BCA Indicators/Model

Highlights Risk assets continue to rise despite a flattening yield curve. Individual investors are more sanguine than institutional investors as stocks make new highs. The S&P 500 is testing the top of a key channel. Will it break out or break down? Bond market sentiment, positioning and technicals today vs. 1994. Feature Risk-on returned to financial markets last week as the S&P 500 hit a new all-time high and oil prices reached a 2-year high. Credit spreads narrowed as well. This occurred despite growing investor angst regarding the flattening yield curve. At 58 basis points, the 2/10 yield curve is still in positive territory, but the recent flattening could be interpreted as heralding a Fed policy mistake. We, too, are concerned. The flattening curve is being driven by the Fed's determination to continue lifting short-term rates even in the face of subdued inflation readings. Our base case outlook sees inflation grinding higher in the coming months, leading to a temporarily steeper curve. Nonetheless, we will re-evaluate our asset allocation if the curve continues to flatten and core inflation remains stuck in a range. BCA expects U.S. stocks to outperform Treasuries in 2018. S&P 500 EPS growth and margins will hold up through mid-year, supported by an above-trend domestic economic expansion in 1H 2018, a dose of fiscal stimulus and accelerating economic activity outside the U.S. Still, many investors are concerned that sentiment and valuations are signaling that a pullback is nigh. Sanguine Sentiment Our technical and sentiment indicators are not flashing red as in previous bear markets, but neither are they giving an all-clear for U.S. equity investors. Sentiment levels are a bigger concern than technical indicators and investors should monitor both for signs of an equity sell-off. BCA's U.S. equity sentiment indicator is elevated, although not at an extreme (Chart 1). Remarkably, in contrast to previous market troughs, individual Investors (panel 2) are more sanguine than either financial advisors (panel 3) or traders (panel 4). Bullishness among traders is at a 10-year high. Typically, after a long bull run, institutions are more cautious about equities than the oft-maligned individual investor. Several other sentiment surveys illustrate the divergence in sentiment between institutions and individuals. As per the American Institute of Individual Investors, the percentage of small investors who are bearish (Chart 2, 35%, panel 2) is in the middle of a 30-year range while the percentage of bulls (29%, panel 3) is at the low end. Moreover, Chart 3 shows the gap in the expectation between households and professionals on future stock market returns (as tallied by the Yale School of Management's International Center for Finance) and on buying the dips (panel 4). That said, individuals and institutions are more aligned on the likelihood of a stock market crash in the next six months. None of the three sentiment indicators from the Yale survey are at an extreme. Chart 1Overall Sentiment Levels Elevated##BR##But Not At Extremes Chart 2Individuals Are Not##BR##Overly Bullish Active managers have reduced equity risk since the beginning of Q4 (Chart 4). At 61%, the average equity exposure of institutional investors surveyed by the NAAIM1 is at the lowest level since May 2016 and is nearly half the 102% exposure at the start of 2017. The March 2017 reading was the highest since 2007, just before the S&P 500 peak in October 2007. Chart 3Gap Between Individual##BR##And Institutional Investors Chart 4Active Managers Still##BR##Overweight Equities... Similarly to previous bear markets, BCA's equity speculation index moved into "high speculation" territory in early 2017 and remains so as the year ends and range bound on average at somewhat lower levels. Net speculative positions of S&P 500 stocks are in balance, however, and do not signal that market risk-taking is rampant (Chart 5). Moreover, the dispersion of equity volatility of new high and lows of the S&P500 is quite wide, ranging from over 20% to below 5%, over previous historical periods since 1994. Although volatility is not a leading indicator of future equity market returns, good or bad, the current low level of volatility, especially over the short-term, 6 months to 1-year, may be longer-lasting, having peaked from over 15% only since early 2016 and now closer to 5%. Longer-term volatility, for example, based on 2-, 3- and 4-years, still remains above 10%. It is not unusual for both short-term and long-term volatility to eventually converge, as seen in post-bear market phases, especially in the mid-2000s (Chart 6). Chart 5Speculation High, But Not At Extremes Chart 6Equity Vol Remains Low Warning Signs From Technicals? On balance, the technical indicators we monitor do not suggest that the market is stretched. Chart 7 shows that the S&P 500 is testing the top end of the 2009-2017 recovery trend channel. A failure to break out of the channel may result in some near-term consolidation for U.S. equities. However, a definitive break above 2616 would imply another upleg for stocks. The escalating advance/decline line is also in a bullish trend (Chart 7). The other technical indicators we monitor fall into two categories. Some are elevated, but not at extremes. Others are still in the middle of the range and are not a concern. The S&P 500 is 6% above its 200-day moving average, in the upper end of its post-2000 range, which is well below the recent highs set in 2009, 2011 and 2013. The S&P's distance from its 50-day MA is in a similar position (Chart 8, panels 1 and 2). BCA's composite technical measure is in the middle of the 2007-2017 range, and is not a concern (Chart 9, panel 5). Moreover, the percent of NYSE stocks above their 10- and 30-week highs are midway in their recent range. Furthermore, new highs minus new lows is at neutral lows (Chart 6, panel 2). Chart 7Breakout...Or Breakdown##BR##At Top Of Channel? Chart 8S&P Not Elevated Vs.##BR##Moving Averages Chart 9U.S. Stocks Not##BR##Overextended Bottom Line: Neither sentiment nor technical indicators are flashing red, although the fact that institutional managers are heavily overweight stocks is worrying. We continue to recommend stocks over bonds in the next 12 months, but acknowledge that risks to BCA's stance are climbing. Investors should be prudent with risk assets, paring back any maximum overweight positions and holding some safe-haven assets within diversified portfolios. BCA's U.S. Equity Strategy service maintains a positive technical stance on the energy sector2 and notes that technicals in the consumer discretionary sector look washed out.3 BCA downgraded consumer discretionary from overweight to neutral on September 25, 2017 despite the attractive technical backdrop of the sector. Is It 1994 - Again? BCA's U.S. Bond Strategy service puts fair value on the U.S. 10-year Treasury at 2.69%,4 and rates may climb as high as 3.0% in 2018 if inflation returns to the Fed's 2.0% target. Fundamentals (elevated inflation, above-trend U.S. growth, a more aggressive Fed) support our bond view. However, what does the technical picture in the bond market tell investors? Charts 10 and 11 show the sentiment and technical indicators for the bond market in 2017 and 1994. The duration positioning of portfolio managers in late 2017 matches the situation in 1994. At 100%, portfolio duration is the highest since March 21, 2017. This positioning implies that the market is vulnerable to a spike in rates, as it was in 1994 when the Fed's 75-basis point rate hike in February caught the market off guard. In October 1994, portfolio duration was 103%. While BCA views a Fed policy mistake as a risk to our bullish equity call in 2018, a 1994-style surprise from the Fed is unlikely. In 1994, the Fed's policy intentions were opaque, at best. Since then, the Fed has become increasingly transparent and frequently seeks a "buy-in" from the market before boosting rates. Chart 10Bond Market Positioning,##BR##Sentiment And Technicals In 1994.... Chart 11...And In##BR##2017 The 10-year Treasury yield is currently in an uptrend as it was in early 1994. Today, yields have climbed 80 bps off their post-Brexit lows in mid-2016. The 10-year yield troughed in October 1993 at 5.19%, and rose 60 bps before the Fed's shock rate hike in early 1994. However, in 1994 yields were only beginning to enter the second decade of what would become a 35-year fall in bond yields. BCA's view is that the 1.57% yield in June 2016 marked the end of that multi-year decline. The bond market in late 2017 is as oversold as the bond market was in early 1994, although it took different paths to get to the same juncture. According to BCA's Composite Bond Indicator, the bond market in late 1993 and early 1994 was working off a deeply overbought position. However, by early 1994, bonds were modestly oversold. BCA's bond measure was deeply oversold in late 2016 and early 2017, but shifted into overbought territory in the summer. Today, bonds are modestly oversold. Panel 4 of Charts 10 and 11 show that Fed rate hikes were not priced in at the end of 1993 and in early 1994; today, a few increases are priced in. Investors were net purchasers of bond funds in 1993 and 1994, which is the same as the current situation. In 1993, however, investors were shedding bond funds while individuals are now adding to their bond positions. Bottom Line: Several sentiment and technical indicators in the bond market echo the scenario in 1994. Nonetheless, 25 years of increased Fed transparency means it would be unlikely that the market will be surprised by the Fed's next rate increase. Still, with a new Fed Chair, a record number of vacancies on the Fed's Board and an unprecedented unwinding of its balance sheet, a policy misstep by the Fed would threaten BCA's position on the economy, equities and bonds in 2018. A bigger risk may be that the bond market is still priced for the low inflation environment to persist. Accordingly, if there is an upside surprise on inflation, bonds could be hit hard on a re-assessment of the Fed's rate path. John Canally, CFA, Senior Vice President U.S. Investment Strategy johnc@bcaresearch.com 1 National Association of Active Investment Managers. 2 Please see BCA Research's U.S. Equity Strategy Weekly Report "Invincible", published November 6, 2017. Available at uses.bcaresearch.com. 3 Please see BCA Research's U.S. Equity Strategy Weekly Report "Resilient", published September 25, 2017. Available at uses.bcaresearch.com. 4 Please see BCA Research's U.S. Bond Strategy Portfolio Allocation Summary "Into The Fire", published November 7, 2017. Available at usbs.bcaresearch.com.

GAA DM Equity Country Allocation Model Update The GAA DM Equity Country Allocation model is updated as of October 31st, 2017. There are no significant changes in country allocations, but minor changes are the reductions in the overweight of Germany, Sweden and Switzerland in favor of Spain and Italy, which were already overweight, and Australia which was underweight, as shown in Table 1. As shown in Table 2 and Chart 1, Chart 2 and Chart 3, the overall model underperformed its benchmark by 73 bps in October, largely due to the underperformance (110 bps) of Level 2 model, resulting from the large underweight of Japan, which was the best performer in October. The underweight of Australia and Canada worked very well too, but not enough to offset the overweight in the euro zone countries. The strength of the USD against the euro also hurt the performance. Since going live in January 2016, the overall model has outperformed the benchmark by 247 bps, largely from the allocation among the 11 non-U.S. countries, which have outperformed their benchmark by 599 bps. Chart 1GAA DM Model Vs. MSCI World Chart 2GAA U.S. Vs. Non U.S. Model (Level1) Chart 3GAA Non U.S. Model (Level 2) Table 1Model Allocation Vs. Benchmark Weights Table 2Performance (Total Returns In USD) Please see also on the website http://gaa.bcaresearch.com/trades/allocation_performance. For more details on the models, please see the January 29th, 2016 Special Report, "Global Equity Allocation: Introducing the Developed Markets Country Allocation Model." http://gaa.bcaresearch.com/articles/view_report/18850. Please note that the overall country and sector recommendations published in our Monthly Portfolio Update and Quarterly Portfolio Outlook use the results of these quantitative models as one input, but do not stick slavishly to them. We believe that models are a useful check, but structural changes and unquantifiable factors need to be considered too in making overall recommendations. GAA Equity Sector Selection Model The GAA Equity Sector Selection Model (Chart 4) is updated as of October 31st, 2017. Chart 4Overall Model Performance Table 3Allocations Table 4Performance Since Going Live The growth component in the model has turned cautious on the global recovery. The aggregate cyclical sector overweight has been reduced to 2.5% from 8% last month. However, cyclical sectors such as energy, materials and industrials have seen an increase in overweight driven by favorable liquidity and momentum backdrop. On the other hand, financials and technology have been downgraded to underweight. Finally, as a result of the bearish outlook from the growth component, the model has turned overweight on utilities. For more details on the model, please see the Special Report "Introducing The GAA Equity Sector Selection Model," July 27, 2016 available at https://gaa.bcaresearch.com. Xiaoli Tang, Associate Vice President xiaoli@bcaresearch.com Aditya Kurian, Research Analyst adityak@bcaresearch.com

Special Report

Highlights On Black Monday, October 19, 1987, equity bourses around the world plunged amid cascading bouts of selling, recording some of their largest single-day losses of the twentieth century. The plunge, exacerbated by derivatives transactions, and transmitted swiftly around the world, marked the first contemporary global financial crisis. BCA clients were well prepared. The Bank Credit Analyst steadily warned of increasing stock market vulnerabilities across all of 1987 even as it correctly predicted that the S&P 500 would most likely soar before eventually cracking. The Federal Reserve's immediate all-out effort to contain the damage ushered in a new central bank template for responding to quaking markets and helped give rise to the Greenspan put. While we do not fear a repeat of Black Monday, the U.S. equity market's long-term prospects are dramatically less appealing than they were in 1987. Investors should be prepared for an extended stretch of public market returns that pale beside the ones earned over the last 30-plus years. Feature 30 years ago today, Black Monday erupted around the world, reaching its nadir in New York, where relentless waves of selling drove the major indexes down 20%. The contagion had spread in a rapid relay from Hong Kong to Europe and then to New York, before fetching up in Auckland and other Asia-Pacific exchanges as Black Tuesday. The event was the centerpiece of what turned out to be sharp, albeit relatively brief, bear markets around the world (Charts 1 and 2). Confounding nearly every observer, however, the crash did not amount to much in a broader economic context and financial markets quickly regained their footing, with global equities vaulting to new highs in the '90s1 amidst speculative excesses that made the '80s' mania look demure. Chart 1Great Runs... Chart 2...And Sudden Stops Like all serious investors, BCA researchers are students of history. Black Monday was the first modern global financial crisis, and its 30th anniversary affords us the chance to study its run-up and aftermath for insights into future dives. It also gives us the chance to return to BCA's extensive archives and see how our forebears assessed conditions in real time. Their ex-ante analysis and forecasts were stellar, and reinforce the robustness of our approach. Their lagging ex-post performance highlights the need for investors to maintain a flexible mindset that can accommodate all possibilities. From Fear To Greed Black Monday marked the definitive end of a historically potent bull market (Table 1) that began, as the best ones do, in revulsion. Business Week's August 1979 cover story trumpeting the death of equities has become notorious, but the S&P 500 didn't bottom for three more years, during which it lost a quarter of its inflation-adjusted value. All told from the end of September 1968 to the end of July 1982, the S&P tumbled 62.5% in real terms (Chart 3). Inflation took a heavy toll on real growth over the 55 quarters of U.S. stocks' lost decade and a half (Chart 4, top panel), but the economy had expanded nonetheless, and stocks emerged from the ashes of the Volcker double-dip recession with a lot of ground to make up. Table 1A Bull With Speed And Stamina Chart 3A Lost Decade And A Half ... Chart 4...Despite Steady, If Unspectacular, Real Growth The ensuing five-year bull market (Chart 5, top panel) unfolded in two phases: the first, which burst out of the gate on a sudden repricing before taking a full year to catch its breath, had the support of earnings growth (Chart 5, middle panel) and re-rating; the second, which went on without pause for two and a half years, was all about re-rating (Chart 5, bottom panel). It finally ended in late August 1987, when skeptical investors could no longer stomach big gains derived entirely from multiple expansion, and stocks began to retreat in earnest in October, sliding 5% and 9% in the two weeks before Black Monday. Proximate triggers included sickly trade data, a competitive devaluation threat and proposed tax legislation that stood to make corporate takeovers a good deal more costly. The first two factors pushed the dollar down and yields up, as investors fretted that the Fed would be forced to raise rates (Chart 6), and the last pulled the plug on runaway speculation in takeover targets. Chart 5A Two-Act Bull Market Chart 6Be Careful What You Wish For The Echo Chamber, ... There is career safety in numbers, but portfolio danger. As the late Barton Biggs put it, there's no investment so good that it can't be destroyed by too much capital. Portfolio insurance may not have even been a good idea, as it didn't amount to anything more than a portfolio-sized stop-loss order, souped up with computer software and derivatives contracts. But by the fall of 1987, its widespread adoption had turned it into a very bad one. Portfolio insurance was developed in the late '70s by two finance professors who sought a method that would allow investors to participate in equity market gains while limiting their downside exposure. When stocks began to decline in the direction of a set downside limit, the portfolio insurance program would reduce net equity exposure via the sale of index futures. Once the market recovered and the program determined the coast was clear, it would unwind the futures positions. Although the technique had its flaws on a micro scale - futures trading wasn't costless, and there was considerable potential for whipsawing - it was doomed at the aggregate level because the index futures market wasn't deep enough to accommodate all the selling pressure that would be unleashed by a significant correction. ... Or, From Wall Street To LaSalle Street And Back Again There was more to Black Monday than portfolio insurance - the event was global, and the technique was not a factor on other bourses - but it helped to create a self-reinforcing spiral between the cash market in New York and the futures market in Chicago. Heavy selling of stocks in New York triggered heavy selling of index futures in Chicago, as insured portfolios sold futures to mitigate their direct cash exposures. The selling redounded back to New York as the futures buyers on the other side of the trade sold the underlying stocks to balance out their long futures positions2 and opportunistic investors seized the chance to front-run the mechanical portfolio insurers.3 The new sales pushed share prices even lower in New York, triggering more index futures selling in Chicago, and cinching the vicious circle. The View From Peel Street BCA, safely removed from the madding crowd in Montreal, foresaw something quite like the crash. The September 1986 and 1987 editions of our annual New York conferences bore the respective titles, "The Escalation in Debt and Disinflation: Prelude to Financial Mania and Crash?" and "Phase II in the Escalation of Debt, Disinflation and Market Mania: Prelude to Financial Crash?" Throughout all of 1987, the monthly Bank Credit Analyst warned of the U.S. equity market's increasing vulnerability and recommended that investors reduce exposure in a disciplined fashion ahead of the inevitable bust. The investment policy recommendation, issued in accord with prudent money management principles, differed from BCA's market forecast, which was for robust, potentially parabolic, gains before the bull market ended. BCA was not trying to have it both ways: it has long been a central tenet of our work that one's investment strategy can - and regularly should - be distinct from one's market forecast. We do not attempt to squeeze every last drop out of a bull or a bear market. Empirical evidence makes it abundantly clear that no one can consistently call tops or bottoms. In the words of turn-of-the-century trading legend Jesse Livermore: "One of the most helpful things that anybody can learn is to give up trying to catch the last eighth - or the first. These two are the most expensive eighths in the world.4" The opening paragraph of the March 1987 Bank Credit Analyst, published six months before the market peak, summarizes our ongoing advice: [I]nvestors who are overexposed should reduce positions to a level comfortable to ride out what will likely become a much more volatile phase of the secular bull market in stocks. ... At some point, it is likely that the U.S. stock market will experience a 1962-type correction - a sharp decline which comes out of the blue as a result of extreme overvaluation and excessive speculation. As then, it is unlikely to be associated with a credit crunch, as almost all post-war bear markets have been. ... At present, there is nothing in the data, either fundamental or technical, which suggests that such a shakeout is imminent. However, the key for investors in this bull market is to have positions which are sufficiently comfortable so that they can ride out sudden, dramatic corrections and participate in the long upward rise, which we feel has much further to go. (pp. 3-4) Eighteen months before the August 25th peak, the March 1986 Bank Credit Analyst's Section III was titled, "The Coming Financial Mania," and its strategy prescriptions were much more aggressive, even as it acknowledged the risks: Increasing volatility should be expected both because of the still lingering risks prevailing and the dramatic price movements in recent months. Hence, conservative investors should not overtrade. To fully capitalize on the ongoing revaluation of financial assets, it is important not to lose positions as a result of the necessary sharp corrections which will be experienced along the way. The stock and bond market potential over the next 2-3 years remains extraordinary. (p.11) The great dilemma for investors is, of course, how aggressively to play the game during the latter stages. The fascination, excitement and danger is the knowledge that vast fortunes are easily made right up to the end, but there is no reliable method to get out just before the crash. [...] Frequently the bubble goes on much longer and prices go far higher than anyone can imagine [...]. Yet, the vulnerabilities grow proportionately to the power of the manic phase. (p.26) Investment strategy in [a manic] environment must be based on the historically observed phenomenon that price appreciation generally accelerates to a climax or blowoff and that the hidden risks grow exponentially with price rises. Therefore, investors must constantly guard against the natural tendency to become increasingly greedy and careless in valuation standards as prices rise. (p.41) As good as BCA's near- and intermediate-term calls were in the run-up to the '87 crash, our longer-term calls were even better. We repeatedly argued that disinflation would be a secular trend, and that it would power secular bull markets in bonds and equities. Three decades on, with the Barclays Aggregate Index, the Barclays High Yield Index and the S&P 500 having produced real annualized total returns of 5%, 9.3% and 7.6%, respectively, the call has been vindicated (Table 2). As BCA foresaw, the harsh monetary medicine administered by the Volcker Fed to slay the inflation dragon has paid hefty market dividends. Table 2A Great Three Decades For Financial Assets The Trouble With The Austrians For all that BCA achieved ahead of Black Monday, and as correct as our long-term calls from the '80s turned out to be, it must be acknowledged that we missed the boat on getting back into equities after the crash. Part of the miss is understandable: one wouldn't expect the strategist with the most prescient call ahead of a downturn to be the first one to identity the beginning of the subsequent rally. The best investors are the ones with the supplest minds, however, and the BCA archives reveal a bias that may have gotten in the way of embracing more bullish near-term outcomes. To wit, one cannot read the 1988 and 1989 Bank Credit Analysts, and indeed, our original leaders' output, without detecting strong sympathies for the Austrian School of Economics (Box 1). BOX 1 An Austrian's Lonely Lot The Austrian School of Economics most saliently parts company with neoclassical economics in its adamant opposition to government intervention and its fraught relationship with credit. Instead of intervening to counter business cycles, Austrians would prefer to let busts run their course so as to cleanse the economy of the excesses embedded in booms. They occupy the Mellonian, purge-the-rottenness-out-of-the-system end of the continuum in opposition to the Debt Supercycle's unconditional forgiveness. Austrians regard banking and credit with some measure of suspicion, as Austrian Business Cycle Theory holds that artificially low interest rates are the raw material of destabilizing booms. Encouraged by central bankers seeking to steer an economy out of recession with a bare minimum of discomfort, borrowers take on debt to invest in projects that may not be able to pay their own way were it not for intervention. Once rates rise after policy accommodation fades, the economy slows and the extent of the malinvestment is revealed. The Debt Supercycle prescribes more of the hair of the dog to alleviate the suffering from malinvestment. The debt overhang is thereby never eliminated; it instead continues to silt up, requiring larger and larger interventions. Unchecked, the degree of intervention required to keep the plates spinning will eventually exceed capacity. This analysis is logically sound, but it so thoroughly contradicts the reigning orthodoxy that an investor who becomes emotionally invested in it is at risk of serially tilting at windmills. There is nothing wrong with the Austrian School per se. We rather like its outsider status, and actively seek heterodox inputs and perspectives so as to stay out of the ruts of the well-worn consensus path. Even its pessimistic bent has its uses; investors are surely exposed to enough cheerleading. Its prescriptions are so bracing, however, that a little goes a long way and real-world users should handle them with care. A popular pair of You Tube videos of actors portraying Keynes and Hayek issuing dueling raps about their respective ideologies (Keynes: I want to steer markets/Hayek: I want them set free!) provide an entertaining example of the Austrian-inspired investor's dilemma. Keynes, drink after drink in hand, is the exuberant life of the party, while the sallow Hayek stares into the bottom of his glass, unable to capture any other partygoers' attention. The simple conceit animating the video - Keynesianism is fun; Austrians are dour scolds - resonates deeply with elected officials. Voters love free drinks, but hate being told to eat their vegetables. The Austrian School, therefore, is a poor guide to the path that policy is likely to take. It also has the problematic effect of introducing an element of moral judgment into what should be a purely objective sphere. Investors should have a laser-like focus on what is most likely to happen and should strive to suppress extraneous notions about what should happen. The Debt Supercycle is a brilliantly incisive way of viewing the interaction between constituents' desires and officials' incentives, and has predicted the long-run direction of policy to a T. Only someone with a focus on money flows, informed by exposure to Austrian Business Cycle Theory, could have come up with it. In the hands of BCA editors in the late '80s, however, it seemed to feed a desire to see the American economy get its comeuppance. Setting aside that desire for punishment - and value judgments altogether - is the clearest way that we could have done better in the aftermath of the crash 30 years ago, when BCA essentially sat out the December '87 - July '90 equity bull market. We should strive to be dispassionate and unbiased observers of the economy and markets. After all, the process illustrated by the Debt Supercycle concept has surely helped put the wind at equities' back throughout the postwar era (Chart 7). Making sense of it without decrying it could help us to provide even better counsel. Chart 7Equity Investing Is An Optimists' Game Then And Now Does 2017 look like 1987? Is another crash lurking just around the corner? Our answers are "no," and "no." We think the resemblances between then and now are merely superficial. The good news is that the probability of a Black Monday-style crash is remote, and we think that even a run-of-the-mill bear market is not likely until our most reliable recession leading indicators, which are still dormant, begin to flash red.5 While that view may come as a short-term relief, 1987's long-term market outlook was vastly superior. While both today's bull market and the '82-'87 bull market began with forward earnings multiples at multi-year lows, the trough multiple in 1982 was in the low sixes, nearly two standard deviations below the mean (Chart 8). Even though it more than doubled by the August '87 peak, it only just reached what is now the mean level for the entire series. This bull market has seen the S&P 500's forward multiple rise to a full standard deviation above the mean. Valuation is not everything, of course. It is a lousy short-term indicator and only issues a reliable intermediate-term signal at extremes. Long-term returns correlate closely with the cyclically-adjusted P/E ("CAPE"), however, and it is currently at levels only previously reached ahead of the 1929 and 2000 peaks (Chart 9). The frothy CAPE portends a tepid long-run U.S. equity outlook. Chart 8Not A Lot Of Room To Grow Chart 9Not The Stuff Of Secular Rallies Both of the bull markets emerged from the ashes of nasty recessions (Chart 10), but the periods' primary economic threats were polar opposites, as were the policy settings adopted to counteract them. The Volcker Fed tightened monetary conditions to the point of pain in the early '80s, plunging the economy into a double-dip recession for the express purpose of eradicating the scourge of double-digit inflation (Chart 11). After the financial crisis, on the other hand, the clear and present danger was the potential for the credit bust to trigger a deflationary spiral. The Bernanke Fed pursued unprecedentedly accommodative policy in response. Chart 10Similarly Nasty Recessions ... Chart 11... But Opposite Inflation Backdrops The policy measures of the early '80s were an example of swapping near-term pain for long-term gain, and they set the stage for secular rallies in financial assets that continue to this day. Once inflation was removed from the equation, interest rates had to fall, and they did so for 35 years. The extraordinary accommodation in the wake of the crisis was an attempt to stave off hysteresis, which boils down to mitigating near-term pain as an insurance policy against long-term pain.6 It may well have worked, but there is no such thing as a free lunch, and the Fed's exertions have likely pulled forward much of the bond and stock markets' future returns. Black Monday And The Fed Put Before the October 20th open, the Fed issued the following statement: The Federal Reserve, consistent with its responsibilities as the Nation's central bank, affirmed today its readiness to serve as a source of liquidity to support the economic and financial system. Although it was only 30 words long, the statement packed a punch. It signaled the Fed's willingness to fulfill its function as the lender of last resort and may also have prodded skittish banks into fulfilling their responsibilities as intermediaries. Behind the scenes, the Federal Reserve Banks of New York and Chicago were doing their utmost to keep the system functioning. New York Fed president Corrigan was twisting lenders' arms to keep credit flowing so the crash would not infect the banking system and the real economy.7 Meanwhile, the Chicago Fed wasn't letting the letter of the law keep it from "help[ing to] engineer a solution" when one of the biggest derivatives market participants "ran short of cash.8" The statement, and the vigorous offstage exertions, countered the Fed's determinedly low profile. These were the days, after all, when monetary policy actions were still regarded as something akin to state secrets. Wall Street firms employed "Fed watchers," who were charged with studying the tea leaves to determine if the Fed had adjusted policy. As late as January 1990, the Bank Credit Analyst could devote an entire Section III to the question, "Has the Federal Reserve Eased?" Some of Alan Greenspan's comments in his memoir may reflect after-the-fact boasting or burnishing, but Black Monday can be viewed as a policy watershed. After it, the Fed's conduct of monetary policy has become transparent to the point of oversharing. More meaningfully for investors, it marked the origin of the "Greenspan Put," the widespread notion among market participants that the Fed would do its best to ward off or mitigate financial market downdrafts. Are ETFs The New Portfolio Insurance? Responsibility for the crash cannot be precisely apportioned among factors, but all post-mortem analyses agree that portfolio insurance played a leading role. While it may well have proven harmless if pursued on a modest scale by a limited number of players, it morphed into a destabilizing force once a critical mass of investors embraced it. On Black Monday, it became a paradox of safety akin to the paradox of thrift: prudent and rational when practiced by one individual, but a metastasizing disaster when followed by a crowd. A reasonable roadmap for someone trying to spot parallels between then and now is to identify market products that may have become overly popular. Wall Street's tendency to wring every last drop out of financing innovations, coupled with investors' tendency to move in herds, can lead to excesses. The latest innovation to achieve wild popularity is the ETF. Is it possible that ETFs could exert the same destabilizing influence as portfolio insurance if investors' ardor for them suddenly cools? We think not. As our Global ETF Strategy service has argued, the claims about passive investing's dangers are overheated.9 The notion that index tracking is undermining price discovery disregards the power of incentives. Passive investing strikes us as the best cure for passive investing: if so many people are pursuing it that index-trackers begin to drown out active investors, the prospective returns to active investing will soar and money will rotate out of index-tracking strategies in sufficient quantity to correct the imbalance. Chatter about a passive bubble also fails to consider the source of fund flows into index-tracking ETFs. The oft-repeated statement, "so much money is flowing into ETFs that it's distorting prices across the board," does not hold up to scrutiny. Away from Japan and Switzerland, where QE purchases of ETFs are being funded with new yen and franc notes, ETFs are not being purchased with new investment capital that has materialized out of thin air. They are being purchased with existing investment capital that has merely been reallocated away from actively managed mutual funds (Chart 12). Chart 12Mirror Image Bubbles are always the result of speculative, excess-profit-seeking activity. Index-tracking ETFs are vehicles intended to deliver market returns. They are the opposite of a get-rich-quick scheme; they're the instrument investors turn to when they give up on quick riches. We do not worry that ETFs are the object of a bubble, or that they are in any way analogous to portfolio insurance in the fall of 1987. Investment Implications Black Monday was a one-off event that remained contained within the financial markets despite widespread fears that it would spread to constrict the broader financial system and the real economy. A lot has changed in 30 years, but the collision of algorithms, derivatives and global pressures squarely places it in our time. It is entirely possible that its elements could come together to create another massive single-day drop. A key difference between future single- or intra-day swoons, and the ones that have already occurred since the crisis, is that they will arrive while the Fed is tightening policy at the margin. The future swoons, then, may not be as likely to disappear quickly without leaving much of a mark. It may go too far to say that market infrastructure is vulnerable, but it would be too optimistic to assume that it has kept pace with the advances in rapid-fire trading and the increasing prevalence of algorithms. It may make sense for investors with less tolerance for risk to maintain an extra cash buffer to protect against swoons and to ensure that they have dry powder to exploit them when they materialize. We remain constructive on the global economy, however, and our house view recommends overweighting risk assets while maintaining below-benchmark duration within bond portfolios. We sympathize with investors who lament that nothing in the public markets is cheap, but synchronized global acceleration remains intact. None of our models are warning of imminent danger. We therefore remain fully invested but vigilant, seeking out signs that the long bull market may be running out of steam. After reviewing our shortcomings in the aftermath of Black Monday, however, we will seek with an open mind and will not attenuate our efforts by awaiting the rapture of a final reckoning, when the sheep and the goats will be separated according to their virtue. The whole point of policy makers' efforts to engineer a rising tide is to keep the goats, and the broader economy, from harm. Doug Peta, Senior Vice President Global ETF Strategy dougp@bcaresearch.com 1 Except in New Zealand, where Black Tuesday popped a bubble of such notable excess that the MSCI New Zealand Index today trades at less than two-thirds of its September 1987 high, and Japan, where the mania lasted until December 1989 and the MSCI Japan Index is still nearly 40% below its all-time high. 2 Index arbitrageurs would have followed the same pattern, but they were sidelined by delayed price quotes and the failure of the NYSE's automated order execution system, which kept them from accurately identifying and exploiting true arbitrage opportunities. 3 Portfolio insurance was no secret - it was estimated that $90 billion of assets were following the strategy - and its potential to amplify selling pressures in a vicious circle had been the subject of a widely followed Wall Street Journal column published a week before the crash. 4 Lefevre, Edwin. Reminiscences of a Stock Operator, John Wiley & Sons, Inc.: Hoboken (NJ), pp. 57-8. Until 1997, the prices of NYSE-listed stocks were quoted in eighth-of-a-dollar increments. 5 For details on the interaction between recessions and equity bear markets, please see the August 16, 2017 Global ETF Strategy Special Report, "A Guide to Spotting and Weathering Bear Markets," available at etf.bcaresearch.com. 6 Hysteresis is the process by which a negative cyclical phenomenon, if left unchecked, can evolve into a secular phenomenon. 7 Greenspan, Alan. The Age of Turbulence: Adventures in a New World, Penguin (New York): 2007, p.108. Greenspan disavowed knowledge of the details, but suggested that Corrigan, "the Fed's chief enforcer," "bit off a few earlobes" while encouraging bankers to keep in mind that, "'if you shut off credit to a customer just because you're a little nervous about him, but with no concrete reason, he's going to remember that'." 8 Greenspan, p. 110.

Special Report

Highlights Since the release of our currency hedging report on September 29, 2017,1 we have received an overwhelming positive response from clients around the globe. We thank our clients for their appreciation of our research. Instead of answering client requests individually, we have decided to publish this follow-up report, in which we apply the same methodology to analyze both static and dynamic hedging strategies to hedge a global equity portfolio for the remaining three home currencies (Swiss franc, Swedish krona and Norwegian krone) in our nine-currency global equity universe. For investors based in Switzerland and Sweden, BCA's dynamic hedging framework, based on the proprietary currency indicators from BCA's Foreign Exchange Strategy (FES) service,2 has also outperformed all the static hedging strategies on a risk-adjusted basis since 2001. For Norway-based investors, however, BCA's dynamic hedging strategy does not generate consistently superior performance. Using static hedging, we find that the Swiss franc, together with U.S. dollar and Japanese yen, maintain their "safe-heaven currency" status, in the sense that CHF-, JPY- and USD-based investors should fully hedge foreign-currency exposure to minimize risk. However, our proposed dynamic hedging can achieve a better return/risk profile with less than 100% hedging. Over a four-year moving performance cycle (in line with how most portfolio managers are evaluated), BCA's dynamic hedging adds little career risk to portfolio managers in Switzerland and Sweden, compared to the "least regret" 50% static hedging, but the same cannot be said for Norwegian PMs. We recommend global equity investors based in the U.S., U.K., euro area, Japan, Canada, Australia, Switzerland and Sweden to use the BCA dynamic hedging framework to manage their foreign currency exposure. For Norwegian investors, we suggest "the least regret" 50% static hedging. Feature Dynamic Hedging Vs. Static Hedging We apply the same methodology as described in the previously published Special Report 3 to hedge an identical global equity portfolio into CHF, SEK and NOK using static and dynamic hedging strategies. As shown in Chart I-1, BCA's dynamic hedging strategy, based on the proprietary Intermediate-Term Timing Model (ITTM)4 indicators from the Foreign Exchange Strategy service, outperforms all static hedging strategies on a risk-adjusted basis for the CHF and SEK portfolios, in line with our findings for the other six home currencies. However, the same is not true for the NOK portfolio. Chart I-1Identical Investment, But Different Risk/Return Profiles The Swiss Perspective: On a static-hedging basis, the Swiss franc holds its "reserve currency" status as classified by Campbell et al,5 in the sense that risk-minimizing Swiss-based investors should fully hedge foreign currency exposure. Unlike the other two "safe-haven" home currencies, the USD and JPY, for which a higher hedge ratio results in lower risk and lower return in both the 16-year period from 2001 and the 41-year period form 1976, the CHF-based portfolio has achieved higher return/lower risk in the 16-year period from 2001 as the hedge ratio increases. The ITTM-based dynamic hedging outperforms the best static hedging (100%) in the shorter period, but the simple momentum-based dynamic hedging is inferior to the best static hedging (90%) in the longer period (Chart I-1, top two graphs and Tables II-1 and II-2). Chart I-2Little Career Risk For Swiss ##br##And Swedish Portfolio Managers The Swedish Perspective: On a static-hedging basis, the SEK-based portfolio behaves in a similar way to the euro-based portfolio in both the shorter and longer periods. In the shorter period from 2001, a higher hedge ratio results in higher returns, albeit gradually, but risk decreases until the hedge ratio hits 30% and then starts to increase such that the full hedge has the highest risk. In the longer period from 1976, a higher hedge ratio results in a lower return, while risk decreases until the hedge ratio hits 70% and then starts to rise, such that the unhedged portfolio has the highest risk and the fully hedged portfolio has the lowest return. On a risk-adjusted basis, the best static hedge ratio is 50% for both the shorter and longer periods. Both the ITTM-based dynamic hedging and the momentum-based dynamic hedging are superior to the best static hedge ratio of 50% (Chart I-1, middle 2 graphs and Table II-3 and II-4). The Norwegian Perspective: On a static-hedging basis, the NOK-based portfolio behaves like the GBP-based portfolio in the longer period from 1976, with return increasing and risk decreasing as hedge ratio increases, but it behaves like the euro- and SEK-based portfolios in the shorter period from 2001. On a risk-adjusted return basis, both the ITTM-based and momentum-based dynamic hedging strategies underperformed the best static hedge which is about 80% hedged (Chart I-1, bottom 2 graphs and Tables II-5 and II-6). Little Career Risk for Swiss and Swedish Portfolio Managers: As shown in Chart I-2, on a rolling four-year basis, the ITTM-based dynamic hedging strategy has outperformed the best static hedging strategy for CHF portfolio (which is 100%) and the best static hedging strategy for SEK portfolio (which is 50%). For the NOK portfolio, however, neither the ITTM-based dynamic strategy, nor the "best static hedging" strategy (which is 80%) can consistently outperform the "least regret" 50% hedging strategy. Equal Playing Field: In theory, if hedges were effective, then an identical global investment should have similar returns for all investors, no matter which home currency they hold. While neither the static hedging strategies nor the momentum-based dynamic hedging approach pass this criteria, BCA's ITTM-based dynamic hedging approach has indeed achieved this: it levels out the playing-field for all investors globally. As shown in Chart I-3, in the period from March 2001 to August 2017, if left unhedged, the same global investment exhibits very different annualized returns for investors in different home currencies, with CHF investors at the low end at around 2.8%, and GBP investors at the high end at around 7%. With BCA's ITTM-based dynamic hedge, however, returns for all investors are similar, no matter which currency is their home currency. Chart I-3BCA Dynamic Hedging Strategy Levels Out The Playing Field Bottom Line: We have back-tested the efficacy of BCA's proprietary currency indicators from the Foreign Exchange Strategy team's Intermediate-Term Timing Models to dynamically hedge a global investment portfolio into nine different home currencies. These indicators have proven to add significant value to eight out of the nine home currencies. Granted, back-tests show good results by default. But our FES team will strive to ensure that these indicators continue to work well going forward. We recommend global equity investors based in the U.S., U.K., euro area, Japan, Canada, Australia, Switzerland and Sweden to use BCA's ITTM currency indicator-based dynamic hedging framework to manage their foreign currency exposure. For Norway-based global equity investors, we suggest the "least regret" 50% static hedging. Xiaoli Tang, Associate Vice President xiaolit@bcaresearch.com Appendix 1: Dynamic Hedging For Three Home Currencies 1.1 The Swiss Perspective Correlations: For Swiss investors, foreign currencies in aggregate have generally been positively correlated with foreign equities and domestic equities (Chart II-1). In addition, the Swiss franc has strengthened over time, especially after 1999. This explains why, on a static basis, the fully hedged portfolio generates the best risk/return profile. (Table II-1 and Table II-2). Chart II-1Swiss Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Table II-1Risk/Return Profile For Global Equities In CHF (2001-2017) Table II-2Risk/Return Profile For Global Equities In CHF (1976-2017) Historical Performance: Since 2001, ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in CHF. The risk is slightly higher than the best static hedging (which is 100%), but the return is over 200 bps higher, resulting in a 40% increase in the risk-adjusted return (Table II-1). In addition, this is achieved with far fewer hedging transactions than the fully hedged strategy as shown in Chart II-2 panel 2. Over the longer period from 1976, the optimal static hedge ratio is about 90%, almost fully hedged as well, as shown in Table II-2. Chart II-2Swiss Perspective: Dynamic Vs. Static Hedging On a 60-month rolling basis, as shown in Chart II-2, the ITTM-based dynamic risk/return profile also prevails. Current State: Currently our indicators show that Swiss investors should not hedge any foreign currency. Chart II-3 shows how the Swiss investors should have hedged their exposure in U.S. dollar. Chart II-3Swiss Perspective: MSCI U.S. Index Dynamically Hedged 1.2 The Swedish Perspective Correlations: For Swedish investors, foreign currencies in aggregate have little correlation with domestic equities as the average correlation from 1980 is almost 0. This overall average can be misleading, however, as evidenced by the rolling 60-month correlation, which was positive before 1998 and then was negative until recently, and is now in the positive territory again (Chart II-4). This is a typical case where dynamic hedging would outperform static hedging, because the latter assumes constant mean and covariance for the chosen time period (Tables II-3 and II-4) Chart II-4Swedish Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Table II-3Risk/Return Profile For Global Equities In SEK (2001-2017) Table II-4Risk/Return Profile For Global Equities In SEK (1976-2017) Historical Performance: Since 2001, ITTM-based dynamic hedging has produced the highest risk-adjusted return in SEK for a global portfolio. The risk profile looks similar to that of the 50% hedged portfolio, but return is much higher, resulting in a 35% increase in the risk-adjusted return (Table II-3). Over the longer period, the optimal static hedge ratio is also 50%, as shown in Table II-4. On a five-year rolling basis, as shown in Chart II-5, the ITTM-based dynamic risk/return profile also prevails. Chart II-5Swedish Perspective: Dynamics Vs. Static Hedging Current State: Currently Sweden-based investors should be hedging only their exposure in Norwegian krona. Chart II-6 shows how the Swedish investors should have hedged their exposure in Canadian dollar. Chart II-6Swedish Perspective: MSCI Canadian Index Dynamically Hedged 1.3 The Norwegian Perspective Correlations: For Norway-based investors, foreign currencies in aggregate have a slightly negative correlation with domestic equities as the average correlation from 1980 is -0.12. This overall average can be misleading, however, as evidenced by the rolling 60-month correlation, which was above this long-run average before the Great Financial Crisis (GFC), but has been in negative territory ever since. On the other hand, the correlations between foreign currencies and foreign equities, and between foreign equities and domestic equities, have also gone though some regime changes (Chart II-7). Chart II-7Norwegian Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Historical Performance: Since 2001, ITTM-based dynamic hedging has produced 7% lower risk-adjusted return for the global portfolio in NOK compared to the best static hedging strategy of 80% (Tables II-5). In the longer period from 1976, the momentum-based dynamic also underperformed the 80% static hedging strategy by 3% on a risk-adjusted return basis (Tables II-6). Table II-5Risk/Return Profile For Global Equities In NOK (2001-2017) Table II-6Risk/Return Profile For Global Equities In NOK (1976-2017) On a five-year rolling basis, as shown in Chart II-8, the ITTM-based dynamic risk/return profile also looks less attractive. Chart II-8Norwegian Perspective: Dynamic Vs. Static Hedging Why does dynamic hedging not work? We do not have a good understanding on this yet. Looking at the individual currency pairs, we notice that our indicators work very well for CAD/NOK, SEK/NOK and JPY/NOK, but not for other pairs, especially during the period between 2011 and 2016 when NOK was strong against most of these currencies. Chart II-9 and Chart II-10 show how JPY/NOK and USD/NOK should have been hedged based on our indicators. The former worked very well, while the latter failed terribly in the period between 2013 and 2016. Chart II-9Norwegian Perspective: MSCI Japanese Index Dynamically Hedged Chart II-10Norwegian Perspective: MSCI U.S. Index Dynamically Hedged 1 Please see Global Asset Allocation and Foreign Exchange Strategy joint Special Report "Currency Hedging: Dynamic Or Static? - A Practical Gide For Global Equity Investors," dated September 29, 2017. 2 Please see Foreign Exchange Strategy Special Report, "In Search Of A Timing Model", dated June 22, 2016 3 Please see Global Asset Allocation and Foreign Exchange Strategy joint Special Report "Currency Hedging: Dynamic Or Static? - A Practical Gide For Global Equity Investors," dated September 29, 2017. 4 Please see Foreign Exchange Strategy Special Report, "In Search of A Timing Model", dated June 22, 2016 5 Campbell, J., K. de Medeiros and L. Viceira, 2010, "Global Currency Hedging," Journal of Finance LXV, 87-122

Special Report

Dear Client, There is no regular report this week. Instead, I am sending you a Special Report written by colleague Mark McClellan, who examines global equity valuations from a bottom-up perspective using our Equity Trading Strategy (ETS) platform. I discussed the intellectual underpinnings for the ETS model in 2015. In addition, if you haven't done so already, please take a moment to listen to our latest webcast, where I survey the global macro landscape, drawing on the material published in our Quarterly Strategy Outlook. Best regards, Peter Berezin, Chief Global Strategist Highlights The performance of Japanese stocks relative to the U.S. has been dismal over the past couple of decades, and the same is true for Europe in the post-Lehman period. However, both the Japanese and European economies are performing impressively this year, profit growth is accelerating and margins are rising. This suggests that there could be some "catch up" for both markets, at least in local-currency terms. Standard valuation measures based on index data also suggest that Eurozone and Japanese stocks are cheap compared to the U.S. Nonetheless, these markets almost always trade at a discount, due to a persistent lackluster profit performance. In this Special Report, we approach the issue from a bottom-up perspective, utilizing the powerful analytics provided by BCA's exciting new Equity Trading Strategy (ETS) platform. The ETS software allows us to compare companies across markets on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach adjusts for structural valuation gaps between these markets and avoids the problems of index construction. Investors can have greater confidence that they will make money on a 12-month horizon by taking a position when the new bottom-up indicators reach +/-1 standard deviations over- or under-valued, although technical information should be taken on board to sharpen the timing. The +/-2 sigma level gives clear buy/sell signals irrespective of fundamental or technical factors. The bottom-up valuation indicators will not replace our top-down versions that are based on index data, but rather will be considered together when evaluating relative value. European stocks are near fair value relative to the U.S. at the moment, while Japan is modestly cheap. We favor the European and, especially, Japanese markets over the U.S., due to policy divergence and the view that EPS has more room to expand in the former two economies. Feature Chart 1European And Japanese Stocks Have Lagged... Japanese equities have been perennial underperformers versus the U.S. for most of the past 2-3 decades in both local- and common-currency terms (Chart 1). The simultaneous bursting of the equity and land bubbles in the 1990s ushered in a prolonged period of deflation in wages and consumer prices. There was a ray of light in the early years of Abenomics, when the aggressive three-arrow approach appeared to be finally lifting the Japanese economy out of a Secular Stagnation. Yen weakness contributed to a surge in earnings-per-share (EPS) in absolute terms and relative to the U.S. Equity multiples rose between 2012 and 2015. Unfortunately, Abe's honeymoon with equity markets faded in 2016 (Chart 2). A bout of yen strength, collapsing inflation expectations, weakening business confidence and a lack of progress on structural reforms caused investors to question the upside potential for Japanese corporate top-line growth. While European indexes have fared better than Japanese stocks relative to the U.S. over the past 25 years as a whole, the post-Lehman period has been particularly tough for European corporate profitability and relative equity market performance. The U.S. total return index has more than doubled its pre-recession peak according to Thomson Reuters/Datastream data, while the Eurozone total return index is only 10% above the previous high-water mark when expressed in U.S. dollars (Chart 2). The yawning return gap between the two equity markets was almost entirely due to earnings as market multiples have moved largely in sync. Earnings-per-share generated by U.S. companies now exceed the pre-recession peak by about 23%. In contrast, earnings produced by their Eurozone peers are a whopping 42% below their peak (common-currency). That said, the earnings backdrop now appears to be shifting. The strengthening global recovery is turbocharging EPS growth in Europe and Japan, where the corporate sector is more leveraged to global growth than is the case in the U.S. Eurozone domestic demand is also hot. Japan is still struggling with deflation, but the economy is performing well and the corporate sector is benefiting from this year's yen pullback. Japanese EPS is surging in both yen and dollar terms. Finally, both Europe and Japan appear cheap versus the U.S. by traditional valuation metrics. Based on index data, these two markets trade at a hefty discount across most of the main valuation measures (Chart 3). This is the case even for normalized measures such as price-to-book. However, these two markets have almost always traded at a discount to the U.S. Chart 2...Due To Depressed Fundamentals Chart 3Europe And Japan Trade At A Discount There are many possible explanations for the persistent valuation gap, including differences in accounting standards, discount rates and sector weights. The wider use of stock buybacks in the U.S. also favors American equity valuations. But most important are historical differences in underlying corporate fundamentals. U.S. companies on the whole have been significantly more profitable over the years based on return on equity and operating margins (Charts 4 and 5). Until recently, U.S. companies have also tended to have lower leverage relative to Europe and Japan, and a higher interest coverage ratio than Europe. Better profitability metrics in the U.S. are not solely an artifact of sector weighting either. Operating margins are lower in Europe and Japan even after applying U.S. sector weights to the other two markets (Chart 6). Chart 4RoE Is Consistently Lower In Japan And Europe Chart 5U.S./Europe/Japan Comparison Chart 6U.S./Europe/Japan Comparison (U.S. Sector Weights) Why the European and Japanese corporate sectors have been profit underachievers is beyond the scope of this paper. U.S. companies reaped most of the benefit from productivity gains over the past 25 years, with the result that the capital share of income soared while the labor share collapsed. European and Japanese companies were less successful in squeezing down labor costs. This raises the question of whether European and Japanese stocks are, in fact, cheap relative to the U.S. Measuring Value Our monthly Bank Credit Analyst publication developed top-down valuation indicators that adjust for different sector weights and persistent differences in the underlying profit fundamentals. These indicators are based on index data, and have a good track record for providing profitable buy/sell signals.1 In this Special Report, we take a bottom-up approach that utilizes the powerful analytics provided by BCA's Equity Trading Strategy (ETS) platform.2 The software allows us to compare companies on a head-to-head basis and rank them based on a wide range of characteristics. The bottom-up approach avoids the problems of index construction when trying to gauge valuation across countries. The web-based platform uses over 27 quantitative factors to rank approximately 10,000 individual stocks in 23 countries, allowing clients to find stocks with winning characteristics at the global level. Users can rank and score individual equities to support a broad set of investment strategies and apply macro and sector views to single-name investments. The ETS approach has an impressive track record.3 Historically, the top-decile of stocks ranked using the "BCA Score" methodology has outperformed stocks in the bottom decile by over 25% a year. The BCA Score includes 27 factors when ranking stocks, including sentiment and momentum. However, since we are interested in developing a valuation metric in this paper, we focus on five valuation measures in the ETS database: trailing P/E, forward P/E, price-to-book, price-to-sales and price-to-cash flow. We combined all of the Eurozone and U.S. companies that have total assets of greater than $1 billion into one dataset. The ETS platform then ranked the stocks from best to worst on a daily basis (i.e., cheapest to most expensive), using an equally-weighted average of the five valuation measures. The average score for U.S. stocks is subtracted from the average score for European stocks, and then divided by the standard deviation of the series. This provides a valuation metric that fluctuates roughly between +/- 2 standard deviations. This approach inherently adjusts for structural valuation gaps. We then used the same methodology to construct bottom-up valuation indicators for Japan relative to the U.S. Chart 7 presents the resulting bottom-up indicators for Europe and Japan, along with our top-down valuation measure. A high reading indicates that European or Japanese stocks are cheap relative to the U.S., while the opposite is true for low readings. Chart 7Top-Down And Bottom-Up Valuation Indicators The underlying bottom-up data extend back to 2000. However, the bursting of the tech bubble in the early 2000's caused major shifts in relative valuation among sectors that skew the indicator when constructed using the entire data set. A cleaner indicator emerges when using only the data from 2005. As with any valuation indicator, it is only useful when it reaches extremes. We calculated the historical track record for a trading rule that is based on critical levels of over- and under-valuation. For example, we calculated the (local-currency basis) excess returns over 3-, 6-, 12- and 24-month horizons generated by (1) overweighting European or Japanese stocks when that market was one and two standard deviations cheap versus the U.S. market, and (2) overweighting the U.S. when the European or Japanese market was one and two standard deviations expensive (Tables 1 and 2). Table 1Eurozone Vs. U.S. Value Indicator: Trading Rule Returns And Batting Average Table 2Japan Vs. U.S. Value Indicator: Trading Rule Returns And Batting Average The trading rule returns are best in the case of Europe when the indicator reached two standard deviations cheap or expensive, providing average returns of almost 11 percent over 12 months. The trading rule returns when the indicator reached +/-1 standard deviation are lower, but still respectable at roughly 3% on 12- and 24-month horizons. The results are even better for the Japan trading rule (Table 2). Excess returns are 14% and 35%, respectively, over 12 and 24-month horizons after the indicator reaches +/-2 standard deviations. The results are very impressive even when using +/-1 standard deviation as the trigger point. Tables 1 and 2 also present the trading rules' batting average. That is, the number of positive excess returns generated by the trading rule as a percent of the total number of signals. For the European indicator, the batting average ranged from 50% on a 3-month horizon to 68% over 12 months when buy/sell signals are triggered at +/- 1 standard deviation. The batting average is much higher (80-100%) using +/- 2 standard deviations as a trigger point, although there were only five months over the entire sample when the indicator reached this level. The batting average is even better for the Japanese indicator. Sharpening The Buy/Sell Signals We then augmented the valuation analysis by adding information on company fundamentals, such as EPS growth and profit margins, among others. The ETS software ranked the companies after equally-weighting the valuation and fundamental factors. However, this approach yielded poor results in terms of the trading rule. This is because, for example, when European stocks reached undervalued levels relative to the U.S., it is usually because the European earnings fundamentals have underperformed those of the U.S. companies. Thus, favorable value is offset by poor fundamentals when scored by the ETS model, muddying the message provided by valuation alone. We also tried including some technical indicators to see if they could add information on timing. Chart 8 compares the valuation indicator discussed above to an enhanced indicator that includes both value and technical factors. Tables 3 and 4 provide the excess returns and batting averages for a trading rule based on the enhanced indicator. Chart 8Bottom-Up Indicators: Value, And Value Plus Technical Table 3Eurozone Vs. U.S. Value And Technical Indicator: Trading Rule Returns And Batting Average Table 4Japan Vs. U.S. Value And Technical Indicator: Trading Rule Returns And Batting Average It turns out that including some technical information does add value, but only in the case of Europe when using +/-1 standard deviation as the trigger point for trades. Both the excess returns and batting average to the trading rule improve. However, this is not the case when using +/-2 sigma. In the case of Japan, including technical information detracts from excess returns for both trigger points. Investment Conclusions Our new ETS platform provides investors with a unique way of picking stocks by combining top-down macro themes with company-specific information. It also allows us to develop valuation tools that avoid some of the pitfalls of index data by comparing stocks on a head-to-head basis. Investors can be fairly confident that they will make money on a 12-month horizon by taking a position when the bottom-up valuation indicators reach +/-1 sigma over- or under-valued. The +/-2 sigma valuation level gives clear buy/sell signals irrespective of fundamental or technical factors for both Europe and Japan. The bottom-up valuation indicators will not replace our top-down versions, but rather will be considered together when evaluating relative value. At the moment, both the top-down and bottom-up versions suggest that European stocks are roughly fairly valued relative to the U.S. market. Japanese stocks are on the cheap side based on both indicators, but neither one exceeds +1 sigma. This means that investors cannot make the allocation decision based on value alone. Valuation indicators need to be at extremes to have any predictive power. Our global equity strategists recommend overweighting Eurozone stocks versus the U.S. on a currency-hedged basis, although not because of valuation. On the plus side, the economy is flying high and there are no warning signs that this is about to end. There is hope for structural reform in France after Macron's election win this year. We give Macron's proposed labor market reforms high marks. Many doubt that these reforms will see the light of day, but our geopolitical team believes that investors are underestimating the chances. The German election in September poured cold water on recent enthusiasm regarding accelerated European integration. This is because Merkel will likely have to deal with a larger contingent of Euroskeptics in the grand coalition that emerges in the coming months. However, we do not expect political developments in Germany to be a headwind for the Eurozone stock market. On the negative side, this year's euro bull phase will take a bite out of earnings. Euro strength so far this year will lop three to four percentage points off of EPS growth by the middle of next year. Our model suggests that this will be overwhelmed by the robust economic expansion at home and abroad, but profit growth will diminish heading into year-end and will likely trail that in the U.S. and Japan over the next six months (local-currency basis). Still, a lot of the negative impact of the currency on profits may already be discounted. The bullish case versus the U.S. is more compelling for the Nikkei, at least in local-currency terms. Valuation is modestly attractive and Japanese earnings are highly geared to economic growth at home and abroad. Japanese EPS is in an uptrend versus the U.S. in both local and common currencies. We do not expect to see a peak in EPS growth until mid-2018, a good six months after the expected top in the U.S. Moreover, an Abe win in the October 22 election would mean that policy will remain highly reflationary in absolute terms and relative to the U.S. However, overweight positions in both the European and Japanese bourses should be currency hedged because the dollar is likely to appreciate over the next 6-12 months due to monetary policy divergences. Mark McClellan, Senior Vice President The Bank Credit Analyst markm@bcaresearch.com 1 Please see The Bank Credit Analyst Special Report, "Are Eurozone Stocks Really That Cheap?" dated July 2016. 2 Please see Equity Trading Strategy Special Report, "Introducing ETS: A Top Down Approach To Bottom-Up Stock Picking," dated December 2, 2015. 3 For more information, please see Equity Trading Strategy Special Report, "Making Money with ETS," dated January 20, 2016 Strategy & Market Trends Tactical Trades Strategic Recommendations Closed Trades

Highlights It is often argued that the U.S. dollar is expensive, but models do not offer a unanimous picture. The U.S. current account, exports share, and cyclical inflation do not point to an obvious dollar overvaluation either. Without a clear valuation signal, the dollar will continue to trade off rate differentials. An increasing body of evidence points toward a rebound in U.S. inflation. As such, U.S. rates are likely to move up relative to the rest of the world, lifting the USD over the next 12 months. Feature We are sending you a shorter regular bulletin this week as we are also publishing a follow up to our joint Special Report titled, "Currency Hedging: Dynamic Or Static? - A Practical Guide For Global Equity Investors," released with the Global Asset Allocation team two weeks ago. In this follow-up, my colleague Xiaoli Tang expands on the same methodology, testing various FX-hedging strategies for international investors - but this time looking at portfolios based in the CHF, the SEK, and the NOK. In this week's regular bulletin, we take a closer look at the U.S. dollar's valuations. The consensus view is that the dollar is expensive. We explore how this claim stacks up against the facts. At this juncture, the U.S. economy is not exhibiting some of the key consequences typical of an economy burdened by an expensive currency. Valuation Models The main argument used by some investors to show that the U.S. dollar is expensive is the traditional purchasing power parity model. This indicator does indeed flag a large 17% overvaluation for the greenback (Chart I-1). However, this is only one metric based on producer price indices. We also like to look at measures that focus on the true determinant of competitiveness: the cost of labor. When we deflate the U.S. dollar's exchange rate using unit labor costs, the dollar is neither a screaming sell nor a screaming buy. It is in line with its long-term average (Chart I-2). The same IMF real effective exchange rate model based on unit labor costs also shows the euro as fairly valued. Thus, on this metric, valuations do not seem to provide a compelling argument to go long or short the dollar, which challenges the universally bearish take on the dollar's perceived overvaluation. Chart I-1An Argument For An###br## Expensive USD Chart I-2But Not All Valuation Approaches ##br##Are That Clearcut We can also double-check the result of this metric using our own long-term fair value model, which incorporates long-term relative productivity trends. This model tries to capture the so-called Balassa-Samuelson effect. This effect is an empirical observation that countries with superior long-term labor productivity trends tend to experience a secular upward bias on their real exchange rates. The perceived overvaluation of the U.S. dollar may in fact be an illusion, because when the Balassa-Samuelson effect is taken into account, the dollar currently trades in line with its fair value (Chart I-3). Chart I-3Another Global Approach With USD At Fair Value Bottom Line: Valuing currencies is always an exercise to be approached with plenty of circumspection. It is easy to look at simple PPP models and argue that the dollar looks very expensive. However, when one takes into account labor market costs and productivity trends, the dollar seems fairly valued. A Look At The Symptoms Chart I-4The U.S. Current Account##br## Shows Little Dollar Strain Models are only as good as their inputs. It is important to try to corroborate their insights with economic reality. An expensive currency should produce three major outcomes: the country's current account position should be deteriorating, its market share of global exports should be falling, and it should be experiencing deep deflationary pressures relative to the rest of the world. Let's begin with the current account. Despite a 17% increase in the U.S. dollar since 2014, the U.S. current account has remained stable (Chart I-4). It is undeniable that this reflects an improvement in the energy trade balance of the U.S., itself a byproduct of the shale revolution. Nonetheless, it also highlights that there is little balance-of-payments strains in the U.S. In fact, the move away from energy imports in itself should point to a higher level of equilibrium for the dollar. The export share of the U.S. also does not point to too much stress created by the dollar bull market. As Chart I-5 illustrates, in contrast to the early 1980s or late 1990s-early 2000s, U.S. exports has been faring well when compared to the rest of the world. This exercise needs to be conducted by comparing U.S. exports to the rest of the world excluding China. China has been grabbing global market share from everyone for 30 years. As an aside, the continued rise of China, as well as its still-large current account surplus of more than US$155 billion, supports the idea that the RMB is indeed cheap and remains attractive on a long-term basis - a message also flagged by our long-term fair value model for the CNY (Chart I-6). Chart I-5Growing U.S. Market Share Chart I-6The Yuan Is Clearly Cheap Finally, there is little evidence that the U.S. dollar is depressing U.S. inflation on a cyclical basis. Changes in financial conditions can temporarily redistribute inflationary pressures between the U.S. and the rest of the world, but an expensive dollar should depress U.S. inflation for an extended period of time on a global relative basis. An expensive U.S. dollar makes the U.S. uncompetitive, and should force some degree of internal adjustment on the U.S. economy. So far, the two-year moving average of U.S. core inflation relative to the OECD does not show the same kind of swoon as in the 1980s or late 1990s. In fact, even after this year's inflation slowdown in the U.S., American inflation remains in an uptrend relative to the rest of the OECD (Chart I-7). One source of worry remains the U.S. net international investment position (NIIP). The U.S.'s NIIP currently stands at -41% of GDP, and despite stabilizing for the past two years, has been in a pronounced downtrend over the past 35 years. Historically, countries like Switzerland or Japan with strong NIIPs have tended to experience long-term upward pressure on their exchange rates, while those with poor NIIPs such as South Africa tend to experience negative secular trends, even in real terms. For the time being, what keeps the negative impact of the NIIP on the USD at bay is that the U.S. continues to earn a positive net income - despite negative net assets abroad (Chart I-8). This reflects the willingness of investors to hold the U.S. dollar for its reserve currency status. For the time being, with a lack of alternative to challenge the U.S. dollar's reserve status, the NIIP should not represent a key hurdle for a few more years. Chart I-7The U.S. is Not Experiencing##br## An Internal Devaluation Chart I-8The Exorbitant ##br##Privilege Bottom Line: The U.S. economy is currently exhibiting few of the signals that would be associated with an expensive dollar: the current account remains well behaved, the country is not losing export market shares to its main competitors, and U.S. inflation remains well behaved relative to the rest of the OECD on a cyclical basis. A key risk remains the U.S.'s net international investment position, but so long as the USD can maintain its unchallenged role as the key reserve in the global financial system, the U.S. is likely to continue to run an income surplus vis-à-vis the rest of the world. So What? When it comes to the FX space, long-term valuations only become binding constraints when they are in the extreme. Right now, there is enough conflicting evidence to suggest that if the dollar is indeed expensive, it is not expensive enough to flash a bright sell signal. In this case, the U.S. dollar's dynamics are likely to be dominated by interest rate differentials. Interest rate curves outside of the U.S. seem currently fairly priced, but this is not the case in the U.S. Thus, with only two full hikes priced in over the next 24 months, one needs to see upside for U.S. interest rates if one is to be bullish on the greenback. Despite last month's very poor employment numbers, a consequence of hurricanes Harvey and Irma, the labor market remains strong enough to justify the Federal Reserve's desire to hike rates. The ISM surveys also remains very strong, with the headline numbers and new order components pointing toward robust growth. The only factor that could impede the Fed is inflation. On this front, we remain optimistic that inflation will not deteriorate much further and that, in fact, it is likely to pick up over the next six months, giving the Fed a green light to increase rates in line with its own forecast: First, in the past, we have highlighted that velocity of money - based on the money of zero maturity and nominal GDP - has been a very reliable leading indicator of inflation over the past 20 years, and is pointing toward a rebound in core inflation measures toward year-end.1 Moreover, the easing in U.S. financial conditions over the past 18 months also points toward upside risks to both U.S. growth and inflation. Second, the strength in the Prices-Paid component of both ISM surveys further increases our optimism. Moreover, the recent vigor of the Supplier Delivery subcomponent - a measure of bottlenecks in the system - also points to pipeline inflationary pressures. It is true that some of the recent spike is most likely skewed by the devastating impact of the hurricanes, but this improving trend began much earlier this year. Historically, a combined improvement in both the Prices-Paid and the Supplier Delivery components of the ISM survey tends to provide long leads on core inflation (Chart I-9). Third, the New York Fed has recently started publishing an underlying inflation trend estimate. This measure has also been rebounding sharply, hitting its highest level in 10 years, also pointing toward higher core inflation (Chart I-10). Chart I-9Pipeline Inflationary Pressures##br## Are Growing In The U.S. Chart I-10Underlying Inflationary ##br##Pressures Are Growing Fourth, the behavior of inflation itself is somewhat encouraging. While the recent core PCE year-over-year numbers have been disheartening, the three-month annualized rate of change has picked up robustly. Historically, this has also led to turning points in the year-on-year number (Chart I-11). Finally, there are signs of underlying vigor in wages. Last week's U.S. average hourly earnings number clicked in at 2.9%.It was likely overinflated by the effect of the hurricanes, which have temporarily dropped workers in low-paid industries out of the sample used by the U.S. Bureau of Labor Statistics to compute this data. However, the median average hourly earnings across the key sectors covered by the BLS has been in an uptrend since the beginning of the year (Chart I-12), pointing to some faint but real early signs of rising underlying wage growth. Moreover, while much ink has been spilled regarding whether or not the Philips curve is flat, there remain a well-defined tight relationship between the U.S. employment cost index (ECI) and the level of employment-to-population ratio in the U.S. (Chart I-13). Our view that employment growth will likely continue to tick in north of 120,000 jobs for the next 12 months, implies further improvement in the employment-to-population ratio, and thus a growing ECI. This will both support household income and consumption as well as our inflation view. Chart I-11Sequential Inflation Pointing ##br##To A Turning Point Chart I-12Cross-Sectional Median ##br##Of Wages Improving Chart I-13The Cross-Sectional Median##br## Of Wages Improving Bottom Line: With no clear message from long-term valuation, the key driver of the dollar is likely to remain interest rate differentials. At this point, U.S. interest rates need U.S. inflation to be able to rise by more than what is implied in the OIS curve and lift the dollar. Signs continue to accumulate that U.S. inflation is likely to turn the corner over the next six months, thanks to an easing in U.S. financial conditions and the pick-up in the velocity of money: the Prices-Paid and Supplier Deliveries components of the ISM have hooked up significantly, the NY Fed's underlying inflation measure is strong, the sequential growth rate in core inflation is improving, and there are growing signs that wage growth in the U.S. is picking up. Mathieu Savary, Vice President Foreign Exchange Strategy mathieu@bcaresearch.com 1 Please see Foreign Exchange Strategy Weekly Report titled "Fade North Korea, And Sell The Yen", dated August 11, 2017, or Foreign Exchange Strategy Weekly Report, titled "Conflicting Forces For The Dollar", dated September 8, 2017, available at fes.bcaresearch.com Trades & Forecasts Forecast Summary Core Portfolio Closed Trades

Highlights Year One Performance: The GFIS recommended model bond portfolio returned 1.1% (hedged into USD) in its first year of existence, slightly underperforming the custom benchmark index by -2bps. Our bearish duration tilts were a drag on performance, while our overweights to U.S. corporate debt were a major contributor. Risk Management Lessons: The maximum overweight to low-beta, but low-yielding, Japanese Government Bonds was a drag on performance by reducing the portfolio yield. This highlights the classic bond management trade-off between controlling portfolio risks, like duration or tracking error, and maximizing sources of return, like interest income. Future Drivers Of Returns: Over the next 6-12 months, we expect the model portfolio returns to again benefit mostly from our below-benchmark duration stance (as global bond yields grind higher) and from our overweight stance on U.S. corporates (as the U.S. economy maintains a solid pace of growth). Feature In September of 2016, we introduced a new element to the BCA Global Fixed Income Strategy (GFIS) service - our recommended model bond portfolio.1 This represented a bit of a departure from the usual macroeconomic analysis and forecasting of financial markets that has been the hallmark of BCA. Yet we felt that it was important to add an actual portfolio, with specific allocations and weightings, given the needs and constraints faced by our readers. With so many of our clients being traditional fixed income managers (or multi-asset managers) who measure investment performance versus benchmark indices, we felt that it was important to have a way to communicate our views within a framework akin to what they deal with each day. Even for clients who are not professional bond managers, the model portfolio can be useful as a way to express how much we prefer one bond market (or sector) versus others. It also gives us a forum to discuss portfolio management issues as an addition to the macro analysis. So far, the reception from clients to this new addition to the GFIS service has been a warm one, and we look forward to additional feedback in the months and years ahead. With the model portfolio just passing its first birthday, we are dedicating this Weekly Report to an overview of the final Year One performance numbers. We will evaluate our winning and losing recommendations, look back at the lessons learned as the model portfolio framework has evolved, and identify what we expect will be the biggest drivers of performance in Year Two based on our current views. Year One Model Portfolio Performance: Winners & Losers Chart 1GFIS Model Portfolio Performance The GFIS model portfolio produced a total return of 1.09% (hedged into U.S. dollars) over first full year since inception on September 20, 2016 (Chart 1). This essentially matched the performance of our custom benchmark index, with the model portfolio lagging by a mere -2bps.2 In terms of the breakdown between government bonds and credit (spread product), the former underperformed the benchmark by -18bps while the latter outperformed by +16bps. A more traditional period to evaluate investment performance is on a calendar year-to-date basis. We also show the 2017 year-to-date (YTD) numbers in Chart 1, measured from January 1st to October 3rd. Over that time period, the total returns are much higher - the model portfolio has returned 2.78%, lagging the index by -6bps. This higher absolute return is mostly due to the strong outperformance of corporate bond markets and the decline in government bond yields seen since March. Broadly speaking, that breakdown of returns lines up with what were our largest strategic market calls: to be underweight overall portfolio duration and overweight U.S. corporate bond exposure (bottom panel). This is obviously a welcome property to see in our returns, which we hope will always line up with our desired tilts! When looking at the detailed decomposition of the returns on the government bond side of the portfolio (Table 1), however, a few points stand out: Table 1A Detailed Breakdown Of The GFIS Model Portfolio The underperformance on the government bond side of the portfolio (Chart 2) came from underweight positions at the long-end (maturities beyond seven years) of yield curves in the U.S. (-4bps), U.K. (-5bps), Germany (-5bps) and, most notably, France (-18bps). Chart 2GFIS Model Portfolio Government Bond Performance Attribution By Country The underweight position in Italy, across the curve, generated another -7bps of underperformance, although this was paired against an overweight to Spanish government bonds that positively contributed to returns (+3bps). Overweights to bonds in the middle and shorter ends of yields curves (maturities less than seven years) positively contributed to returns in the U.S. (+6bps), Germany (+2bps) and France (+2bps). Our significant overweight to Japanese government bonds, intended as a way to reduce portfolio duration by increasing exposure to a market with a low beta to global bond yields, also helped boost performance (+8bps). The conclusion? By concentrating our recommended duration underweights on longer-maturity bonds, and raising the weightings on shorter-maturity government debt, we imparted a bearish curve steepening bias on top of the reduced duration exposure. It is no surprise that our recommended government bond allocations underperformed during the bull-flattening move in global yield curves seen earlier this year. By contrast, the returns on the credit (spread) product allocations within the GFIS model portfolio tell a more positive story (Chart 3): Chart 3GFIS Model Portfolio Spread Product Performance Attribution The outperformance came from our overweight allocations to U.S. Investment Grade (IG) corporate debt, focused on Financials (+14bps) and Industrials (+4bps), and U.S. High-Yield (HY), concentrated on Ba-rated (+13bps) and B-rated (+8bps) bonds. U.S. Mortgage-Backed Securities (MBS) were a laggard during the first year of the model bond portfolio (-12bps), which largely came from an ill-timed tactical move to overweight in the 4th quarter of 2016. More recently, our underweight stance on MBS has been only a modest drag on the total return of the portfolio since the peak in U.S. bond yields back in March. Our decisions to reduce exposure to Euro Area IG (-5bps) and HY (-2bps) corporate debt earlier in the year, and our more recent decision to downgrade Emerging Market (EM) sovereign (-1bp) and corporate debt (-4bps), were both small negative contributors to performance. Summing it all up, our spread product allocations performed well because of the overweight to U.S. IG and HY corporates. The underweights in Euro Area and EM credit were set up as relative value allocations versus U.S. equivalents, so the underperformance versus the benchmark should be viewed against the substantial outperformance from U.S. corporates. The MBS underperformance was small on a YTD basis, but we see an opportunity for that to soon turn around, as we discuss later. Bottom Line: The GFIS recommended model bond portfolio returned 1.1% (hedged into USD) in its first year of existence, slightly underperforming the custom benchmark index by -2bps. Our bearish duration tilts were a drag on performance, while our overweights to U.S. corporate debt were a major contributor. Lessons Learned On Risk Management As the first year of the GFIS model portfolio progressed, we added elements to the framework to help us manage the overall risk of the portfolio. Specifically, we began to include a tracking error calculation to show the relative volatility of the portfolio to its benchmark.3 When we first introduced that tracking error back in April, we were running far too little risk in the portfolio given the relatively modest position sizes (Chart 4). Rather than be an "index hugger", we decided to increase the sizes of all our relative tilts (Chart 5), and the tracking error rose accordingly from a mere 25bps to over 60bps. This is still below the 100bps limit that we decided to impose on the relative volatility of the model portfolio, but we were comfortable not running less-than-maximum risk given that valuations on many spread products were not extraordinarily cheap. The time to max out a risk budget is early in the credit cycle when spreads are wide, not when the cycle is far advanced and spreads are relatively tight. Yet one lesson that was learned in Year One was that too much focus on tracking error can result in lost opportunities to boost the performance of the portfolio. As part of our strategic call to maintain a below-benchmark overall duration stance, we upgraded Japan to maximum overweight in the model portfolio back on July 4th.4 With Japanese Government Bonds (JGBs) having such a low beta to yield changes in the overall Developed Markets (Chart 6), adding more Japan exposure was a way to get more defensive on duration in a way that would also boost our desired tracking error (since we were adding more of an asset less correlated to the other government bonds in the portfolio). Chart 4Tracking Error Of##BR##The Model Portfolio Chart 5Allocations Between##BR##Government Bonds & Spread Product Chart 6Are JGBs The##BR##Optimal Duration Hedge? Yet by increasing the allocation to low-beta JGBs, we were also adding exposure to "no-yield" JGBs. The overall yield of the model portfolio suffered as a result, fully offsetting the bump to the portfolio yield from the increase in allocations to spread product in April (Charts 7 & 8). With the benefit of hindsight, increasing the allocation even more to something like U.S. HY corporate bonds would have a been a more prudent way to redirect government bond exposure to a low-beta market that would have boosted the overall portfolio yield (Chart 9). Chart 7Too Much Japan##BR##In The Portfolio ... Chart 8... Offsetting The Yield Pick-Up##BR##From Spread Product Chart 9There Is Not Enough Yield##BR##In The Model Portfolio Going forward, we will pay more attention to managing the portfolio yield more actively as another piece of our model bond portfolio framework that can help boost expected returns. Bottom Line: The maximum overweight to low-beta, but low-yielding, Japanese Government Bonds was a drag on performance by reducing the portfolio yield. This highlights the classic bond management trade-off between controlling portfolio risks, like duration or tracking error, and maximizing sources of return, like interest income. The Outlook For The Next Year Looking towards the next twelve months, the biggest expected drivers of returns in our model bond portfolio are expected to come from the following allocations: Below-benchmark overall duration exposure: We are sticking to our guns on the future direction of global bond yields, which have more room to rise over the next 6-12 months. The coordinated global economic upturn is showing little sign of slowing, with leading indicators still rising and pointing to upward pressure on real bond yields (Chart 10). At the same time, inflation expectations in the developed economies remain too low relative to current levels of inflation (bottom panel). Thus, we expect government bond yield curves to bear-steepen as central banks will respond slowly to the rise in inflation. This will benefit the steepening bias we have in the model portfolio from the underweights in longer maturity buckets in the U.S., Europe and the U.K. (Chart 11). Chart 10Future Drivers Of Performance:##BR##Below-Benchmark Duration Chart 11An Unexpected##BR##Bull Flattening This Year Overweight U.S. corporate bonds (both IG and HY): Looking over the indicators from our U.S. Corporate Bond Checklist, the backdrop is not yet pointing to a period of expected underperformance for U.S. corporates (Chart 12). While balance sheet fundamentals do appear stretched, as indicated by our Corporate Health Monitor (2nd panel), the overall stance of U.S. monetary conditions is neutral (3rd panel), while bank lending standards are not yet restrictive (4th panel). We expect the Fed to deliver another 25bp rate hike in December, and at least another 2-3 hikes in 2018, which will shift monetary conditions into more restrictive territory. A very rapid rise in the U.S. dollar would worsen this trend, but we expect only a moderate grind higher in the greenback as the Fed slowly delivers additional rate hikes and non-U.S. growth remains robust. While the solid global economic backdrop should benefit all growth-sensitive assets like corporate debt, we see more attractive relative valuations on U.S. corporates versus Euro Area or EM equivalents. The upcoming tapering of asset purchases by the European Central Bank (ECB) also represents a major risk to Euro Area corporate debt, as the ECB will be slowing the pace of its corporate bond buying. One other sector that can potentially boost the portfolio performance in Year Two versus Year One is U.S. MBS. Our colleagues at our sister service, U.S. Bond Strategy, now see MBS valuations as looking attractive to other U.S. spread product like IG corporates (Chart 13).5 The relative option-adjusted spreads (OAS) on MBS and U.S. IG are a good leading indicator of the relative performance of the two asset classes and current spread levels should lead to a better return profile for MBS over IG. Another factor benefitting MBS is the continued rising trend in U.S. bond yields (and mortgage rates) that we expect over the next 6-12 months, which will reduce mortgage prepayments that would weigh on MBS returns (bottom panel). Chart 12Future Drivers Of Performance:##BR##Overweight U.S. Corporates Chart 13Upgrade U.S. MBS##BR##To Neutral This week, we are upgrading our MBS allocation to neutral from underweight in our model portfolio. However, given that our allocations to U.S. corporates are already fairly significant, we are choosing to "fund" the MBS upgrade by lowering our weighting on U.S. Treasuries (see the model portfolio allocations on Page 14). Bottom Line: Over the next 6-12 months, we expect the model portfolio returns to again benefit mostly from our below-benchmark duration stance (as global bond yields grind higher) and from our overweight stance on U.S. corporates (as the U.S. economy maintains a solid pace of growth). We are also now more constructive on valuations on U.S. MBS, thus we are upgrading our allocation to neutral at the expense of U.S. Treasuries. Robert Robis, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com 1 Please see BCA Global Fixed Income Model Special Report, "Introducing Our Recommended Global Fixed Income Portfolio", dated September 20th, 2016, available at gfis.bcaresearch.com. 2 The GFIS model portfolio custom benchmark index can most simply be described as the Barclays Global Aggregate Index, but with allocations to global high-yield corporate debt replacing very highly-rated spread product. We believe this to be more indicative of the typical internal benchmark used by global multi-sector fixed income managers. 3 Please see BCA Global Fixed Income Strategy Special Report, "Adding A Risk Management Framework To Our Model Bond Portfolio", dated June 20th 2017, available at gfis.bcareseach.com. 4 Please see BCA Global Fixed Income Strategy Weekly Report, "Central Banks Are Now Playing Catch-Up", dated July 4th 2017, available at gfis.bcaresearch.com. 5 Please see BCA U.S. Bond Strategy Weekly Report, "Dollar Watching: Yet Another Debate", dated October 10th 2017, available at usbs.bcaresearch.com. The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Appendix - Selected Sectors From The GFIS Model Portfolio Recommendations Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

GAA DM Equity Country Allocation Model Update The GAA DM Equity Country Allocation model is updated as of September 29th, 2017. The model sharply reduced its allocation to the U.K. to a bare minimum in response to the tightening in liquidity condition as the Bank of England warned of a rate hike in "coming months." The funds are reallocated to the Spain and Germany. Other smaller changes are the reductions in Italy and Australia in favor of Sweden and Switzerland, as shown in Table 1. As shown in Table 2 and Charts 1, 2 and 3, the overall model outperformed its benchmark by 44 bps in September. Both level 1 and level 2 models performed well, with level 2 outperforming its benchmark by 63 bps and level 1 outperforming its benchmark by 9 bps, as the underweight in Australia, U.S. and Japan versus the overweight in Italy, Germany and Netherland worked very well. Since going live in January 2016, the overall model has outperformed the benchmark by 341 bps, largely from the allocation among the 11 non-U.S. countries, which has outperformed its benchmark by 743 bps. Chart 1GAA DM Model Vs. MSCI World Chart 2GAA U.S. Vs. Non U.S. Model (Level1) Chart 3GAA Non U.S. Model (Level 2) Table 1Model Allocation Vs. Benchmark Weights Table 2Performance (Total Returns In USD) Please see also on the website http://gaa.bcaresearch.com/trades/allocation_performance. For more details on the models, please see the January 29th, 2016 Special Report, "Global Equity Allocation: Introducing the Developed Markets Country Allocation Model." http://gaa.bcaresearch.com/articles/view_report/18850. Please note that the overall country and sector recommendations published in our Monthly Portfolio Update and Quarterly Portfolio Outlook use the results of these quantitative models as one input, but do not stick slavishly to them. We believe that models are a useful check, but structural changes and unquantifiable factors need to be considered too in making overall recommendations. GAA Equity Sector Selection Model The GAA Equity Sector Selection Model (Chart 4) is updated as of September 30, 2017. Chart 4Overall Model Performance Table 3Allocations Table 4Performance Since Going Live The model continues to be optimistic on global growth as seen by an increasing allocation to cyclical sectors. Additionally, the model has also reduced its underweight on consumer discretionary stocks, which is currently the only cyclical sector to have a below-benchmark allocation. Finally, the biggest shift was a downgrade in utilities from overweight to underweight. This was primarily driven by momentum. For more details on the model, please see the Special Report "Introducing The GAA Equity Sector Selection Model," July 27, 2016 available at https://gaa.bcaresearch.com. Xiaoli Tang, Associate Vice President xiaoli@bcaresearch.com Aditya Kurian, Research Analyst adityak@bcaresearch.com

Special Report

Highlights In this report, we analyze both static and dynamic hedging strategies to hedge an identical global equity portfolio to six home currencies: the U.S. dollar, the British pound, the euro, the Japanese yen, the Canadian dollar and the Australian dollar. We propose an easy to implement dynamic hedging framework based on the proprietary currency indicators from BCA's Foreign Exchange Strategy (FES) service. It has outperformed all the static hedging strategies since 2001 on a risk-adjusted return basis. In addition, it levels out the playing field for all investors as the hedged returns are quite similar, no matter what their home currencies are. Among the static hedging strategies, the "least-regret" hedge ratio of 50% has lived up to its reputation, as it has reduced risk by more than 50% without severely jeopardizing returns. Over a four-year moving performance cycle (in line with how most portfolio managers are evaluated), the proposed dynamic hedging adds little career risk to portfolio managers compared to the "least regret" 50% static hedging, while provides superior returns most of the time. A global equity portfolio's currency exposure should be managed in a centralized currency overlay under the supervision of the Chief Investment Officer, so that equity and currency managers can both fully utilize their expertise in their respective field, and the organization can manage currency risk more efficiently at the total portfolio level. Feature Dynamic Hedging Outperforms Static Hedging We have received many client requests asking about currency hedging in global equity portfolios with different home currencies. This is not surprising given how currency movements have overwhelmed global equity portfolio returns of late. For example, this year the U.S. dollar's weakness against the major currencies has beefed up returns for U.S. investors who did not hedge their foreign exposure (Table I-1). Table I-1Year-To-Date Currency Movements Have Overwhelmed Equity Returns GAA's equity country allocation recommendations are by default unhedged on USD basis. When a hedge is required we make the recommendation explicit. There have been many conflicting views on whether global equity investors should hedge foreign currency exposure, and if so, what proportion of the exposure should be hedged. Perold and Schulman (1988) suggest fully hedging foreign currency exposure, because significant risk reduction can be achieved without a significant loss of return.1 This is based on their famous "free lunch" claim that average currency returns are zero over the long-term. At the other extreme, Froot (1993) suggests that foreign currencies should not be hedged at all for long-term investors, because purchasing power parity holds in the long term and exchange rates are mean reverting.2 There are many other views in between, including the universal hedge ratio proposed by Black (1989), which is estimated to be 77%.3 Campbell et al (2010) classify currencies into reserve currencies (USD, EUR and CHF), commodity currencies (AUD and CAD) and neutral currencies (JPY and GBP), and suggest that risk-minimizing investors should hold reserve currencies while hedging out commodity currencies, based on the correlations between the currencies and equity markets.4 Realizing the limitations of "static hedging," which assumes constant correlation, mean and variance, many academic researchers have explored "dynamic hedging models" that allow the mean and covariance to be time-varying by using complex econometric modelling techniques.5,6,7 These academic dynamic hedging strategies have shown improvement over the 50% static hedge, but it's very difficult for investors without a strong quantitative background to understand these very complex approaches. In fact, with all these confusing academic recommendations floating around, the "least regret" hedge ratio of 50% has been quite popular among practitioners.8 To help our clients better understand the role of currency management in global equity portfolios, we have joined forces with BCA's Foreign Exchange Strategy (FES) service to de-mystify the currency hedging process, and to propose a simple framework for clients with six different home currencies to dynamically hedge their foreign currency exposure based on the FES team's proprietary Intermediate-Term Timing Model (ITTM) indicators.9 These indicators have been used in their regular weekly publications. The ITTM-based dynamic hedging strategy is back-tested from 2001 because the ITTM indicators only date back to 2001 (See methodology). We have also back-tested a simple trend-following momentum-based dynamic strategy to see how a dynamic hedging methodology works in a longer period from 1976. For comparison we have back-tested ten different static hedging strategies that employ fixed hedge ratios across currencies and time. The test results not only clarify much of the confusion about static hedging, they also demonstrate that on a risk-adjusted return basis, BCA's ITTM-based dynamic hedging strategy has outperformed all static hedging strategies for all investors with six different home currencies since 2001 (Chart I-1). Even in the very long run of 41 years from August 1976, the simple momentum-based dynamic hedging outperforms the static strategies for investors with five home currencies, with only the AUD portfolio being worse off (Chart I-2). Chart I-1Identical Investment, But Different Risk/Return Profiles (3/2001-8/2017) Chart I-2Identical Investment, But Different Risk/Return Profiles (8/1976-8/2017) Now let's clarify some of the confusion: Does static hedging reduce portfolio risk at little expense of lowering returns? The answer is yes only for USD and JPY portfolios. For both the 41-year period from 1976 and the shorter 16-year period from 2001, a higher hedge ratio results in lower risk with slightly lower return. So fully hedged would be the "optimal" strategy for risk-minimizing investors with USD and JPY as home currencies (Panel 1 and 2 on left side in both Chart I-1 and Chart I-2). For the U.K. portfolio, the answer is a partial yes, because the "optimal hedge ratio" for risk-minizing investors, around 60%-80%, does produce the lowest risk, but the paths to that "optimal" point are totally opposite when different time periods are chosen. In the short period (Chart I-1, bottom left panel), it follows the same pattern as that for U.S. and Japanese investors up to the optimal point. As the hedge ratio increases, however, return drops and risk increases! In the longer period from 1976, as the hedge ratio increases, return increases and risk decreases until the "optimal point," after which risk and returns both increase (Chart I-2, bottom left panel). In the period from 2001, the AUD, CAD and EUR portfolios share a similar pattern to that of the GBP portfolio in the period from 1976. The AUD portfolio behaved the same in both the shorter period (Chart I-1, top right) and the longer period (Chart 2, top right), while the EUR and CAD portfolios behaved differently in the longer-term period from the shorter period. Overall, the CAD portfolio's "optimal" hedge ratio is around 0-30%, the AUD portfolio around 30-60% while the EUR portfolio is around 40% in the shorter period but around 90% in the longer period (Chart I-1 and Chart I-2). It is also worth noting that even though both the CAD and AUD are commodity currencies, AUD investors benefit significantly from hedging, while CAD investors' risk/return profile does not change much. This is because 1) currency returns for AUD investors do not average to zero in the long term due to positive carry, and 2) correlations between foreign currencies, foreign equities and domestic equities are not constant over time or across currencies (Appendix 2, Chart II-1 and Chart II-4) How does the "least regret" 50% static hedge do? The 50% hedge ratio fares quite well compared to the "optimal" static hedge ratio in terms of risk reduction for all portfolios in both periods, because more than half of the total risk reduction (the highest minus the lowest) occurs around the 50% hedge (Chart I-1 and Chart I-2). How does BCA's ITTM-based dynamic hedge do in terms of risk reduction? The ITTM produces better risk-adjusted returns for each portfolio than all the static hedges. In terms of risk, it generates the lowest risk for the EUR and GBP portfolios and is comparable to the 50% static hedge for other portfolios, while it generates a much higher return than all static hedges (Chart I-1 and Chart I-2). How does the BCA Dynamic Hedge (ITTM) compare to the 50% static hedge over time? Chart I-3 shows that the dynamic hedge consistently outperformed the 50% static hedge for all six home currencies since 2001 without significantly increasing hedging transactions, compared to the fixed 50% hedge ratio for all currency pairs. The dash line in each panel of Chart I-3 corresponds to the market cap-weighted aggregate of hedge ratios of foreign currencies for each home currency. On average, they are comparable to 50%. Most portfolio managers are measured on a moving four-year performance cycle. Chart I-4 shows that the ITTM-based dynamic hedging strategy has outperformed the 50% static hedging most of the time on a moving four-year basis, with only CAD, USD and JPY managers suffering brief drawdowns. Chart I-3BCA Dynamic Hedging Adds ##br##Value For All Investors Chart I-4Little Career Risk For ##br##Portfolio Managers In theory, if hedges were effective, then an identical global equity portfolio should have similar returns for all investors, no matter what home currency they hold. While neither the static hedging strategies nor the momentum-based dynamic hedging approach could pass this criteria, BCA's ITTM-based dynamic hedging approach does. It levels out the playing field for all investors globally. As shown in Chart I-5, in the period from March 2001 to August 2017, if left unhedged, the same global investment exhibits very different annualized returns for investors in different home currencies, with AUD investors at the low end at around 3.2%, and GBP investors at the high end at around 7%. With BCA's ITTM-based dynamic hedge, however, returns for all investors are very similar, no matter which currency is their home currency. Chart I-5BCA Dynamic Hedging Strategy Levels Out The Playing Field The BCA Dynamic Hedging Methodology To back-test all the hedging strategies, we use an identical global equity portfolio that is a market cap-weighted aggregate of nine countries/regions based on MSCI U.S., Japan, the euro area, U.K., Canada, Australia, Switzerland, Sweden and Norway. This universe accounts for about 97% of the current MSCI World index (Chart I-6). The history of the euro area before 1987 (when MSCI data were available) is calculated as a market cap-weighted aggregate of MSCI Germany, France, Spain, Italy, Austria, Belgium and the Netherlands. Chart I-6GAA Global Equity Universe We evaluate the same global portfolio for investors with six different home currencies: USD, JPY, GBP, EUR, CAD and AUD. The Measurement for Hedging Efficacy: All the academically claimed "optimal" hedge ratios, static or dynamic, are based on risk-minimizing in a mean-variance optimization framework. For practitioners, however, "not meeting return objectives" is a much higher risk priority than minimizing portfolio risk. This is why we advocate risk-adjusted return as our measurement for hedge efficacy. The Objective: We aim to find the hedging strategies that outperform the widely used 50% static hedging strategy on a risk-adjusted return basis. At BCA, our philosophy has always been not to strive to find the optimal solution, but to try our best to find a feasible solution that best meets the stated objective. Interestingly, this approach has produced superior results for the GBP and EUR portfolios, achieving the combination of "highest return - lowest risk." It also levels out the playing field for all investors by generating similar hedged returns for all investors, no matter what home currency they hold (Chart I-1 and Chart I-5). The Proprietary Currency Indicators: The indicators from Foreign Exchange Strategy service's Intermediate Term Timing Model (ITTM) are built under the assumptions that the uncovered interest rate parity (UIP) holds, and the fair value of a currency pair is a function of the real rate differential (using an average of two-year and 10-year real rates), Junk OAS (a proxy for global risk appetite), commodity prices and past-year trends (52-week moving average). When a currency pair deviates from its fair value to an extreme, then the likelihood for a trend reversal is high.9 The Simple Rule-Based Dynamic Hedging: For each home currency, we evaluate the other eight foreign currencies individually based on each pair's ITTM indicator. When a foreign currency overshoots fair value above the upper band of its historical range, then the currency is a short, and the short position is kept until the foreign currency value touches the lower band of its historical range that's below the fair value. The Implementation: We use one-month forward contracts and re-balance on a monthly basis. The gain or loss of the underlying equity index during the month is not hedged, but converted back to the home currency at the month-end spot rate. For history before 2001, we use one-month interest rates to calculate hedged returns (Please see Appendix 1 hedged return calculations using forwards and also interest rates). Example: Chart I-7 illustrates how a Canadian investor uses the ITTM indicator to decide when to hedge JPY exposure when they invest in the MSCI Japan equity index. Chart I-7Canadian Investor: Japanese Index Dynamically Hedged The top panel shows the hedging signal for JPY (solid line) versus our proprietary ITTM indicator for JPY/CAD exchange rate (dash line). The upper and lower band are set as 7% above fair value and 9% below fair value; the bottom panel shows the relative performance of the MSCI Japan hedged in CAD versus the MSCI Japan unhedged in CAD. Currently CAD investors should hedge their JPY exposure based on the ITTM indicator. Some Suggestions For Asset Allocators Use a centralized currency overlay portfolio to manage foreign currency exposure. The overlay portfolio should be managed by currency specialists (either in-house or using external managers) under the supervision of the CIO. The objective of the overlay portfolio is to manage currency exposure based on the underlying assets of the organization's Total Portfolio, such that the risk/return profile of the total portfolio is improved against its benchmark. Global equity portfolio manager performance should be measured on an unhedged basis in their respective home currency. Some have argued that a fully hedged benchmark such as the MSCI All Country Total Return Index in local currencies should be used to measure the performance of a global equity portfolio manager. We strongly disagree, even though in theory it does not really matter what benchmark is used. MSCI's "total return indexes in local currencies" for global equity aggregates are theoretical in nature. They cannot be replicated in the real world because they are calculated on a daily basis using the previous closing day's exchange rate to calculate foreign equity return10 - such that the "local return" for each foreign equity index is perfectly hedged. Unfortunately this "perfect hedge" is humanly impossible, because as shown in Formula (1) in Appendix 1, only when both the spot rate and forward rate at the time t are equal to the spot rate at time t+1, can the local returns be fully captured! Xiaoli Tang, Associate Vice President xiaolit@bcaresearch.com 1 Perold, A and E. Schulman, 1988, "The free lunch in currency hedging: Implications for investment policy and performance standards", Financial Analyst Journal 44, 45-50. 2 Froot K., 1993, "Currency hedging over long horizons", NBER working paper 4355 3 Black, F., 1989, "Universal hedging: optimizing currency risk and reward in international equity portfolios", Financial Analyst Journal 45, 16-22 4 Campbell, J., K. de Medeiros and L. Viceira, 2010, "Global currency hedging", Journal of Finance LXV, 87-122 5 Gagnon, L., T., McCurdy, and G., Lypny, 1998, " Hedging foreign currency portfolios", Journal of Empirical Finance 5, 197-220 6 Hautsch, N., and J. Inkmann, 2003, "Optimal hedging of the currency exchange risk exposure of dynamically balanced strategic asset allocations", Journal of Asset Management 4, 173-189 7 Brown, C., J., Dark, and W., Zhang, "Dynamic currency hedging for international stock portfolios", 2012, Ph.D. dissertation, University of Melbourn 8 Michenaud, S., and B., Solnik, 2008, "Applying regret theory to investment choices: Currency hedging decisions", Journal of International Money and Finance 27, 677-694 9 Please see BCA Foreign Exchange Strategy Special Report titled "In Search Of A Timing Model", June 22, 2016 available at fes.bcaresearch.com 10 "Note on index calculation in local currency", MSCI Bara Index Calculation Methodologies, p19, May 2010 11 MSCI uses 2 business days as described in "MSCI Index Methodology: MSCI Hedged Indexes", July 2013 Appendix 1 We use one-month forward contracts and re-balance on a monthly basis. The gain or loss of the underlying equity index during the month is not hedged, but converted back to the home currency at the month-end spot rate. Before reliable forward contract rates were available (Jan 2001), we use one-month interest rates instead to calculate hedged returns. Please note that for simplification we have ignored the bid-ask spread in all the quotations, and we do not take into account the time lag required to implement a hedge.11 In practice, however, these are valid considerations. Appendix 2: Dynamics Hedging For Six Home Currencies 2.1 The Australian Perspective Correlations: For Aussie investors, foreign currencies in aggregate have a negative correlation with the domestic equity index (especially in the period from 2001), and a mostly positive but low correlation to the unhedged foreign equities. (Chart II-1). So hedging away foreign currency exposure would increase total risk of the global portfolio. This is why a fully hedged portfolio has the highest risk (Chart II-2, Table II-1). Historical Performance: Since 2001, ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in AUD. The return profile looks similar to the fully hedged portfolio, but risk is much lower (Table II-1). Over the longer period, the optimal static hedge ratio is about 70%, which is actually quite close to the 50% hedge, as shown in Table II-2. On a five-year rolling basis, as shown in Chart II-2, the ITTM-based dynamic risk/return profile also prevails. Current State: Currently AUD investors should only be hedging their exposure in Swiss francs and U.S. dollars, the two "safe haven" currencies. Actually, our indicators show a close to 100% hedge of the Swiss franc, as shown in Chart II-3, which did hurt the risk/return profile during the GFC period despite an outstanding full-period performance. Table II-1Risk/Return Profile For Global Equity In AUD (2001-2017) Table II-2Risk/Return Profile For Global Equity In AUD (1976-2017) Chart II-1Australian Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-2Australian Perespective: Dynamic Vs. Static Hedging Chart II-3Australian Perspective: Swiss Index Dynamically Hedged 2.2 The Canadian Perspective Correlations: For Canadian investors, foreign currencies on aggregate have a negative correlation with the country's domestic equity index, but the correlation with unhedged foreign equities has oscillated in both positive and negative territory. The correlation between Canadian equities and unhedged foreign equities has been always positive, in the range of 0.4-0.8 (Chart II-4). So hedging away foreign currency exposure will increase total risk within a global portfolio. This is why a fully hedged portfolio has the highest risk (Chart II-5, Table II-3 and Table II-4). Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest annualized return for the global portfolio in CAD without significantly increasing volatility from the unhedged scenario. In terms of risk-adjusted return, dynamic hedging has outperformed the best static hedging scenario (40% hedged) by about 46% (Table II-3). On a five-year rolling basis, as shown in Chart II-5, the dynamic risk/return profile also prevails. Current State: Currently Canadian investors should be hedging all foreign currencies except for the Australian dollar. Chart II-6 shows how CAD investors dynamically hedge their exposure to the Swedish krona. Table II-3Risk/Return Profile For Global Equity In CAD (2001-2017) Table II-4Risk/Return Profile For Global Equity In CAD (1976-2017) Chart II-4Canadian Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-5Canadian Perspective: Dynamic Vs. Static Hedging Chart II-6Canadian Perspective: Swedish Index Dynamically Hedged 2.3 The Japanese Perspective Correlations: For Japanese investors, correlations among foreign currencies, foreign equities and domestic equities seem to have gone through regime changes since the GFC, as shown in Chart II-7. These regime changes in correlations explain the evolving risk/return profile of the static hedges in Chart II-8. Before the GFC, full hedge and 50% hedge had similar risk profiles, but after the GFC full hedge reduced total risk significantly due to the sharp increase in correlations. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in JPY, with a much higher return and slightly higher risk (Table II-5). Fully hedged has the lowest risk in both periods, and 50% is close to the "optimal" static hedge in both periods as well (Table II-5 and II-6). On a five-year rolling basis, as shown in Chart II-8, the ITTM-based dynamic risk/return profile. The 50% static hedge has similar risk profile to the ITTM-based dynamic hedge, but with lower returns. Current State: Currently JPY investors should be only hedging their exposure in Swiss francs and U.S. dollars. Chart II-9 shows how Japanese investors should have hedged CAD exposure. Table II-5Risk/Return Profile For Global Equity In JPY (2001-2017) Table II-6Risk/Return Profile For Global Equity In JPY (1976-2017) Chart II-7Japanese Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-8Japanese Perspective: Dynamic Vs. Static Hedging Chart II-9Japanese Perspective: Canadian Index Dynamically Hedged 2.4 The U.S. Perspective Correlations: For U.S. investors, correlations among foreign currencies, foreign equities and domestic equities seem to have gone through regime changes since the GFC, as shown in Chart II-10. These regime changes in correlation explain the evolving risk/return profile of the static hedges in Chart II-11. Before the GFC, full hedge and 50% hedge had similar risk profiles, but after the GFC full hedge exhibited significantly reduced total risk due to the sharp increase in correlations. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in USD, with a much higher return and slightly lower risk compared to the 50%, which is the best static hedge for the period (Table II-7). Fully hedged had the lowest risk in both periods, but the return profiles of the static hedges were very similar, supporting the widely held belief that in the long run currency returns are close to zero (Table II-7 and II-8). On a five-year rolling basis, as shown in Chart II-11, the dynamic risk/return profile is close to that of the 50%. Current State: Currently U.S. investors should only be hedging their exposure in euros (Chart II-12). Table II-7Risk/Return Profile For Global Equity In USD (2001-2017) Table II-8Risk/Return Profile For Global Equity In JPY (1976-2017) Chart II-10U.S. Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-11U.S. Perspective: Dynamic Vs. Static Hedging Chart II-12U.S. Perspective: EMU Index Dynamically Hedged 2.5 The Euro Perspective Correlations: For investors that call the euro home currency, the correlation and foreign equities has been positive. The correlation between foreign currencies and domestic equity index, however, has been changing signs over time, currently sitting at near zero! (Chart II-13). This explains the shape of the risk/return profile in Chart II-14 and Table II-9. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in euros, which has a higher return and lower risk than the best static hedge of 80% (Table II-9). Over the longer period, the optimal hedge ratio is about 50% hedge, which is actually quite close to those between 40-70%, as shown in Table II-10. On a five-year rolling basis, as shown in Chart II-14, the dynamic risk/return profile definitely prevails. Current State: Currently euro investors should not be hedging any foreign exposure, based on our indicators. Chart II-15 shows how euro investors should have hedged JPY exposure over time since 2001. Table II-9Risk/Return Profile For Global Equity In Euro (2001-2017) Table II-10Risk/Return Profile For Global Equity In Euro (1976-2017) Chart II-13Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-14Euro Area Perspective: Dynamic Vs. Static Hedging Chart II-15Euro area Perspective: Japanese Index Dynamically Hedged 2.6 The British Perspective Correlations: For British investors, correlations between foreign currencies and domestic equities have gone through several regime changes over time in both positive and negative territory, as shown in Chart II-16. The positive correlation between foreign and domestic equities has also increased over time. This would definitely make static hedging worse off (Table II-11 and II-12). Historical Performance: Not surprisingly, since 2001, dynamic hedging has produced the highest risk-adjusted return for the global portfolio in pounds, which has a higher return and lower risk than the best static hedge of 20% (Table II-11). Over the longer period, the optimal hedge ratio is about 90% hedge, Table II-12. On a five-year rolling basis, as shown in Chart II-17, the dynamic risk/return profile prevails, as it shares similar risk as the 50% hedge but with a higher return. Current State: Currently the British investors should not be hedging CAD and Swedish krona, while all six other currencies should be hedged. Chart II-18 shows how U.K. investors should have hedged their AUD exposure over time since 2001. Table II-11Risk/Return Profile For Global Equity In GBP (2001-2017) Table II-12Risk/Return Profile For Global Equity In GBP (1976-2017) Chart II-16Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-17Dynamic Vs. Static Hedging Chart II-18U.K. Perspective: Australian Index Dynamically Hedged

Special Report

Dear client, I am on the road this week, attending BCA's New York Conference and teaching the BCA Academy. In lieu of a regular report, we are sending you a Special Report, a cooperation with our Global Asset Allocation service. In this piece, my colleague Xiaoli Tang tests the benefits of various currency hedging strategies for global equity portfolios. In addition to traditional hedging rules, Xiaoli deploys the Intermediate-Term Timing Models developed by the Foreign Exchange Strategy team to build dynamic hedging strategies, which result in superior risk/reward profiles. I trust you will find this report interesting and informative. Best regards, Mathieu Savary Highlights In this report, we analyze both static and dynamic hedging strategies to hedge an identical global equity portfolio to six home currencies: the U.S. dollar, the British pound, the euro, the Japanese yen, the Canadian dollar and the Australian dollar. We propose an easy to implement dynamic hedging framework based on the proprietary currency indicators from BCA's Foreign Exchange Strategy (FES) service. It has outperformed all the static hedging strategies since 2001 on a risk-adjusted return basis. In addition, it levels out the playing field for all investors as the hedged returns are quite similar, no matter what their home currencies are. Among the static hedging strategies, the "least-regret" hedge ratio of 50% has lived up to its reputation, as it has reduced risk by more than 50% without severely jeopardizing returns. Over a four-year moving performance cycle (in line with how most portfolio managers are evaluated), the proposed dynamic hedging adds little career risk to portfolio managers compared to the "least regret" 50% static hedging, while provides superior returns most of the time. A global equity portfolio's currency exposure should be managed in a centralized currency overlay under the supervision of the Chief Investment Officer, so that equity and currency managers can both fully utilize their expertise in their respective field, and the organization can manage currency risk more efficiently at the total portfolio level. Feature Dynamic Hedging Outperforms Static Hedging We have received many client requests asking about currency hedging in global equity portfolios with different home currencies. This is not surprising given how currency movements have overwhelmed global equity portfolio returns of late. For example, this year the U.S. dollar's weakness against the major currencies has beefed up returns for U.S. investors who did not hedge their foreign exposure (Table I-1). Table I-1Year-To-Date Currency Movements Have Overwhelmed Equity Returns GAA's equity country allocation recommendations are by default unhedged on USD basis. When a hedge is required we make the recommendation explicit. There have been many conflicting views on whether global equity investors should hedge foreign currency exposure, and if so, what proportion of the exposure should be hedged. Perold and Schulman (1988) suggest fully hedging foreign currency exposure, because significant risk reduction can be achieved without a significant loss of return.1 This is based on their famous "free lunch" claim that average currency returns are zero over the long-term. At the other extreme, Froot (1993) suggests that foreign currencies should not be hedged at all for long-term investors, because purchasing power parity holds in the long term and exchange rates are mean reverting.2 There are many other views in between, including the universal hedge ratio proposed by Black (1989), which is estimated to be 77%.3 Campbell et al (2010) classify currencies into reserve currencies (USD, EUR and CHF), commodity currencies (AUD and CAD) and neutral currencies (JPY and GBP), and suggest that risk-minimizing investors should hold reserve currencies while hedging out commodity currencies, based on the correlations between the currencies and equity markets.4 Realizing the limitations of "static hedging," which assumes constant correlation, mean and variance, many academic researchers have explored "dynamic hedging models" that allow the mean and covariance to be time-varying by using complex econometric modelling techniques.5,6,7 These academic dynamic hedging strategies have shown improvement over the 50% static hedge, but it's very difficult for investors without a strong quantitative background to understand these very complex approaches. In fact, with all these confusing academic recommendations floating around, the "least regret" hedge ratio of 50% has been quite popular among practitioners.8 To help our clients better understand the role of currency management in global equity portfolios, we have joined forces with BCA's Foreign Exchange Strategy (FES) service to de-mystify the currency hedging process, and to propose a simple framework for clients with six different home currencies to dynamically hedge their foreign currency exposure based on the FES team's proprietary Intermediate-Term Timing Model (ITTM) indicators.9 These indicators have been used in their regular weekly publications. The ITTM-based dynamic hedging strategy is back-tested from 2001 because the ITTM indicators only date back to 2001 (See methodology). We have also back-tested a simple trend-following momentum-based dynamic strategy to see how a dynamic hedging methodology works in a longer period from 1976. For comparison we have back-tested ten different static hedging strategies that employ fixed hedge ratios across currencies and time. The test results not only clarify much of the confusion about static hedging, they also demonstrate that on a risk-adjusted return basis, BCA's ITTM-based dynamic hedging strategy has outperformed all static hedging strategies for all investors with six different home currencies since 2001 (Chart I-1). Even in the very long run of 41 years from August 1976, the simple momentum-based dynamic hedging outperforms the static strategies for investors with five home currencies, with only the AUD portfolio being worse off (Chart I-2). Chart I-1Identical Investment, But Different Risk/Return Profiles (3/2001-8/2017) Chart I-2Identical Investment, But Different Risk/Return Profiles (8/1976-8/2017) Now let's clarify some of the confusion: Does static hedging reduce portfolio risk at little expense of lowering returns? The answer is yes only for USD and JPY portfolios. For both the 41-year period from 1976 and the shorter 16-year period from 2001, a higher hedge ratio results in lower risk with slightly lower return. So fully hedged would be the "optimal" strategy for risk-minimizing investors with USD and JPY as home currencies (Panel 1 and 2 on left side in both Chart I-1 and Chart I-2). For the U.K. portfolio, the answer is a partial yes, because the "optimal hedge ratio" for risk-minizing investors, around 60%-80%, does produce the lowest risk, but the paths to that "optimal" point are totally opposite when different time periods are chosen. In the short period (Chart I-1, bottom left panel), it follows the same pattern as that for U.S. and Japanese investors up to the optimal point. As the hedge ratio increases, however, return drops and risk increases! In the longer period from 1976, as the hedge ratio increases, return increases and risk decreases until the "optimal point," after which risk and returns both increase (Chart I-2, bottom left panel). In the period from 2001, the AUD, CAD and EUR portfolios share a similar pattern to that of the GBP portfolio in the period from 1976. The AUD portfolio behaved the same in both the shorter period (Chart I-1, top right) and the longer period (Chart 2, top right), while the EUR and CAD portfolios behaved differently in the longer-term period from the shorter period. Overall, the CAD portfolio's "optimal" hedge ratio is around 0-30%, the AUD portfolio around 30-60% while the EUR portfolio is around 40% in the shorter period but around 90% in the longer period (Chart I-1 and Chart I-2). It is also worth noting that even though both the CAD and AUD are commodity currencies, AUD investors benefit significantly from hedging, while CAD investors' risk/return profile does not change much. This is because 1) currency returns for AUD investors do not average to zero in the long term due to positive carry, and 2) correlations between foreign currencies, foreign equities and domestic equities are not constant over time or across currencies (Appendix 2, Chart II-1 and Chart II-4) How does the "least regret" 50% static hedge do? The 50% hedge ratio fares quite well compared to the "optimal" static hedge ratio in terms of risk reduction for all portfolios in both periods, because more than half of the total risk reduction (the highest minus the lowest) occurs around the 50% hedge (Chart I-1 and Chart I-2). How does BCA's ITTM-based dynamic hedge do in terms of risk reduction? The ITTM produces better risk-adjusted returns for each portfolio than all the static hedges. In terms of risk, it generates the lowest risk for the EUR and GBP portfolios and is comparable to the 50% static hedge for other portfolios, while it generates a much higher return than all static hedges (Chart I-1 and Chart I-2). How does the BCA Dynamic Hedge (ITTM) compare to the 50% static hedge over time? Chart I-3 shows that the dynamic hedge consistently outperformed the 50% static hedge for all six home currencies since 2001 without significantly increasing hedging transactions, compared to the fixed 50% hedge ratio for all currency pairs. The dash line in each panel of Chart I-3 corresponds to the market cap-weighted aggregate of hedge ratios of foreign currencies for each home currency. On average, they are comparable to 50%. Most portfolio managers are measured on a moving four-year performance cycle. Chart I-4 shows that the ITTM-based dynamic hedging strategy has outperformed the 50% static hedging most of the time on a moving four-year basis, with only CAD, USD and JPY managers suffering brief drawdowns. Chart I-3BCA Dynamic Hedging Adds ##br##Value For All Investors Chart I-4Little Career Risk For ##br##Portfolio Managers In theory, if hedges were effective, then an identical global equity portfolio should have similar returns for all investors, no matter what home currency they hold. While neither the static hedging strategies nor the momentum-based dynamic hedging approach could pass this criteria, BCA's ITTM-based dynamic hedging approach does. It levels out the playing field for all investors globally. As shown in Chart I-5, in the period from March 2001 to August 2017, if left unhedged, the same global investment exhibits very different annualized returns for investors in different home currencies, with AUD investors at the low end at around 3.2%, and GBP investors at the high end at around 7%. With BCA's ITTM-based dynamic hedge, however, returns for all investors are very similar, no matter which currency is their home currency. Chart I-5BCA Dynamic Hedging Strategy Levels Out The Playing Field The BCA Dynamic Hedging Methodology To back-test all the hedging strategies, we use an identical global equity portfolio that is a market cap-weighted aggregate of nine countries/regions based on MSCI U.S., Japan, the euro area, U.K., Canada, Australia, Switzerland, Sweden and Norway. This universe accounts for about 97% of the current MSCI World index (Chart I-6). The history of the euro area before 1987 (when MSCI data were available) is calculated as a market cap-weighted aggregate of MSCI Germany, France, Spain, Italy, Austria, Belgium and the Netherlands. Chart I-6GAA Global Equity Universe We evaluate the same global portfolio for investors with six different home currencies: USD, JPY, GBP, EUR, CAD and AUD. The Measurement for Hedging Efficacy: All the academically claimed "optimal" hedge ratios, static or dynamic, are based on risk-minimizing in a mean-variance optimization framework. For practitioners, however, "not meeting return objectives" is a much higher risk priority than minimizing portfolio risk. This is why we advocate risk-adjusted return as our measurement for hedge efficacy. The Objective: We aim to find the hedging strategies that outperform the widely used 50% static hedging strategy on a risk-adjusted return basis. At BCA, our philosophy has always been not to strive to find the optimal solution, but to try our best to find a feasible solution that best meets the stated objective. Interestingly, this approach has produced superior results for the GBP and EUR portfolios, achieving the combination of "highest return - lowest risk." It also levels out the playing field for all investors by generating similar hedged returns for all investors, no matter what home currency they hold (Chart I-1 and Chart I-5). The Proprietary Currency Indicators: The indicators from Foreign Exchange Strategy service's Intermediate Term Timing Model (ITTM) are built under the assumptions that the uncovered interest rate parity (UIP) holds, and the fair value of a currency pair is a function of the real rate differential (using an average of two-year and 10-year real rates), Junk OAS (a proxy for global risk appetite), commodity prices and past-year trends (52-week moving average). When a currency pair deviates from its fair value to an extreme, then the likelihood for a trend reversal is high.9 The Simple Rule-Based Dynamic Hedging: For each home currency, we evaluate the other eight foreign currencies individually based on each pair's ITTM indicator. When a foreign currency overshoots fair value above the upper band of its historical range, then the currency is a short, and the short position is kept until the foreign currency value touches the lower band of its historical range that's below the fair value. The Implementation: We use one-month forward contracts and re-balance on a monthly basis. The gain or loss of the underlying equity index during the month is not hedged, but converted back to the home currency at the month-end spot rate. For history before 2001, we use one-month interest rates to calculate hedged returns (Please see Appendix 1 hedged return calculations using forwards and also interest rates). Example: Chart I-7 illustrates how a Canadian investor uses the ITTM indicator to decide when to hedge JPY exposure when they invest in the MSCI Japan equity index. Chart I-7Canadian Investor: Japanese Index Dynamically Hedged The top panel shows the hedging signal for JPY (solid line) versus our proprietary ITTM indicator for JPY/CAD exchange rate (dash line). The upper and lower band are set as 7% above fair value and 9% below fair value; the bottom panel shows the relative performance of the MSCI Japan hedged in CAD versus the MSCI Japan unhedged in CAD. Currently CAD investors should hedge their JPY exposure based on the ITTM indicator. Some Suggestions For Asset Allocators Use a centralized currency overlay portfolio to manage foreign currency exposure. The overlay portfolio should be managed by currency specialists (either in-house or using external managers) under the supervision of the CIO. The objective of the overlay portfolio is to manage currency exposure based on the underlying assets of the organization's Total Portfolio, such that the risk/return profile of the total portfolio is improved against its benchmark. Global equity portfolio manager performance should be measured on an unhedged basis in their respective home currency. Some have argued that a fully hedged benchmark such as the MSCI All Country Total Return Index in local currencies should be used to measure the performance of a global equity portfolio manager. We strongly disagree, even though in theory it does not really matter what benchmark is used. MSCI's "total return indexes in local currencies" for global equity aggregates are theoretical in nature. They cannot be replicated in the real world because they are calculated on a daily basis using the previous closing day's exchange rate to calculate foreign equity return10 - such that the "local return" for each foreign equity index is perfectly hedged. Unfortunately this "perfect hedge" is humanly impossible, because as shown in Formula (1) in Appendix 1, only when both the spot rate and forward rate at the time t are equal to the spot rate at time t+1, can the local returns be fully captured! Xiaoli Tang, Associate Vice President xiaolit@bcaresearch.com 1 Perold, A and E. Schulman, 1988, "The free lunch in currency hedging: Implications for investment policy and performance standards", Financial Analyst Journal 44, 45-50. 2 Froot K., 1993, "Currency hedging over long horizons", NBER working paper 4355 3 Black, F., 1989, "Universal hedging: optimizing currency risk and reward in international equity portfolios", Financial Analyst Journal 45, 16-22 4 Campbell, J., K. de Medeiros and L. Viceira, 2010, "Global currency hedging", Journal of Finance LXV, 87-122 5 Gagnon, L., T., McCurdy, and G., Lypny, 1998, " Hedging foreign currency portfolios", Journal of Empirical Finance 5, 197-220 6 Hautsch, N., and J. Inkmann, 2003, "Optimal hedging of the currency exchange risk exposure of dynamically balanced strategic asset allocations", Journal of Asset Management 4, 173-189 7 Brown, C., J., Dark, and W., Zhang, "Dynamic currency hedging for international stock portfolios", 2012, Ph.D. dissertation, University of Melbourn 8 Michenaud, S., and B., Solnik, 2008, "Applying regret theory to investment choices: Currency hedging decisions", Journal of International Money and Finance 27, 677-694 9 Please see BCA Foreign Exchange Strategy Special Report titled "In Search Of A Timing Model", June 22, 2016 available at fes.bcaresearch.com 10 "Note on index calculation in local currency", MSCI Bara Index Calculation Methodologies, p19, May 2010 11 MSCI uses 2 business days as described in "MSCI Index Methodology: MSCI Hedged Indexes", July 2013 Appendix 1 We use one-month forward contracts and re-balance on a monthly basis. The gain or loss of the underlying equity index during the month is not hedged, but converted back to the home currency at the month-end spot rate. Before reliable forward contract rates were available (Jan 2001), we use one-month interest rates instead to calculate hedged returns. Please note that for simplification we have ignored the bid-ask spread in all the quotations, and we do not take into account the time lag required to implement a hedge.11 In practice, however, these are valid considerations. Appendix 2: Dynamics Hedging For Six Home Currencies 2.1 The Australian Perspective Correlations: For Aussie investors, foreign currencies in aggregate have a negative correlation with the domestic equity index (especially in the period from 2001), and a mostly positive but low correlation to the unhedged foreign equities. (Chart II-1). So hedging away foreign currency exposure would increase total risk of the global portfolio. This is why a fully hedged portfolio has the highest risk (Chart II-2, Table II-1). Historical Performance: Since 2001, ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in AUD. The return profile looks similar to the fully hedged portfolio, but risk is much lower (Table II-1). Over the longer period, the optimal static hedge ratio is about 70%, which is actually quite close to the 50% hedge, as shown in Table II-2. On a five-year rolling basis, as shown in Chart II-2, the ITTM-based dynamic risk/return profile also prevails. Current State: Currently AUD investors should only be hedging their exposure in Swiss francs and U.S. dollars, the two "safe haven" currencies. Actually, our indicators show a close to 100% hedge of the Swiss franc, as shown in Chart II-3, which did hurt the risk/return profile during the GFC period despite an outstanding full-period performance. Table II-1Risk/Return Profile For Global Equity In AUD (2001-2017) Table II-2Risk/Return Profile For Global Equity In AUD (1976-2017) Chart II-1Australian Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-2Australian Perespective: Dynamic Vs. Static Hedging Chart II-3Australian Perspective: Swiss Index Dynamically Hedged 2.2 The Canadian Perspective Correlations: For Canadian investors, foreign currencies on aggregate have a negative correlation with the country's domestic equity index, but the correlation with unhedged foreign equities has oscillated in both positive and negative territory. The correlation between Canadian equities and unhedged foreign equities has been always positive, in the range of 0.4-0.8 (Chart II-4). So hedging away foreign currency exposure will increase total risk within a global portfolio. This is why a fully hedged portfolio has the highest risk (Chart II-5, Table II-3 and Table II-4). Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest annualized return for the global portfolio in CAD without significantly increasing volatility from the unhedged scenario. In terms of risk-adjusted return, dynamic hedging has outperformed the best static hedging scenario (40% hedged) by about 46% (Table II-3). On a five-year rolling basis, as shown in Chart II-5, the dynamic risk/return profile also prevails. Current State: Currently Canadian investors should be hedging all foreign currencies except for the Australian dollar. Chart II-6 shows how CAD investors dynamically hedge their exposure to the Swedish krona. Table II-3Risk/Return Profile For Global Equity In CAD (2001-2017) Table II-4Risk/Return Profile For Global Equity In CAD (1976-2017) Chart II-4Canadian Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-5Canadian Perspective: Dynamic Vs. Static Hedging Chart II-6Canadian Perspective: Swedish Index Dynamically Hedged 2.3 The Japanese Perspective Correlations: For Japanese investors, correlations among foreign currencies, foreign equities and domestic equities seem to have gone through regime changes since the GFC, as shown in Chart II-7. These regime changes in correlations explain the evolving risk/return profile of the static hedges in Chart II-8. Before the GFC, full hedge and 50% hedge had similar risk profiles, but after the GFC full hedge reduced total risk significantly due to the sharp increase in correlations. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in JPY, with a much higher return and slightly higher risk (Table II-5). Fully hedged has the lowest risk in both periods, and 50% is close to the "optimal" static hedge in both periods as well (Table II-5 and II-6). On a five-year rolling basis, as shown in Chart II-8, the ITTM-based dynamic risk/return profile. The 50% static hedge has similar risk profile to the ITTM-based dynamic hedge, but with lower returns. Current State: Currently JPY investors should be only hedging their exposure in Swiss francs and U.S. dollars. Chart II-9 shows how Japanese investors should have hedged CAD exposure. Table II-5Risk/Return Profile For Global Equity In JPY (2001-2017) Table II-6Risk/Return Profile For Global Equity In JPY (1976-2017) Chart II-7Japanese Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-8Japanese Perspective: Dynamic Vs. Static Hedging Chart II-9Japanese Perspective: Canadian Index Dynamically Hedged 2.4 The U.S. Perspective Correlations: For U.S. investors, correlations among foreign currencies, foreign equities and domestic equities seem to have gone through regime changes since the GFC, as shown in Chart II-10. These regime changes in correlation explain the evolving risk/return profile of the static hedges in Chart II-11. Before the GFC, full hedge and 50% hedge had similar risk profiles, but after the GFC full hedge exhibited significantly reduced total risk due to the sharp increase in correlations. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in USD, with a much higher return and slightly lower risk compared to the 50%, which is the best static hedge for the period (Table II-7). Fully hedged had the lowest risk in both periods, but the return profiles of the static hedges were very similar, supporting the widely held belief that in the long run currency returns are close to zero (Table II-7 and II-8). On a five-year rolling basis, as shown in Chart II-11, the dynamic risk/return profile is close to that of the 50%. Current State: Currently U.S. investors should only be hedging their exposure in euros (Chart II-12). Table II-7Risk/Return Profile For Global Equity In USD (2001-2017) Table II-8Risk/Return Profile For Global Equity In JPY (1976-2017) Chart II-10U.S. Perspective: Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-11U.S. Perspective: Dynamic Vs. Static Hedging Chart II-12U.S. Perspective: EMU Index Dynamically Hedged 2.5 The Euro Perspective Correlations: For investors that call the euro home currency, the correlation and foreign equities has been positive. The correlation between foreign currencies and domestic equity index, however, has been changing signs over time, currently sitting at near zero! (Chart II-13). This explains the shape of the risk/return profile in Chart II-14 and Table II-9. Historical Performance: Since 2001, the ITTM-based dynamic hedging has produced the highest risk-adjusted return for the global portfolio in euros, which has a higher return and lower risk than the best static hedge of 80% (Table II-9). Over the longer period, the optimal hedge ratio is about 50% hedge, which is actually quite close to those between 40-70%, as shown in Table II-10. On a five-year rolling basis, as shown in Chart II-14, the dynamic risk/return profile definitely prevails. Current State: Currently euro investors should not be hedging any foreign exposure, based on our indicators. Chart II-15 shows how euro investors should have hedged JPY exposure over time since 2001. Table II-9Risk/Return Profile For Global Equity In Euro (2001-2017) Table II-10Risk/Return Profile For Global Equity In Euro (1976-2017) Chart II-13Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-14Euro Area Perspective: Dynamic Vs. Static Hedging Chart II-15Euro area Perspective: Japanese Index Dynamically Hedged 2.6 The British Perspective Correlations: For British investors, correlations between foreign currencies and domestic equities have gone through several regime changes over time in both positive and negative territory, as shown in Chart II-16. The positive correlation between foreign and domestic equities has also increased over time. This would definitely make static hedging worse off (Table II-11 and II-12). Historical Performance: Not surprisingly, since 2001, dynamic hedging has produced the highest risk-adjusted return for the global portfolio in pounds, which has a higher return and lower risk than the best static hedge of 20% (Table II-11). Over the longer period, the optimal hedge ratio is about 90% hedge, Table II-12. On a five-year rolling basis, as shown in Chart II-17, the dynamic risk/return profile prevails, as it shares similar risk as the 50% hedge but with a higher return. Current State: Currently the British investors should not be hedging CAD and Swedish krona, while all six other currencies should be hedged. Chart II-18 shows how U.K. investors should have hedged their AUD exposure over time since 2001. Table II-11Risk/Return Profile For Global Equity In GBP (2001-2017) Table II-12Risk/Return Profile For Global Equity In GBP (1976-2017) Chart II-16Domestic And Unhedged Foreign Equities Vs. Foreign Currencies Chart II-17Dynamic Vs. Static Hedging Chart II-18U.K. Perspective: Australian Index Dynamically Hedged Trades & Forecasts Forecast Summary Core Portfolio Closed Trades