Global

Highlights Rising Bond Yields: Global risk assets are discounting a V-shaped economic recovery. With economic data starting to revive as more economies emerge from virus-related shutdowns, bond yields are showing signs of following suit. Duration Strategy: Even with global yields showing signs of a cyclical bottom, we continue to recommend a neutral duration stance. Central banks will remain highly accommodative given the lack of inflationary pressures after the deep COVID-19 recessions. There are still significant risks in the coming months from a potential second wave of coronavirus after economies reopen, worsening US-China relations and domestic US sociopolitical turmoil. Duration Proxy Trades: Given those lingering uncertainties, we prefer to focus on “duration-lite” trades in the developed economies, like overweighting inflation-linked government bonds versus nominals as inflation expectations will drift higher over the next 6-12 months. Feature Dear Client, Next week, instead of publishing a regular Weekly Report, we will hold a webcast on Tuesday, June 16 at 10:00 am ET, discussing our latest views on global fixed income markets. The format will be a short presentation, followed by a Q&A session. We hope you will join us, armed with interesting questions. Kind regards, Rob Robis, Chief Fixed Income Strategist Chart of the WeekBond Yields Bottoming, But Backdrop Not Yet Bearish Bond yields around the world awoke from their COVID-19 induced slumber last week, responding to a growing body of evidence indicating that global growth has bottomed. Over a span of four days, benchmark 10-year government bond yields rose in the US (+20bps), Germany (+13bps), Canada (+20bps), China (+14bps), Japan (+4bps), Mexico (+13bps) and the UK (+12bps). There is potential for yields to continue drifting higher over the next few months, as more countries reopen from virus-related shutdowns. The bounce already seen in survey data like manufacturing and services PMIs, as well as economic sentiment measures like the global ZEW index, should soon translate into real improvements in activity data. This comes at a time when rising commodity prices, most notably oil, suggest that depressed inflation expectations can lead bond yields higher. The cyclical bottom for global yields has likely passed, based on the improvement already seen in our own Global Duration Indicator (Chart of the Week). However, economic policy uncertainty remains elevated as devastated economies try to reopen from lockdowns. In addition, our Central Bank Monitors continue to indicate pressure on policymakers to keep interest rates as low as possible to maintain easy financial conditions as easy as possible. Tighter monetary policies remain a distant prospect, given very high unemployment rates. The cyclical bottom for global yields has likely passed, based on the improvement already seen in our own Global Duration Indicator. Amid those uncertainties, we recommend maintaining a neutral strategic (6-12 months) and tactical (0-6 months) stance on overall duration exposure in fixed income portfolios. Instead, we prefer focusing on lower volatility trades that will benefit from improving global growth and policy reflation, like going long inflation-linked bonds versus nominal government debt throughout the developed markets with breakevens looking too low on our models. Why Are Bond Yields Rising Now? We see five main reasons why global bond yields have started to move higher: 1) Investor risk aversion is declining There has been a sharp recovery in global risk appetite since late March, diminishing the demand for risk-free global government debt. In the US, the S&P 500 is up 43% from its March lows, while the NASDAQ index is back to the all-time highs reached before the coronavirus turned into a global pandemic (Chart 2). US corporate debt has also performed well since the March 23rd peak in spreads, with investment grade and high-yield spreads down -227bps and -564bps, respectively. Non-US assets are also flying, with emerging market (EM) equities up 29% and EM USD-denominated corporate debt up 14% in excess return terms over US Treasuries since the March trough. Even severely lagging assets like European bank stocks are showing a pulse, up 38% since the lows of May 15. Commodity prices are also improving, led not only by gains in oil after the April crash by recoveries in the prices of growth-sensitive commodities like copper (+17%) and lumber (+42%). Add it all up, and the message is clear: investors now prefer risk to safety, which has tempered the demand for government bonds. The flipside of the boom in risk appetite is weakening prices for safe haven assets (Chart 3). The price of gold in US dollar terms is down -4% from the 2020 high on May 20, while the euro price of gold is down –6%. Safe haven currencies like the Japanese yen and Swiss franc have underperformed, while interest rate volatility measures like the US MOVE index and long-dated euro swaption volatility are back to the pre-coronavirus lows. Chart 2Risk Assets Are Booming Worldwide Chart 3Safe Haven Trades Losing Luster Add it all up, and the message is clear: investors now prefer risk to safety, which has tempered the demand for government bonds that helped drive yields lower when risk assets were tanking in late February and March. 2) Global growth is improving One of the reasons for the improvement in investor risk appetite is belief that the world economy has exited from the severe COVID-19 global recession. While timely real data is still coming in slowly given reporting lags, there has been a notable bounce in survey data in many countries. PMIs for both manufacturing and services climbed higher in May (Chart 4). The expectations components of economic confidence measures like the ZEW indices have also recovered the losses seen in February and March. Data surprises have also been increasingly on the positive side of late in China, Europe and the US, including the shocking 2.5 million increase in US employment in May. However, the US unemployment rate remains very high at 13.3%, indicating abundant spare capacity that will likely take years, not months, to work off – a problem that most of the world will continue to deal with post-recession. 3) Central bank liquidity is booming The other main reason for the boom in risk asset performance that has started to put upward pressure on bond yields is the extremely accommodative stance of global monetary policy. This is occurring through 0% policy rates in the developed economies but, even more importantly, the aggressive expansion of central bank balance sheets through quantitative easing (QE). The Fed has its foot firmly on the monetary accelerator, with year-over-year growth in its balance sheet of 87% (Chart 5). The European Central Bank (ECB) is no slouch, though, with its balance sheet up 19% from a year ago and having expanded its Pandemic Emergency Purchase Program (PEPP) by another €600 billion last week. Chart 4Signs Of Life In The Global Economy Chart 5'QE Forever' Driving Money From Bonds To Risk Assets The combined annual growth of the central bank balance sheets for the “G4” (the Fed, ECB, Bank of Japan and Bank of England) is now up to 26%. The rate of G4 balance sheet expansion has been a reliable leading indicator of global risk asset performance since the 2008 financial crisis (with about a 12-month lead), and the current boom in “liquidity” suggests that the current rise in global equity and corporate bond markets can continue over the next year. Easing global financial conditions are now returning to levels that should support economic growth in the coming months, helping to mitigate (but not eliminate) the potential credit stresses from companies that have suffered during the COVID-19 recession. This recovery remains fragile, however, and policymakers will continue to maintain an extremely dovish policy bias – even with significant fiscal stimulus measures also in place to help economies climb out of recession. This suggests that the current rise in global bond yields is not the start of a new bond bear market driven by expectations of tighter monetary policies. The current rise in global bond yields is not the start of a new bond bear market driven by expectations of tighter monetary policies. Chart 6Global Bond Sentiment Is Still Very Bullish 4) Bullish sentiment for bonds is at extremes From a contrarian perspective, another factor helping put a floor underneath bond yields is investor sentiment towards fixed income, which remains bullish. The widely followed ZEW survey of economic forecasters also contains a question on the expected change in bond yields over the next year. The latest read on the surveys shows a net balance still expecting lower bond yields in the US, Germany, the UK and Japan, nearing levels seen prior to the end of the recessionary bond bull markets in the early 2000s and after the 2008 financial crisis (Chart 6). In addition, the Market Vane survey of bullish sentiment on US Treasuries is nearing past cyclical peaks, suggesting limited scope for new bond buyers that could drive US yields to new lows. 5) Inflation expectations are moving higher Finally, global yields are rising because the inflation expectations component of yields has started to move higher. The hyper-easy stance of monetary policy is playing a role here. Market-based inflation expectations measures like the breakevens on inflation-linked bonds (or CPI swap rates) are a vote of confidence by investors in the “appropriateness” of policy settings. The fact that inflation expectations are now drifting higher suggests that bond markets now believe that central banks are now "easy" enough to give inflation a shot at rising sustainably as growth recovers. Global yields are rising because the inflation expectations component of yields has started to move higher. Chart 7Oil Prices & Breakeven Inflation Rates Are Both Recovering This move higher in inflation expectations can continue in the coming months, particularly with global oil prices likely to move even higher. Our colleagues at BCA Research Commodity & Energy Strategy are quite bullish on oil prices, forecasting the benchmark Brent oil price to rise to around $50/bbl by the end of 2020 and continuing up to $78/bbl by the end of 2021. Such an outcome would push up market-based inflation expectations, and likely put more upward pressure on nominal bond yields, given the strong correlation between oil and inflation breakevens in the developed economies that has existed over the past decade (Chart 7). Bottom Line: Global risk assets are discounting a V-shaped economic recovery. With economic data starting to revive as more economies emerge from virus-related shutdowns, bond yields are showing signs of following suit. Duration Strategy For The Next Few Months The trends in growth, inflation and financial conditions all suggest bond yields can continue to drift higher over at least the next 3-6 months. Yet given the potential for a negative shock from a second wave of coronavirus infection, or geopolitical uncertainties in a volatile US election year, a below-benchmark global duration stance is not yet warranted. This is especially true with unemployment rates in most countries remaining elevated even as growth rebounds from recession, forcing central banks to maintain a very dovish policy posture. Our “Risk Checklist” that we have been monitoring to move to a more aggressive recommended investment stance on global spread product – the US dollar, the VIX and the number of new COVID-19 cases - can also be helpful in helping us determine when to shift to a more defensive bias on global duration. On that note, the Checklist still argues for a neutral duration stance, rather than positioning for a big move higher in yields. The US dollar has started to soften, but remains at a very high level relative to interest rate differentials (Chart 8). A weaker greenback is a source of global monetary reflation, primarily through changes in the prices of commodities and other traded goods that are denominated in dollars, but also by helping alleviate funding pressures for companies that have borrowed heavily in US dollars (especially in the emerging world). The dollar is also an “anti-growth” currency that appreciates during periods of slowing global growth, and vice versa, so some depreciation should unfold as more of the world economy emerges from lockdown (middle panel). The VIX index – a measure of investor uncertainty - continues to climb down from the massive surge in February and March, now sitting at 26 after peaking around 80. This is the one part of our Risk Checklist that argues for reducing duration exposure now. We prefer trades that will benefit from the combination of continued global policy reflation and growing investor risk appetite. We call these “duration-lite” trades. The daily number of new reported cases of COVID-19 (using data from the World Health Organization) has come down dramatically in Europe, but in the US the decline in new cases has stalled over the past month – a worrisome sign as the country continues to reopen amid mass protests in major cities (Chart 9). New cases outside the US and Europe are rapidly moving higher, however, primarily in major Latin American countries like Brazil and Mexico. This suggests that while there is a concern about a “second wave” of coronavirus later in the year, the risks from the first wave are far from over. Chart 8Still Not Much Reflationary Push From A Weaker USD Chart 9The COVID-19 Threat Has Not Gone Away Instead of shifting to a below-benchmark recommended stance on overall portfolio duration too soon in the cycle, we prefer trades that will benefit from the combination of continued global policy reflation and growing investor risk appetite. We call these “duration-lite” trades. Specifically, we like owning inflation-linked government bonds versus nominal debt, while also positioning for steeper government yield curves (on a duration-neutral basis). Longer-dated breakeven inflation rates within the major developed markets are becoming increasingly correlated to both the level of 10-year government bond yields (Chart 10) and the slope of the 2-year/10-year yield curve (Chart 11). Chart 10Rising Inflation Expectations Will Lead To Higher Bond Yields ... Chart 11... And Steeper Yield Curves In terms of country selection for these trades, we look to the valuations on inflation-linked bond breakevens from our modeling framework that we introduced back in late April.1 In that framework, we model 10-year breakevens as a function of oil prices, exchange rates and the long-run trend in realized inflation. Chart 12Global Inflation Breakevens Look Cheap On Our Models In Chart 12, we show the deviation of 10-year inflation breakevens from the model-implied fair value, shown both terms of standard deviations and basis points. The “cheapest” breakevens from our models are for inflation-linked bonds in Italy and Canada, although almost all counties (outside of the UK) have breakevens to look far too low. This suggests that global bond investors should consider a multi-country portfolio of inflation-linked bonds versus nominal paying equivalents – or in countries where the inflation-linked bond markets are small and illiquid, duration-neutral yield curve steepeners - as a more efficient way to play for a continuation of the current reflationary global backdrop without taking duration risk. Bottom Line: Even with global yields showing signs of a cyclical bottom, we continue to recommend a neutral duration stance. Given the lingering uncertainties about a second wave of coronavirus, and the rising political and social tensions in the US only five months before the presidential election, we prefer to focus on “duration-lite” trades in the developed economies - like overweighting inflation-linked government bonds versus nominals as inflation expectations will drift higher over the next 6-12 months.   Robert Robis, CFA Chief Fixed Income Strategist rrobis@bcaresearch.com Footnotes 1 Please see BCA Research Global Fixed Income Strategy Weekly Report, "Global Inflation Expectations Are Now Too Low", dated April 28, 2020, available at 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

  Dear Client, In lieu of our regular report this week, we are sending you a Special Report from my colleague Jonathan LaBerge. Jonathan examines the global effectiveness of recent pandemic containment measures to judge both the odds of a second infection wave and what policy responses are likely to be effective in countering one were it to occur. In addition, I will take part in a discussion on the longer-term outlook for inflation alongside my colleagues Robert Robis and Robert Ryan in a live webcast this Friday, June 5 at 8:00 AM EDT (1:00 PM BST, 2:00 PM CEST, 8:00 PM HKT). Best regards, Peter Berezin, Chief Global Strategist   Highlights In this report we examine the effectiveness of COVID-19 containment measures across 30 of the largest global economies to determine which measures best explain cross-country “success” at fighting the pandemic. Our findings are generally consistent with the recommendations of health experts today and the historical experience of the Spanish flu. The speed at which measures were deployed appears to have been a very important factor contributing to success, and the most economically-damaging measures seem to have been among the most effective in combating the spread of the disease. This underscores that fighting a secondary infection wave, were one to occur, would be enormously costly even if more effective and less blunt containment measures succeeded at preventing uncontrolled spread. Equity investors are thus making a risky bet in extrapolating early reopening. We recommend a tactically neutral allocation towards equities versus bonds, within the context of a cyclically-overweight stance. Feature Global equities have rallied 38% from their March 23 low, and remain only 9% below this year’s high. The rally in stocks reflects, in part, the very aggressive response that has occurred from both fiscal and monetary authorities. But it also reflects the view that pandemic containment measures have succeeded in controlling the spread of the disease in the western world, and that developed countries will be able to continue to progressively roll back containment measures and restart their stalled economies. In this report we investigate the effectiveness of recent pandemic containment measures across 30 of the largest economies in the world, based on data sourced from the Oxford COVID-19 Government Response Tracker. The goal of the report is to determine which of the measures best explain cross-country “success” at fighting the pandemic, in order to judge both the likelihood of a second wave of COVID-19 infections and what policy responses are likely to be effective in countering one were it to occur. Our findings are generally consistent with the recommendations of health experts today and the historical experience of the Spanish flu. The speed at which authorities responded to COVID-19 appears to be among the most important factors contributing to the relative success among countries in combating the pandemic, and the most economically-damaging containment measures (school and workplace closures, event cancellation, and travel restrictions) appear to have been among the most effective in combating the pandemic, depending on the measure of success in question. For investors, this underscores that fighting a secondary infection wave, were one to occur, would be enormously costly even if more effective and less blunt containment measures succeeded at preventing uncontrolled spread. In this regard, it appears that equity investors are making a risky bet in extrapolating early reopening, arguing for a tactically neutral allocation towards equities versus bonds within the context of a cyclically-overweight stance. Measuring Government Responses To COVID-19 Chart 1The Forcefulness Of Government Responses To COVID-19 Over Time The COVID-19 Government Response Tracker (OxCGRT) was launched by the University of Oxford’s Blavatnik School of Government in late-March as a tool to track and compare policy responses of governments tackling the coronavirus outbreak. The tracker originally included 11 indicators of how governments were responding: 7 measures of closures & containment, 2 measures tracking the economic response, and 2 measures of investment in health care and vaccine research. The original 7 measures of closures & containment were combined into an aggregate measure dubbed the Government Response Stringency Index (Chart 1), which has been widely cited over the past two months. OxCGRT was updated in late-April, and the changes included both new indicators (bringing the total to 18) and amendments to the way in which some of the original indicators were defined. Table 1 provides a list of the closure/containment and health indicators as well as their definitions, along with the codes or scores used to denote the different levels of stringency in each indicator. In the charts shown in this report, indicator values are shown rescaled to be between 0 – 100; as an example, a score of 2 out of 4 for international travel would show up as a rescaled value of 50. Table 1Description Of The Oxford Government Response Stringency Index Components Pertaining To Closures & Containment And Health Among the measures shown in Table 1, we test the power of 11 indicators (all 8 closure/containment and 3 public health measures) to explain the cross-country “success” of 30 countries in fighting the COVID-19 pandemic. “Success” is defined in three ways: limiting the magnitude of the virus’ spread (peak per capita cases and fatalities), limiting the time to the peak in new cases and fatalities, and the speed at which new cases and fatalities decline following a peak. Finally, we make our own adjustment to OxCGRT’s indicators by penalizing targeted measures rather than providing a bonus to general measures (as is done when calculating the official Stringency Index). We then combine these adjusted measures into our own index using the same equally-weighted methodology as employed when calculating the official Stringency Index. Our adjusted index has a somewhat stronger relationship with our three measures of success than OxCGRT’s Stringency Index, validating our approach to reducing the score of any given indicator by half when the measure is targeted rather than general in nature. Explaining The Magnitude Of The Cross-Country Spread For the 30 countries included in our analysis, Chart 2 illustrates the relationship between the degree of the virus’ spread (both in terms of current per capita confirmed cases and fatalities) and the average level of our adjusted stringency index in the early phase of each country’s outbreak. Chart 2At First Blush, Stringency Does Not Appear To Predict The Ultimate Magnitude Of The Spread… Given the persistent findings from epidemiological models that efforts to contain pandemics must occur quickly, we define this early phase as the first six weeks following the day in which a country reached five confirmed cases. This point was reached in late-January in most Asian countries, Australia, the US, and some parts of Europe, and in February for almost all of the remaining developed economies that we examined. Emerging market economies reached this point in the first half of March. The chart makes it clear that the average stringency of containment measures during this early phase has little power to predict the ultimate magnitude of the spread when considering all 30 countries. However, the charts also highlight that several European countries – Belgium, Italy, the Netherlands, Spain, and Switzerland – are significant outliers, especially in terms of per capita fatalities. Chart 3 removes these outlying nations, and underscores that the degree to which the virus ultimately spread across countries is better explained for the remaining countries by the early stringency of their response. The higher the stringency of the measures, the lower the current (or peak) number of per capita cases and deaths. Chart 3…But The Relationship Is Stronger After Excluding Outlying European Nations Table 2 shows the R-squared values for each of the individual measures, including and excluding the five outlying countries noted above. The first noteworthy point from Table 2 is that the indicators appear to have a better ability to predict fatalities than confirmed cases, which is true for nearly all of the indicators across all three measures of success that we examine in this report. To us, this emphasizes a point that has become apparent over the past two months, namely that the meaningfulness of confirmed case data varies significantly across countries due to differences in testing practices and availability. To use the parlance of global macro analysts, confirmed cases are “soft data,” whereas fatalities (and hospitalizations) represent “hard data.” Of course, as is also the case in global macro analysis, the soft data tends to lead the hard data, which helps explain why confirmed cases of the disease will remain an important leading indicator for the US economy until they largely capture asymptomatic and mild cases that are not likely to lead to hospitalization or death. Table 2School Closures, Canceling Public Events, And International Travel Restrictions Seem To Explain Lower Fatalities The second noteworthy point is that while none of the measures have particularly strong predictive power for all countries, several do when the outliers are removed. Importantly, strong restrictions on international travel during the early outbreak phase show up as being the most important predictor of reduced fatalities per capita, followed closely by school closures. The cancellation of public events, public information campaigns, and domestic travel restrictions also appear to be relatively important predictors. Notably, cross-country differences in testing policies and the comprehensiveness of contact tracing do not seem to explain the variation in per capita cases and fatalities. As a final point on the magnitude of the spread of COVID-19, it is not immediately clear why the five European countries that we highlighted have been such sizeable outliers in the direction of higher per capita cases and fatalities despite seemingly stringent measures. At present, we have two theories: Given the importance of early and strong restrictions on travel highlighted in Table 2, it is possible that the efficacy of these restrictions has somehow been lower in these countries. It is possible that higher cases and fatalities in these countries can be explained by differences in the management of nursing homes and other elder care facilities, a factor that is not directly measured in the OxCGRT data. On the latter point, data from the Canadian province of Quebec underscores the impact of managing (or mismanaging) long-term care facilities. While Chart 3 highlights that Canada’s experience as a whole appears to be reasonably well-explained by the fairly low stringency of its response during the early phase of its outbreak, the province of Quebec has incurred a particularly high per capita fatality rate that is on par with the outlying European countries that we noted. Chart 3…But The Relationship Is Stronger After Excluding Outlying European Nations Chart 4Mismanaging Elder Care Facilities Significantly Affects The Fatality Rate Chart 4 presents a breakdown of cumulative Quebec COVID-19 fatalities by place of residence, which clearly demonstrates the impact of public and private nursing homes on the overall fatality rate. The death toll in publicly-funded homes has been particularly high, which even after accounting for the higher proportion of elderly residents in these types of facilities points to mismanagement and/or inadequate funding as key drivers of the disease’s spread (and thus fatalities given that nursing home residents face high risk from the disease). Similar dynamics may exist in the European countries that we cited, which could help explain their outlier status. Evidence On Hastening A Peak, And Post-Peak Decline, In New Cases & Fatalities Chart 5 presents the relationship between the amount of time needed to reach a peak in new cases and fatalities and the measure we used to predict the virus’ spread: the average level of our adjusted stringency index in the early phase of each country’s outbreak. Chart 5Stringent Early Measures Shorten The Time To A Peak In New Cases And Fatalities Table 3School Closures, Workplace Closures, And Canceling Public Events Seem To Explain Lower Days To Peak Cases And Fatalities The charts support the argument that stringent early measures shorten the number of days to a peak in new cases and fatalities. Table 3 presents the predictive power of the individual measures, which highlights some differences in the effectiveness of the measures to hasten the time to peak compared with their ability to predict the ultimate magnitude of the spread of the disease: Workplace closures appear to be somewhat better, and school closures somewhat worse, at predicting the speed at which countries reached a peak in new cases and fatalities than they were at predicting the ultimate magnitude of the spread. Closing public transport and restricting domestic travel were modestly successful at predicting the degree of spread but have essentially no power to predict the variation in the time to peak. Finally, restricting international travel was the strongest predictor of the degree of spread but also had essentially no power to predict the amount of time needed to reach a peak in new cases. As noted above, we use the average level of our adjusted stringency index as a predictor for both the prevalence of the disease and the time needed to reach a peak in new cases and deaths. Since it is an average of a given period of time, this variable measures a combination of the stringency of the restrictions as well as how early they were deployed. To test the relative importance of the severity of the measures versus the speed at which they occur, we apply the same approach as in Chart 5 but we replace OxCGRT’s score of each measure’s value for each country with a dummy variable (0 for no measures or 1 for any measures, again rescaled to be between 0 - 100) while retaining the penalty for targeted measures described above. In simple terms, we abstract from whether the severity of the measures is low or high and instead focus simply on whether any measures were applied and when. Chart 6 and Table 4 present the results. With the exception of a country’s testing policy’s ability to predict confirmed cases, the charts and table show that there is little difference between the full indicators and the dummy versions. This suggests that where general (rather than targeted) measures to reduce the spread of the virus have been effective, they have been so because of the speed of their deployment rather than their strictness. Chart 6The Deployment Speed Of Containment Measures Seems More Important Than Their Strictness Table 4Small Differences Between The Full Measures And Those Focused Only On The Existence Of Any Response Finally, Chart 7 and Table 5 present the ability of the various (full) indicators to predict how successful various countries have been at reducing new cases and fatalities following a peak. The chart and table underscore that the measures have not been particularly successful at explaining the degree to which countries have reduced new cases and deaths, with the exception of two measures: the cancellation of public events and contact tracing. And in the case of the latter, the prevalence of contact tracing appears to help explain greater reductions in new cases, but only weakly explains reduced fatalities. Chart 7Overall, Stringency Does Not Predict Success At Reducing New Cases And Fatalities Following A Peak Table 5Canceling Public Events And Contact Tracing Appear To Have Some Success At Hastening A Post-Peak Decline Key Takeaways Correlation does not necessarily imply causation, and thus the ability of a particular containment measure to predict differences in COVID-19 outcomes across countries may not always reflect the effectiveness of the measure. Nonetheless, there are several important takeaways from the evidence provided above: The speed at which authorities responded to COVID-19 appears to be among the most important factors contributing to the relative success among countries in combating the pandemic. Implementing general rather than targeted measures does seem to be important, but beyond that stringency does not appear to be the key driver differentiating outcomes across countries. Worryingly, the most economically-damaging containment measures (school and workplace closures, event cancellation, stay-at-home orders, and travel restrictions) appear to have been among the most effective measures in combating the pandemic, depending on the measure of success in question. While containment measures appear to have succeeded in bringing about a peak in new cases and fatalities in most developed economies, the evidence shown above highlights how long painful measures need to be in place in order to have an impact. For example, Chart 8 illustrates the relationship between workplace closure and the time to peak in new fatalities; with the exception of Hong Kong, even in the countries that acted quickly and forcefully to close workplaces it took over a month to reach a peak. To the extent that global policymakers are aware of the relative importance of restrictions on international travel in limiting the ultimate spread of the disease in the countries we examined, that may suggest that international borders will remain closed or severely restricted for some time or will reoccur very quickly if evidence of a secondary infection wave were to emerge later this year. Chart 8Reaching A Peak In Fatalities Is Costly Even For Countries That Act Quickly There is another important insight for investors that is not immediately apparent from our work, but emerges when we examine two surprising findings. Looking closely, investors can infer that the public’s awareness and fear of the disease has contributed to successes in combating COVID-19, in ways that are not readily captured by the OxCGRT’s data. To us, the key observation is that both closures of public transportation and the forcefulness of stay-at-home orders showed themselves to be mediocre-to-poor predictors, when it seems straightforward to see that reduced crowding on buses and subways and physical distancing have very likely helped reduce the spread of the virus. This seeming discrepancy is likely resolved by the fact that the public acted themselves to take these measures out of fear of getting sick, meaning that cross-country differences in rules pertaining to these measures have not been especially relevant in predicting outcomes. Chart 9 supports this point by highlighting that subway ridership in New York city fell well before the city issued a mandatory stay at home order, as did the median year-over-year growth rate in US seated restaurant diners before the first stay at home order was issued in the US (Chart 10). Chart 9New York City Subway Ridership Fell Well Before The Stay-At-Home Order Was Issued Chart 10Diners Started Avoiding Restaurants Before Any Stay-At-Home Orders Were Issued Chart 11Sweden Is A Very Big Outlier In Terms Of Where The Stringency Of Its Measures Peaked The importance of public behavior in altering the spread of COVID-19 is also evident from Sweden’s experience, albeit in a different way than has been commonly discussed. The charts above highlighted that Sweden has indeed been somewhat of an outlier in terms of its experience with COVID-19 relative to the stringency of its early response, but there have been a few other countries with similar experiences. Where Sweden has been a very significant outlier is the level at which its Stringency Index peaked, at least compared with other advanced economies (Chart 11). And yet, Sweden appears to have achieved a peak in new fatalities based on the data available today. Swedish policymakers have cited the country’s high levels of social trust and cohesion as part of the reason why more strict measures were not absolutely necessary to prevent an uncontrollable/exponential spread of the disease, and we see no reason to doubt that this has been an important, if not crucial, factor – Scandinavian countries have long ranked highly on these types of characteristics. Investment Conclusions The first important point for investors is that our findings are generally consistent with the recommendations of health experts today and the historical experience of the Spanish flu. One of the key lessons of the Spanish flu is that removing or relaxing measures too early can lead to a renewed rise in mortality rates,1 and it thus seems clear that the reopening of economies before the first wave of infections has fully dissipated increases the odds of a second wave. In this regard, it appears that US equity investors are making a risky bet in extrapolating early reopening. Second, the fact that the public's behavior can significantly alter COVID-19 outcomes across countries has both potentially positive and negative implications for the odds of a secondary infection wave and for near-term economic growth. For the economy, it implies the possibility of sustainable economic reopening alongside a controllable risk of renewed spread if the public can be convinced to treat the ongoing risk of the disease very seriously without fearing it. This outcome may be more likely if mask wearing in public – a variable not captured in OxCGRT’s data – becomes and remains widespread in advanced, western economies. At the same time, it may also suggest that the “Swedish approach” of accepting higher fatalities in exchange for lighter containment measures within the context of a “controlled” spread of the disease may not be possible in other countries, if Sweden does indeed enjoy higher levels of social trust and cohesion compared with other countries and if these factors have been key in preventing the disease from spreading there at an exponential rate. Finally, our analysis has underscored that fighting a secondary infection wave, were one to occur, would be enormously costly even if more effective and less blunt containment measures succeeded at preventing uncontrolled spread. Given this, and the higher risk of increased infections introduced by economic reopening and the recent widespread protests in the US, we recommend that investors remain tactically neutral equities versus bonds within the context of a cyclically-overweight stance.   Jonathan LaBerge, CFA, Vice President Special Reports jonathanl@bcaresearch.com   Footnotes 1    Please see Global Asset Allocation Special Report, “Lessons From The Spanish Flu,” dated May 20, 2020. Global Investment Strategy View Matrix Current MacroQuant Model Scores