Global

Highlights Duration: The upturn in bond yields is not yet confirmed by our preferred global growth indicators. We anticipate that a reduction in trade uncertainty during the next few months will cause our indicators to rebound. But until then, investors should view the bond sell-off as tenuous. Yield Curve: Expect modest 2/10 steepening during the next few months, as the Fed keeps rates low even as economic growth improves. Steepening will show up in real yields, not in the TIPS breakeven inflation curve. The 2/10 slope will stay in a range between 0 bps and 50 bps for the next 6-12 months. Yield Curve Strategy: The 5-year Treasury note looks expensive compared to the rest of the yield curve, and historical correlations suggest it will rise the most if the Fed delivers fewer rate cuts than are currently expected. We recommend that investors short the 5-year bullet versus a duration-matched 2/30 barbell. Await Confirmation Bond yields look like they might be bottoming. The 2-year and 10-year Treasury yields are up 10 bps and 31 bps, respectively, since the 2/10 slope briefly inverted in late August (Chart 1). We are cautiously optimistic that the growth revival getting priced into Treasury yields will materialize. However, it’s vital to note that the yield rebound is not yet confirmed by the economic data. Even timely global growth indicators like the CRB Raw Industrials index remain downbeat (Chart 1, bottom panel). If global growth measures don’t bottom soon, then Treasury yields are certain to fall back. Chart 1Yields Are Ahead Of The Data We do expect the economic data to follow bond yields higher. We noted in last week’s report that the weakness in US economic data is concentrated in survey measures (aka “soft” data), while measures of actual economic activity (aka “hard data”) are holding up well.1    For example: The ISM Manufacturing survey is below its 2016 trough, but the year-over-year growth rate in industrial production is well above 2016 levels (Chart 2, top panel). Capacity utilization also remains elevated (Chart 2, bottom panel). New orders for core capital goods are holding firm, even with CEO confidence at its lowest since 2009 (Chart 2, panel 2). Employment growth remains strong, despite the employment component of the ISM Non-Manufacturing survey being just above the 50 boom/bust line (Chart 2, panel 3). Chart 2Will "Soft" Data Rebound? Our interpretation of the divergence is that uncertainty about the US/China trade war is weighing on sentiment and holding survey measures down. If that uncertainty is removed, survey measures will quickly rebound and converge with the “hard” data. On that front, we think it’s very likely that trade uncertainty diminishes during the next few months. The US and China have already agreed to an informal “phase one deal” that will require China to buy $40-$50 billion of US agricultural goods while the US delays the October 15 tariff hike. Odds are that President Trump will also delay the planned December 15 tariff hike and probably roll back some existing tariffs.2 The reason is that while Trump’s overall approval rating has been consistently low; until recently, he had been receiving high marks for his handling of the economy (Chart 3). But his economic approval rating took a tumble this summer and, as we head toward the 2020 election, he desperately needs an economic boost and/or policy victory to push up his numbers. We already see some tentative signs of a rebound in the regional Fed manufacturing surveys. A tactical retreat on trade should improve sentiment and cause survey data to move higher, alongside bond yields. And in fact, we already see some tentative signs of a rebound in the regional Fed manufacturing surveys (Chart 4). October figures are out for the New York, Philadelphia, Richmond, Kansas City and Dallas surveys, and they have all diverged positively from the national ISM. Chart 3It's Trump's Economy Chart 4Some Optimism From Regional Surveys Bottom Line: The upturn in bond yields is not yet confirmed by our preferred global growth indicators. We anticipate that a reduction in trade uncertainty during the next few months will cause our indicators to rebound. But until then, investors should view the bond sell-off as tenuous. Yield Curve: Macro Drivers We noted in the first section that the 2/10 Treasury slope has steepened sharply since it briefly broke below zero in late August. In this section, we consider whether this 2/10 steepening might continue. To do this we run through the main macro drivers of the yield curve. The Fed Funds Rate Traditionally, there is a very tight correlation between the fed funds rate and the slope of the curve (Chart 5). Fed tightening puts upward pressure on the curve’s front-end relative to the back-end, leading to a bear-flattening. Conversely, Fed easing drags the front-end down relative to the long-end, leading to bull-steepening. Chart 5The Fed's Yield Curve Control The traditional pattern broke down between 2009 and 2015 when the fed funds rate was pinned at zero. This period saw many episodes of bear-steepening and bull-flattening. But since the funds rate has been off zero, the traditional correlation has begun to re-assert itself. Our base case outlook calls for one more 25 bps rate cut tomorrow, followed by an extended on-hold period. This scenario might be expected to impart some mild steepening pressure to the curve, except for the fact that the front-end is already priced for 53 bps of easing during the next 12 months, significantly more than we expect. Our base case outlook calls for one more 25 bps rate cut tomorrow, followed by an extended on-hold period. If our base case scenario is incorrect, and growth continues to deteriorate, forcing the Fed to cut rates all the way back to zero. Then we would expect some initial bull-steepening, followed by bull-flattening as the funds rate approaches the zero bound. Wage Growth Wage growth is another excellent yield curve indicator, mainly because it helps determine the direction of the fed funds rate. Stronger wage growth causes the Fed to tighten and the curve to flatten. On the flipside, wage growth is a less effective indicator during Fed easing cycles, when it tends to lag changes in the funds rate (Chart 6). In fact, while wage growth is tightly correlated with the 2/10 slope, it lags changes in the slope by about 12 months (Chart 6, panel 2). Chart 6Wages Lead Tightening, But Lag Easing The upshot is that if the economy heads toward recession, then wage growth will not be a timely indicator of Fed rate cuts. However, if recession is avoided and wages continue to accelerate (Chart 6, bottom 2 panels), strong wage growth will limit how accommodative the Fed can be as it seeks to re-anchor inflation expectations. As such, persistently strong wage growth will limit the amount of curve steepening that can occur. Inflation Expectations The Fed’s need to re-anchor inflation expectations in a range consistent with its target is the main reason to forecast curve steepening. At present, the 10-year TIPS breakeven inflation rate is a mere 1.66%, well below the 2.3%-2.5% range that the Fed would consider “well anchored”. One might conclude that if the Fed succeeds in driving this rate higher, it will impart significant steepening pressure to the curve. However, we must also note that the 2-year TIPS breakeven inflation rate is even lower than the 10-year rate (Chart 7). Given our view that long-dated inflation expectations adapt only slowly to the actual inflation data, we would expect both the 2-year and 10-year breakevens to rise in tandem, exerting some modest flattening pressure on the curve.3 Chart 7Any Steepening Will Come From Real Yields Ironically, if the Fed is successful in re-anchoring long-dated inflation expectations, we expect it will cause the yield curve to steepen, but through its impact on real yields. At present, the 2-year and 10-year real yields are 0.37% and 0.14%, respectively. The act of holding rates steady for long enough to re-anchor inflation expectations will exert downward pressure on the 2-year real yield, while the 10-year real yield will rise in response to an improved growth outlook. The Fed’s goal of re-anchoring inflation expectations will likely lead to some curve steepening, but through the real component of yields, not the inflation component. The Neutral Rate The neutral rate – the fed funds rate that is neither inflationary nor deflationary – is a major wild card when it comes to the yield curve. Right now, the median Fed estimate calls for a neutral rate of 2.5%, while the market is pricing-in an even lower rate of 2%, at least according to the 5-year/5-year forward Treasury yield (Chart 8). Neutral rate estimates have been revised lower during the past few years, exerting significant flattening pressure on the yield curve. In theory, if we reach an inflection point where neutral rate estimates are revised higher, it would lead to substantial curve steepening. One thing to watch to help predict movement in neutral rate estimates is the gold price.4 Gold performs well when the market perceives monetary policy as increasingly accommodative, either because the Fed is cutting rates or because the assumed neutral rate is rising. The 2013 drop in gold foreshadowed downward revisions to the Fed’s neutral rate estimate (Chart 8, bottom panel). A further increase in gold, especially once the Fed stops cutting rates, would send a strong signal that current neutral rate estimates are too low. Monetary policy arguably exerts its greatest economic impact through the housing market. Investors can also watch the housing market for clues about the neutral rate. Monetary policy arguably exerts its greatest economic impact through the housing market. If housing activity starts to wane, it can be a strong signal that interest rates are too high. Last year, housing activity started to flag once the mortgage rate moved above 4% (Chart 9). If 4% proves to be the ceiling on mortgage rates, it would mean that the Fed’s current neutral rate estimate is roughly correct. However, home prices have moderated since last year, and new construction has started to focus more on the low-end of the market, where supply remains scarce.5 This shift in focus from homebuilders has caused the price of new homes to fall considerably (Chart 9, bottom panel), a supply side re-adjustment that could make the housing market more resilient in the face of higher rates. Chart 8Tracking The Neutral Rate: Gold Chart 9Tracking The Neutral Rate: Housing An upward re-assessment of the neutral rate would impart steepening pressure to the yield curve, but only if it occurs quickly, before the Fed has time to deliver offsetting rate hikes. However, we think it’s more likely that any increase in neutral rate estimates will occur gradually, alongside Fed tightening. In that case, a roughly parallel upward shift in the yield curve would be the most likely outcome. Verdict Considering all of the above factors, we would look for some modest 2/10 curve steepening during the next few months. The steepening will be driven by the Fed’s desire to re-anchor long-dated inflation expectations, a desire that will result in them keeping rates steady (apart from one more cut tomorrow), even as economic growth improves. As noted above, this steepening will show up in real yields, not in the TIPS breakeven inflation curve. That being said, strong wage growth and overly dovish market rate cut expectations will ensure that any steepening is well contained. We expect the 2/10 slope to stay in a range between 0 bps and 50 bps for the next 6-12 months. Yield Curve Strategy Chart 10Treasury Yield Curve When thinking about how to position a Treasury portfolio for our expected yield curve outcome, we first look at the value proposition offered by different Treasury maturities. Chart 10 shows the Treasury yield curve, and also each maturity’s 12-month rolling yield. The rolling yield is simply the combination of each maturity’s 12-month yield income and the price impact of rolling down the curve. It can be thought of as the return you would earn holding each bond for 12 months in an unchanged yield curve environment. The first thing that sticks out in Chart 10 is that the 5-year note offers poor value. We also note that the curve steepens sharply beyond the 5-year maturity point, so maturities greater than 5 years benefit a lot from rolldown. The simple intuition from Chart 10 is confirmed by our butterfly spread models.6  Chart 11shows that the 5-year bullet looks very expensive relative to a duration-matched barbell portfolio consisting of the 2-year and 10-year notes. In fact, with only a few exceptions, bullets are expensive relative to barbells across the entire Treasury curve (see Appendix). Chart 11Bullets Are Very Expensive All else equal, bullets tend to outperform barbells when the yield curve steepens. However, given current valuations, it would take a lot of steepening for bullets to outperform barbells during the next few months. Chart 12Yield Curve Correlations Further, Chart 12 shows that the front-end of the yield curve – out to about the 5-year/7-year point – tends to steepen when our 12-month discounter rises, while the long-end of the curve – beyond the 7-year point – tends to flatten. Given that our 12-month discounter is currently -53 bps, meaning that the market is priced for 53 bps of rate cuts during the next year, we expect it will rise during the next few months. This should exert the most upward pressure on the 5-year/7-year part of the curve. We have been recommending that investors play the curve by going long a 2/30 barbell and shorting the 7-year bullet. But given the significant rolldown advantage in the 7-year compared to the 5-year, we amend that recommendation this week. We now recommend that investors short the 5-year bullet and go long a duration-matched barbell consisting of the 2-year and 30-year maturities. Bottom Line: The 5-year Treasury note looks expensive compared to the rest of the yield curve, and historical correlations suggest it will rise the most if the Fed delivers fewer rate cuts than are currently expected. We recommend that investors short the 5-year bullet versus a duration-matched 2/30 barbell. Appendix Table 1Butterfly Strategy Valuation: Raw Residuals In Basis Points (As of October 25, 2019) Table 2Butterfly Strategy Valuation: Standardized Residuals (As of October 25, 2019) Ryan Swift U.S. Bond Strategist rswift@bcaresearch.com Footnotes 1 Please see U.S. Bond Strategy Weekly Report, “Crisis Of Confidence”, dated October 22, 2019, available at usbs.bcaresearch.com 2 For further details on BCA’s outlook for US/China trade negotiations please see Geopolitical Strategy Weekly Report, “How Much To Buy An American President?”, dated October 25, 2019, available at gps.bcaresearch.com 3 For further details on how inflation expectations adapt to the actual inflation data please see U.S. Bond Strategy Weekly Report, “Adaptive Expectations In The TIPS Market”, dated November 20, 2018, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, “A Signal From Gold?”, dated May 1, 2018, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, “The Long Awkward Middle Phase”, dated July 2, 2019, available at usbs.bcaresearch.com 6 For details on our butterfly spread models please see U.S. Bond Strategy Special Report, “Bullets, Barbells And Butterflies”, dated July 25, 2017, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Special Report

Highlights No, it’s not: We expect negative rates to remain the exception rather than the rule. A growing body of evidence suggests that negative rates may be doing more harm than good. Stronger global growth is likely to lift inflation over the next few years, thus making the debate around negative rates increasingly irrelevant. Contrary to conventional wisdom, there is scant evidence that structural forces related to globalization, automation, weak trade unions, and demographics are holding back inflation. Asset allocators should overweight global equities during the next 12-to-24 months, while maintaining a short duration bias in fixed-income portfolios.  A more defensive stance towards equities may be necessary starting in 2022. Just A Matter Of Time? Chart 1A Spike In Negative-Yielding Debt There is nearly $14 trillion of negative-yielding debt outstanding today (Chart 1). While most of this debt has been issued in the euro area and Japan, many investment professionals believe that negative yields will eventually become the norm in the U.S. and other developed economies. The rationale for this belief is easy to understand: The current expansion, like all past expansions, will inevitably end (in many investors’ minds, it already has). Once a recession is afoot, central banks will try to ease monetary policy even more than they already have. The Fed has cut rates by more than five percentage points on average during past recessions (Chart 2). Even a mild recession could see U.S. rates fall to zero. Once rates reach zero, pushing them into negative territory could become the logical next step. Chart 2Will The U.S. Join The Negative Rate Club After The Next Recession? It is a compelling argument. However, it rests on two assumptions. The first is that negative rates are an effective tool against an economic downturn. That is far from clear. Second, the argument presupposes that the forces which have pushed some countries to adopt negative rates will endure until the next recession. To those who see the current expansion as very “late stage” and regard the persistence of low interest rates as largely structural in nature, this is a perfectly plausible assumption. However, as we discuss later on, it is probably flawed. The Merits (Or Lack Thereof) Of Negative Rates In theory, negative rates could incentivize banks to loan out excess funds in order to avoid paying interest on reserves. It could also boost demand for credit. In practice, banks have been reluctant to force depositors to pay interest on their savings. Instead, they have absorbed the cost of negative rates through lower net interest margins. At a time when some banks are still struggling to shore up their balance sheets, the introduction of negative rates may have perversely resulted in less lending. Labor market slack has diminished significantly around the world. Some policymakers have slowly come around to the conclusion that negative rates may be doing more harm than good. Most senior Fed officials have rejected negative rates as an effective policy tool. Japanese and European officials have been more supportive of negative rates. The ECB even cut rates further into negative territory in September. However, ECB officials have acknowledged the harm done to the banking system by introducing a tiering system that shields a portion of excess bank reserves from negative deposit rates. The Swedish Riksbank, an early pioneer of negative rates, has even gone as far as to warn that “if negative nominal interest rates are perceived as a more permanent state, the behavior of agents may change and negative effects may arise.” Groundhog Day Judging by today’s low level of bond yields, it is easy to conclude that deflationary forces are just as powerful as they were a decade ago. There are, however, at least two important differences between now and then. First, the deleveraging cycle has ended in most developed economies. As a share of GDP, U.S. nonfinancial private-sector debt has risen over the past four years. Even in Japan, private debt levels have moved off their lows. The ratio of private debt-to-GDP has been broadly flat in the euro area, with rising debt levels in France offsetting falling leverage in Italy and Spain (Chart 3). Second, labor market slack has diminished significantly around the world. The unemployment rate in the G7 has fallen from a peak of 8.4% in 2009 to 4.2%. It is currently a full percentage point below its pre-recession low of 5.2% set in 2007 (Chart 4). Chart 3Deleveraging Has Ended In Most Developed Markets Chart 4Falling Unemployment Rate Across Developed Markets Some have argued that disguised joblessness is distorting the official unemployment statistics. While this was a major problem earlier in the recovery, it is much less of a concern today. In the U.S., the share of the working-age population that wants a job, but is not actively looking for one, is smaller than in 2007 (Chart 5). Whither The Phillips Curve? Falling unemployment has pushed up wage growth. Indeed, for all the talk about how the Phillips curve is dead, the “wage version” of the curve – which is how William Phillips originally formulated the concept – is very much alive and well (Chart 6). Chart 5U.S. Labor Market Slack Has Diminished Chart 6Falling Unemployment Has Pushed Up Wage Growth Chart 7Rising Labor Share Of Income Occurring Alongside Labor Market Tightening What is true is that the “price version” of the Phillips curve – the one that compares unemployment with price inflation – still looks very flat in most countries. This is another way of saying that rising nominal wages have mainly translated into higher real wages, with an accompanying increase in labor’s share of income (Chart 7). Workers tend to spend more of their incomes than companies. If the share of national income flowing to workers continues to rise, aggregate demand will increase. Unless supply expands in tandem, shortages of goods and services will arise, leading to higher inflation. Getting Close To The Kink There is considerable theoretical and econometric evidence suggesting that the Phillips curve is kinked.1  When slack is plentiful, modest declines in spare capacity have little effect on inflation. When slack disappears altogether, however, inflation can surge. This was certainly what happened during the 1960s. Chart 8 shows that U.S. core inflation was remarkably stable at around 1.5% in the first half of the decade. It was only in 1966 that inflation took off, rising to nearly 4% in less than two years. Core inflation proceeded to make its way to over 6% in 1970, a full three years before the first oil shock. The U.S. unemployment rate was two percentage points below NAIRU in 1966. By most estimates, the unemployment rate today is still a bit less than a point below its full employment level. Thus, an inflationary breakout is not imminent. This is confirmed by a wide variety of leading indicators for inflation (Chart 9). Chart 8Inflation Took Off In The 1960s Amid An Overheated Economy Chart 9An Inflation Breakout Is Not Imminent... Nevertheless, U.S. inflation has begun to firm at the margin (Chart 10). Trimmed mean inflation, which according to one Fed study does a better job of tracking underlying inflationary trends than more conventional measures, has been running at over 2% for much of the past 12 months.2  The median item in the CPI basket is rising by about 3%. Inflation has been slower to accelerate outside the U.S., partly because there is still more slack abroad. Nonetheless, embryonic signs of inflation are emerging. The deflationary pressures which plagued countries such as Spain have receded (Chart 11). Prices in Japan have been rising since 2014, albeit at a slower pace than the Bank of Japan is targeting (Chart 12). Chart 10... But Inflation Is Firming At The Margin Chart 11Deflationary Pressures Have Receded in Spain Chart 12Prices In Japan Have Been Rising Since 2014... Albeit At A Slower Pace Than The BoJ's Target The Myth Of Structurally Low Inflation Will structural forces contain the extent to which inflation rises even if unemployment continues to decline? Perhaps, but we would not bet on it. While globalization, automation, weak trade unions, and demographics are often cited as structural deflationary forces, the importance of these factors is greatly exaggerated. Globalization Conceptually, the disinflationary force stemming from globalization should be a function of the degree to which globalization is increasing. Yet, as Chart 13 illustrates, the ratio of global trade-to-GDP has been flat for over a decade. Correspondingly, the share of U.S. imports from emerging markets has stabilized at below 25%. Chart 13AGlobalization Has Peaked Chart 13BGlobalization Has Peaked A variety of studies have concluded that slack abroad has only a minimal effect on U.S. inflation.3 This is not surprising. The lion’s share of GDP consists of services, which are not easily tradeable. Imports account for only 14.8% of U.S. GDP. Many imported goods also have U.S. substitutes, which means that a large appreciation in the dollar is often necessary to induce Americans to shift purchases abroad.  Automation The belief that faster productivity growth is necessarily deflationary involves a fallacy of composition. Yes, above-average productivity gains in one sector of the economy will cause prices in that sector to decline relative to other prices. But falling prices will also boost real incomes, leading to more spending. Rising spending will lift prices elsewhere in the economy. Chart 14Globally, Productivity Growth Has Been Falling For Over A Decade Chart 15Steadier Prices For Computer Hardware And Software In Recent Years In any case, the whole narrative about how faster productivity growth is deflationary seems rather antiquated considering that productivity growth has been quite weak in most of the world for over a decade (Chart 14). Consistent with this, the price deflator for electronic goods has been falling a lot less rapidly in recent years than it has in the past (Chart 15). Chart 16Retail Sector Profit Margins Are Strong What about the so-called Amazon effect? The problem with the claim that online shopping is undermining corporate pricing power is that outside of department stores, profit margins in the retail sector remain quite high (Chart 16). In fact, recent productivity growth in the U.S. distribution sector has actually been slower than in the 1990s, a decade which produced large productivity gains stemming from the displacement of “mom and pop” stores with “big box” retailers such as Walmart and Costco. Trade Unions The declining influence of trade unions is often cited as a reason for why inflation will remain subdued. There are a number of problems with this argument. First, unionization rates in the U.S. peaked in the mid-1950s, more than a decade before inflation began to accelerate. Second, while the unionization rate continued to decline in the U.S. during the 1980s and 1990s, it remained elevated in Canada. Yet, this did not prevent Canadian inflation from falling as rapidly as it did in the United States (Chart 17). Chart 17Inflation Fell In Canada, Despite A High Unionization Rate Chart 18Higher Inflation Led To More Inflation-Indexed Wage Contracts, Not The Other Way Around The widespread use of inflation-linked wage contracts in the 1970s also appears to have been a consequence of rising inflation rather than the cause of it (Chart 18).   Demographics Demographics has undoubtedly been a deflationary force for most of the past 40 years. Slower population growth reduced spending on everything from houses to refrigerators, thus sapping demand from the economy. The influx of women into the labor force also boosted the available supply of goods and services, while the increase in the share of the population in their prime earning years – ages 30-to-50 – raised savings. Chart 19The Worker-To-Consumer Ratio Has Peaked Globally Now that baby boomers are starting to retire, however, they are transitioning from being savers to dissavers. Chart 19 shows that the ratio of workers-to-consumers has begun to decline globally as the post-war generation leaves the labor force. As more people stop working, aggregate savings will fall. The shortage of savings will put upward pressure on the neutral rate. If central banks drag their feet in raising policy rates in response to an increase in the neutral rate, monetary policy will end up being too stimulative. As economies overheat, inflation will pick up. It Shouldn’t Be Hard There are many hard problems in the world. Finding a cure for cancer is hard. Reconciling general relativity with quantum mechanics is hard. In contrast, getting people to spend money should not be hard. People like to consume! Just give them money and they will spend it. If they don’t spend enough of the money that they receive, just give them some more. So why has raising demand proven to be so difficult in many countries? The answer is that central banks have been asked to do too much. Fiscal policy should have been a lot more stimulative. If there is one potential benefit of negative rates, it is that they could incentivize governments to loosen fiscal policy by cutting taxes and/or raising spending. After all, if you can get paid to issue debt, why not do it? In an age of brewing political populism, the temptation to run larger budget deficits will grow. Central banks will indulge governments by keeping rates low. The path to higher rates is lined with lower rates. As economies eventually overheat, inflation will rise, thus allowing central banks to finally move away from negative rates. Real rates will stay low, but nominal rates will increase in line with higher inflation. Of course, if inflation eventually gets too high, central banks will be forced to step on the brakes. We do not see that happening in the next two years, but it could occur later on. Thus, asset allocators should overweight equities during the next 12-to-24 months, while maintaining a short duration bias in fixed-income portfolios. A more defensive stance towards equities may be necessary starting in 2022.   Peter Berezin Chief Global Strategist peterb@bcaresearch.com   Footnotes 1 Jeremy Nalewaik, “Non-Linear Phillips Curves with Inflation Regime-Switching,” Federal Reserve Board (Divisions of Research & Statistics and Monetary Affairs) (August 2016); and Anil Kumar and Pia Orrenius, “A Closer Look at the Phillips Curve Using State Level Data,” Federal Reserve Bank of Dallas, Working paper No. 1409 (May 2015). 2 Jim Dolmas and Evan F. Koenig, “Two Measures Of Core Inflation: A Comparison,” Federal Reserve Bank Of Dallas, Working Paper No. 1903 (February 25, 2019). 3 Please Jane Ihrig, Steven B. Kamin, Deborah Lindner, and Jaime Marquez, “Some Simple Tests of the Globalization and Inflation Hypothesis,” Board of Governors of the Federal Reserve System (International Finance Discussion Papers No. 891) (April 2007); Janet. L. Yellen, 'Panel discussion of William R. White “Globalisation and the Determinants of Domestic Inflation”,' Presentation to the Banque de France International Symposium on Globalisation, Inflation and Monetary Policy (March 2008); and Fabio Milani, “Global Slack And Domestic Inflation Rates: A Structural Investigation For G-7 Countries,”Journal of Macroeconomics, (32:4) (2010). Strategy & Market Trends MacroQuant Model And Current Subjective Scores Tactical Trades Strategic Recommendations Closed Trades

Special Report

I am on the road this week, so instead of our regular weekly report, we are sending you an update of our long-term fair value models. I hope to report any insights I have gained next week. Regards, Chester Ntonifor  Highlights Our long-term FX models are not sending any strong signals right now, with the U.S. dollar at fair value.  The cheapest currencies are the yen, the Norwegian krone and Swedish krona. The priciest currencies are the South African rand and the Saudi riyal. Feature This week we are updating our long-term FX models, part of a set of technical tools we use to help us navigate FX markets. Included in these models are variables such as productivity differentials, terms-of-trade shocks, net international investment positions, real rate differentials, and proxies for global risk aversion. These models cover 22 currencies, incorporating both G-10 and emerging market FX markets. The models are not designed to generate short- or intermediate-term forecasts. Instead, they reflect the economic drivers of a currency's equilibrium. Their main purpose is to provide information on the longevity of a currency cycle, depending on where we are in the economic cycle. For all countries, the variables are highly statistically significant, and of the expected signs. Together with other currency models we maintain in-house, these help us guide currency strategy, while providing a crosscheck when we might be offside. U.S. Dollar Chart 1The U.S. Dollar Is Close To Fair Value The uptrend in the dollar that has been in place since 2011 has lifted it only as far as the neutral zone. This is the biggest risk to our cyclical bearish dollar view. The big driver behind the uptrend has been interest rate differentials. If U.S. interest rates continue to roll over relative to their G-10 counterparts, this will lower the greenback’s fair value (Chart 1). The Euro Chart 2The Euro Is Trading At A Discount The euro is cheap by one standard deviation below its fair value. Historically, when the euro has hit its fair value bands, it has tended to mean-revert. The big driver lifting the euro’s fair value is the cumulative current account. Our bias is that the R-star for the euro area could start to head higher in the coming quarters, which will further lift its fair value (Chart 2). The Yen Chart 3The Yen Is Still Undervalued The yen is cheap by most relative price measures. The latest uptick in the yen’s fair value is driven by an appreciation in the gold-to-oil ratio, a measure of risk aversion (Chart 3). We believe the yen sits in a beautiful spot at the current economic juncture. Further deterioration in economic data will lead to higher risk aversion and a higher fair value. Meanwhile, a pickup in economic activity will still keep the fair value rising from a current account perspective.  The British Pound Chart 4GBP Grinding Higher Towards Its Fair Value The pound is cheap by most model measures, including our fundamental models. Downside in the pound has tended to capitulate around 1.5 standard deviations below fair value, even during the ERM crisis. Of course, the latest down leg has been politically driven, since the economic fair value of the pound has not really shifted by much (Chart 4). The Canadian Dollar Chart 5The Canadian Dollar Is Slightly Overvalued The fair value for the Canadian dollar has been falling since the 2011 peak in the commodity cycle. This still leaves the CAD slightly above fair value today (Chart 5). Meanwhile, the current account deficit has narrowed but remains quite wide by historical standards, which does not bode well for the CAD’s long-term fair value. The Australian Dollar Chart 6Aussie At Fair Value The recent drop in the Australian dollar has nudged it slightly below its fair value. However, like the Canadian dollar, the fair value of the Aussie has been dropping in recent years on the back of depressed commodity prices. Given the growing importance of liquified natural gas in Australia’s export mix, we believe terms of trade will remain a tailwind for the Australian currency over the longer term (Chart 6). The New Zealand Dollar Chart 7The Kiwi Has Been Fluctuating Around Its Fair Value The New Zealand dollar is currently at fair value, similar to its antipodean neighbor. Like other commodity currencies, its fair value has fallen in recent years. The catalyst has been the drop in commodity prices, along with the fall in relative real rates (Chart 7). The Swiss Franc Chart 8The Swiss Franc Is Not Expensive The Swiss franc is not as cheap as the yen, but our fundamental models show it as undervalued. The biggest driver in the rise of the franc’s fair value has been the structural trade surplus. The rise in the gold-to-oil ratio has further helped boost the fair value of the exchange rate (Chart 8). The Swedish Krona Chart 9The Krona Is Cheap The Swedish krona is one of the cheapest currencies in our universe, together with the Norwegian krone. The key model inputs for the Swedish krona are interest rate differentials and relative productivity trends. So, while the fair value of the krona has been falling for several years, the currency is still massively undershooting this fair value (Chart 9). The Norwegian Krone Chart 10The Krone Is Cheap Too The Norwegian krone is the cheapest it has been in the history of our model. More interestingly, the fair value has actually risen in recent years as the exchange rate has nosedived (Chart 10). The big driver in lifting the fair value has been the rise in crude oil prices. Within the commodity complex, the Norwegian krone is the most attractive. The Chinese Yuan Chart 11The Yuan Is Not Expensive The Chinese yuan is currently at one standard deviation below fair value. The yuan’s fair value has been mostly rising during the entire history of our model. This is driven predominately by higher relative productivity (Chart 11). We lie in the camp that there will be no significant devaluation in the RMB, in part because the exchange rate is already cheap. The Brazilian Real Chart 12The Brazilian Real Is Slightly Overvalued The Brazilian real is slightly above fair value, according to our fundamental models. Meanwhile, the fair value has been falling since 2011, in line with other commodity currencies (Chart 12). The current account component of the model should start to rise if reforms in Brazil lead to better productivity and improved competitiveness. The Mexican Peso Chart 13The Mexican Peso Is Now Above Fair Value The Mexican peso is trading a nudge above fair value. Over the last few years, opposing forces in the model have kept the fair value roughly flat. On one hand, the rising gold-to-oil ratio has been negative, as the peso is a cyclical currency. On the other hand, the cumulative current account has started to improve and bond yield differentials remain positive (Chart 13). The Chilean Peso Chart 14The Chilean Peso Is At Fair Value The fair value of the Chilean peso has been roughly flat for many years. This has also been the case for the real effective exchange rate, with fluctuations between half a percent of one standard deviation around fair value (Chart 14). This suggests the peso is mainly a trading currency, especially versus other emerging markets. The Colombian Peso Chart 15The Colombian Peso Is Depressed The Colombian peso is cheap, and is also one of our favorite petrocurrencies. The reason is that it has one of the strongest correlations to oil prices among commodity currencies, even though that correlation has been weakening (Chart 15). The South African Rand Chart 16The South African Rand Is Above Its Fair Value The South African rand is now trading slightly above its fair value (Chart 16). The correlation between precious metals prices and the South African rand has gradually weakened, largely due to domestic supply constraints and shrinking mining production. Meanwhile, the current account deficit continues to widen. This has gradually eroded the rand’s fair value. The Russian Ruble Chart 17The Russian Ruble Is Not Cheap The Russian ruble is now sitting around 0.5 standard deviations above its fair value (Chart 17). We are positive on oil, which will boost the fair value of petrocurrencies, including the Russian ruble. Meanwhile, real interest rates are at relatively high levels in Russia, even though the model’s results do not provide significant explanatory power. We are currently long RUB/EUR in our petrocurrency basket, with the Russian ruble being the best-performing petrocurrency. The Korean Won Chart 18The Korean Won Has Cheapened Further The Korean won has underperformed this year, and is now trading at a non-negligible discount to its fair value (Chart 18). Meanwhile, the fair value of the Korean won has been rising over the years. This has been partly driven by an increasing current account surplus – at least up until the trade war began. The fair value also tends to benefit from risk flare-ups. The Philippine Peso Chart 19The Philippine Peso Has Appreciated The Philippine peso has increased by 5% against the U.S. dollar year-to-date. However, despite its recent appreciation, the peso is still trading at a 6% discount to its long-term fair value (Chart 19). The Philippine peso is one of the few currencies whose REER tends to have well-defined and long cycles that last five-to-eight years. It will be important to watch if the recent appreciation is the start of a new trend. The Singapore Dollar Chart 20The Singapore Dollar Is Still Overvalued The Singapore dollar is another currency whose REER tends to have long cycles, probably a feature of the managed float (Chart 20). The Singapore dollar is a defensive currency, and so the decline in other emerging market currencies has made it slightly expensive. The Hong Kong Dollar Chart 21The Hong Kong Dollar Is Overvalued The HKD’s REER has been rising in recent years, meaning inflation in Hong Kong has been outpacing that of other regions (Chart 21). This has made the HKD expensive, according to our models. However, the fair value has been on an uptrend in recent years, in part driven by rising relative productivity. The Saudi Riyal Chart 22The Saudi Riyal Is Expensive The fair value of the Saudi riyal has been falling for quite a while on declining relative productivity (Chart 22). This has made the riyal incrementally expensive. However, it may take much more stretched valuations before greater tensions arise in the peg. Trades & Forecasts Forecast Summary Core Portfolio Tactical Trades Limit Orders Closed Trades

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

Highlights Shifting Trends: The factors that have driven bond yields lower throughout 2019 – slowing growth, rising uncertainty, demand for safe assets and dovish monetary policy expectations – have all started to turn in a more bond-bearish direction. Duration & Country Allocation Strategy: Maintain a moderate below-benchmark stance on aggregate bond portfolio duration. Favor lower-beta countries with central banks that are more likely to stay relatively dovish as global yields drift higher, like core Europe, Australia and Japan. Credit Allocation Strategy: Stay overweight corporate bonds versus government debt in the U.S. and Europe, both for investment grade and high-yield. Maintain just a neutral stance on EM USD-denominated spread product, but look to upgrade if global growth improves further and the USD begins to weaken. Feature Chart of the WeekBond Yields Sniffing A Turn In Global Growth? It has been fifty days (and counting) since the 2019 low for the benchmark 10-year U.S. Treasury yield was reached on September 3. The year-to-date low for the benchmark 10-year German bund yield was seen six days before that on August 28. Yields have risen by a healthy amount since those dates, up +34bps and +37bps for the 10yr Treasury and Bund, respectively. This has occurred despite the significant degree of bond-bullish pessimism on global growth and inflation that can be found in financial media reporting and investor surveys. The fact that yields are now steadily moving away from the lows suggests that the 2019 narrative for financial markets – slowing global growth, triggered by political uncertainty and the lagged impact of previous Fed monetary tightening and China credit tightening, forcing central banks to turn increasingly more dovish – is no longer correct. If that is true, yields have more near-term upside as overbought government bond markets begin to “sniff out” a bottoming out of global growth momentum (Chart of the Week). In this Weekly Report, we take a look at the changing state of the factors that fueled the sharp decline in bond yields in 2019. We follow that up with a review of all our current recommended investment positions on duration, country allocation and spread product allocations in light of recent developments. We conclude that maintaining a below-benchmark duration exposure, while favoring lower-beta countries in sovereign debt and overweighting corporate debt in the U.S. and Europe, is the most appropriate fixed income strategy for the next 6-12 months. The timing of the bottoming of yields in the major developed markets (DM) should not be surprising, given the more bond-bearish turn of reliable leading directional yield indicators. Yields Are Rising At The Right Time, For The Right Reasons Chart 2Bond-Bullish Growth & Inflation Factors Are Turning The timing of the bottoming of yields in the major developed markets (DM) should not be surprising, given the more bond-bearish turn of reliable leading directional yield indicators. The diffusion index of our global leading economic indicator (LEI), which leads the real (ex-inflation expectations) component of DM bond yields by twelve months, is at an elevated level (Chart 2). At the same time, the slowing of the annual rate of growth in the trade-weighted U.S. dollar, which leads 10-year DM CPI swap rates by around six months, is signaling that bond yields have room to increase from the inflation expectations side. Finally, the rising trend of positive data surprises for the major DM countries is also pointing to higher yields. Breaking it down at the country level, the pickup in DM 10-year bond yields since the 2019 lows has been widespread (Charts 3 & 4). The range of yield increases is as low as +16bps in Japan, where the Bank of Japan (BoJ) is pursuing a yield target, to +46bps in Canada where the economy and inflation are both accelerating. Chart 3Pricing Out Some Expected Rate Cuts … Chart 4… Across All Developed Markets The increase in yields has also occurred alongside reduced expectations for easier monetary policy. Our 12-month discounters, which measure the expected change in short-term interest rates priced into Overnight Index Swap (OIS) curves, show that markets have partially priced out some (but not all) expected rate cuts in all major DM countries. The Three Things That Have Changed For Global Bond Markets So what has changed to trigger a reduction in rate cut expectations and an increase in global yields? The bond-bullish narrative that we refer to in the title of this report can be broken down into the following three elements, which have all turned recently: Slowing global growth (now potentially bottoming) Chart 5Global Growth Bottoming Out Current global growth is still trending lower, when looking at measures like manufacturing PMIs or sentiment surveys like the global ZEW index. Forward-looking measures like our global LEI, however, have been moving higher in recent months, suggesting that a bottom in the PMIs may soon unfold (Chart 5). We investigated that improvement in our global LEI in a recent report and concluded that the move higher was focused almost exclusively within the emerging market (EM) sub-components that are most sensitive to improving global growth.1 This fits with the improvement shown in the OECD LEI for China, a bottoming of the annual growth rate of world exports, and the general acceleration of global equity markets – the classic leading economic indicator. Rising political uncertainty (now potentially fading) The U.S.-China trade war (including the implications for the upcoming 2020 U.S. presidential election) and the U.K. Brexit saga have been the main sources of bond-bullish political uncertainty over the past several months. Yet recent developments have helped reduce the odds of the most negative tail risk outcomes, providing a bit of a boost to global bond yields. The U.S. and China have agreed (in principle) to a “phase one” trade deal that, at a minimum, lowers the chances of a further escalation of the trade dispute through higher tariffs. Meanwhile, the momentum has shifted towards a potential final Brexit agreement between the U.K. and European Union that can avoid an ugly no-deal outcome. Our colleagues at BCA Research Geopolitical Strategy believe that developments are likely to continue moving away from the worst-case scenarios, given the constraints faced by policymakers.2 U.S. President Donald Trump is now in full campaign mode for the 2020 elections and needs a deal (of any kind) to deflect criticism that his trade battle with China is dragging the U.S. economy into recession. Already, there has been a sharp decline in income growth for workers in swing states that could vote for either party’s candidate in next year’s election (Chart 6). Trump cannot afford to lose voters in those states, many of which are in the U.S. industrial heartland (i.e. Ohio, Michigan) that helped put him in the White House. In other words, he is highly incentivized to turn down the heat on the trade war or else face a potential loss next November. While these political uncertainties have not been fully resolved by these latest developments, the shift in momentum away from worst-case scenarios has likely been enough to reduce the safe-haven bid for DM government bonds, helping push yields higher. Meanwhile, China is facing a slowing economy and rising unemployment, but with reduced means to fight the downtrend given high private sector debt that has impaired the typical response between easier monetary conditions and economic activity (Chart 7). While the Chinese government does not want to be seen as caving in to U.S. pressure on trade policy, its desire to maintain social stability by preventing a further rise in unemployment from the trade war provides a powerful incentive to try and ratchet down tensions with the U.S. Chart 6Political Reasons For Trump To Retreat On Trade In the U.K., a no-deal Brexit is an economically painful and politically unpopular outcome that would severely damage the re-election chances of Prime Minister Boris Johnson and his Conservative party. Thus, even a hard-line Brexiteer like Johnson must respond to the political constraints forcing him to try and get a Brexit deal done (Chart 8). Chart 7Economic Reasons For China To Retreat On Trade Chart 8Political Reasons To Retreat On A No-Deal Brexit While these political uncertainties have not been fully resolved by these latest developments, the shift in momentum away from worst-case scenarios has likely been enough to reduce the safe-haven bid for DM government bonds, helping push yields higher. Bull-flattening pressure on yield curves (now turning into moderate bear-steepening) The final leg down in bond yields in August had a technical aspect to it, fueled by the demand for duration and convexity from asset-liability managers like European pension funds and insurance companies. Falling yields act to raise the value of liabilities for that group of investors, forcing them to rapidly increase the duration of their assets to match the duration of their liabilities (the technique used to limit the gap between the value of assets and liabilities). That duration increase is carried out by buying government bonds with longer maturities (and higher convexity), but also through the use of interest rate derivatives like long maturity swaps and swaptions. The end result is a bull flattening of yield curves (both for government bonds and swaps) and a rise in swaption volatility (i.e. the price of swaptions). Those dynamics were clearly in play in August after the shocking imposition of fresh U.S. tariffs on Chinese imports early in the month. Bond and swaption volatilities spiked, and bond/swap yield curves bull-flattened, in both Europe and the U.S. (Chart 9). That effect only lasted a few weeks, however, and volatilities have since declined and curves have steepened. This suggests that the “convexity-buying” effect has run its course and is now starting to work in the opposite direction, with asset-liability managers looking to reduce the duration of their assets as higher yields lower the value of their liabilities. This is putting some upward pressure on longer-maturity global bond yields. Chart 9Signs Of Reduced Convexity-Related Bond Buying Chart 10Bull-Flattening Yield Curve Pressures Easing Up A Bit Chart 11Fed & ECB Actions Should Help Steepen Up Curves The steepening seen so far must be put in context, however, as yield curves remain very flat across the DM world (Chart 10). Term premia on longer-term bonds remain very depressed, although those should start to increase as global growth stabilizes and the massive safe-haven demand for global government debt begins to dissipate. Some pickup in inflation expectations would also help impart additional bear-steepening momentum to yield curves – a more likely result now that the Fed and ECB have both cut interest rates and, more importantly, will start provide additional monetary easing by expanding their balance sheets (Chart 11). Bottom Line: The factors that have driven bond yields lower throughout 2019 – slowing growth, rising uncertainty, demand for safe assets and dovish monetary policy expectations – have all started to turn in a more bond-bearish direction. Reviewing Our Recommended Bond Allocations In light of these shifting global trends described above, the fixed income investment implications are fairly straightforward: Yields are rising around the world, suggesting that the current move is a shift higher driven by non-country-specific factors like more stable future global growth prospects. Duration: A moderate below-benchmark overall duration stance is warranted for global fixed income portfolios, with yields likely to continue drifting higher over at least the next six months. A big surge in yields is unlikely, as central banks will need to see decisive evidence that global growth is not only bottoming, but accelerating, before shifting away from the current dovish bias. Given the reporting lags in the economic data, such evidence is unlikely to appear until the first quarter of 2020 at the earliest. Yet given how flat yield curves are across the DM government bond markets, the trajectory of forward rates is quite stable relative to spot yield levels, making it much easier to beat the forwards by positioning for even a modest yield increase. Country Allocation: Yields are rising around the world, suggesting that the current move is a shift higher driven by non-country-specific factors like more stable future global growth prospects. In that case, using yield betas to the “global” bond yield is a good way to consider country allocation decisions within a fixed income portfolio. We looked at those yield betas in an August report, using Bloomberg Barclays government bond index data for the 7-10 year maturity buckets of individual countries and the Global Treasury aggregate (Chart 12).3 The rolling 3-year betas were highest in the U.S. and Canada, making them good countries to underweight within a global government bond portfolio in a rising yield environment. The yield betas were lowest in Japan, Germany and Australia, making them good overweight candidates. The U.K. was a unique case of having a relatively high historical yield beta prior to the 2016 Brexit referendum and a lower yield beta since then - making the U.K. allocation highly conditional on the resolution of the Brexit uncertainty. Spread Product Allocation: The backdrop described in this report, where global growth is bottoming out but where central banks maintain a dovish bias, is a perfect sweet spot for global spread product like corporate bonds and Peripheral European government debt. Thus, an overweight stance on overall global spread product versus governments is warranted. The backdrop described in this report, where global growth is bottoming out but where central banks maintain a dovish bias, is a perfect sweet spot for global spread product like corporate bonds and Peripheral European government debt. With regards to our current strategic fixed income recommendations and model bond portfolio allocations, we already have much of the positioning described above in place. We are below-benchmark on overall duration, underweight higher-beta U.S. Treasuries; overweight government bonds in lower-beta Germany, France, Japan and Australia (Chart 13); overweight investment grade corporate bonds in the U.S., euro area and U.K.; and overweight high-yield corporate bonds in the U.S. and euro area. Chart 12Favor Lower-Beta Government Bond Markets There are areas where our positioning could change, however. Chart 13Lower-Beta Laggards Should Start To Outperform In terms of government bonds, we are currently overweight the U.K. and neutral Canada. A final Brexit deal would justify a downgrade of Gilts to at least neutral, if not underweight, as the Bank of England has signaled that rate hikes would be justified if the Brexit uncertainty was resolved. A downgrade of higher-beta Canadian government debt to underweight could also be justified, although the Bank of Canada is not signaling that a change in monetary policy (in either direction) is warranted. For now, we will hold off on any change to our U.K. stance, as it is now likely that there will be another extension of the Brexit deadline beyond October 31. As for Canada, we remain neutral for now but will revisit that stance in an upcoming Weekly Report. With regards to spread product, we are only neutral EM USD-denominated sovereign and corporate debt, as well as Spanish sovereign bonds; and underweight Italian government debt. An EM upgrade to overweight would require two things that are not yet in place: a weaker U.S. dollar and accelerating Chinese economic growth. Chart 14Stay Overweight Corporates In The U.S. & Europe As for Peripheral governments, we have preferred to be overweight European corporate debt relative to sovereign bonds in Italy and Spain. The recent powerful rally in the Periphery, however, has driven the spreads over German bunds in those countries down to levels in line with corporate credit spreads (Chart 14). We will maintain these allocations for now, but will investigate the relative value proposition between euro area Peripheral sovereigns and corporates in an upcoming report. Bottom Line: Maintain a moderate below-benchmark stance on aggregate bond portfolio duration. Favor lower-beta countries with central banks that are more likely to stay relatively dovish as global yields drift higher, like core Europe, Australia and Japan. Stay overweight corporate bonds versus government debt in the U.S. and Europe, both for investment grade and high-yield. Maintain just a neutral stance on EM USD-denominated spread product, but look to upgrade if global growth improves further and the USD begins to weaken. Robert Robis, CFA Chief Fixed Income Strategist rrobis@bcaresearch.com   Footnotes 1 Please see BCA Research Global Fixed Income Strategy Weekly Report, “What Is Driving The Improvement In The BCA Global Leading Economic Indicator?”, dated October 2, 2019, available at gfis.bcaresearch.com. 2 Please see BCA Research Geopolitical Strategy Weekly Report, “Five Constraints For The Fourth Quarter”, dated October 11, 2019, available at gps.bcaresearch.com. 3 Please see BCA Research U.S. Bond Strategy/Global Fixed Income Strategy Weekly Report, “Where’s The Positive Carry In Bond Markets?", dated August 20, 2019, available at usbs.bcaresearch.com and gfis.bcaresearch.com. Recommendations The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Special Report

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

Special Report

Highlights As an introduction to a series of BCA Special Reports on the investment consequences of climate change, we review the science around the subject and suggest a framework for analyzing its implications. The scientific consensus is that global warming is a reality and most likely human-induced. However, the uncertainty around the magnitude of the impact of climate change is large. The consequences of climate change are delayed, uncertain and global. But, for investors, the prudent course of action is to accept the scientific consensus – and the impact it will have on policymakers – and hedge or invest appropriately.  Feature Chart 1Climate Change Global Perception Bank of England Governor Mark Carney has called climate change “the tragedy of the horizon.” It is now perceived as a major threat across the globe (Chart 1). As such, it is essential to assess its macro and market consequences. In this introduction to our Climate Change Special Series, we review the existing literature and suggest a framework to assess the market relevance of this phenomenon. Going forward, we will produce a series of market-driven reports designed to help investors both mitigate the risk to their portfolios and identify opportunities arising from climate change. We intend to cover topics such as green financing, energy, and the geopolitical aspect of climate change, just to cite a few.  These reports will incorporate both quantitative and qualitative analysis to generate actionable investment recommendations. What Is Climate Change? Climate science is not new. The initial understanding of the effect of heat-trapping gases on global temperature dates back to Joseph Fourier’s early 1800s study of planetary temperature. Subsequent research showed the importance of the greenhouse effect, a phenomenon whereby greenhouse gas molecules (e.g. CO2, CH4, N2O) absorb infrared radiation emitted from Earth before reemitting it in all directions, including back to the Earth’s surface, thus making it harder for this energy to leave the planet. This excess of energy stored in the planet, above its normal energy balance, causes temperature increases. The distribution of environmental damages caused by global warming will not be uniform around the world. The rate of warming and other climate changes will differ across regions due to climate processes and feedbacks linked to local conditions.1 Regardless, up to 14% of the global population will experience above 2°C (3.6°F) warming – a level seen by scientists as a trigger for permanent damages and changes – even if the increase in global mean surface temperature (GMST) were limited to 2°C (3.6°F) by 2100 (CarbonBrief, 2018). The consequences of climate change are delayed, uncertain and global.  Even under the maximum policy effort scenario, studies assign 60% odds to an increase greater than 2°C (3.6°F) (Nordhaus, 2018). The longer policymakers, companies and investors delay tackling this issue, the less likely the world will stay below the 2°C threshold and the more rapid and abrupt the transition to a low-carbon economy will eventually be. A sudden transition will be more disruptive to the economy and damaging to investors. Defining The Issue: The Earth’s Atmosphere As A Global Common The Earth’s atmosphere - specifically its function as a sink for CO2 and other greenhouse gases (GHG) - falls within the problem of the global commons.2 It is a natural resource requiring global cooperation for its sustainable use and provision. Problematically, the consequences of climate change are delayed, uncertain and global.  Delayed because the burden of climate change policies mainly falls on current generations, whereas the benefits of lower climate damage accrue to future generations, leading every generation to think it can survive the issue and let the next generations handle it. Uncertain because the list of harms from climate change lengthens with the advance in climate-science studies. We learn more and more about the extent to which human activities are at fault and the extent of the damage that will befall the planet. Global because it does not matter whether the emissions take place in China, Europe, or the U.S. since GHG mix immediately once in the atmosphere. In that sense, it is a collective-action problem in which every country’s interest is to shift the abatement costs onto its neighbor. The global aspect is crucial. The optimal emission level of one country does not follow the global social optimal. Hence, every country has an incentive to emit as much GHG as possible now, before any consequences occur (Combes, 2016). What We Know So Far: Historical Data Both climate-alarmist and climate-denier groups have captured the public debate.This polarization clouds the underlying facts about current trends and the difference between what is unlikely, likely, or very likely to happen. The resulting lack of consensus will lead to over- or under-adaptation by the various economic agents, depending on their interests. FACT 1: GLOBAL WARMING IS A REALITY Anthropogenic Greenhouse Gas Emissions - Emissions of carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) have risen steadily since the industrial revolution and at a brisk pace relative to the previous 12,000 years (Chart 2). Chart 2GHG Global Emissions Global Mean Surface Temperature - It rose by an estimated 1°C (1.8°F) from 1901 to 2016. According to NASA data, the 10 warmest years recorded in the past 139 years all occurred after 2005 (Chart 3). Chart 3Global Land And Ocean Temperature Global Mean Sea level - It has risen by an estimated 20.3cm (8 inches) since 1900 due to the expansion of waters and meltwater from shrinking ice sheets. Almost half of this rise happened in the last 25 years (Chart 4). Glacier and Ice Sheet - The melting of ice sheets will reduce the earth’s reflectivity, accelerating the warming process (Chart 5). The record low of sea ice extent in the Arctic and Antarctic was observed in 2012 and 2017, respectively. Chart 4Global Mean Sea Level Chart 5Glacier And Ice Sheet Precipitation - Historical changes in precipitation are much more volatile and region-specific than temperature and sea level changes. Moreover, there is a lack of data covering the period before 1951, which leads to low confidence in estimates of precipitation for this period and medium confidence post-1951. Annual average precipitation for global land areas increased slightly over the period 1901–2008, and the magnitude of observed changes varies across different datasets (Hartmann, 2013). Extreme Weather Events - These are defined, in a meteorological sense, as events at the “edges of the complete range of weather experienced in the past.” The frequency and severity of extreme weather events has been linked to global warming (Table 1) (Scott, 2016). Table 1Extreme Weather Events (1950 - Present) FACT 2: CLIMATE CHANGE IS HUMAN-INDUCED The Intergovernmental Panel on Climate Change (IPCC) – considered the world’s most authoritative scientific body on climate change – concluded in 2013 that the probability that global warming was human-induced was at least 95% (Table 2).  Table 2Evolution Of The Assessments Of Human Influence On Climate Change Chart 6Global Warming & Global GHG Emissions Since the late nineteenth century, GHG emissions – mainly CO2 – and global land and ocean mean temperature have shared a common steep upward trend (Chart 6). A recent study by Mann et al. estimates that in the absence of GHG emissions, the odds that 13 out of the 15 warmest years ever measured would all have happened in the current century are extremely small.3 More recently, a report by the National Academies of Sciences, Engineering, and Medicine (NASEM) concluded that “[I]n many cases, it is now possible to make and defend quantitative statements about the extent to which human-induced climate change has influenced either the magnitude or the probability of occurrence of specific types of events or event classes.”  According to most recent peer-reviewed studies, at least 97% of actively publishing climate scientists now accept human-caused climate warming (Cook, 2016). While science is not a matter of popular vote, this level of consensus among experts suggests that for investors the most prudent course of action is to accept the scientific consensus and hedge or invest appropriately. Projections & Assumptions Chart 7Global Emissions Projections Climate economics deals with conditional projections based on unknown probability distributions, implying a high level of uncertainty. The level of confidence around the nearer segments of the projections is relatively elevated. Conversely, at the far end of the projected period, by 2100 for most studies, the uncertainty increases drastically. According to the United Nations Environment Programs’ 2018 Emissions Gap report,  the 2°C (3.6°F) target drafted in the Paris Agreement in 2015 would require global emissions to be capped at 40 gigatons of CO2 equivalent by 2030. Throughout our Climate Change Special Series, we will rely on the following assumptions based on the IPCC Fifth Assessment Report (AR5) and the summary estimates from around 150 academic papers, the majority of which were published in 2018 (CarbonBrief, 2018).  Anthropogenic Greenhouse Gas Emissions - Global emissions rose in 2017 and are now ~14 GtCO2e above the required level by 2030. Current pledges are insufficient to meet the Paris Agreement’s long-term temperature goals (Chart 7). Key factors driving changes in anthropogenic GHG emissions are mainly economic and population growth. Projections of greenhouse gas emissions vary over a wide range, depending on both socio-economic development and climate policy – which are fundamentally uncertain. Climate economics deals with conditional projections based on unknown probability distributions, implying a high level of uncertainty. The majority of models indicate that scenarios meeting levels similar to RCP2.6 (a scenario that aims to keep global warming likely below 2°C (3.6°F) above pre-industrial temperatures) are characterized by substantial net negative emissions by 2100, on average 2 GtCO2e per year.   Chart 8Global Mean Surface Temperature Projections Global Mean Surface Temperature - Under all assessed emission scenarios, surface temperature is projected to rise over the twenty-first century. The change over the 2016-2035 period will be very similar to 1986-2005, and will likely be in the range of 0.3°C to 0.7°C (0.5°F to 1.3°F). Beyond that, the mean temperature rise across IPCC scenarios for 2046-65 and 2081-2100 is estimated to be 1.4°C (2.5°F) and 2.2°C (4°F), respectively (Chart 8). These estimates imply that there will be more frequent hot and fewer cold temperature extremes over most land areas on daily and seasonal timescales. Global Mean Sea Level - It has been established that the likelihood sea levels will rise in more than 95% of the ocean area is very high. Under all IPCC scenarios, the rate of sea level rise will very likely exceed the observed rate during 1971-2010. About 70% of the coastlines worldwide are in fact projected to experience sea level change within +/- 20% of the global mean. Precipitation - There are likely more land regions where the number of heavy precipitation events has increased than where it has decreased. Recent detection of increasing trends in extreme precipitation and discharge in some catchments implies greater risks of flooding at regional scale (medium confidence). These changes will not be uniform, with high latitudes and the equatorial Pacific more likely to experience an increase in annual mean precipitation while many mid-latitude and subtropical dry regions are likely to experience a decrease in mean precipitation. It remains a challenge to determine long-term trends in precipitation for the global oceans. Extreme Weather Events - Projections on extreme weather events can only infer the probability distribution of such events, i.e. more or less likely to happen. With a 1°C (1.8°F) additional warming, risks from extreme weather events are high (medium confidence from IPCC). More importantly, we can say with high confidence that these risks increase progressively with further warming.     Embracing Uncertainty The uncertainty around the magnitude of the impact of climate change is large. Yet, bounded uncertainty is informational. We can extract the following important, actionable conclusions: Projections for economic variables are relatively more uncertain than for geophysical variables. The link between GHG emissions and rising temperature is more certain than the level of emissions, output, and damages (Nordhaus, 2018).  Therefore, the largest uncertainty comes from economic growth and the level of emissions. We do not rely on estimates of global GDP impacts. On the other hand, it is easier to build scenarios for geophysical variables and obtain investment-relevant information from these projections. Simulating the path of future emission allows us to map this onto future temperature, sea level, and extreme weather variations. Economic models suggest that the higher the uncertainty, the larger the weights on low-probability/high-impact scenarios. This implies a positive risk premium due to risk aversion and favors stricter mitigation policies as insurance to shattering outcomes. As climate models are fine-tuned and continuously point to large damage uncertainty, the desired strength of policy could increase. Win-Win or “no-regrets” investments are the most likely at first.4 The Kaya Identity provides a simple framework to project future GHG emissions to visualize the uncertainty associated with different assumptions. The identity links future emissions to observable macroeconomic variables (see the Appendix for more details): F = P * (G/P) * (E/G) *(F/E) Where F denotes global CO2 emissions from human sources, P represents global population, G equals global GDP, and E is global energy consumption. The identity provides a useful framework for policymakers. To reduce emissions, there needs to be a reduction in one or more of the identity's components. Altering demographic trends and reducing global GDP per capita are very unlikely to happen given the damaging impact it could have – both for individuals and politicians’ careers!  At a global level, this leaves us with energy efficiency and carbon intensity of energy as the only key and viable options to reduce CO2 emissions.  Why Does It Matter To Investors? Markets are probably still underpricing climate-related risks because the effects only materialize gradually and in the long term – exceeding most investors’ investment horizon. Investors such as pension funds, insurers, wealth managers, and endowments need to be responsive to the threat posed by climate change. They typically have multi-decade time horizons, with portfolio exposure across the global economy. Their increasing interest in Environmental, Social, and Governance (ESG) measures fits well within this context.5 It reflects a need for more transparency and more stringent investing standards. Determining which firms or sectors will either win or lose the “green race” will be of the outmost importance to investors. Businesses are still navigating the financial and operational implications of climate change. To some extent, this can already be assessed based on the readiness of firms and sectors to adapt to a green economy – looking at the number of environmental technology patent applications, for example. Markets are probably still underpricing climate-related risks. The financing needed to mitigate climate change represents yet another opportunity for investors. Green bonds and sustainability-linked debt instruments are more widespread than ever. Sustainable debt issuance reached record levels last year, with a total of $260 billion issued, according to Bloomberg New Energy Finance data. Year-to-date issuance has nearly reached $180 billion.  Green bonds offer two main benefits to issuers:  corporate branding that sends a strong signal to the market of their commitment to climate change, and a wider investor base. Our series of market-driven reports are intended to both identify the risks and opportunities arising from climate change in order to help investors mitigating the risk to their portfolios. They will rely on the simple framework we present below. Climate Change Framework In future reports in our Climate Change Special Series, we will summarize our findings using a comprehensive analytical framework developed by Batten (2018) to assess the impact of climate change via physical and transition risks with respect to the type of shock induced by each type of risk. Physical Risks Physical risks are the most visible and immediate source of risk to investors and the financial sector. They can be defined as those risks that arise from the interaction between climate-related events and human and natural systems, including their ability to adapt— e.g. the volatility in food prices following a drought or a flood.6 An increase in climate-induced physical risks – such as heat waves, floods and storm – will have a direct effect on insurers. If these risks are uninsured, the deterioration of the balance sheets of affected households and corporations is likely to hurt the banking system. Electrical utilities, real estate and transportation infrastructure are other physical assets at risk of capital losses. Transition Risks Chart 9Public Opinion Of Policy Options To Tackle Climate Change Transition risks can be defined as the risks of economic dislocation and financial losses associated with the transition to a lower-carbon economy. Detrimental effects manifest themselves through three possible channels: Reduced production and consumption of high carbon products, especially energy produced using fossil fuels, potentially leading to stranded assets. Improvement in the energy efficiency of existing products and processes – energy intensity. Moving to low-carbon energy production – that is reducing carbon intensity. Lower energy intensity and carbon intensity, highlighted in the Kaya Identity above, can be achieved through technological innovation. The relationship between climate change and policy or regulatory framework is manifold, as policymakers will need both to respond to the consequences of climate change and to shape future GHG emissions. The primary responsibility for strategic planning rests with governments, which have a variety of policy options at their disposal (Chart 9).  Table 3 provides a useful template to link both physical and transition risks to the type of shocks they can induce, and importantly, how it can ultimately turn into financial and geopolitical risks. Table 3A Simple And Useful Template To Summarize Our Findings Climate change can impact demand (from investment, consumption or trade) or supply (labor, capital stock, technology or other inputs). For example, transition risks such as distortions from asymmetric climate policies across countries could directly impact trade or investment (FDI). This is what is commonly referred to as the pollution haven hypothesis, which states that more stringent environmental regulations induce polluting industries to relocate to countries with relatively lax environmental regulations. Ensuing reports in the Climate Change Special Series will include this template as a mean to summarize our findings. APPENDIX The Kaya Identity And Uncertainty Feedback Loop7 Diagram 1The Uncertainty Feedback Loop The Kaya Identity links observable macroeconomic and demographic variables to GHG emissions: CO2 = P * (Y / P) * (E / Y) * (CO2 / E) Where denotes P global population, Y global GDP, and E primary energy consumption.    It highlights the large degree of uncertainty around the macroeconomic impact on GHG emissions – especially at the end of the forecast period when additional uncertainty emanates from the feedback loop illustrated in Diagram 1. Historical Trend In CO2 Emissions From 1990 to 2014 CO2 emissions growth was 2.1% p.a.8: Global CO2 emissions during this period were pushed higher by population growth (1.3% p.a.) and rising rates of GDP per capita (1.9% p.a.). This was partly offset by declining energy intensity (-1.3% p.a.) (Chart 10). Chart 10Kaya Identity Components: Global Level The extent of the impact of these variables on CO2 emissions is region-specific. Therefore, when the identity is expressed at an aggregate and global level, it can lead to inaccuracies in long-term scenario analysis since it does not account for dependencies across the variables and does not differentiate between high population growth in countries with low vs. high GDP per capita growth, or between high GDP per capita growth from countries with high vs. low carbon intensity energy sources.  Using The Kaya Identity To Project Future GHG Emissions Population - The UN currently expect the population to grow by an average 0.4% p.a. through 2100 in its medium variant scenario. GDP per capita - The OECD projects GDP per capita will grow 2.2% p.a. between 2018 and 2060. Energy Intensity - We assume a 1.5% p.a. decline in energy intensity over the 2018-2100 period – the trend over the past decade. Carbon Intensity - In line with scenario B2 of the IPCC Special Report on Emissions Scenarios (SRES), we assume a 0.4% p.a. Combined, this leads to a 21% increase in CO2 emission by 2050, and a 63% increase by 2100. Accounting for other scenarios for each component results to a wide range of potential cumulative CO2 emissions; a median temperature between 2.6°C and 4.8°C by 2100 (Table 4). It is noteworthy that a rise in temperature above 2°C by 2100 is almost certain under all these scenarios. Table 4Scenarios Using The Kaya Identity Emission Reduction Possibilities Table 5Policy Approach Per Factor To reduce CO2 emissions, policies aimed at reducing the growth rate of one or more of the Kaya Identity’s components will be needed (Table 5). Assuming a constraint-free world, reducing average population and income growth rates to 0% from the projected 0.4% and 2.1% would reduce cumulative emission by 60% in 2100 vs. the baseline. Economic growth is the main driver of emissions growth. For instance, post-GFC, Europe’s emissions have been subdued due to poor economic growth. However, the constraints on these variables exist and are binding. These are not the area of focus to tackle climate change. Consequently, this leaves energy efficiency and carbon intensity of energy as the only viable options to reduce GHG emissions. In order to avoid breaching the 2°C target, the IPCC estimates CO2 concentration needs to be capped below 400 ppm by 2100. This can only be achieved by significant improvements to energy efficiency. Economic theory suggests that given that energy is a cost of production, energy efficiency will continue to improve. However, the required pace of reduction in energy intensity surpasses the incentive provided by the price mechanism. The externalities of an energy intensive economy are delayed and uncertain. Thus, these are not fully included in the cost-benefit analysis of investing in new technology. As a result, policies aimed at reducing the carbon intensity of global energy input will be an important source of CO2 reduction. This includes decreasing the carbon intensity of fossil fuels – e.g. switching coal to natural gas and developing carbon capture and storage technology – and reducing the share of fossil fuels in the energy mix – e.g. switching fossil fuel energy to renewables. We will expand on alternative sources of energy in a subsequent report. Importantly, the policy response should differ between regions. The drivers of emissions are heterogeneous and policies should fit the regional reality. The Kaya Identity can also be applied at the country or regional level. Chart 11The Kaya Identity Applied At The Country Level U.S. - Elevated income growth offset by increasing energy efficiency (Chart 11, panel 1). China - Robust income growth drove CO2 emissions higher (Chart 11, panel 2). Europe - Falling energy intensity and carbon intensity led to a decline in emissions (Chart 11, panel 3). References Fourier, J. (1827). Mémoire sur les Températures du Globe Terrestre et des Espaces Planétaires, Mémoires de l’Académie Royale des Sciences, 7, 569-604. ‘Global’ warming varies greatly depending where you live, published by CarbonBrief on July 2, 2018. Nordhaus, William (2018). Projections and Uncertainties about Climate Change in an Era of Minimal Climate Policies, American Economic Journal: Economic Policy, 10(3): 333-360. Edenhofer, O. et al. (2015), The Atmosphere as a Global Common, The Oxford Handbook of the Macroeconomics of Global Warming. Hardin, Garrett (1968), The Tragedy of the Commons, Science 162, no. 3859: 1243–1248. Jean-Louis Combes et al. (2016), A review of the economic theory of the commons, Revue d’économie du développement, Vol 27.  Climate Science as Culture War, Stanford Social Innovation Review (Fall 2012). The Fourth National Climate Assessment: Volume 2 Impact, Risks, and Adaptation in the United States, U.S. Global Change Research Program (2018) and Climatic Research Unit temperature database Hartmann et al. (2013), Observations: Atmosphere and Surface. In: Climate Change 2013: The Physical Science Basis, Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Scott, P. (2016), How climate change affects extreme weather events, Science 352(6293):1517-1518. Mann et al. (2016), The Likelihood of Recent Record Warmth, Scientific Reports 6:19831. Fischer, E. M., and R. Knutti, Anthropogenic Contribution to Global Occurrence of Heavy-Precipitation and High-Temperature Extremes, Nature Climate Change 5 (April 27, 2015): 560. Cook et al. (2016), Consensus on Consensus: A Synthesis of Consensus Estimates on Human-Caused Global Warming, Environmental Research Letters 11, 4:048002. The impacts of climate change at 1.5C, 2C and beyond, CarbonBrief (2018). The Emissions Gap Report 2018, United Nations (2018). Batten, Sandra (2018), Climate change and the macro-economy: a critical review, Bank of England Staff Working Paper No. 706. Robert S.J. Tol (2019), Climate Economics: Economic Analysis of Climate, Climate Change and Climate Policy, Cheltenham, U.K. Edward Elgar Publishing Limited. Hugo Bélanger Senior Analyst HugoB@bcaresearch.com Jeremie Peloso Research Analyst JeremieP@bcaresearch.com Footnotes 1 For instance, Canada is estimated to be warming at twice the global rate. 2 The term “global commons” is used to define common resources or environmental issues crossing national boundaries. They have either no well-defined property right (no individual or nation has private control of their use) or lack an international enforcement mechanism to control their use (Edenhofer, 2015). The market failures associated with common pool resources (CPR) were popularized in Garret Hardin’s famous 1968 paper “Tragedy of the Commons”. 3 The likelihood is between 1 in 5,000 and 1 in 170,000 chances. 4 No-regret strategies are cost-effective under multiple climate change and policy response scenarios. Win-win actions provide beneficial externality while contributing to adaptation to various climate change scenarios. Under uncertainty, these strategies are the most likely to be implemented to begin the adaptation process rather than a riskier wait-and-see approach. Please see “Examples of ‘no-regret’, ‘low-regret’ and ‘win-win’ adaptation actions,” published by climate exchange. It is available at climatexchange.org.uk. 5 Please see Global Asset Allocation Special Report, “ESG Investing: No Harm, Some Benefit,” dated November 21, 2018, and available at gaa.bcaresearch.com 6 Please see BCA Special Reports, “Agriculture In The Age Of Climate Change,” dated October 23, 2019, and available at bca.bcaresearch.com 7 This section is largely inspired from Robert S.J. Tol (2019), Climate Economics: Economic Analysis of Climate, Climate Change and Climate Policy, Cheltenham, U.K. Edward Elgar Publishing Limited. 8 Lowercase letters denote annual growth rates of each component.