Equities

Special Report

When we first set out to create ETS, we leveraged the vast amount of research on stock market anomalies and started with factors that had a proven track-record. We also wanted to minimize the risk of data mining, and as a result, we decided to keep things simple. We stuck with established metrics, and avoided venturing down the long, windy path of custom metrics and adjustments: the more parameters that are included, the easier it is to find spurious results. We also opted for simple, linear mappings – lower is better, or higher is better. When deviating from this approach, it is easy to lose sight of the big picture and spend time optimizing minute details that will not have a large impact on the overall performance of the model. Fast-forward to 2019. We have a stable model that has performed well out of sample, and we are now comfortable taking a second pass over our factors. We can now analyze each model input in detail to see if we can make any improvements. The first factor under the microscope is the Altman Z-Score. This metric was first developed by Edward I. Altman in 1968 to estimate the probability of bankruptcy over a two-year horizon. It uses various items from a firm’s income statements and balance sheet and combines them using a set of weights to yield the so-called Altman Z-Score.1 Since this metric relies on estimated coefficients for a specific set of data, and given that these coefficients date back to the 60s, we wanted to investigate the possibility of using a more modern model of distress in place of the original Altman-Z score.  Our research process led us to review the academic literature, replicate the most promising results internally, and finally, to implement our own variation of the Altman-Z score given our unique dataset.   A Proxy For Bankruptcy In order to assess the ability of a distress model to predict insolvency, it is necessary to have a historical dataset of failure events. If not, it is impossible to measure the success of a given model. Failure can be defined in several ways, the most obvious of which is a Chapter 7 or Chapter 11 bankruptcy filing, at least in the U.S. To ensure that we captured the maximum possible failure events for our global set of stocks, we avoided the Chapter 7/11 definition and developed a simple market-cap based proxy for bankruptcy. We define the failure event of a firm as the point in time where the market-cap drops to 1% of its historical maximum. We also enforce that this event must occur after the maximum to prevent false failures at the beginning of a firm’s trading history. We found that this definition signals failure at the appropriate time for a selection of large-cap firms known to have experienced distress; a notable example is Enron Corp. (Chart 1). In general, we deem this a reasonable proxy for bankruptcy since the price of shares will drop substantially upon news of insolvency. Ultimately, even though it may, in some cases, flag firms that are technically not bankrupt, for our purposes, we are happy to avoid buying firms that plunge below 1% of their previous valuation peak. Academia To The Rescue? A review of the modern finance literature suggested that the Altman Z-Score, while methodologically sound, still had room for improvement as a distress measure. We anticipated that using a more refined distress model, we could amplify the effect of the existing Altman-Z Factor in ETS. In light of this, our first attempt at replacing the Altman-Z score involved running a logistic regression (logit) model presented in Campbell et al. (2010)2 on the ETS universe of stocks and constructing a new factor based on the output from this model. Following the authors’ methodology, we constructed a set of eight variables (three accounting ratios, and five market-based variables) and computed the probability of failure on our entire historical dataset. The probabilities were then percentile-ranked to construct a score from 0-100%, where firms with a score of 0% correspond to those with the highest probability of failure. This ranking system is consistent with the other ETS factors where lower scores are bad and higher scores are good. Although the model by Campbell et al. worked well for predicting subsequent failure (as judged by our bankruptcy proxy measure), the metric fell short performance-wise when tested within our universe of stocks. In brief, our decile-spread metric3 yielded a negative number, indicating poor separation of returns between deciles. Based on the above, we sought to construct our own distress model using similar methods, but applied to the universe of valid ETS firms. There are several advantages to this approach: First, it prevents errors in the construction of the explanatory variables. For instance, when replicating models found in academic literature, we cannot be sure that we are constructing our variables in the exact same way as reported, and hence we cannot have complete confidence in the accuracy of the computed probabilities of failure. Second, it frees us from the shackles of a complicated model, and we can seek to reduce our own model down to something more computationally tractable, and avoid overfitting the data.  Finally, it allows us to base our model on global stock data, as opposed to the U.S.-centric CRSP/COMPUSTAT dataset used by many academic institutions. Bringing It “In-House” The construction of our own distress model began with an analysis of the variables defined in Campbell et al. First, we wanted to determine which variables showed the strongest deviation between success and failure groups. From the calculation of summary statistics and histograms, the variables that showed the most prominent separations between success and failure groups were Net Income to Market Total Assets (NIMTA), Total Debt to Market Total Assets (TDMTA), monthly excess return (EXRET), and the 3-month standard deviation of daily returns (SIGMA) (Chart 2). Detailed descriptions of the variables are provided in the Appendix. We computed NIMTA, TDMTA, EXRET, and SIGMA at monthly frequencies with the goal of constructing a logit model that would predict failure 12 months into the future. The logit model implies that at time t, the probability of firm i failing in the next 12 months is given by:   where y represents the sequence of success (0) and failure (1) events for each firm at each point in time, x represents the four explanatory variables, and α, β are the parameters to fit. The parameters were fit on a rolling basis to avoid look-ahead bias when constructing a trading strategy based on the outcome of the model. Therefore, at each year-end, we used the trailing success/failure data up to that point to estimate parameters. The fitting yielded intuitive results, suggesting that higher TDMTA, SIGMA and lower NIMTA, EXRET indicate a higher probability of failure (see parameter list in Appendix Table 1). Another positive characteristic of these parameters is that they remain relatively stable throughout time, alluding to the robustness of the model. From visual inspection of the parameters, we see that EXRET has the strongest effect on failure probability, followed by SIGMA, TDMTA, and NIMTA. It is not surprising that EXRET has the strongest effect since it is based on a 12-month trailing weighted average. Therefore, we can expect this variable to capture the prolonged period of descent that firms typically undergo before reaching a bankruptcy event. Indeed, firms in our failure group experienced, on average, a failure event 5 years after reaching peak market cap. As for the remaining variables, it is natural to expect increased volatility and subpar financial statements to signal imminent failure. Model accuracy was tested by looking at its predictions on an individual and aggregate basis. Reassuringly, we found individual examples of large firms with known failure events showing the expected behavior (Chart 3). However, the more significant result is that on average, the model predicts a high probability of failure for the group of true failure firms (N=3213) relative to a random sample of true success firms (Chart 4). Armed with a successful distress model, we then translated the predictions into an ETS factor, giving it a score from 0-100%. Using a relative change condition in the parameters as a guide,4 we chose the factor inception date to be year-end 2004. Since our stock data begins in 1995, this allows for a sufficient “burn-in” time for the distress model to stabilize. The latter is important given the spike in U.S. Chapter 11 filings starting in 2001, which is also captured by our bankruptcy proxy (Chart 5).  With the factor constructed, we replaced the existing Altman-Z factor in the full model, keeping the same initial weight.  As expected, this replacement led to an improvement in the overall model according to our standard performance metrics, which are described in a previous report.5 Specifically, we observed a 6% improvement in the decile ordering metric (DECORD) and a small (<1%) reduction in mean absolute difference in BCA Score (MAD), leading to a 7% increase in our overall risk-reward ratio (DECORD / MAD). The new factor, simply named Distress on the platform, also performs well when examined in isolation. To test this, we ran the equivalent of an equal-weighted, daily-rebalanced backtest on ten deciles separated by the Distress score. Calculating the Sharpe ratio on the decile returns, we observe an increasing trend in the performance of the factor as a function of decile (Chart 6). Therefore, with this factor we obtain the desired separation in performance between deciles on a risk-adjusted basis.  What About Machine Learning? Disclaimer: This section involves a small digression on machine learning One statistical challenge with our success/failure classification problem is that the number of “success” events (i.e. no indication of bankruptcy) greatly outnumbers the failure events. Indeed, with our bankruptcy proxy, approximately 1% of the 3.5 million firm-month observations in our dataset are marked as failures. Despite the small failure rate, classic logistic regression can still provide good results; however, we wished to explore the effectiveness of machine learning techniques in tackling this type of classification problem. The use of machine learning (ML) was motivated by an analogous problem outside the world of stocks, namely the detection of credit card fraud. One can imagine that out of the massive number of credit card transactions registered daily, only a small fraction is actually fraudulent. Now, suppose that for a given period we obtained transaction data that includes characteristics such as dollar amount, time of day, and whether or not the transaction was normal or fraudulent. We could then use this prior data to train an ML model to recognize fraudulent transactions when presented with new, out-of-sample data. A specific class of ML models known as “autoencoders” have been shown to be effective in handling this task (See Box 1 for more details). Given that the credit-card problem is directly analogous to our distress problem, we set up a simple autoencoder to test whether it could detect firm failure with a higher degree of accuracy than the logit model. As inputs to the model, we used the same eight variables from Campbell et al. Perhaps surprisingly, we found that the best version of our autoencoder model did not surpass the accuracy of the logit model, at least in terms of maximizing precision and recall simultaneously.6 Furthermore, the predictions of the autoencoder were less effective at separating winners from losers after being transformed into a factor score between 0-100%. This, of course, does not mean that ML techniques are hopeless. It simply means that, for this particular set of parameters and model structure, the autoencoder was not able to provide additional predictive value over the simpler, more computationally efficient logit model.     Box 1 Autoencoders The rise of machine learning has brought with it a slew of jargon that seems intimidating at first, but can actually be understood without extensive training in mathematics or computer science. The term “autoencoder” may very well fall into this category. To get right into it, an autoencoder is a type of neural network tasked with learning a compressed representation of the original input or “training” data. The flow of data through an autoencoder is shown below (Box Diagram 1). The goal of the autoencoder is to optimize the weights connecting nodes in the network so that the difference between the input and output is minimized. However, due to the action of the “reduction” or “encoding” phase of the autoencoder, the output, i.e., the “reconstructed” or “decoded” data, will always be a compressed approximation of the original input. The latter is important for classification problems because it means that post-training, any data presented to the network that does not resemble the original training data will likely be poorly reconstructed, and flagged as anomalous.  Going back to our problem of identifying distressed firms, let’s step through how we use an autoencoder to distinguish between “success” and “failure” groups. Separate the data (i.e., the observations for all accounting and market-based variables) into success and failure groups according to the bankruptcy proxy measure. Train the autoencoder to reconstruct the success group. Each variable of interest corresponds to a node in the “input” layer of the network. Feed the autoencoder the failure group data and record the error in reconstructing this data. Set an error threshold (a number) for determining when a firm should be marked a failure vs. success. Generate predictions based on this threshold. Analyze the accuracy of these predictions using metrics such as precision and recall. Future Work A key priority of the ETS team is to use modern techniques in computing to uncover new sources of alpha for our clients. With the recent explosion of advancements in data science and machine learning, we are excited about leveraging the knowledge from these fields to improve the ETS model. We believe that this report provides a glimpse into the style of factor analysis we expect to conduct in the near future. As for the distress model developed here, we are content with the performance of the simple logistic regression for now, but we will continue to examine whether modern classification techniques can help improve accuracy.  One method in particular that piques our interest is so-called “Gradient Boosting”, which has made big waves in the data science community for its use in winning competitions on Kaggle.com.7 Overall, we will continue to monitor the model’s ability to exploit stock-market anomalies and adjust accordingly, whether that means adjustments to existing factors, or the development of new ones. In addition, we will strive to keep ETS an integral part of your asset management workflow via new features on the web platform. As always, we hope you find the recent additions valuable, and please do not hesitate to reach out with any comments or questions.   Spencer Moran, Senior Analyst Equity Trading Strategy spencerm@bcaresearch.com   Appendix Variable Definitions Here we provide the definitions of the variables used in the distress model. The denominator in NIMTA and TDMTA represents “Market Total Assets” and is given by Market Cap. plus Total Debt.  The formulas are as follows: EWM is a 12-month exponential weighted moving average used to incorporate more history into the variable but provide greater weight to more recent observations. STD is an annualized, trailing three-month standard deviation. The symbols R, Rm, and r represent monthly returns, monthly market returns, and daily returns, respectively. Parameters Here we provide the parameters (aka a and b’s) obtained from the rolling fit of the distress model. Recall that parameters are estimated at the end of each year, and then these parameters are used to compute predictions for the following year to avoid look-ahead bias. Footnotes 1      The original Z-score formula was as follows: Z =   1.2X1 + 1.4X2 + 3.3X3 + 0.6X4 + 1.0X5, where: X1 : working capital / total assets X2 : retained earnings / total assets X3 : earnings before interest and taxes / total assets  X4 : market value of equity / book value of total liabilities X5 : sales / total assets 2  Campbell, John Y., Jens Dietrich Hilscher, and Jan Szilagyi. 2011. Predicting financial distress and the performance of distressed stocks. Journal of Investment Management 9(2): 14-34. 3   The decile-spread is a metric we use to evaluate the performance of a trading system. Essentially, it looks at the sum of the differences between the annual compounded growth rates of each score decile. Ideally, the deciles should be in order, and the upper deciles should outperform the lower deciles. This metric captures both these aspects, and summarizes them in a single number. 4       The factor inception date was chosen to be the date when the relative change in the magnitude of the parameter set (treated as a vector) fell below 1%. 5      Please see Equity Trading Strategy, “ETS Goes Global,” dated October 31, 2016, available at ets.bcaresearch.com 6      Precision is defined as the ratio of actual failures to all failures predicted by the model. Meanwhile, recall is the ratio of actual failures to the sum of actual failures plus false successes predicted by the model. 7      Kaggle is a popular online community (recently acquired by Google) where members compete to solve open problems in data science and machine learning.

Highlights Portfolio Strategy The path of least resistance is higher for the broad equity market on the back of a reflationary impulse and a less dogmatic Fed. Now that the SPX forward EPS bar has been lowered to the ground, upward surprises loom, especially if the third catalyst we have been highlighting in recent research materializes: a positive resolution to the U.S./China trade spat. The recent M&A fever, a less dogmatic Fed that has suppressed the 10-year Treasury yield and a pick up in the U.S. credit impulse can serve as catalysts to unlock excellent value in the S&P biotech index. Upgrade to overweight. A profit margin squeeze on the back of soft pharma pricing power, weak operating conditions and a race to buy out biotech stocks to build up drug pipelines warn that the derating phase has just began for the S&P pharma index. Downgrade to underweight. Recent Changes Boost the S&P biotech index to overweight today. Trim the S&P pharma index to underweight today. Table 1 Featured The S&P 500 has been flirting with its 200 day moving average and once it categorically clears this hurdle there are high odds that previous resistance will turn into support. The next important level is 2,800, as we highlighted in recent research, a level where the SPX failed numerous times last year.1 Encouragingly, the character of the market has changed from December’s extreme daily weakness to this year’s significant daily resilience. As we first posited on January 18, while everyone is looking for a retest to re-enter the equity market, we already had the retest in December and are now in a slingshot recovery eerily similar to the 2016 and 1998 episodes.2 Importantly, what has changed since the post-December Fed meeting carnage is that the bond market has completely priced out Fed hikes for 2019 and the 10-year Treasury yield is 15bps lower. Chart 1 highlights this reflationary backdrop for U.S. stocks. Our proprietary Reflation Gauge (RG, comprising oil prices, interest rates and the U.S. dollar) is probing levels last hit in 2012. Historically, our RG and equity momentum have been joined at the hip and the current message is to expect a rebound in the latter. Chart 1Heed The Reflation Message The latest ISM manufacturing survey also corroborates the signal from our RG. The jump in the ISM new orders-to-inventories ratio underscores that the rebound in stocks has further to run (bottom panel, Chart 1). Granted, a lot rests on EPS and in order for stocks to propel to fresh all-time highs later this year, as we expect, profits will have to deliver. On that front, despite recent steep downward EPS revisions across the board, we believe the level of quarterly EPS will hit fresh all-time highs in the back half of the year, carrying stocks into uncharted territory (Chart 2). As a reminder, BCA’s view remains that the U.S. will avoid recession in 2019. Chart 2Joined At The Hip One key profit driver that has put pressure on recent earnings releases and will continue to weigh on internationally-exposed P&Ls is the greenback. With a delayed effect, the first two quarters of this year should bear the brunt of last year’s steep U.S. dollar climb, but that effect will reverse in the back half of 2019. Not only is the greenback inversely correlated with the SPX, but also with the global manufacturing PMI (trade-weighted U.S. dollar shown inverted and advanced, Chart 3). Chart 3Dollar The Reflator... Thus, the greenback is a key macro variable that we are closely monitoring. On that front, global U.S. dollar based liquidity is one of the most important determinants/drivers of global growth. The longer U.S. dollar liquidity gets drained, the more downward pressure it will put on SPX momentum and SPX EPS (Chart 4). Once U.S. dollar based liquidity starts to get replenished at the margin, it can serve as a catalyst for a global growth recovery. A Fed tightening cycle pause and recent acknowledgment that the balance sheet asset roll off is important and the Fed stands ready to tweak it, are a net positive for at least a trough in global U.S. dollar liquidity. Chart 4...But Watch Global Dollar Liquidity Adding it up, the path of least resistance is higher for the broad equity market on the back of a reflationary impulse and a less dogmatic Fed. Now that the SPX forward EPS bar has been lowered to the ground, upward surprises loom, especially if the third catalyst we have been highlighting in recent research materializes: a positive resolution to the U.S./China trade spat.3 This week we make a couple of subsurface changes to a defensive sector; these changes do not alter our recommended benchmark allocation to the overall sector. Biotech’s Gain Is... Biotech stocks have been the center of attention recently as the BMY/CELG deal put the whole sector in play, and today we are boosting exposure to overweight in the S&P biotech index. We doubt the merger mania is over and we continue to believe that more mega deals are in store, either intra or inter-industry, with Big Pharma hungry and in a hurry to replenish their drug pipeline. While this is not the sole reason for an above benchmark allocation, 50-60% M&A deal premia are a boon for investors (Chart 5). Chart 5M&A Frenzy From a long-term macro perspective biotech stocks have been the primary beneficiaries of the 35-year bond bull market. In other words, the multi-decade grind lower in the U.S. Treasury yield has been synonymous with biotech outperformance (10-year U.S. Treasury yield shown inverted, Chart 6). Chart 6Biotech Equities And Rates Move In Opposite Direction The Fed’s recent monetary policy U-turn is a welcome development and these high growth stocks will benefit from the 55bps fall in the 10-year Treasury yield since the early-November peak. In addition, another macro tailwind is working in the S&P biotech index’s favor. The resurgent U.S. credit impulse is unambiguously bullish for this health care index that excels when margin debt availability is rising and liquidity is plentiful (bottom panel, Chart 7). Chart 7Revving Credit Impulse Says Buy Biotech Stocks Surprisingly, the sell-side community does not share our enthusiasm on any of these positive catalysts. Relative profit growth is forecast to be nil in the next year. In the coming five years, biotech stocks are expected to trail the overall market’s profit growth by 4%/annum (middle panel, Chart 8). This is extremely pessimistic and a first in the 24-year history of the I/B/E/S data set, and it is contrarily positive. Relative revenue growth forecasts are also grim for the upcoming 12 months and both revenue and profit forecasts present low hurdles to overcome (fourth panel, Chart 8). Chart 8Analysts Have Thrown In The Towel With regard to technicals and valuations, investors are doubtful that biotech stocks can stage a playable turnaround. Cyclical momentum remains moribund, printing below the zero line. Meanwhile, the S&P biotech index trades at a 25% discount to the SPX forward P/E and well below the historical mean (second & bottom panels, Chart 8). Chart 9 shows that biotech stocks are also cheap on a relative dividend yield basis. The S&P biotech index has been so oversold that it now sports a dividend yield higher than the S&P 500. Nevertheless, there is one key risk we are closely monitoring. Biotech initial public offerings are at all-time highs, with private equity and venture capital funds rushing for the exit doors. This is worrisome as it offsets the supply reduction owing to the M&A fever and has historically coincided with biotech relative share price peaks (Chart 10). Chart 9Compelling Relative Value Chart 10Watch This Risk Netting it all out, the recent M&A fever, a less dogmatic Fed that has suppressed the 10-year Treasury yield and a pick up in the U.S. credit impulse can serve as catalysts to unlock excellent value in the S&P biotech index. Bottom Line: Boost the S&P biotech index to overweight today. The ticker symbols for the stocks in this index are: BLBG: S5BIOT – ABBV, AMGN, GILD, BIIB, CELG, VRTX, REGN, ALXN, INCY. …Pharma’s Pain In mid-2017 we went underweight the S&P pharma index and booked healthy gains roughly a year later when we lifted exposure to neutral. Since then, Big Pharma has enjoyed a reprieve on the back of congressional inaction and the fact that the Trump Administration’s drug pricing wrath was less severe than initially feared. However, the time has come to trim the S&P pharma index to underweight. Chart 11 shows that pharmaceutical companies have been nearly uninterruptedly raising prices for the past four decades. Higher selling prices have been synonymous with higher profits and thus higher share prices. Chart 11Margin Trouble But, something happened in the new millennium. Relative performance peaked as pharma embarked on a mega M&A boom in the late-1990s with the Pfizer/Warner Lambert deal breaking all-time industry M&A records. Why? Because profit margins crested and have never reclaimed their previous zenith (top and middle panels, Chart 11). Neither have relative share prices. Worryingly, pharma prices have hit a wall during the past four years and can barely keep up with overall inflation, despite still being opaque (bottom panel, Chart 11). As both Democrats and Republicans are united to bring down health care costs in general and drug prices in particular, pharma profits will likely suffer a secular downdraft. The implication is that, as pharma revenues erode they will deal a blow to profits. Consequently, the outlook for relative share prices is dim. Importantly, pharma executives have not been frugal enough to offset the soft pricing power backdrop. Headcount has been expanding consistently since 2012 and a wide gap has opened up relative to industry selling price inflation, akin to the one in the mid-2000s that suppressed relative share prices (Chart 12). Chart 12Pricing Power Pressure Similar to the M&A boom of the late-1990s, there has been a global pharma M&A race with multiple deal announcements in the past few months, underscoring that the industry is not standing still. As Big Pharma CEOs compete to outdo their peers and buy drug pipelines mostly in the biotech space (Chart 5), they will continue to degrade the industry balance sheet (third panel, Chart 12). Our strategy is to overweight the hunted (biotech) and avoid the hunters (Big Pharma). On the operating front, a supply check reveals that pharma wholesale and manufacturing inventories are growing, whereas shipments are on the verge of contraction. Pharma industrial production has petered out and industry productivity gains are waning (Chart 13). This deteriorating operating backdrop will weigh on relative profits. Chart 13Deteriorating Operating Metrics... With regard to the macro front, a vibrant U.S. economy – with the ISM manufacturing survey ticking higher and the labor market firing on all cylinders – suggests that defensive pharma relative profits will resume their downtrend (bottom panel, Chart 13). Tack on the U.S. dollar’s reversal since the November peak and defensive pharma equities will remain under pressure (second panel, Chart 14). Chart 14...But EPS Bar Is On The Floor Nevertheless, there are three risks to our negative S&P pharma view. First, the M&A fever dies down and there are no additional purchases of biotech outfits. Second, Congress and the President drag their feet and fail to agree on new hawkish pharma pricing legislation. Finally, sell-side analysts have thrown in the towel and maybe most of the bad news is reflected in bombed out relative profit and sales growth estimates (third & fourth panels, Chart 14). In sum, a profit margin squeeze on the back of soft pharma pricing power, weak operating conditions and a race to buy out biotech stocks to build up drug pipelines warn that the derating phase (bottom panel, Chart 14) has just began for the S&P pharma index. Downgrade to underweight. Bottom Line: Trim the S&P pharma index to underweight. The ticker symbols for the stocks in this index are: BLBG: S5PHAR – JNJ, PFE, MRK, LLY, BMY, ZTS, AGN, MYL, NKTR, PRGO. Health Care Remains In The Neutral Column Despite these two subsurface health care sector moves, our overall exposure to the S&P health care sector remains intact at neutral. Please look forward to reading our upcoming research where we will be updating the S&P managed health care, S&P health care facilities and S&P health care equipment subsectors.   Anastasios Avgeriou, Vice President U.S. Equity Strategy anastasios@bcaresearch.com Footnotes 1      Please see BCA U.S. Equity Strategy Weekly Report, “Trader’s Paradise” dated January 28, 2019, available at uses.bcaresearch.com. 2      Please see BCA U.S. Equity Strategy Insight Report, “Don’t Bet On A Retest” dated January 18, 2019, available at uses.bcaresearch.com. 3      Please see BCA U.S. Equity Strategy Weekly Report, “Dissecting 2019 Earnings” dated January 22, 2019, available at uses.bcaresearch.com. Current Recommendations Current Trades Size And Style Views Favor value over growth Favor large over small caps

Highlights Stay tactically overweight to equities for the time being. Close the overweight to industrial commodities versus equities. The financials, basic resources, and industrials equity sectors can continue to outperform for a few months longer. EM can also continue to outperform DM for a few months longer. Overweight Germany’s DAX versus German bunds. The second half of the year is going to be much tougher than the first half. Feature Chart of the WeekPessimism Was Overdone: The Classical Cyclicals And EM Are Rebounding Locked In An Intimate Embrace Last week, we highlighted a frustrating truth: for the past 16 months the broad equity market has been on a journey to nowhere. Yet the journey has been far from boring. There have been exciting detours of 10-15 percent in both directions, albeit these moves have been short-lived, lasting no more than three months at a time. The same truth applies to the broad bond market: for the past sixteen months the global long bond yield – defined here as the average of the yields on the 30-year German bund yield and 30-year T-bond – has also ended up going nowhere. On this journey too, there have been exciting detours of up to 50 basis points in both directions, but these moves have also lasted no more than three months before retracing. It follows that for the past 16 months, the strategic allocation to equities, bonds and cash has had zero impact on investment performance, but the tactical allocation to the asset classes has had a huge impact. Yet here’s the thing: the sharp tactical moves in the bond market and in the stock market have been intimately embraced. When the global long bond yield has approached the top of its range, it has catalysed a sharp sell-off in equities; and when the bond yield has approached the bottom of its range, it has catalysed a sharp rally in equities (Chart I-2). In fact, over the past 16 months, asset allocation has boiled down to a very simple trading rule based on the global long bond yield: above 2.2 percent, sell equities; below 1.95 percent, buy equities. Today, the yield stands at 1.85 percent, suggesting a tactically overweight stance to equities. Chart I-2The Sharp Tactical Moves In The Bond Market And Stock Market Are Intimately Connected The Persistent Trends Are In Sectors Some investors cannot shift their portfolios quickly enough to exploit the tactical opportunities in the markets. They need trends that persist for at least six months to a year. The good news is that these more persistent trends do exist, but to find them you have to look at equity sectors, and specifically the classically cyclical sectors (Chart of the Week). The financials and basic resources sectors were in strong relative downtrends through most of 2018; but for the last four months these classically cyclical sectors have flipped into very clear uptrends (Chart I-3 and Chart I-4). The same is true for industrials, albeit the end of the downtrend has happened more recently (Chart I-5). Chart I-3Financials Are Rebounding Chart I-4Basic Resources Are Rebounding Chart I-5Industrials Are Rebounding For the avoidance of doubt, technology is not a classically cyclical sector because the sales of technology products – particularly to consumers – are relatively insensitive to short-term fluctuations in the economy. In fact, the relative performance of technology is an almost perfect mirror-image of financials (Chart I-6). Chart I-6The Technology Sector Is Not A Classical Cyclical Neither is the chemicals sector a classical cyclical. Given that raw material prices are an input cost for chemical manufacturers, the chemicals sector can underperform when raw material prices are rising in a cyclical up-oscillation (Chart I-7). It follows that the three true classically cyclical sectors are: financials, basic resources and industrials. Chart I-7The Chemicals Sector Is Not A Classical Cyclical What if your investment process does not allow you to invest in sectors and benefit from their well-defined and longer trends? The good news is that you can play these same trends through regional and country stock market indexes. We refer readers to previous reports for the details, but the crucial message is that regional and country relative performances stem from nothing more than the stock markets’ defining sector skews combined with sector relative performances.1 This revelation of what truly drives regional and country relative performance is bittersweet. It is sweet because it simplifies an investment process that can be very complicated. But it is also bitter because it highlights that the investment industry is still replete with unnecessary layers of complexity. Still, just to drive home the point, we would like the charts to do the talking. The relative performance of financials, the relative performance of Italy’s MIB, and the relative performance of Emerging Markets (EM) versus Developed Markets (DM) are all effectively one and the same story (Chart I-8 and Chart I-9). Chart I-8One And The Same Story: Financials And Italy... Chart I-9...And Financials And EM Versus DM What Are The Markets Telling Us, And Do We Agree? Another very common question we get is: what is our forecast for economic growth and profits growth? For example, two questions on everyone’s lips right now are: can Germany avoid a technical recession, and what is our forecast for Germany’s growth from here? These are indeed important questions, but for investors they are not the most important questions. Financial markets are a discounting mechanism. So for investors, the most important question should always be: what is discounted in the current market price, and is that too optimistic or too pessimistic? Over-optimism and over-pessimism on the economy are especially important for the classically cyclical sectors because their profits have a very high operational gearing to their sales: a small change in the sales outcome has a huge impact on the profit outcome and, therefore, the price.  If the price is discounting a booming economy and what actually transpires is that the economy grows modestly, then a seemingly benign outcome of respectable growth will paradoxically cause the price to slump. Conversely, if the price is discounting a very pessimistic outcome and what actually transpires is anything better than the ultra-pessimism, then even a bad outcome will paradoxically cause the price to soar. In this regard, the recent underperformance of Germany’s DAX versus German bunds is at an extreme not far from that during the euro sovereign debt crisis in 2011-12 (Chart I-10). So the important question for investors is: will the actual economic outcome transpire to be as extreme as that? Our answer is that the extreme underperformance of the DAX versus bunds is discounting an overly pessimistic outcome, and on that basis the correct stance is to be overweight the DAX versus bunds.   Chart I-10Overly Pessimistic: The DAX Versus Bunds Turning to the classical cyclicals, these sectors have rebounded because their embedded assumptions for growth reached peak pessimism in October. Since then, the pessimism has abated at the margin because of improving short-term impulses from Chinese stimulus, lower global bond yields, and sharply lower energy prices. Given that positive (and negative) impulse phases reliably tend to last for six to eight months, our expectation is that this tailwind for the classical cyclical sectors – financials, basic resources, and industrials – can continue for a few months longer. Which means that the outperformance of EM versus DM can also continue for a few months longer. In terms of asset allocation, long industrial commodities versus equities worked very powerfully at the end of last year, but the relative merits of the two asset classes are now more evenly balanced. Hence, we are now closing this position in profit. Finally, our major concern is for later in the year when the aforementioned improving short-term impulses will inevitably fade, and even potentially reverse. Bear in mind that the impulses arise from the short-term changes in credit flows, bond yields, and the oil price. It follows that to recreate these positive impulses for later in the year, bond yields and/or the oil price have to keep falling. This is not our base case, so enjoy the positive impulses while they last! As the year progresses the investment environment is going to get much tougher. Fractal Trading System* The sharp underperformance of the Nikkei 225 versus the Hang Seng is at the limit of tight liquidity that has signaled all of the recent trend reversals in this relative position. Accordingly, this week’s recommended trade is to go long the Nikkei 225 versus the Hang Seng. Set a profit target of 4.5 percent with a symmetrical stop-loss. We now have seven open positions. For any investment, excessive trend following and groupthink can reach a natural point of instability, at which point the established trend is highly likely to break down with or without an external catalyst. An early warning sign is the investment’s fractal dimension approaching its natural lower bound. Encouragingly, this trigger has consistently identified countertrend moves of various magnitudes across all asset classes. Chart I-11 The post-June 9, 2016 fractal trading model rules are: When the fractal dimension approaches the lower limit after an investment has been in an established trend it is a potential trigger for a liquidity-triggered trend reversal. Therefore, open a countertrend position. The profit target is a one-third reversal of the preceding 13-week move. Apply a symmetrical stop-loss. Close the position at the profit target or stop-loss. Otherwise close the position after 13 weeks. Use the position size multiple to control risk. The position size will be smaller for more risky positions. * For more details please see the European Investment Strategy Special Report “Fractals, Liquidity & A Trading Model,” dated December 11, 2014, available at eis.bcaresearch.com Dhaval Joshi, Senior Vice President Chief European Investment Strategist dhaval@bcaresearch.com Footnote 1 Please see the European Investment Strategy Weekly Report “Oil, Banks, And Bonds: The Oddities Of 2018”, dated November 29, 2018 available at eis.bcaresearch.com Fractal Trading System Recommendations Asset Allocation Equity Regional and Country Allocation Equity Sector Allocation Bond and Interest Rate Allocation Currency and Other Allocation Closed Fractal Trades Trades Closed Trades Asset Performance Currency & Bond Equity Sector Country Equity Indicators Bond Yields Chart II-1Indicators To Watch - Bond Yields Chart II-2Indicators To Watch - Bond Yields Chart II-3Indicators To Watch - Bond Yields Chart II-4Indicators To Watch - Bond Yields Interest Rate Chart II-5Indicators To Watch - Interest Rate Expectations Chart II-6Indicators To Watch - Interest Rate Expectations Chart II-7Indicators To Watch - Interest Rate Expectations Chart II-8Indicators To Watch - Interest Rate Expectations

Special Report

Highlights China’s recently released pro-auto-consumption policy will lead to a moderate 5-8% recovery in auto sales/production this year. However, the impact from the stimulus will be much less than the previous two episodes in 2009 and 2016. The value of Chinese auto sales is likely to increase by RMB 200 billion to 350 billion, which is about 0.2-0.4% of the country’s nominal GDP in 2018. New-energy cars will continue to gain market share with supportive policies. Meanwhile, domestic brand car manufacturers will likely benefit most from the upcoming recovery in the Chinese auto market, while American car producers will benefit the least. We recommend preparing to go long Chinese auto stocks in the domestic market in absolute terms, subject to the terms of a trade agreement with the U.S. In addition, we continue to overweight domestic consumer discretionary stocks versus the benchmark, and versus domestic consumer staples. Feature China is the world’s largest car producer and consumer – its domestic sales account for about 30% of global auto sales (Chart 1, top panel). The country experienced a 3% contraction in auto sales and production through last year, the first year of negative annual growth in 28 years. The contraction rapidly accelerated into the double digits over the past few months (Chart 1, bottom panel). Chart 1Chinese Auto Industry: Policy Stimulus = Recovery In 2019 As the auto sector is an important driver of China’s economic growth, whenever the industry has shown signs of weakness, the central government has typically implemented a series of supportive policies designed to stimulate the domestic auto market. The authorities successfully did this in 2009-2010 and 2016-2017. Late last month, they again announced a set of pro-auto-consumption policies. The question going forward is how effective these measures will be in boosting auto sales. We believe the recovery will be rather moderate compared with the 2009-2010 and 2016-2017 episodes. Chances are that the growth of auto sales and production will recover to 5-8% in 2019. As a result, we recommend preparing to go long Chinese auto stocks in absolute terms, subject to the terms of a trade agreement with the U.S. Cyclical And Secular Forces Shaping Auto Sales A comparison of the current auto market to the one that prevailed in 2009 and 2016 is helpful to gauge the extent of the strength of the pending auto sales recovery expected this year. Box 1 shows the recently released pro-auto-consumption plan by the Chinese government, which focuses on six aspects, including promoting auto replacement, NEV sales, auto sales in rural areas, pick-up truck sales, development of the second-hand car market, and auto sales in cities that have restricted auto sales policies.   BOX 1: China’s Stimulus Package For Domestic Auto Industry The recently released pro-auto-consumption plan by the Chinese government includes: Promoting auto replacement: Providing subsidies to consumers who scrap their older, higher-polluting cars for new, lower-emission or zero-emission cars; Encouraging NEV sales: Providing subsidies to advanced NEV sales and giving more privileges to new energy trucks; Promoting auto sales in rural areas: Providing subsidies to rural residents who scrap their tricycles to buy a truck with cylinder capacity equal or less than 3.5 tons, or a passenger car with cylinder capacity equal or less than 1.6L; Promoting pick-up truck sales: Widening access areas within cities for pick-up trucks; Accelerating the development of the second-hand car market: Allowing second-hand car trades across different cities and provinces; Loosening auto sales restrictions in cities that have restricted auto sales policies. Regarding the amount of subsidies, the government did not provide details.   Putting it all together, we believe that this time the impact from the stimulus will be much more muted than the previous two episodes in 2009 and 2016. First, there is no sales tax reduction measure in this round of stimulus. The most important driver for the auto market recovery in 2009 and 2016 was a sales tax reduction in passenger cars with cylinder capacity equal to or less than 1.6L from 10% to 5% (Chart 2). However, this time, there is no such cut. While the government is maintaining zero sales tax on new energy vehicles (NEV), the sales tax on all automobiles remains at 10% this year. Chart 2The Lessons From The 2009 And 2016 Episodes Second, domestic pent-up demand for automobiles is much lower than it was in both 2009 and 2016. The car ownership rate, defined as the number of passenger cars per 1000 households, has risen significantly to 453 in 2018 (Chart 3). This means that nearly half of Chinese households already own at least one car as of 2018. In comparison, the car ownership rate was only 91 in 2008 and 318 in 2015. Chart 3Less Pent-Up Demand For Autos In 2019 Than Before Third, Chinese households’ debt levels have surged in the past few decades, constraining their ability to purchase cars and other goods (Chart 4, top panel). While many investors compare the cross-country household debt burden relative to GDP, Chinese household debt has already risen to nearly 120% of households’ disposable income, surpassing the U.S. (Chart 4, bottom panel). Chart 4Increasing Households' Debt Burden Constrains Ability To Buy A Car Fourth, while the recent stimulus packages aim to promote auto sales in rural areas, the difficulty of getting auto loans is much higher for the average rural household than for the average urban household, as the former generally have much lower income levels. In addition, peer-to-peer lending, which has become a major source of auto loans in recent years due to lower lending standards compared with banks, has collapsed since last year (Chart 5). With tightening regulations, the difficulty of acquiring auto loans through peer-to-peer lending is currently higher than before. Chart 5Rising Difficulty To Get An Auto Loan Lastly, there has been a structural decline in consumers’ willingness to buy cars due to increasing traffic congestion, limited parking space and more advanced public transportation. Moreover, more mature car rental markets and the rising use of car-sharing services have also helped reduce the need to buy a car, to some extent. This is a major difference from 2009-2010 and 2016. In Chart 6, both falling households’ marginal propensity to consume and declining consumption loan growth suggest a decreasing willingness to consume among Chinese consumers. Chart 6Chinese Consumers: Falling Willingness To Consume With all the aforementioned cyclical and structural forces in place, the impact on domestic auto sales from the recent stimulus package will be smaller in 2019 than in 2009 and 2016. That said, these policies will still be supportive, and likely sufficient to lift auto sales from contraction back to positive growth this year. Estimating the magnitude of the impact remains challenging, however, due to lingering uncertainty about the size of government subsidies. Based on all six measures listed in Box 1, the scale of subsidies provided by the government will be the major determinant for auto sales growth in China in 2019. In general, the bigger the subsidies, the stronger the push on auto sales. In 2009, both the central government and local government provided subsidies for stimulating auto sales. This time, while the financing sources could still be both central and local governments, local governments’ ability to finance auto consumption stimulus is diminishing due to their much higher debt levels and weaker revenues from land sales than in the past. For now, our view is that the impact from the stimulus will be much less significant than the previous two episodes in 2009 and 2016. Auto sales growth was 4.7% and 3% in 2015 and 2017, respectively. With recently announced stimulus, we expect the growth will be higher than in those years. Bottom Line: We expect that the growth of Chinese auto sales/production volumes will rebound to 5-8% this year, much slower than the 45% growth seen in 2009 and 14% growth in 2016. With a similar growth rate in value terms, Chinese auto sales are likely to increase by RMB 200 to 350 billion, which is about 0.2-0.4% of the country’s 2018 nominal GDP. The Winners And Losers At 5-8%, growth will be equivalent to a 1.5-2 million-unit increase in domestic auto sales. This will lead to a similar increase in auto production, as most cars are domestically produced. In terms of fuel use, automobiles can be classified as gasoline cars, diesel cars and new-energy cars. Chart 7 shows that gasoline cars currently hold 84% market share. In terms of brand, automobiles can be categorized as Chinese brands, Japanese brands, German brands, American brands, Korean brands and others. Chart 8 shows their market structure, with Chinese brands currently accounting for 42% of total market share. As the Chinese auto market is set to have a moderate recovery this year, which kinds of cars will benefit most, and which will benefit least? Even though China plans to gradually reduce its subsidies on NEVs to zero in 2021, several factors suggest that NEVs will still be the biggest winner, taking more market share from both gasoline and diesel cars. The government is aiming to increase the NEV market share from 4.5% currently to 20% by 2025. Assuming total sales rise to 32 million units in 2025 from current levels of 28 million (about 2% annual growth), this would imply that NEV sales will surge to 6.4 million units from 1.3 million currently, which is equal to 26% annual growth over the next seven years (Chart 9). Chart 9NEV Sales: Plenty Of Upside In addition to governments continuing subsidies, the sales tax on NEVs will be held at zero until the end of 2020, a big advantage over non-NEV vehicles, which carry the 10% sales tax. In addition, in cities that have license restrictions on car sales or have time or area restrictions on on-road autos, NEVs are not constrained by such policies, which is an attractive privilege for car buyers to consider. For example, in Shanghai, it costs over 80,000 RMB to buy a license plate for a non-NEV car if the potential buyer is lucky enough to be selected by random draw. In comparison, buying a NEV allows the buyer to have a free license plate. Current NEVs can achieve recharge mileage of 300-450 kilometers, with a price of RMB 100,000 to RMB 150,000 per unit. While the recharge mileage is sufficient for most daily use, prices are no longer substantially higher than prices for traditional gasoline or diesel cars. Major global and local NEV producers are expanding their production in China. For example, Tesla last month started building its mega electric car manufacturing plant in Shanghai, which will initially produce 250,000 cars per year, and eventually ramp up to half a million. This will be about five times the number of vehicles the company currently produces in the U.S. Most NEVs that have been sold in China are Chinese-brand NEVs. However, with China further opening up its auto sector and allowing more foreign NEV producers to invest and produce cars in China, Chinese NEV producers will face increasing competition and may lose some market share to foreign NEV producers. Meanwhile, Chinese NEV-related supportive policies will likely benefit both local and foreign NEV producers as the government is determined to develop the domestic NEV market and encourage NEV sales. That said, local producers will still enjoy slightly more favorable policies than foreign ones. Given that the government is promoting smaller-engine passenger car sales in rural areas and encouraging the replacement of old diesel cars with NEVs, sales and production of gasoline cars may also increase slightly, while diesel cars are likely to rise the least. In terms of brand, Chinese and American brands lost share to Japanese and German brands last year. We believe Chinese brands will benefit most from this year’s government-led auto market recovery for two reasons (Chart 10, top panel): Chart 10Chinese Brands Will Benefit Most From This Year’s Policy Stimulus The authorities will likely favor local brand producers in terms of benefitting from the subsidies they give to car buyers. In addition, local brand cars in general have lower prices than foreign brands, which could be the most attractive feature for price-sensitive rural residents. In the meantime, as the government encourages local auto replacement, this may benefit Japanese and German brands (Chart 10, second and third panels), as buyers with replacement needs will likely upgrade their cars to ones of higher quality and better reputation. Among American cars, while we are positive on American NEV car sales in China, we still expect American cars to continue to lose market share due to weakening sales of American non-NEV car sales (Chart 10, bottom panel). American cars are generally more expensive than Chinese-brand cars, and they are often perceived as slightly lower quality than either Japanese or German brands. Moreover, the ongoing trade dispute may bias Chinese buyers against buying an American car. Bottom Line: We believe NEV producers and Chinese-brand car producers will benefit most from this year’s government-led auto market recovery. Investment Implications There are several important conclusions that stem from our research. First, while rebounding auto production will likely lift demand for many metals, housing construction is artificially supporting demand and is set to decelerate over the coming year (Chart 11). Consequently, we do not believe that accelerating auto production alone is a license to be long industrial metals over the coming year. Chart 11Weakening Property Market Weighs More On Commodity Market Second, within the equity space, we recommend that global investors prepare to go long domestic auto stocks on an absolute basis after the outcome of the U.S.-China trade talks emerges later this month. Rebounding auto production will likely lead to a cyclical improvement in auto producer earnings, which in combination with deeply oversold conditions bodes well for the 6-12 month outlook (Chart 12). Chart 12Look To Long Domestic Auto Stocks In An Absolute Term U.S. negotiators are seeking increased access to the Chinese auto market, which implies that the outcome of the negotiations carries some event risk for domestic producers (particularly if China’s concessions on this front turn out to be large). But our sense is that we are likely to recommend an outright long position favoring domestic automakers barring a trade deal with deeply negative implications for domestic producer market share. Third, our bullish bias towards Chinese auto producers and our constructive outlook for the home appliance market supports two of our existing trades favoring consumer discretionary stocks. Chart 13 highlights that production and sales volume for several home appliance products is depressed, and stands to benefit from a flurry of policy announcements late last month that were intended to support the industry. Chart 13Home Appliances: Rebound Soon On Stimulus As Well Both auto producers and home appliance manufacturers belong to the consumer discretionary sector, and we recommend maintaining a long domestic consumer discretionary position versus both the domestic benchmark and relative to consumer staples (both trades were initiated on November 141). While domestic consumer discretionary stocks are expensive vs. the domestic benchmark on a P/B basis (Chart 14), the sector’s relative P/E ratio is trading at the very low end of its historical range and the trade has eked out modest positive gains since initiation. Chart 14Remain Overweighting Consumer Discretionary Sector Our long discretionary / short staples trade has faired much worse, down 11% since initiation due to a significant rally in consumer staples stocks (rather than losses in the discretionary sector). We recommend that investors stick with the trade over the coming 6-12 months despite the loss, as Chart 15 highlights that the discretionary / staples trade could not be more extreme in terms of relative performance or valuation. Our bet is that this trade will reverse course in 2019, for a meaningful period, in response to a cyclical tailwind from policy. Chart 15Stay Long Discretionary / Short Staples   Ellen JingYuan He, Associate Vice President Emerging Markets Strategy EllenJ@bcaresearch.com   Footnotes 1 Please see BCA Research’s China Investment Strategy Special Report “Chinese Household Consumption: Full Steam Ahead?”, published November 14, 2018. Available at cis.bcaresearch.com. Cyclical Investment Stance Equity Sector Recommendations

​​​​​​​ Overweight In Monday’s Cyclical Indicator Update, we highlight our cyclical portfolio bent, driven by three core catalysts that we think will take U.S. equities higher. These are: a definitively more dovish Fed, which would help restrain the greenback, a continuation of the earnings juggernaut and a positive U.S./China trade resolution. One cyclical sector that looks particularly attractive is S&P financials. The divergence between the directions for our cyclical macro indicator (CMI) and our valuation indicator (VI) for financials has reached stunning levels. The CMI is accelerating into pre-GFC territory as credit quality, loan growth and unemployment are all in the sweet spot while the VI has fallen to two standard deviations below fair value. Our technical indicator (TI) sends a signal that financials are modestly oversold though this relatively neutral message does not diminish the most bullish signal in our cyclical indicator’s history. Bottom Line: We reiterate our overweight recommendation for S&P financials. Please see Monday’s Cyclical Indicator Update for more details on this as well as our cyclical indicator updates on the other GICS1 sectors and our large cap/small cap style preference.