatively linked. In both cases, we suspect that the relationship is a behavioral rather than a fundamental one. For energy, while rising headline inflation in developed countries is usually associated with rising energy prices, this is not true in the case of China. Chart I-8 highlights that differences between headline and core inflation over the past decade have almost always been driven by rising food prices. This implies that some investors (incorrectly) view energy stocks as a hedge against increases in consumer prices, even if those increases are not driven by rising fuel costs. In the case of real estate, investor expectations of eroding real disposable income and its impact on the housing market are likely the best explanation for the negative link between real estate relative performance and rising headline inflation. Whereas rising core inflation likely reflects a durable improvement in economic momentum (and thus would be positively correlated with income growth), episodes of rising Chinese headline inflation often reflect supply shocks that investors may perceive to be detrimental to household spending power (and thus expected housing demand). Investment Conclusions Our work aimed at explaining historical periods of Chinese investable sector outperformance has three investment implications in the current environment. Cyclicals will probably outperform defensives over the coming year if China strikes a trade deal with the US and the Chinese economy incrementally improves, as we expect. First, within China’s investable market, Chart I-4 illustrated that cyclical stocks are very depressed relative to defensives. Given our view that Chinese investable stocks are likely to outperform their global peers over a 6-12 month time horizon, we would also favor cyclicals to defensives over that period. For investors who are not yet overweight cyclical stocks in China, we would advise waiting for concrete signs that growth has bottomed (which should emerge sometime in Q1) before putting on a long position as we remain tactically neutral towards Chinese versus global stocks. But the key point is that it is highly unlikely that cyclicals will underperform defensives over the coming year if China strikes a trade deal with the US and the Chinese economy incrementally improves, as we expect. Second, the fact that investable health care and utilities stocks have particularly leading properties suggests that they should be monitored closely over the coming few months. A technical breakdown in the relative performance of these sectors would be an important sign that market participants are anticipating a bottoming in China’s economy, which may give investors a green light to position for a bullish cyclical stance. For now, both of these sectors continue to outperform (Chart I-9), supporting our decision to remain tactically neutral towards Chinese stocks. Third, the heightened negative sensitivity of industrials and positive sensitivity of consumer staples to monetary policy suggests that the relative performance trend between the two sectors may serve as a reflationary barometer for China’s economy. Chart I-10 shows that industrials outperformed staples last year once the PBOC shifted into easing mode, and anticipated the recovery in the pace of credit growth. However, industrials soon began to underperform staples, which also seems to have anticipated the fact that the recovery in credit was set to be less powerful than what has occurred during previous cycles. The fact that the relative performance trend is off its recent low is notable, and may suggest that China’s existing reflationary stance will be sufficient to stabilize economic activity if a trade deal with the US is indeed finalized in the near future. Chart I-9Key Defensive Sectors Are Still Outperforming, Supporting Our Neutral Tactical Stance Chart I-10Industrials Vs. Staples Anticipated That Easing Would Only Be Measured As a final point, BCA Research's China Investment Strategy service will aim to use our newly developed sector outperformance probability models to make more active equity sector recommendations in the future. These recommendations will not mechanically follow the models; rather, we plan to use the models as a stand in for what typically would be expected given the macro and financial market environment, and as a basis to investigate “abnormal” relative performance. We hope you will find these models to be a helpful quantification of the risk versus return prospects of allocating among China’s investable sectors. As always, we welcome any feedback that you may have about our approach.   Energy Chart II-1 Table II-1   Unsurprisingly, our energy sector model highlights that periods of energy outperformance are strongly linked to periods of rising crude oil prices. However, what is surprising is that periods of accelerating headline inflation in China are even more closely linked to periods of energy sector outperformance than episodes of rising oil prices, and that these periods of accelerating inflation are not generally caused by rising energy prices. The lack of a clear economic rationale for this relationship implies that some investors (incorrectly) view energy stocks as a hedge against increases in consumer prices, even if those increases are largely driven by rising food prices. The model also highlights that periods of strong undervaluation have historically been significant in predicting future energy sector outperformance, with a lag of roughly 8 months. The probability of energy sector outperformance has fallen sharply according to our model, but for now we continue to recommend a long absolute energy sector position on a 6-12 month time horizon. BCA’s Commodity & Energy Strategy service expects oil prices to trade at $70/barrel on average next year,5 Chinese headline inflation continues to rise, and we noted in our October 2 Weekly Report that energy stocks are heavily discounted.6 Barring a durable decline in oil prices below $55/barrel, investors should continue to favor China’s energy sector. Materials Chart II-2 Table II-2 Our model highlights that the materials sector is one of the clearest plays on accelerating industrial activity within the investable universe. Among the macro variables that we tested, periods of investable materials outperformance are strongly positively linked with periods when our BCA Activity Index and our leading indicator for the index have been rising. Periods of materials sector outperformance have also been positively correlated with prior periods of oversold technical conditions and rising broad market stock prices, underscoring that materials are a strongly pro-cyclical sector. We currently maintain no active relative sector trades, but our model suggests that investors should be underweight the investable materials sector relative to the broad investable index. Industrials Chart II-3 Table II-3 Periods of industrial sector outperformance have historically been positively correlated with relative industrial sector earnings, broad market stock prices, and prior oversold technical conditions. They have been negatively correlated with periods of tight monetary policy, rising core inflation, and prior overbought technical conditions. Since 2010, periods of industrial sector performance have led periods of tight monetary policy by 8 months, the longest lead of relative equity performance to any macro variable that we tested in our model (and the longest lead that we allowed). Industrial sector performance has also been strongly negatively linked with periods of rising core inflation. These findings, and the fact that our Activity Index and its leading indicator have not been highly successful at predicting periods of industrial sector outperformance, strongly suggest that industrials, while pro-cyclical, are primarily driven by expectations of easy monetary policy. We noted in an August 2018 Special Report that state-owned enterprises have become substantially leveraged over the past decade,7 and in a more recent report we highlighted that industries such as machinery have experienced a significant deterioration in corporate health over the past decade.8 This helps explain why industrial sector performance is so negatively impacted by tight policy. Our model suggests that the best time to be overweight industrial stocks is the early phase of an economic rebound, when Chinese stock prices are rising but market participants are not yet expecting tighter policy. These conditions may present themselves sometime in Q1, but probably not over the coming 0-3 months. Consumer Discretionary Ex-Internet & Direct Marketing Retail Chart II-4 Table II-4 Besides materials, China’s investable consumer discretionary sector has historically been the most positively associated with coincident and leading measures of industrial activity. Rising core inflation is also highly positively related to consumer discretionary outperformance, which may reflect improved pricing power for the sector. The strong link with industrial activity is in contrast to depictions of China’s consumer sector as being less correlated to money & credit trends than the overall economy, and is supportive of our view that industrial activity forms one of the three pillars of China’s business cycle.9 We ended the estimation period of our model as of December 2018, in order to avoid including the distortive effects of last year’s changes to the global industry classification standard (which resulted in Alibaba’s inclusion and overwhelming representation in the investable consumer discretionary sector). As such, the results of our model apply today to consumer discretionary stocks ex-internet & direct marketing retail. For now, the absence of an uptrend in our Activity Index and in core inflation is signaling underperformance of discretionary stocks outside of internet & direct marketing retail. Outperformance this year largely reflects a significant advance in consumer durable and apparel: by contrast, automobiles & components have underperformed the broad market by roughly 14% year-to-date. Consumer Staples Chart II-5 Table II-5 Historically, periods of consumer staples outperformance have been predicted by a falling Activity Index, periods of tight monetary policy, and over/undervalued conditions. The impact of monetary policy is particularly heavy in the model, suggesting that consumer staples are somewhat the mirror image of industrials in terms of the impact of leverage on relative equity performance. This too is supported by our August 28 Special Report,10 which noted that corporate health for the food & beverage sector was the strongest among the sectors we examined. However, the model failed to capture what has been very significant staples outperformance this year, highlighting the occasional limits of a rule-of-thumb approach to sector allocation. Investable consumer staples are reliably low-beta compared with the broad market, and we are not surprised that investors have strongly favored the sector this year amid enormous economic and policy uncertainty. An eventual improvement in economic activity, coupled with fairly rich valuation, should work against consumer staples stocks sometime in the first quarter of 2020. Investors who are positioned in favor of China-related assets should also be watching closely for any signs of a technical breakdown in the relative performance trend of investable staples. Health Care Chart II-6 Table II-6 Among the macro variables tested in our model, periods of health care outperformance are negatively related to coincident and leading measures of industrial activity and strongly negatively related to rising core inflation.  Health care outperformance is also strongly negatively related to periods of rising broad market stock prices, and positively related to prior oversold technical conditions. These results clearly signify that investable health care is a defensive sector, to be owned when the economy is slowing and when investable stocks in general are trending lower. Our model suggests that health care stocks are likely to continue to outperform, as they have been since the beginning of the year. A substantive US/China trade deal that meaningfully reduces economic uncertainty remains the key risk to health care outperformance over a 6- to 12-month time horizon. Financials Chart II-7 Table II-7 Our model highlights that periods of financial sector outperformance over the past decade have been negatively associated with periods of rising core inflation (a strong relationship), and with periods of rising index earnings. Oversold technical conditions have also helped explain future episodes of financial sector outperformance. The link between core inflation and the outperformance of financials appears to represent a behavioral rather than a fundamental relationship. When modeling periods of rising financial sector relative earnings, the trend in broad market EPS is more predictive than that of core inflation, highlighting that the latter’s explanatory power is due to investor behavior. The results of our model, and the fact that core inflation leads Chinese index earnings, suggests that financials are fundamentally counter-cyclical and that investors see rising Chinese core inflation as confirmation that an economic expansion is underway (and that broad market earnings are likely to rise). Our model is currently predicting financial sector outperformance, but investable financials have modestly underperformed since the beginning of the year. This appears to have been caused by the underperformance of financial sector earnings this year as overall index earnings growth has decelerated, contrary to what history would suggest. We suspect that the ongoing shadow banking crackdown is related to financial sector earnings underperformance, and we would advise against an overweight stance towards investable financials until signs of improving relative earnings emerge. Banks Chart II-8 Table II-8 Our model shows that periods of banking sector outperformance are more linked to macro variables than has been the case for the overall financial sector. Specifically, bank performance is negatively correlated with leading indicators of economic activity and rising core inflation, and especially negatively correlated with periods of tight monetary policy. Banks have also typically outperformed following periods of oversold technical conditions. Similar to financials, bank earnings are typically counter-cyclical, but relative bank earnings have not been good predictors of relative bank performance over the past decade. Still, the negative association of relative stock prices with leading economic indicators, rising core inflation and rising interest rates underscores that investors should normally be underweight banks if they expect overall Chinese stock prices to rise. Also similar to the overall financial sector, our model is currently predicting outperformance for bank stocks, but investable banks have underperformed year-to-date. The shadow banking crackdown is also likely impacting investable bank earnings, leading to a similar recommendation to avoid bank stocks until relative earnings look to be trending higher. “Tech+”   Chart II-9 Table II-9 Our technology model has worked well at predicting periods of tech sector outperformance over the past several years, particularly from 2015 – 2017. The model suggests that, in addition to being negatively related to prior overbought conditions, periods of technology sector outperformance are associated with improving growth conditions, easy monetary policy, and rising prices. In other words, tech stocks are a growth & liquidity play. Owing to last year’s changes to the GICS, the results of our model apply today to Chinese investable internet & direct marketing retail, the media & entertainment industry group (within the new communication services sector), and the now considerably smaller information technology sector (the sum of which could be considered the “tech+” sector). The model has been predicting tech sector outperformance since May (in response to easier monetary policy), which has occurred for the official information technology sector. However, the BAT (Baidu, Alibaba, and Tencent) stocks are only up fractionally in relative terms from their late-May low. Our expectation that China’s economy is likely to bottom in Q1 means that we may recommend upgrading “tech+” stocks relative to the investable benchmark in the coming months. Telecom Services Chart II-10 Table II-10 Our model for telecommunication services (now a level 2 industry group within the communication services sector) illustrates that telecom stocks have historically been counter-cyclical. Periods of telecom outperformance have been negatively associated with periods of rising core inflation, rising broad market stock prices, and rising broad market EPS. It is notable that telecom services stocks are driven more by cycles in overall stock prices than by cycles in economic activity. This suggests that investors tend to focus on the fact that telecom stocks are reliably low-beta compared with the overall investable market, causing out(under)performance of telecoms when the broad market is falling(rising). Similar to financials & banks, telecom stocks have not outperformed this year, in contrast to what our model would suggest. Earnings also appear to be the culprit, with the level of 12-month trailing earnings having fallen nearly 10% since the summer. China Mobile accounts for a sizeable portion of the telecom services index, and the company’s recent earnings weakness seems to be due to depreciation charges stemming from forced investment on 5G spending (mandated by the Chinese government). Our sense is that this will have only a temporary effect on telecom services EPS, meaning that investors should continue to expect the sector to behave in a counter-cyclical fashion over the coming year. Utilities Chart II-11 Table II-11 The early performance of our utilities model was mixed, as it generated several false sell signals during the 2011 – 2013 period despite recommending, on average, an overweight stance. However, over the past five years, the model has performed extremely well in terms of explaining periods of relative utilities performance. The model highlights that utilities are straightforwardly counter-cyclical. The relative performance of utilities stocks is positively related to its relative earnings trend, and negatively related to economic activity, rising core inflation, and broad market stock prices.  Consistent with a decline in the overall MSCI China index, the model has correctly predicted utilities outperformance this year. We expect utilities to underperform over a 6-12 month time horizon, but would advise against an aggressive underweight position until hard evidence of a bottom in Chinese economic activity emerges. Real Estate Chart II-12 Table II-12 Our model for the relative performance of investable real estate has been among the most successful of those detailed in this report, which is somewhat surprising given the macro factors that the model shows drive real estate performance. While periods of relative real estate performance are modestly (negatively) associated with periods of tight monetary policy, rising headline inflation is the most important macro predictor of real estate underperformance. Among market factors driving performance, real estate stocks reliably underperform when broad market EPS are trending higher, and they historically outperform for a time after becoming relatively undervalued. Real estate relative performance is also strongly linked to periods of rising house prices, but the former tends to significantly lead the latter. Given that core inflation has better predicted episodes of tight monetary policy than headline inflation, investor expectations of eroding real disposable income is likely the best explanation for the negative link between real estate relative performance and rising headline inflation. Whereas rising core inflation likely reflects a durable improvement in economic momentum (and thus would be positively correlated with income growth), episodes of rising Chinese headline inflation often reflect supply shocks that investors may perceive to be detrimental to household spending power (and thus expected housing demand). Beyond the negative link between higher inflation and interest rates on investable real estate performance, the strong negative association with broad market earnings underscores that investors treat real estate as a defensive sector. We thus expect real estate stocks to continue to outperform in the near term, but underperform over a 6-12 month time horizon.   Jonathan LaBerge, CFA Vice President jonathanl@bcaresearch.com   Footnotes 1. Please see China Investment Strategy, "Six Questions About Chinese Stocks," dated January 16, 2019. 2. Please see Federal Reserve Bank of New York, The Yield Curve as a Leading Indicator at https://www.newyorkfed.org/research/capital_markets/ycfaq.html 3. This is despite frequent concerns among investors that the PBOC is inclined to tighten in response to detrimental supply shocks. 4. Please see China Investment Strategy, "Messages From BCA’s China Industry Watch," dated August 28, 2019. 5. Please see Commodity & Energy Strategy, "Policy Uncertainty Lifts USD, Stifles Global Oil Demand Growth," dated October 17, 2019. 6. Please see China Investment Strategy, "China Macro & Market Review," dated October 2, 2019. 7. Please see China Investment Strategy, "Chinese Policymakers: Facing A Trade-Off Between Growth And Leveraging," dated August 29, 2018. 8. Please see China Investment Strategy, "Messages From BCA’s China Industry Watch," dated August 28, 2019. 9. Please see China Investment Strategy, "The Three Pillars Of China’s Economy," dated May 16, 2018. 10. Please see China Investment Strategy, "Messages From BCA’s China Industry Watch," dated August 28, 2019. Cyclical Investment Stance Equity Sector Recommendations

Special Report

Highlights Rising recession risk, shaky economic fundamentals, and absence of positive yielding assets motivate us to reexamine which assets can be counted on to protect a portfolio in the future. We analyze 10 safe havens on four different dimensions: consistency, versatility, efficiency, and costs. Using this framework, we examine the historical performance of each safe haven and provide an outlook on their likely effectiveness over the next decade. We conclude that U.S. TIPS and farmland should provide the best portfolio protection. Cash, U.S. Treasuries and gold are other good alternatives. Meanwhile, U.S. investment-grade bonds, global ex-U.S. bonds, silver, and currency futures are likely to be poor protection choices. Feature For most investors, capital preservation is the most important goal when managing money. However, how to go about it remains a difficult question.  Investing in safe havens can be painful during bull markets, as their returns are usually lower than those of equities. Moreover, economic, political, and financial regimes change over time, which means that an asset that protected your portfolio in the past might not do so in the future. Therefore, it becomes good practice to review one’s safety measures periodically, even if one does not think that a crash is imminent. The current environment in particular, is a propitious time to review safe havens given that: Chart I-1A Great Time To Review Safety Measures A key recession signal is flashing red: The yield curve inverted in the United States in August (Chart I-1 – top panel). An inversion of the yield curve does not necessarily imply a recession, but historically it has been a very reliable signal of one, given that it indicates that monetary policy is too tight for the economy. Structural risks are rising: Rich equity valuations in the U.S. and high leverage levels elsewhere are signs that the pillars supporting this bull market might be fragile (Chart I-1 – middle panel). In addition, protectionism and populism, forces that BCA has long argued are here to stay, threaten to upend the regime of free trade that has benefited equities since the 1950s.1 Yields are near all-time lows: Historically, investors have been able to endure bear markets by hiding in safe assets with positive yield, as these assets will normally provide a reliable cash flow regardless of the economic situation. However, these type of assets are increasingly hard to find, particularly in the government bond space, where 50% of developed country bonds have negative yields (Chart I-1 – bottom panel). Considering these factors, how should investors protect their portfolios in the next decade? To answer this question, we analyze 10 safe havens divided into five broad asset classes: Nominal government bonds: U.S. Treasuries and global ex-U.S. government bonds. Other fixed income: U.S. investment-grade credit and U.S. TIPS.2 Currencies: yen futures and Swiss franc futures. Precious metals: gold futures and silver futures. Other assets: farmland and U.S. cash. We look at historical performance since 1973 for all safe havens except for global ex-U.S. bonds and farmland. For these assets, we look at performance since 1991 due to limited data availability. We mainly look at quarterly returns in order to compare illiquid assets to publicly traded ones. We do not consider each safe haven in isolation, but rather as an addition to equities within a portfolio. Specifically, we explore our safe haven universe relative to the MSCI All Country World equity index from the perspective of a U.S. investor. For our non-U.S. clients, we will release a report from the perspective of other countries if there is sufficient interest. Importantly, we do not look only at historical performance. We also examine whether there is a reason to believe that future returns will be different from past ones, by analyzing how the properties of each safe haven might have changed. When evaluating each safe haven, we focus on four properties: Consistency: a safe haven should generate consistent positive returns during periods of negative equity performance, with returns increasing with the severity of the equity drawdown. Versatility: safe havens should perform well across different types of crises. Efficiency: a safe haven should produce enough upside during crises, so only a small allocation to the safe haven is necessary to reduce losses. Costs: drag to portfolio overall performance (opportunity costs) should be as small as possible. Readers who wish to see just our overall conclusions should read our Summary Of Results section below. For our analysis of how safe havens have performed in the past, please see the Historical Performance section. Finally, for our analysis of how we expect the performance of safe havens to change, please see our Outlook section. Summary Of Results The Best Safe Havens U.S. TIPS should be an excellent safe haven to protect a portfolio in the next decade. While TIPS might not be as cheap to hold as they have been in the past, upside potential remains strong, which means that a moderate allocation can provide substantial protection to an equity portfolio. Moreover, U.S. TIPS are one of the best hedges against crises triggered by rising rates and inflation, which in our view are the biggest structural risks that asset allocators face. Farmland could also be a great safe haven for investors who have the ability to allocate to illiquid assets given that it is the cheapest safe haven in terms of portfolio drag. However, investors should be aware that the current low yield could potentially affect its performance during crises. Good Alternatives Cash can be a good alternative to protect an equity portfolio, given its outstanding performance during equity drawdowns caused by inflation. Moreover, its opportunity costs should decrease relative to the past. However, investors should take into account that the efficiency of cash at the current juncture is poor, which means that a relatively large allocation is needed in order to achieve meaningful portfolio protection. A portfolio with a 30% allocation to Treasuries historically provided the same downside protection as a portfolio with a 44% allocation to gold. We also like gold futures as a safe haven since they offer some of the most attractive opportunity costs. In addition, their upside is greater than that of most safe havens due to their negative correlations with real rates. However, gold’s volatility makes it an unreliable asset, which prevents us from placing it higher in the safe haven hierarchy. Historically, U.S. Treasuries have been one of the best safe havens to hedge an equity portfolio. Will this performance continue in the future? We do not think so. While yields are still high enough to provide plenty of upside potential, they have fallen to the point where they have increased the opportunity costs of U.S. Treasuries and reduced their consistency. The Rest Global ex-U.S. bonds have very limited upside due to their low yields. Meanwhile U.S. investment-grade credit remains at risk from poor corporate balance sheets, compounded by the fact that credit no longer has an attractive yield cushion. Currencies like the yen and the Swiss franc will continue to be unreliable and very expensive safe havens. Finally, while silver’s costs and reliability could improve, its high cyclicality relative to other safe havens will make silver a poor protection choice. Historical performance Consistency How did safe havens perform when equities lost money? To assess consistency, we plot the performance of each safe haven during all quarters when global equities had losses (Chart I-2). Cash and farmland were the only assets to have positive returns during every equity drawdown. U.S. Treasuries and U.S. TIPS were also very consistent, and had the additional advantage that their returns tended to increase as equity losses worsened. Global ex-U.S. bonds, while not as consistent, generated positive returns most of the time. Chart I-2Safe Haven Returns During Drawdowns In Global Equities On the other hand, investment-grade bonds, the yen, the Swiss franc, gold, and silver were much more inconsistent. In general, even though these assets had larger positive returns than other assets, they were prone to deep selloffs concurrent with equity drawdowns. Silver was the worst of all safe havens, being mostly a negative return asset during quarters of negative equity performance. Versatility How did the type of crisis affect the performance of safe havens? We classify crises according to their catalyst into the following four categories: bursts of U.S. asset bubbles (tech bubble, 2008 housing crisis), ex-U.S. crises (1998 EM crisis, European debt crisis), flash crashes/political events (1987 Black Monday, 9/11 terrorist attack),  rate/inflation shocks (1974 oil crisis, 1980 Fed shock) and others (every other equity drawdown we could not classify).3  We look at the performance of seven safe havens since 1973 (Chart I-3A) and of all 10 since 19914 (Chart I-3B): Chart I-3ASafe Haven Return During Different Type Of Crisis (1973 - Present) Chart I-3BSafe Haven Return During Different Type Of Crisis (1991 - Present)   During bursts of U.S. asset bubbles, U.S. Treasuries were the most effective hedge in both sample periods, followed by U.S. TIPS and farmland. Corporate bonds, cash, gold, and the Swiss franc also had positive returns, though they were small. Finally, the yen and silver had negative returns. During crises happening outside of the U.S., U.S. Treasuries were once again the best option. U.S. TIPS, yen futures, farmland, gold, and U.S. investment-grade bonds also provided strong returns.  Meanwhile, global ex-U.S. bonds and cash provided relatively weak returns, while both the Swiss franc and silver accrued losses. During flash crashes/political events, the Swiss franc had the best performance followed by global ex-U.S. bonds, though in general all safe havens but silver provided positive returns. Rate/inflation shocks were the most difficult type of crisis to hedge. Cash and U.S. TIPS were by far the best performers. Moreover, while U.S. Treasuries were able to eke out a small positive return, all other safe havens lost money during these crises. Efficiency How much allocation to each safe haven was needed to protect an equity portfolio? Chart I-4 show how adding incremental amounts of each safe haven5 to an equity portfolio reduced the overall portfolio’s 10% conditional VaR (the average of the bottom decile of returns).6 Since 1973, U.S. TIPS and U.S. nominal government bonds were the most efficient safe havens, providing the most protection per unit of allocation (Chart I-4 – top panel). Conditional VaR was reduced by almost half when allocating 40% to either Treasuries or TIPS. Cash, U.S. investment-grade, the yen, the Swiss franc, gold, and silver followed in that order. The difference between the safe havens was significant. As an example, a portfolio with a 30% allocation to U.S. Treasuries historically provided the same downside protection as a portfolio with a 36% allocation to U.S. IG credit, a 39% allocation to the yen or a 44% allocation to gold. Meanwhile, there was no allocation to silver which would have provided the same level of protection. When using a sample from 1991, the main difference was the reduced efficiency of cash – the result of lower average interest rates when using a more recent sample. Other than cash, the efficiency of most safe havens remained unchanged: U.S. Treasuries were the best option, followed by U.S. TIPS, farmland, U.S. investment-grade bonds, global ex-U.S. government bonds, cash, the yen, gold, the Swiss franc, and silver in that order (Chart I-4 – bottom panel). Chart I-4Historically, Fixed-Income Assets Were The Most Efficient Safe Havens Costs How do safe haven returns compare to equities? To evaluate opportunity costs, we compare the difference of the historical return of each safe haven versus global equities. Overall, hedging with currencies was extremely costly, as their return was well below that of equities in both samples (Chart I-5). Cash was also an expensive safe haven to hedge with, particularly in the most recent sample. On the other hand, fixed-income assets like U.S Treasuries, investment-grade credit, and U.S. TIPS had very low costs (global ex-U.S. bonds also had cost of around 2% in a limited sample).  Farmland had negative opportunity costs because it outperformed equities during the sample period.7 Chart I-5Historically Fixed Income Assets And Farmland Had The Lowest Opportunity Cost Outlook Chart I-6No More Yield Cushion Chart I-7Silver Has Become Less Cyclical For our outlook, we assess how the four traits under study have changed for all safe havens: Consistency: Will safe havens continue to be reliable in the absence of high coupons? Many of the safe havens in our sample were effective at hedging equities due to their high yield. Even if they had negative capital appreciation, total returns stayed positive thanks to the offsetting effect of the yield return. However, as rates have declined, yield return has also decreased substantially (Chart I-6). Therefore, safe havens, like cash, government bonds, and even farmland will not be as consistent as they were in the past. Credit could be even more vulnerable: the combination of a low yield, and unhealthy fundamentals will turn U.S. corporate bonds into a negative-return asset in the next crisis. Silver might be the lone safe haven to improve its consistency. Industrial use for silver has fallen substantially in the past 10 years, decreasing its cyclical nature (Chart I-7). Thus, while silver might still be an erratic safe haven, it should be more consistent in the future than its historical performance would suggest.   Versatility: What will the next crisis look like? Chart I-8Inflation and Political Crisis Will Plague The 2020s Determining what the next crisis will look like is crucial for safe haven selection. Below we rank the types of crises in order of how likely and severe we think they will be in the future: Inflation/rate shock: We expect inflation to be significantly higher over the next decade. This will be the highest risk for asset allocators in the future. As we explained in our May 2019 report, a change in monetary policy framework, procyclical fiscal policy, waning Fed independence, declining globalization, and demographic forces are all conspiring to lift inflation in the next decade.8 Importantly, we believe that the Fed will be dovish initially, as it cannot let inflation continue to underperform its target after missing the mark for the last 10 years (Chart I-8 – top panel). However, this will cause an inflationary cycle, which will eventually lead the Fed to raise rates significantly and trigger a recession. Political events/flash crashes: Political events will also pose a risk to the markets on a structural basis. The rise of China as a superpower has shifted the world into a paradigm of multipolarity, which historically has resulted in military conflict. Moreover, animus for conflict is not dependent on President Trump. The American public in general feels that the economic relationship with China is detrimental to the United States (Chart I-8 – bottom panel). This means that any president, Democrat or Republican will have a political incentive to jostle with China for economic and political supremacy for years to come. Ex-U.S. crises: We expect Emerging Markets in general, and China in particular, to be among the most vulnerable parts of the global economy as we enter the next decade. Over the last 10 years, China’s money supply has increased four-fold, becoming larger than the money supply of the U.S. and the euro area combined. In addition, corporate debt as a % of GDP stands at 155%, higher than Japan at the peak of its bubble and higher than any country in recorded history (Chart I-9). We rank this type of crisis slightly below the first two because Emerging Market assets are depressed already. Thus, while we believe that there is further downside to come for these economies, some weakness has already been priced in. U.S. asset bubble burst: We believe that there are no systemic excesses in the U.S. economy, making a U.S. asset bubble burst a lesser risk than other types of crises. Although it is true that U.S. corporate debt stands at all-time highs, it is still at a much lower level than in other countries. Moreover, weakness of corporate credit is not likely to have systemic consequences on the economy, given that leveraged institutions like banks and households hold only a small amount of outstanding corporate debt (Chart I-10). Chart I-9EM crises Are Also A Risk Chart I-10A U.S. Corporate Debt Deblacle Will Not Have Systemic Consequences What does this ranking mean in terms of safe haven performance? U.S. TIPS and cash should be held in high regard as they will be some of the only assets that will perform well during an inflation/rate shock. The Swiss franc and global ex-U.S. bonds should be best performers during political crises, although U.S. TIPS could also provide adequate protection. Efficiency: Is there any upside left for safe havens when interest rates are near zero? As yields go below the zero bound it becomes harder for bonds to generate large positive returns. European or Japanese government bonds in particular would need their yields to go deep into negative territory to counteract a large selloff in equities (Table I-1). But can interest rates go that low? We do not think so. The recent auction of German bunds, where a 0%-yielding 30-year bond attracted the weakest demand since 2011, suggests that interest rates in these countries might be close to their lower bound.  On the other hand, though U.S. yields are low, they are still high enough for U.S. Treasuries to provide high returns in case of a crisis. Table I-1No Room For Positive Returns In The Government Bond Space Outside Of The U.S. Low rates also have an effect on the efficiency of U.S. investment-grade bonds, cash, and farmland because their upside during crises does not come from capital appreciation but rather from their yield, (the price of IG credit actually declines during most crisis). As mentioned earlier, their yield has declined substantially compared to the past, which means that a larger allocation will be necessary to counteract a selloff. Chart I-11Switzerland Has A High Incentive To Prevent The Franc From Appreciating The upside of the yen could also be compromised. The Bank of Japan is likely to intervene aggressively in the currency market to prevent the Japanese economy from falling into a deflationary spiral, since it is very difficult for it to lower Japanese rates further. The Swiss franc is even more vulnerable. In contrast to Japan, Switzerland is a small open economy that has to import most of its products (Chart I-11). This means that the Swiss National Bank has a very high incentive to intervene in currency markets during a crisis, given that a rally in the franc could depress inflation severely. What about U.S. TIPS? In contrast to nominal government bond yields or even yields on corporate debt, U.S. real rates are not limited by the zero bound (Chart I-12).  This makes TIPS a more attractive option than other fixed-income assets, since real rates can have much more room for further downside than nominal ones. To be clear, this will only be the case if our forecast of an inflationary crisis materializes. Likewise, since gold is heavily influenced by real rates, it should also offer significant upside during the next crisis.9 Chart I-12Real Rates Have More Downside Potential Than Nominal Ones Costs: Can I afford to hold safe havens in a world of low returns? To provide an outlook for the expected cost of each safe haven, we use the return assumptions from our June Special Report.10 We subtract the expected return on global equities from the expected return for each safe haven to reach an expected cost value. However, three of the safe havens (global ex-U.S. government bonds, the Swiss franc and silver) did not have a return estimate. We compute their expected returns as follows: For the Swiss franc we use the methodology we used for all other currencies in our report. We base the expected return on the current divergence from the IMF PPP value, as well as the IMF inflation estimates. In addition, we add the relative cash rate assumed return for both our yen and Swiss franc estimates, as futures take into account carry return. For global ex-U.S. bonds we take the weighted average of the expected return of the euro area, Japan, U.K., Canada, and Australia government bonds. We weight the returns according to their market capitalization in the Bloomberg/Barclays government bond index. Due to silver’s dual role as an inflation hedge and industrial metal, silver prices are a function of both gold prices and global growth. To obtain a return estimate we run a regression on silver against these two variables and use our growth and gold return estimate to arrive at an assumed return for silver. Chart I-13 shows our results: while their cost will improve, currency futures remain the most expensive hedge. The opportunity cost of precious metals and cash will decrease, making them more attractive options than in the past. Meanwhile, low yields will increase the opportunity costs of most fixed-income assets. Finally, farmland will remain the cheapest safe haven, even with decreased performance. Chart I-13Oportunity Cost For Fixed Income Safe Havens Will Be Higher Than In The Past Juan Manuel Correa Ossa Senior Analyst juanc@bcaresearch.com Appendix A Footnotes 1 Please see Geopolitical Strategy Special Report, "The Apex Of Globalization – All Downhill From Here, " dated November 12, 2014, available at gps.bcaresearch.com. 2 We use a synthetic TIPS series for data prior to 1997. For details on the methodology, please see: Kothari, S.P. and Shanken, Jay A., “Asset Allocation with Inflation-Protected Bonds,” Financial Analysts Journal, Vol. 60, No. 1, pp. 54-70, January/February 2004. 3 For a detailed list of how we classified each equity drawdown, please see Appendix A. 4 The only crises caused by a rate/inflation shock occurred in 1974 and 1980. Thus we have this type of drawdown only in Chart 3A and not in Chart 3B. 5 For yen, Swiss franc, silver and gold futures we assume an allocation to an ETF which follows their performance. Since futures have zero initial costs they cannot be directly compared to traditional assets in terms of percentage allocation. 6 We prefer this measure over VaR given that it captures the properties of the left tail of returns more accurately. 7  While the farmland index subtracts management fees, we recognize that there are costs involved in holding these illiquid assets which are not necessarily captured by the return indices. Thus, the real historical cost of holding farmland was not negative but likely close to zero. 8 Please see Global Asset Allocation Strategy Special Report "Investors’ Guide To Inflation Hedging: How To Invest When Inflation Rises," dated May 22, 2019, available at gaa.bcaresearch.com. 9 Please see Commodity & Energy Strategy Special Report "All that Glitters…And Then Some" dated July 25, 2019, available at ces.bcaresearch.com. 10 Please see Global Asset Allocation Strategy Special Report "Return Assumptions - Refreshed and Refined" dated June 25, 2019,  

The once-reliable negative correlation between gold and the USD was indefinitely suspended beginning in 4Q18 by the pervasive economic uncertainty we identified last week as the culprit holding back global oil demand growth via a super-charged dollar.1 This uncertainty is most pronounced in the U.S. and Europe vis-à-vis gold, and partly explains the performance of safe havens, particularly the USD, which has soared to new heights on a trade-weighted goods basis, and gold (Chart of the Week). So far, gold has held its ground after breaking above $1,500/oz from the low $1,200s in mid-2018, indicating investors are much more concerned about economic risks arising from economic policy uncertainty than inflation and other diversifiable risks gold typically hedges (Charts 2A, 2B). Cyclically we remain positive on gold prices on the back of a lower dollar and rising inflation pressure in the U.S. Chart of the WeekDemand For Safe Havens Soars As Economic Policy Uncertainty Rises Economic policy uncertainty in Europe and the U.S. supports gold prices. Even so, we are putting a $1,450/oz stop-loss on our long gold portfolio hedge to cover tactical risks showing up in our technical indicators. In addition, as is the case with oil demand, if the ceasefire we are expecting in the Sino-U.S. trade war materializes in 1H20 and limited trade – mostly in ags and energy – is forthcoming, demand for safe-haven assets could weaken gold prices at the margin. Fiscal and monetary stimulus globally also could revive economic growth and commodity demand, pushing global yields higher, which would put negative pressure on gold at the margin, as well, given the high correlation between real rates and gold prices. Chart 2AU.S., Euro Economic Uncertainty Correlated With Gold Prices Chart 2BU.S., Euro Economic Uncertainty Correlated With Gold Prices Highlights · Energy: Overweight. Saudi Arabia and Kuwait are on the verge of signing an historic pact to restart production from the Neutral Zone. Kuwait expects to sign the pact within 30 to 45 days. Potential production from the jointly operated fields – Khafji and Wafra – is estimated at ~ 500k b/d. Ramping up production at the Wafra field could take up to 6 months. Importantly, both countries are expected to respect their production quota mandated under the OPEC 2.0 agreement expiring in 1Q20.2 Separately, Chevron’s waiver to operate in Venezuela was extended for three months from the Trump administration this week. · Base Metals: Neutral. Chile copper production was up 1% and 11% y/y in July and August, according to the World Bureau of Metal Statistics. Earlier this week, the Union of workers at Chile’s Escondida copper mine – the world’s largest – held a strike in support of broader protests sparked by the increase of metro fare last Friday. Chile’s President suspended the fare hike on Saturday, but the protests are still ongoing and have now caused 15 deaths.3 · Precious Metals: Neutral. The gold/silver ratio fell 9% since July 2019. Our tactical long spot silver recommendation is up 3% since inception in August 2019, and our strategic long gold position is up 21%. Cyclically, we remain positive on both silver and gold prices, more on this below. A tactical pullback is possible; money managers have started liquidating some of their long gold positions, dropping by 67k contracts from September levels, according to CFTC data. · Ags/Softs: Underweight. According to USDA data, corn and soybean harvest are 30% and 46% complete, lagging behind their respective 47% and 64% five-year average pace. For corn, the USDA rates 54% of the U.S. crop good or excellent, vs. 66% a year earlier. For beans, 56% of the crop is rated good or excellent, vs. 68% last year. Separately, China announced waivers allowing up to 10mm MT of U.S. soybeans to be imported by domestic and international crushing concerns. The waivers are in place until March 2020. Feature The once-reliable negative correlation between gold and the USD will remain muted over the short-term tactical horizon – 3 to 6 months – as economic policy uncertainty continues to stoke global demand for safe havens.4 The once-reliable negative correlation between gold and the USD will remain muted over the short-term. This can be seen in the elevated correlations between the USD’s broad trade-weighted goods index with the Baker-Bloom-Davis (BBD) Economic Policy Uncertainty (EPU) indexes for the U.S. and Europe (Chart 3).5 Rising economic uncertainty – particularly since 4Q18 – has created a rare environment in which both the USD and gold trended up simultaneously and continue to move in the same direction. The implication of this is that gold’s correlation with both the USD and EPU is weaker than before because economic policy uncertainty now is positively correlated with the dollar. Chart 3Strong USD, EPU Correlation Chart 4Correlation of Daily Gold, USD Returns Also Moving Sharply Higher There is a possibility global policy uncertainty could be reduced later this year if the U.S. and China can agree on a trade ceasefire... The typically negative correlation between daily returns of gold and the USD also is weakening, moving toward positive territory (Chart 4), as both the USD and gold trend higher simultaneously (Chart 5).   Chart 5Gold and USD Levels Trending Higher ...If this occurs, the risk premium supporting gold will ease, and markets will once again turn their attention to possible inflationary consequences of the global stimulus. Our short-term technical indicator is signaling an overbought gold market (Chart 6), and our fair-value model indicates gold should be trading ~ $1,450/oz (Chart 7). The latter signal off our fair-value model is less concerning, given the demand for safe-haven assets like the USD and gold now dominates gold’s typical drivers. Chart 6Gold Technical Indicators Signal Overbought Market Chart 7High USD Correlation Throws Off Fair-Value Model However, to be on the safe side, we are placing a $1,450/oz stop-loss on our long-term gold position, which as of Tuesday’s close was up 21% since inception on May 14, 2017. This is a precautionary measure, which recognizes the possibility global policy uncertainty could be reduced later this year if the U.S. and China can agree on a trade ceasefire, and global fiscal and monetary policy are successful in reviving EM income growth, which would revive commodity demand generally, pushing up global bond yields. If this occurs, the risk premium supporting gold will ease, and markets will once again turn their attention to possible inflationary consequences of the global stimulus. During that period, the monetary and fiscal aggregates we track as explanatory variables for gold prices will reassert themselves as the dominant drivers of gold prices (see below). This could produce tension between a falling USD and rising real rates as growth picks up, which would send us to a risk-neutral setting re gold, given the current high correlation between gold and real rates, which should remain strong until the Fed starts hiking rates again, most likely in 2020 (Chart 8). This is part of the reason we are including the stop-loss at $1,450/oz for our existing gold position: During this risky period going into 1H20 economic uncertainty could dissipate, and real rates could rise. Although the USD depreciation would mute these effects, rising real rates would be a risk to gold prices Chart 8Rising Real Rates Could Weaken Gold Prices Economic Uncertainty Dominates Gold’s Fundamentals At present, economic policy uncertainty overwhelms the other factors we typically use as explanatory variables when modeling gold prices. In Table 1, we collect the variables we consider when assessing gold’s fair value. At present, economic policy uncertainty overwhelms the other factors we typically use as explanatory variables when modeling gold prices. This variable broadly falls in the geopolitical risk we regularly account for in our analysis of gold markets. Table 1Fundamental And Technical Gold-Price Drivers If the uncertainty captured by the EPU indexes is resolved, we would expect the dollar to fall and the negative gold-USD correlation to reassert itself and strengthen. Checking off each of these groups, we see: · Demand for inflation hedges remaining muted over the short-term, as inflationary pressures remain weak. In line with our House view, however, we do expect inflation could move higher toward the end of next year and overshoot the Fed’s 2% target for the U.S. This would support gold prices. · Monetary and financial aggregates are working less well as explanatory variables for gold prices in a market dominated by economic policy uncertainty. The USD-gold correlation continues to be disrupted by strong demand for safe-haven assets. As inflation picks up next year, we expect nominal bond yields to rise. Real rates, however, could remain subdued, as long as the Fed is not aggressively raising rates to get out ahead of a possible revival of inflation (Chart 9). Later in 2020, the correlation between rates and gold should be supportive for gold prices – the correlation fades when the Fed tightens, which creates a demand for safe-haven assets like gold. All the same, an increase in real rates would be a risk to gold prices in 1H20. · At present, demand for portfolio-diversification assets via safe-haven assets is a powerful force in gold’s price evolution. It is worthwhile pointing out, however, that if global economic uncertainty is resolved and global growth does rebound, recession fears will diminish, thus reducing the marginal impact of geopolitical shocks. On the other hand, if the uncertainty captured by the EPU indexes is resolved, we would expect the dollar to fall and the negative gold-USD correlation to reassert itself and strengthen. Should that happen, short-term volatility in gold will rise (Chart 10). Chart 9Bond Yields Should Rise As Inflation Revives In 2H20 Chart 10Investors Expect Large Positive Moves In Gold And Silver Prices Investment Implications As India’s and China’s economic growth picks up, we expect income to grow, which would support physical gold demand in EM countries. Over a tactical horizon – i.e., 3 to 6 months – we expect global economic policy uncertainty to remain elevated. Going into 2020 – and particularly in 2H20 – we expect the USD to weaken on the back of global monetary accommodation policies and increased fiscal stimulus. We also are expecting a ceasefire in the Sino-U.S. trade war, which will revive trade somewhat and support EM income growth and commodity demand. These assumptions, which we’ve laid out in previous research, will be bullish cyclical factors supporting commodities generally. Bottom Line: A ceasefire in the Sino-U.S. trade war, coupled with global fiscal and monetary stimulus, will reduce some of the economic uncertainty dogging aggregate demand. This should be apparent in the data in 1H20. As a result, we continue to expect rising EM income growth to be cyclically bullish for commodities generally. This will allow inflation to revive – again, assuming the Fed does not become aggressive in raising rates. Chart 11EM Income Growth Will Support Demand For Gold Net, this will be bullish for gold: As India’s and China’s economic growth picks up, we expect income to grow, which would support physical gold demand in EM countries (Chart 11).   Robert P. Ryan Chief Commodity & Energy Strategist rryan@bcaresearch.com Hugo Bélanger Senior Analyst Commodity & Energy Strategy HugoB@bcaresearch.com   Footnotes 1               Please see our report entitled “Policy Uncertainty Lifts USD, Stifles Global Oil Demand Growth,” published October 17, 2019. It is available at ces.bcaresearch.com. 2              Please see “Kuwait Sees Neutral Zone Oil Pact With Saudis Within 45 Days,” published by Bloomberg.com on October 19, 2019. 3              Please see “Chile lawmakers call for social reforms as protests mount,”  published by reuters.com on October 22, 2019. 4              We expect a ceasefire in the Sino-US trade war to be announced in 1H20, which will defuse – but not eliminate – an important risk for global growth in our analytical framework.  We expect this will allow the relationship between the USD and gold to move back to its previous equilibrium in 1Q20 or 2Q20. 5              For more info on the Baker-Bloom-Davis index, please see policyuncertainty.com   Investment Views and Themes Recommendations Strategic Recommendations Tactical Trades TRADE RECOMMENDATION PERFORMANCE IN 2019 Q3 Commodity Prices and Plays Reference Table Trades Closed in 2019 Summary Of Trades Closed In 2018 Summary Of Trades Closed In 2017 Summary Of Trades Closed In 2016

While our bullish take on refiners got to a slippery start, it has recovered all the losses and this position is now in the black. Factors are falling into place for additional gains in the coming months and we recommend investors stick with this overweight recommendation in pure-play downstream stocks. Domestic refined product consumption remains upbeat and should serve as a catalyst for further relative share prices gains (top panel). Refining product supply dynamics are also moving in the right direction. Gasoline inventories are getting whittled down and should boost beaten down refining relative profit expectations (inventories shown inverted, bottom panel). This firming demand/supply backdrop has already been a boon to refining margins and should continue to underpin relative share price momentum (middle panel). Bottom Line:  Stay overweight the S&P oil & gas refining & marking index. The ticker symbols for the stocks in this index are: BLBG – S5OILR – MPC, VLO, PSX, HFC. Please refer to the most recent Weekly Report1 for more details. Footnotes 1 Please See U.S. Equity Strategy, "Is This It?," dated October 21, 2019.

Bank Credit Analyst

Highlights Given that rising crop yields have been the main vehicle through which global supply of agricultural commodities grew to meet expanding demand, the risks posed to yields due to climate change are non-trivial. The impact of climate change will manifest itself in the form of two simultaneous trends: the gradual rise in temperatures alongside more frequent and severe weather events. While the latter will threaten immediate supply, the former is a slower moving process, and its net negative impact is unlikely to manifest before 2030. The implications of climate change on agriculture producers are non-uniform. Low-latitude countries with economies that are highly dependent on the agriculture sector will suffer most. Expect greater volatility in agriculture prices as the frequency of weather events will raise uncertainty. Feature The steady expansion of global population and rising per-capita calorie consumption has directly translated to growing demand for agricultural products of all types. However, these demand-side pressures increasingly will be met with disruptions to global supply of agricultural commodities, as the impact of climate change raises uncertainty. In any given year, the aggregate decisions of farmers all over the world – i.e., the choice of which crops to plant and how much acreage to dedicate to each crop – determine the supply and market prices of ags. In this competitive market, each farmer attempts to maximize his or her welfare by planting the crops that are expected to yield the greatest profit. Chart 12010/11 Shock Highlights Ag Vulnerability To Weather The collective action of these producers in reaction to perceived demand generally leads to stable prices, especially for staple commodities such as grains and oilseeds, which differ from industrial commodities in that they are not highly correlated with global business cycles. Demand trends are long-term and slow moving, and typically do not result in abrupt price pressures, as farmers have time to adjust and adapt to changing consumer preferences. Unforeseen, weather-induced supply-side shocks, therefore, are the main source of sudden price changes in ag markets. Such a shock was dramatically on display during the drought-induced crop failures in major grain and cereal producing regions in the most recent global food crisis of 2010/11. While this massive supply shock was not the first of its kind (Chart 1, on page 1), it highlighted the vulnerability of ag markets to weather risks and specifically the evolving environment under climate change. A 2019 study quantifies the impact of shifting weather patterns on the agricultural market, finding that year-to-year changes in climate factors during the growing season explain 20%-49% of change in corn, rice, soybean, and wheat yields, with climate extremes accounting for 18%-43% of this variation.1 In theory, the impact can manifest in several ways, sometimes contradictory: Extreme weather events: An increase in the frequency and intensity of droughts or floods which threaten to wipe out crops or reduce yields, creating unpredictable supply shocks. The gradual rise in temperature: Each crop has cardinal temperatures – defined by the minimum, maximum and optimum – that determine its boundaries for growth. Increases in temperatures induced by global warming may push the boundary, reducing yields in some regions. Changes in precipitation patterns: In many areas precipitation is projected to increase – both in short bursts and over longer periods. This will lead to greater soil erosion resulting in deterioration in the quality of soil. In other regions, precipitation will decrease, and drought is expected to become more frequent.2 Moreover, the interaction of these factors – along with other region-specific variables – will amplify the impact on crops: Rising temperatures and greater precipitation will result in greater amounts of water in the atmosphere, producing increased water vapor and greater cloud cover. This will reduce solar radiation, and will harm crop productivity. Elevated atmospheric carbon dioxide and CO2 fertilization: Greater CO2 concentrations brought on by continued growth in air pollution are positive for crops as they stimulate photosynthesis and plant growth. However, the impact differs across crops with plants such as soybeans, rice and wheat set to benefit relatively more than plants such as corn.3 Moreover, elevated atmospheric CO2 levels can help crops respond to environmental stresses and reduce yield losses due to ozone and crop water loss through partial stomatal closure and a reduction in ozone penetration into leaves. Temperature changes and the magnitude and intensity of precipitation impact soil moisture and surface runoff. Indirect effects of climate change – weeds, pests and pathogens – also present challenges as they require changes to management practices and may raise farming costs required. The impact of climate change on agriculture markets is already evident in increasing intensity and frequency of extreme weather events. The confluence of these factors, and the region- and crop-specific nature of these variables, makes it impossible to estimate the impact of evolving climate conditions on ag products with great accuracy. Nevertheless, our research suggests that the impact of climate change on ag markets will create opportunities in this evolving and highly uncertain market. Abrupt Shocks Amid Gradual Warming: The Long And Short View The impact of climate change on agriculture markets is already evident in the increasing intensity and frequency of extreme-weather events such as heatwaves, floods, and droughts. Charts 2A, 2B, and 2C, illustrate the impact of major weather events in crop-producing regions of the U.S. on yields, production and acreage for the crop year in which the events took place. Chart 2AExtreme Weather Events Reduce U.S. Corn Supplies … Chart 2B… Soybean Supplies … Chart 2C… And Wheat Supplies In A Big Way Chart 3Climate-Induced U.S. Supply Shocks Associated With Price Spikes   While the individual losses are a function of the magnitude of the event, the events highlighted translate to a 16%, 10%, and 7% decline in corn, soybean, and wheat yields, respectively. These supply disruptions generally do not extend beyond the event year, as the new crop year offers farmers a clean slate to raise output and maximize profits. Given that the U.S. is a major global supplier of these crops, extreme weather events and the subsequent supply reductions lead to non-negligible price pressures (Chart 3). While crop conditions thus far have failed to deteriorate in trend (Chart 4), greater frequency and intensity of weather events raise the probability of a decline in overall crop and could lower supply.   Chart 4Crop Conditions Have Generally Held Up Expanding the analysis to other major crop-producing regions of the world, we find that once again, extreme-weather events are associated with a decline in yields and production in the corresponding crop year (Chart 5). This exercise also indicates that the impact of droughts is significantly more pronounced than the impact of floods.4 While the weather-induced supply shocks described above are unpredictable, abrupt, and have an immediate impact on output and prices, the gradual warming of temperatures is a slow-moving process. Consequently, the impact will manifest in the form of gradual changes that are difficult to capture and quantify, especially given the mitigating effect of CO2 fertilization – i.e., higher yields resulting from higher CO2 in the atmosphere. Nonetheless, rising temperatures will become a serious risk in crop-planting regions both in the U.S. and globally (Chart 6). While rising temperatures are expected to bring about increasingly more wide-ranging supply disruptions (Chart 7), the net impact over the coming decade is not a clear negative. Chart 5Weather Events, Especially Droughts, Hurt Global Supplies Chart 6Rising Global Temperatures Will Pose A Serious Risk … Chart 7… Especially Above The 2°C Mark One study expects the positive impact of CO2 fertilization on yields to overwhelm the negative effect of rising temperatures over the coming decade (Table 1). Elsewhere, studies forecast different responses, with some predicting incremental yield gains over the coming decade before temperatures rise to levels that overwhelm the benefits of greater CO2. Similarly, according to the FAO’s assessment, the net negative impact of climate change on global crop yields will only become apparent with a high degree of certainty post-2030.5 Table 1Estimates For The Response Of Global Average Crop Yields To Warming And CO2 Changes Over The Next Decades Bottom Line: Given that rising crop yields have been the main vehicle through which global ag supply grew to meet expanding demand, the risks posed to yields due to climate change are non-trivial. Supply disruptions generally do not extend beyond the event year, as the new crop year offers farmers a clean slate to raise output and maximize profits. The impact will manifest itself in the form of two simultaneous trends: the gradual rise in temperatures alongside more frequent and severe weather events. While the latter will threaten immediate supply, the former is a slower moving process, and its net negative impact is unlikely to manifest before 2030. The Winners … And Losers Rising temperatures are expected to result in a negligible impact on ag markets over the coming decade; yet this finding is not uniform across all regions. The FAO study cited above finds that by 2030, the projected impact on crop yields will be slightly net negative in developing countries. However, in developed countries, the effect will be net positive. In terms of global supply, the impact of climate change over the coming decade is expected to remain relatively contained, affecting certain regions at various times without causing major global disruptions. That said, as global warming and extreme weather persist, the ramifications will begin to extend beyond individual regions, and will cause supply shocks on a global scale. In part, this can be explained by a greater potential for net reductions in crop yields in warmer, low-latitude areas and semi-arid regions of the world.6 This non-uniform impact will create relative winners and losers. Producers located in temperate regions – where climate change does not yet pose as serious a threat – are set to profit from their increased role in global supply. Conversely, tropical regions are much more vulnerable to climate change. This is especially true for those whose economies are highly dependent on agriculture (Chart 8). Chart 8Agricultural Economies In Tropical Regions Are Most Vulnerable On net, the overall economies of DM countries – which generally are not economically dependent on agriculture and are located in northern regions – will be relatively more insulated from the impact of climate change on the agriculture sector. Aside from the impact on producers, the implications on consumers are also region-dependent. Clearly the direct impact of climate change on global agriculture will be higher food prices, which directly impacts the food component of inflation generally. As a result, consumers who spend a large share of their income to food – generally consumers in lower income countries – will be hardest hit (Chart 9). Chart 9Higher Food Prices Disproportionately Hurt Consumers In Lower Income Countries In theory, a food supply shock is transitory, and given that food is usually excluded from core inflation gauges targeted by central banks, monetary policy should not react to these price spikes. All the same, aside from this direct impact on inflation, food inflation can also pass-through into other components of the CPI basket, for example through wage pressures or inflation expectations. This would lead to a more persistent impact on core inflation, forcing policy makers to react to these transitory forces, complicating the monetary policy response function for these countries. Given that inflation expectations are less well-anchored in lower income economies and that food makes up a larger share of consumption expenditures in these economies, they are most vulnerable to weather-induced food shocks. Chart 10Subsidies Partially Insulate Against International Shocks In countries where food prices are highly subsidized, the impact of higher global food prices will not immediately translate to higher domestic prices. This explains why there is no one-to-one relationship between global food prices and domestic food prices (Chart 10). Instead, the higher prices are absorbed by the governments, resulting in an expansion in government expenditures. This distorts the local food market, as it prevents demand from adjusting to the higher prices, and could potentially result in an undershoot in inventories that makes global markets even more vulnerable to further supply shocks. Bottom Line: The implications of climate change on ag producers are non-uniform. While higher-latitude regions are set to benefit, at least in the short-run, low-latitude countries with economies that are highly dependent on the agriculture sector will suffer most. On the consumer side, individuals who spend a large share of their income on food are set to suffer most. While consumers in countries that subsidize the crops will be protected from the immediate inflation risk, they may feel a delayed impact due to an increase in budget expenditures needed to cover the larger import bill. Mitigation Efforts While the potential impact of climate change on the agriculture sector can be large, it will be at least partially managed through adoption of mitigation policies (Diagram 1). Diagram 1Adaptation Reduces Vulnerability A key question in determining the extent of this behavior is whether warming temperatures and the increased occurrence and intensity of extreme events will be sufficient to justify a major acceleration of investment in agriculture. These efforts would range from simple management changes on the part of farmers to technological advances that raise the productivity of farming or reduce the vulnerability of farmers to climate change. For example, farmers across the U.S. have been planting corn and soybeans earlier in the spring, resulting in an advancement in planting dates (Chart 11). The earlier planting has also been accompanied by a longer growing season with the average number of days in the season increasing. Farmers are also adapting by altering their decisions on which crops to plant. For example, since soybean and corn are planted in many of the same regions of the U.S., farmers often plant more soybeans than corn when experiencing weather shocks. Chart 11Weather Events, Especially Droughts, Hurt Global Supplies The agriculture sector is also using more efficient machinery that can plant and harvest crops much faster as well as developing heartier seeds and more potent fertilizers. In turn, farmers will alter their decision making by selecting crop varieties or species that are more resistant to heat and drought. Or they will change fertilizer rates, amounts and timing of irrigation, along with other water-management techniques. Farmers also are making wider use of integrated pest and pathogen management techniques, in order to raise the effectiveness of pest, disease, and weed control. Given that the number of firms in the agriculture sector are fewer in developed markets than in the rest of the world, management decisions can be more easily implemented in the former. Farmers across the U.S. have been planting corn and soybeans earlier in the spring, resulting in an advancement in planting dates. On the other hand, emerging market countries where ag output is driven by numerous individual farmers will have a more difficult time implementing policies. Individual farms may not have the means to support themselves, which raises the potential impact of climate change. What is more, climate-change mitigation efforts may require projects, programs, or funds set aside by the government to support these efforts. This is more likely to occur in wealthier developed countries. Bottom Line: Adaptation and mitigation measures on the part of farmers have the potential to reduce the impact of climate change. That said, farmers in richer countries with the funds and institutions in place to support the ag sector likely will fare better. Investment Implications Over the coming decade, the ramifications of climate change are likely to be contained to a regional level. Although global supply will be vulnerable to regional disruptions, the impact will, in part, be mitigated by inventories, which have been rising for years. These stocks will create a buffer against unpredictable supply shocks (Chart 12). Chart 12Higher Inventories Needed To Buffer Against Unpredictable Shocks However, given that the global soybean market resembles an oligopoly with Brazil, the U.S., and Argentina accounting for 81% of global supply, global soybean prices will be more vulnerable to supply events in these regions than other crops (Chart 13). Chart 13Soybeans Most Vulnerable To Shocks Affecting Major Producers At the other end of the spectrum, global wheat markets will be relatively more insulated from isolated weather events impacting any one major producer as each of these regions contributes a relatively small share to global wheat output. This analysis also finds that yields and supply generally recover in the crop year following an extreme climate event. This implies that while the extent of damage from these events can be severe, they are not persistent unless the increasing frequency of extreme events leads to a secular change. Aside from the price impact, the weather and temperature changes will manifest in the form of greater volatility in supply, translating to greater price volatility. Options-implied volatilities for corn, wheat and soybeans have been on a general downtrend since the two major global food scares in 2007/08 and 2010/11 (Chart 14). We expect the trend to reverse going forward as the frequency of weather events will create greater price uncertainty. We summarize the findings of this report in Table 3 (Appendix, on page 16). Chart 14Volatility Will Go Up Roukaya Ibrahim Editor/Strategist RoukayaI@bcaresearch.com Jeremie Peloso Research Analyst JeremieP@bcaresearch.com Amr Hanafy Research Associate AmrH@bcaresearch.com Hugo Bélanger Senior Analyst HugoB@bcaresearch.com Isabelle Dimyadi Research Associate Isabelled@bcaresearch.com Appendix Table 2Extreme Weather Events In The U.S. Table 3Summary Table Footnotes 1 Please see Vogel et al, The effects of climate extremes on global agricultural yields, Environ. Res. Lett 14 054010, 2019. 2 As a consequence of greenhouse gas emissions precipitation is expected to increase in high altitude regions such as much of the U.S. and decrease in subtropical regions such as the southwest U.S., Central America, southern Africa, and the Mediterranean basin. 3 Plants can be broken down into either C3 or C4 based on the way they assimilate atmospheric CO2 into different physiological components. While rising CO2 causes C3 plants to raise the rate of photosynthesis and reduce the respiration rate, C4 plants do not experience a rise in photosynthesis since  photosynthesis is already saturated. For example, studies show that soybean yields increased 12%-15% under 550 ppm vs. 370 ppm CO2 concentrations while corn experienced negligible yield increases. 4 Please see Lesk C., P. Rowhani, and N. Ramankutty, Influence of extreme weather disasters on global crop production, Nature, 529(7584), 84-87, 2016. 5 Please see The State Of Food And Agriculture: Climate Change, Agriculture, And Food Security, Food and Agriculture Organization of the United Nations, 2016. 6 Please see Stevanovic et al., The impact of high-end climate change on agricultural welfare, Sci-Adv 2(8), 2016.

Highlights Portfolio Strategy Soft housing demand, the trough in interest rates, new home price deflation and weak industry employment prospects suggest that an underweight stance is now warranted in the S&P homebuilding index.      Firming demand/supply dynamics, IMO Sulfur 2020 regulations, and bombed out relative profit expectations all signal that further gains are in store for pure-play refining equities. Recent Changes Downgrade the S&P homebuilding index to underweight, today. Table 1 Feature Equities made a run for fresh all-time highs last week, continuing to cheer the trade war “phase one” deal and breathing a big sigh of relief on better-than-expected bank earnings. We doubt a real deal will materialize which would include Intellectual Property and the tech sector. Instead all we got was a trade truce, at best. Larry Kudlow’s recent football analogy is worth repeating: “It's like being on the seven-yard line at a football game…And as a long suffering New York Giants fan, they could be on the seven and they never get the ball to the end zone…When you get down to the last 10 percent, seven-yard line, it's tough". As a reminder, steep tariffs remain in place and there are high odds that the damage already done to global trade is severe enough that it will be months before the emergence of any green shoots. Meanwhile, following up on our “chart of the year candidate” we published two weeks ago, we drilled deeper and discovered two additional economically sensitive indexes that have consistently peaked prior to the SPX in the past three cycles (Chart 1). They now comprise the U.S. Equity Strategy’s Equity Leading Indicator – an equally weighted composite of the S&P Banks index, the Russell 2000 index and the Value Line Geometric index – which signals that the easy money has already been made this cycle in the SPX (Chart 2). Chart 1Three Bulletproof Signals... Chart 2...Combined Into One Leading Equity Indicator Importantly, absent profit growth, it remains extremely difficult for equities to embark on a sustainable fresh leg up by solely relying on multiple expansion. Chart 3 shows our updated Corporate Pricing Power Indicator (CPPI) and it continues to deflate. In fact the steep fall in our CPPI more than offsets the fall in wage growth warning that the margin contraction in the S&P 500 has staying power1 (bottom panel, Chart 3). Drilling beneath the surface, our CPPI is waving a red flag. As a reminder, we calculate industry group pricing power from the relevant CPI, PPI, PCE and commodity growth rates for each of the 60 industry groups we track. Table 2 also highlights shorter term pricing power trends and each industry's spread to overall inflation. Only 42% of the industries we cover are lifting selling prices by more than 1%, and 33% are outright deflating. Worrisomely, only 26% of sectors are raising prices at a faster clip than overall inflation. With regard to pricing power trends, two thirds of the industries we cover are either flat or in a downtrend (Table 2). Chart 3Nil Corporate Pricing Power Table 2Industry Group Pricing Power Gold has jumped to the top of our table galloping at a 26%/annum rate (keep in mind it was deflating in our early July update), and only three additional commodity-related industries made it to the top twenty (Table 2). The disappearance of the commodity complex from the top ranks is consistent with global PPI ills and U.S. dollar strength. This week we update two groups, one early and one deep cyclical. Interestingly, defensive sectors have a healthy showing in the top ten spots with five entries. On the flip side, commodities in general and energy-related industries in particular occupy the bottom of the ranks as WTI crude oil is steeply deflating from the October 2018 peak. Adding it up, corporate sector selling price inflation is sinking in line with depressed inflation expectations. As we posited in our recent profit margin Special Report, profit margins have already peaked for the cycle. We reiterate our cautious overall equity market view on a cyclical 9-to-12 month time horizon. This week we update two groups, one early and one deep cyclical. Cracking Homebuilding Foundations We recommend downgrading the niche S&P homebuilding index to underweight, as most, if not all, positive profit drivers are already reflected in relative share prices. Specifically, the drop in interest rates has been more than accounted for by the year-to-date outperformance in homebuilders. Since the Great Recession, homebuilders have been in clearly defined mini up-and-down cycles, and there are high odds we will soon enter a down oscillation (bottom panel, Chart 4). Interest rates bottomed in early September and there is little additional push they can exert to relative share prices (10-year Treasury yield shown inverted, top panel, Chart 4). Chart 4Relative Gains Are Exhausted Worrisomely, consumers’ expectations to purchase a new home nosedived last month according to The Conference Board’s survey, and that demand softness will weigh on housing starts and ultimately homebuilding revenues (Chart 5). Chart 5Cracks Forming Adding insult to injury, new house selling prices are losing ground to existing home prices, but such discounting is no longer boosting volumes as new home sales market share gains have stalled recently. Already, S&P homebuilding sales are contracting and the risk is that deflation gets entrenched in this construction industry (Chart 6). While the mortgage application purchase index (MAPI) has been rising on the back of the plunge in interest rates, the 30bps rise in the 10-year Treasury yield since September 1 signals that the MAPI has tentatively crested (second panel, Chart 7). Chart 6Contracting Sales Chart 7Margin Trouble Simultaneously, lumber prices are gaining steam and coupled with contracting new home prices signal that homebuilding profits will suffer a setback (middle & fourth panels, Chart 7). This stands in marked contrast to the sell-side community that has been ratcheting up profit estimates for the S&P homebuilding index (bottom panel, Chart 7). Netting it all out, soft housing demand, the trough in interest rates, deflating new home prices and weakening industry employment prospects suggest that an underweight stance is now warranted in the S&P homebuilding index. On the operating front, the labor market is also emitting a distress signal. Job openings in the construction industry are sinking like a stone and residential construction employment growth is flirting with the contraction zone. Historically, the ebbs and flows in construction jobs have moved in lockstep with relative share price performance and the current message is to expect a drawdown in the latter (Chart 8). Most of the indicators we track underscore a challenging homebuilding backdrop in the coming months. However, there is a key risk to our view: interest rates. Were the 30-year fixed mortgage rate to fall further from current levels, it would entice first time home buyers and cushion the blow to homebuilding demand (mortgage rates shown inverted, top panel, Chart 9). Similarly, bankers are willing extenders of mortgage credit and are reporting rising demand for residential real estate loans as a lagged consequence of falling rates. But, our sense is that the easy gains are exhausted and a reversal is in the offing in most of these measures (Chart 9). Chart 8Heed The Labor Market's Message Chart 9Potentially Lower Rates Are A Key Risk Netting it all out, soft housing demand, the trough in interest rates, deflating new home prices and weakening industry employment prospects suggest that an underweight stance is now warranted in the S&P homebuilding index. Bottom Line: Downgrade the S&P homebuilding index to underweight, today. The ticker symbols for the stocks in this index are: BLBG – S5HOME – DHI, LEN, PHM, NVR. Stick With Refiners While our bullish take on refiners got to a slippery start, it has recovered all the losses and this position is now in the black. Factors are falling into place for additional gains in the coming months and we recommend investors stick with this overweight recommendation in pure-play downstream stocks. Encouragingly, refining stocks have been trouncing the overall energy index of late and have resumed their multi-year relative uptrend (top panel, Chart 10). With regard to the export relief valve, U.S. net exports of refined products are on a secular uptrend and surprisingly unaffected by the greenback’s moves (bottom panel, Chart 10). Tack on the soon to be adopted International Maritime Organization (IMO) Sulfur 2020 regulations in maritime transportation fuel, and U.S. refiners that produce lower-sulfur fuel oil are well positioned to outearn the SPX. Chart 10Resumed Uptrend Domestic refined product consumption remains upbeat and should serve as a catalyst to unlock excellent value in this niche energy subgroup (middle panel, Chart 11). In fact, gasoline consumption is expanding anew on the back of rising vehicle miles travelled (bottom panel, Chart 11). Chart 11Solid Demand... Refining product supply dynamics are also moving in the right direction. Gasoline inventories are getting whittled down and should boost beaten down refining relative profit expectations (inventories shown inverted, bottom panel, Chart 12). Importantly, this firming demand/supply backdrop has been a boon to refining margins and should continue to underpin relative share price momentum (middle panel, Chart 12). In terms of what is baked in the cake for this industry, the expected profit growth bar is extremely low and falling and relative value has been fully restored. First in terms of relative valuations, the relative trailing price-to-sales ratio has corrected 35% from the mid-2018 peak (middle panel, Chart 11). On a forward PE ratio basis refiners are extremely appealing compared with the SPX following a near halving in the relative forward PE in the past fifteen months (second panel, Chart 13). Chart 12...Supply Backdrop Is Boosting Crack Spreads  Chart 13Profit Hurdle Is Uncharacteristically Low Second, relative EPS growth has sunk below the zero line both twelve months and five years forward. Such pessimism is overdone and we would lean against sell-side bearishness (bottom panel, Chart 13). Even the refining industry’s net earnings revisions ratio has collapsed, which is contrarily positive (third panel, Chart 13). Adding it all up, firming demand/supply dynamics, IMO Sulfur 2020 regulations, and bombed out relative profit expectations all signal that further gains are in store for pure-play refining equities. Bottom Line: Stay overweight the S&P oil & gas refining & marking index. The ticker symbols for the stocks in this index are: BLBG – S5OILR – MPC, VLO, PSX, HFC.   Anastasios Avgeriou, U.S. Equity Strategist anastasios@bcaresearch.com   Footnotes 1      Please see BCA U.S. Equity Strategy Special Report, “Peak Margins” dated October 7, 2019, available at uses.bcaresearch.com. Current Recommendations Current Trades Size And Style Views Stay neutral cyclicals over defensives   (downgrade alert) Favor value over growth Favor large over small caps (Stop 10%)

Special Report

Today we are also publishing a Special Report titled Chinese Auto Demand: Time For A Recovery? Highlights India is the third-largest world consumer of crude oil. Hence, fluctuations in its oil consumption is a non-negligible factor behind global oil prices. India’s petroleum demand growth is slowing cyclically due to the domestic demand slump and a dramatic drop in vehicle sales. This, combined with China’s ongoing slowdown in petroleum product demand, will have a non-trivial impact on oil prices in the next six months. From a structural perspective, India’s long-term demand growth for petroleum is decelerating as well. Feature India’s petroleum products consumption growth is slowing. Chart 1India Is The World's Third Largest Crude Oil Consumer India is the world’s third-largest consumer of crude oil, guzzling 5% of global consumption (Chart 1). Hence, fluctuations in India’s crude oil/petroleum consumption is a non-negligible factor affecting global oil prices. India’s petroleum products consumption growth is slowing. This comes on top of China’s ongoing petroleum demand deceleration. Together, the two countries account for 19% of the world’s oil intake. Therefore, deceleration in their oil consumption growth will have a considerable impact on the outlook for global oil demand growth. A Pronounced Cyclical Oil Demand Slump Indian petroleum consumption growth has decelerated significantly on the back of slumps in Indian domestic spending and economic activity (Chart 2). Please click on this link for an in-depth analysis on the domestic demand slump in India. Chart 2Indian Petroleum Consumption Growth Has Been Dwindling Specifically, vehicle purchases and industrial sectors have been hit hard. These sectors are critical for Indian petroleum consumption, since transportation demand accounts for 50% and industrial activity for around 25% of total petroleum consumption (Chart 3). Indian vehicle sales have been in freefall. Chart 3Transportation & Industry Guzzle The Most Fuel In India Chart 4Indian Vehicle Sales Are In Deep Contraction Indian vehicle sales have been in freefall. Chart 4 shows passenger car sales are shrinking at 30% and sales of two and three-wheeler units are contracting at 20% from a year ago. Moreover, commercial vehicles and tractor unit sales are falling at annual rates of 35% and 10%, respectively. Chart 5 illustrates that the number of registered vehicles is expanding at a lower rate than before – i.e., its second derivative has turned negative. This signals a further growth slowdown in gasoline and diesel consumption. We use the second derivative in this analysis because registered vehicles are a stock variable. However, we are trying to explain changes in petroleum consumption which is a flow variable. Therefore, the second derivative of a stock variable (the number of registered cars on the road) explains the first derivative of a flow variable (the growth rate of oil consumption). Looking ahead, vehicle sales will remain in the doldrums because of a lack of financing. In particular, the impulse on auto loans issued by commercial banks is negative (Chart 6). Chart 5Slowing Growth Of Vehicles On The Road = Weaker Pace Of Fuel Consumption Chart 6Indian Banks: Negative Vehicle Loan Impulse More worrisome is the ongoing turmoil in India’s non-bank finance sector (NBFCs), which has also significantly hit auto sales. In the past, the NBFC sector played a major role in funding Indian auto purchases. For instance, according to the ICRA, an independent rating agency in India, NBFCs have helped fund the purchases of 65% of two-wheelers, 30% of passenger cars and around 55% of commercial vehicles – both new and used. Given these non-bank finance companies are currently facing formidable funding and liquidity pressures amid rising NPLs (Chart 7), they are being forced to shrink their balance sheets. This is damaging to auto sales. Please click here for an in-depth analysis on the Indian banking and non-bank finance sectors. Chart 7Major Asset-Liability Mismatches Among Indian Non-Bank Finance Sector Chart 8India's Capex Has Been Weak Turning to the industrial sector, overall Indian capital spending has been weak. India’s real gross fixed capital formation has rolled over, the number of capex projects underway is nosediving and both capital goods imports and production are contracting by 7% and 12% on an annual basis (Chart 8). Falling industrial activity has taken a toll on the consumption growth of petroleum products with industrial applications, such as bitumen, naphtha and petroleum coke, etc. The growth rate in demand for these products is dropping — a significant development since they account for 25% of overall petroleum consumption in India.1  Bottom Line: India’s petroleum consumption growth has been slowing drastically from a cyclical perspective. And Moderating Structural Oil Demand Growth It appears there are structural factors at play that will also reduce India’s long-term demand for petroleum. On top of the cyclical demand slowdown, it appears there are structural factors at play that will also reduce India’s long-term demand for petroleum: Chart 9Impressive Efficiency Gains In India's Vehicle Fleet The fuel efficiency of India’s vehicle fleet is markedly improving (Chart 9). Additionally, since 2015-16 the Indian government has been proactively pursuing new emission/fuel efficiency standards. For instance, emissions standards for new passenger vehicles will fall to 4.2 L/100 KM by 2023 down from its current level of 4.6 L/100 KM. This will lead to a 7% reduction in auto fuel consumption. While this is not a large reduction, the government has the scope to implement even stricter standards since Indian car makers are easily meeting these targets. Finally, the Indian government has been aggressively promoting electric vehicles (EVs) as an alternative to traditional autos. It has made the advancement of this sector a priority. Ownership of EVs is currently negligible in India. However, the government is pushing for EVs to make up 30% of vehicle sales by 2030. In addition, it has been providing incentives such as sales tax cuts and subsidies to the sector. Finally, Mahindra and Tata Motors are already establishing a lead in the EV industry and are developing new EV models in collaboration with foreign automakers.  Bottom Line: The pace of India’s structural demand for petroleum will also be downshifting. Oil Inventory Not A Critical Factor Chart 10China: Oil Inventory Drives Oil Imports Inventory accumulation and destocking can play an important role in oil price fluctuations. For example, inventory accumulation plays a key role in driving Chinese crude oil imports (Chart 10). There is a dearth of data on Indian oil inventories to make a strong inference about its de- and re-stocking cycles. However, we have the following observations: India has the capacity to store 5.33 million tons worth of strategic oil reserves - equivalent to around 10 days of its crude oil consumption. It is not clear whether or not these reserves are at full capacity. However, even if we assume they are only 50% full and the government decides to fill its reserves all at once, this would require the importation of an additional 2.67 million tons of oil, equivalent to only 1.2% of Indian crude oil imports and 0.05% of global crude oil demand. This is a negligible amount, and is unlikely to have any impact on global oil prices. Furthermore, while the Indian government is planning to expand its storage capacity by an extra 6.5 million tons, this will only take place in the next six to eight years. Thus, it will not meaningfully affect oil imports in the medium term. Chart 11India: Oil Consumption Drives Oil Imports Finally, India’s crude oil imports are strongly correlated with its petroleum final consumption (Chart 11). Therefore, it is reasonable to assume that Indian consumption – not the oil inventory cycle – is relevant for crude imports, and by extension for oil prices. Bottom Line: India’s petroleum product and crude oil inventory fluctuations are too small to influence the nation’s crude imports and hence global oil prices. Investment Conclusions From a cyclical perspective, Indian final demand for crude oil has been weakening. A major re-acceleration in economic growth and hence oil demand is not imminent. We discuss the outlook for China’s auto sales in a separate report published today. Together India and China consume 19% of world oil, and therefore a deceleration in their oil consumption growth will have a non-trivial impact on the pace of global oil demand growth. Chart 12Expansion Pace Of Vehicles On The Road Has Downshifted In India & China Our estimations for annual growth in cars on the road (excluding 2-wheelers) has dropped to 5.8% in India and 10.5% in China (Chart 12). This entails a slower pace of oil demand growth than in the past. Besides, if one rightly assumes petroleum consumption per car is declining for structural reasons due to technological advancements by car manufacturers and enforcement of stricter efficiency standards by governments, oil consumption growth will be considerably slower going forward relative to the past 20 years. Together India and China consume 19% of world oil, and therefore a deceleration in their oil consumption growth will have a non-trivial impact on the pace of global oil demand growth. This presents a major risk for crude prices in the next 6 months or so. Beyond the cyclical horizon, the long-term demand outlook for oil is also downbeat. Please note that this is the view of BCA’s Emerging Markets Strategy team, and differs from that of BCA’s house view, which is bullish on oil. Chart 13India’s Relative Equities Performance Benefits From Lower Oil Prices In turn, low oil prices are positive for the relative performance of Indian stocks versus the EM equity benchmark (Chart 13). This was among the primary reasons why we upgraded the allocation to this bourse within an EM equity portfolio to neutral from underweight on September 26, 2019. In absolute terms, the outlook for Indian share prices remains downbeat, as discussed in the same report. Finally, to express our negative view on oil prices, we are reiterating our short oil and copper / long gold position recommended on July 11, 2019. Industrial commodities such as copper and oil will continue to underperform gold prices in the medium term (the next six months). Ayman Kawtharani, Editor/Strategist ayman@bcaresearch.com   Footnotes 1      Diesel consumption will also be impacted. While the latter is mostly consumed by the transportation sector in India, diesel does have some industrial applications as well. Equities Recommendations Currencies, Credit And Fixed-Income Recommendations

Pervasive global policy uncertainty continues to fuel USD safe-haven demand. This keeps the Fed’s broad trade-weighted dollar index for goods close to record highs, which continues to stifle oil demand. At present, we do not expect this pervasive uncertainty to dissipate. For this reason, we are lowering our oil-demand growth expectation slightly for this year and next. Our estimate of global supply growth is slightly lower for this year and next, as well; we continue to expect OPEC 2.0 to maintain production discipline and for capital markets to restrain U.S. shale-oil growth.1 Our price forecast for 4Q19 is $66/bbl on average, an estimate that includes a risk premium reflecting continued tension in the Persian Gulf. Our updated supply-demand balances for 2020 reduce our Brent price forecast to $70/bbl versus our earlier expectation of $74/bbl. We continue to expect WTI to trade $4.00/bbl below Brent next year. Highlights Energy: Overweight. The Trump administration likely will not renew Chevron’s waiver to operate in Venezuela when it expires October 25. This raises the likelihood the country’s oil output will fall below 300k b/d, down from the 650k b/d we currently estimate.2 Production could revive next year, if Russian or Chinese firms step in to fill the void. This is not certain, however, as the U.S. is pressing both to end their support for the Maduro regime. Separately, the Aramco IPO could occur as early as November, according to press reports. Base Metals: Neutral. Copper treatment and refining charges in Asia are staging a recovery, clocking in at $56.70/MT at the end of last week, according to Metal Bulletin’s Fastmarkets. The MB index fell to a record low of $49.20/MT in late August. Precious Metals: Neutral. Gold volatility remains elevated – standing at 15.1% p.a. on the COMEX – as markets continue to process news re a partial easing of tensions in the Sino-US trade war. Geopolitical tensions, which now encompass Turkey-US relations, remain elevated. Ags/Softs: Underweight. Uncertainty around a partial deal involving ag exports from the U.S. to China remains high, as negotiators deliberately minimize expectations of a successful outcome. The big sticking point appears to be whether U.S. tariffs on Chinese imports due to kick in in December will be removed. Feature Uncertainty arising from global economic policy risk continues to dominate commodity markets. This has been the case going on three years. While it is ubiquitous, it is difficult to isolate. In earlier research, we noted the tightening of global financial conditions – largely the result of the Fed’s rates normalization policy, which resulted in four rate hikes last year, and China’s deleveraging policy – were responsible for the sharp slowing of oil demand seen in 2H18-1H19.3 Recently concluded research allows us to extend our earlier thesis to account for the effect of pervasive global policy uncertainty over the past three years, which has dominated our analysis of commodity markets generally, oil in particular. To wit: We find a strong, positive correlation between uncertainty, as measured by the Baker-Bloom-Davis Global Economic Policy Uncertainty (GEPU) index, and the Fed's USD broad trade-weighted index for goods (TWIBG) from January 2017 to now (Chart of the Week).4 Chart of the WeekUSD Absorbs Global Policy Uncertainty USD Absorbs Global Uncertainty Sudden policy shifts have, over the past three years, resulted in a steady increase in the level of the GEPU index. Prior to 2017, the correlations between the GEPU index and the USD TWIBG were running at 33% and 63% for the periods 2000 to 2016 and 2010 to 2016, the post-GFC period for y/y returns. However, as right- and left-wing populism gained ground globally and monetary policy generally became more “data dependent” and ad hoc at the Fed, ECB and BoJ, the GEPU and USD TWIBG indices became highly correlated, surpassing 90% (Chart 2).5 This period saw the U.S. become more and more assertive vis-à-vis trade and foreign policy, particularly in re China, Iran and Venezuela, which caused those states to implement their own policy responses. In addition, as monetary policy generally became increasingly accommodative, central banks – and policy analysts – became less certain about the effects of their policies on the broader economy (e.g., the Fed shifting away from rates normalization, the ECB’s re-launching of QE, and the BoJ’s interest-rate targeting regime). Chart 2Co-Movement In GEPU, USD TWIBG Often, commodity markets were forced to adjust to sudden policy changes – e.g., the imposition of trade tariffs against China, or the granting of waivers to Iran’s eight largest importers in November 2018 just before oil-export sanctions were re-imposed. Sudden policy shifts have, over the past three years, resulted in a steady increase in the level of the GEPU index. Increasing uncertainty translated into a steadily increasing USD TWIBG, with safe-haven demand for dollars rising, as the Chart of the Week indicates. To date, we have not decomposed the drivers of monetary conditions, particularly in re central-bank accommodation versus global economic policy uncertainty on the evolution of the USD. The GEPU index hit a record high in August 2019, while the USD TWIBG hit a record in September 2019. It is possible the effects of general policy uncertainty could be cumulative – as earlier uncertainties remain unresolved and new ones are added to the global mix (e.g., US-Turkey foreign-policy tensions now have been added to other geopolitical risks). It is entirely possible global monetary policy easing – particularly from the Fed – is accommodating safe-haven demand accompanying higher uncertainty. If the Fed were to tighten while uncertainty remains elevated the USD could rally sharply and impact commodity demand even more. Persistent USD Strength Lowers Oil Price Forecast Based on our analysis, the effects of the uncertainty we observe in the USD above are transmitted to GDP globally, which feeds through to commodity demand. As the USD strengthens, it raises the local-currency cost of commodities and the cost of servicing USD-denominated debt ex-US. In addition, on the supply side, a stronger dollar lowers local production costs at the margin, which stokes deflation globally.  All else equal, these effects push oil prices lower by reducing demand and increasing supply at the margin. On the back of a stronger USD and persistent uncertainty, we are once again lowering our estimate of global demand growth. This is most pronounced in EM economies (Chart 3), but there are feedback effects into DM in the form of reduced trade volumes, which hits manufacturing economies like Germany harder than service-dominated economies like the US. On the back of a stronger USD and persistent uncertainty, we are once again lowering our estimate of global demand growth to 1.13mm b/d this year and 1.40mm b/d in 2020 (Chart 4). This is down slightly from 1.2mm b/d this year and 1.5mm b/d next year. In line with the U.S. EIA, we also lowered our estimate of 2018 demand, which has the effect reducing the level of demand we expect in 2019 and 2020. Chart 3Local-Currency Oil Costs Are High Chart 4BCA Research Supply-Demand Balances We maintain our expectation fiscal and monetary stimulus globally will revive demand, but, given the deleterious effects of global uncertainty and its effects on demand via the USD, we are moderating our position some, as the downward adjustment to consumption indicates. On the supply side, we expect KSA’s output to be fully restored by November, and for production in the Kingdom to average 9.9mm b/d in October and November. We are expecting overall OPEC 2.0 output growth of 250k b/d on average in the 2Q20 to 4Q20 interval, down from our previous growth estimate of 500k b/d. In the US, we expect shale-oil output to grow 900k b/d in 2020, versus 1.3mm b/d in 2019, which will leave overall U.S. crude output at 13.3mm b/d next year on average, as capital-market constraints continue to act as a governor on total output (Chart 5). Chart 5U.S. Shale-Oil Output Will Remain Capital-Constrained Overall, we expect global supply to finish 2019 at 100.8mm b/d and at 102.3mm b/d next year, which is down slightly from our earlier estimates (Table 1). Even with demand moderating, we expect inventories to continue to draw this year and into 3Q20 before they resume building, as the combination of OPEC 2.0 production discipline and capital markets constrain output (Chart 6). Chart 6OECD Oil Inventories On Track To Draw Table 1 Investment Implications Continued voluntary and involuntary production restraint will allow global inventories to draw despite slightly lower demand. Given our supply-demand expectations, we forecast Brent will trade lower next year, at $70/bbl on average versus our earlier expectation of $74/bbl. This is ~ $10/bbl above the median consensus. We continue to expect WTI to trade $4.00/bbl below Brent next year. Continued voluntary and involuntary production restraint will allow global inventories to draw despite slightly lower demand, which will keep Brent and WTI forward curves backwardated next year (WTI was in a slight carry earlier this week, while Brent was backwardated). We would caution that any resolution of the profound uncertainty currently dogging global markets could unleash pent-up demand that would sharply rally commodities generally, and oil in particular. This could take the form of a broad trade agreement that ends the Sino-US trade war – an unlikely, but not impossible,  turn of events – or an unexpected reduction in tensions in the Persian Gulf, again, unlikely but not impossible. Bottom Line: Resolution of global policy uncertainty would revive commodity demand, as safe-haven USD demand gives way to higher consumer spending, renewed growth in global trade and investment. Until then, uncertainty will continue to hamper commodity demand growth, particularly for oil.   Robert P. Ryan, Chief Commodity & Energy Strategist rryan@bcaresearch.com Hugo Bélanger, Senior Analyst Commodity & Energy Strategy HugoB@bcaresearch.com   Footnotes 1      OPEC 2.0 is the moniker we coined for the producer coalition formed at the end of 2016 to regain control of production following the disastrous market-share war launched by OPEC in 2014, which took Brent prices from above $100/bbl to $26/bbl by early 2016.  The coalition is led by the Kingdom of Saudi Arabia (KSA) and Russia. 2      Please see Venezuelan oil output could be halved without Chevron waiver extension: analysts, posted by S&P Global Platts October 14, 2019.  3      Please see our report entitle Central Bank Easing Key To Oil Prices, published September 5, 2019.  It is available at ces.bcaresearch.com. 4      This GEPU is a monthly GDP-weighted index of newspaper headlines containing a list of words related to three categories – “economy,” “policy” and “uncertainty.”  Newspapers from 20 countries representing almost 80% of global GDP (on an exchange-weighted basis) are scoured monthly to create the index.  Please see GEPU and Baker-Bloom-Davis for additional information. 5      Both series are plotted as percent changes y/y in Chart 2. For the 2017 - 2019 period, the coefficient of determination for this model is 0.81 using a regression of the USD on the GEPU.  There was no statistically significant relationship between them either from 2000 to 2016, or from 2010 to 2016.  Insert SOFTS text here Investment Views and Themes Recommendations Strategic Recommendations Tactical Trades TRADE RECOMMENDATION PERFORMANCE IN 2019 Q3 Commodity Prices and Plays Reference Table Trades Closed in 2019 Summary of Closed Trades