Global

Special Report

Highlights We update our assumptions for the likely 10-15 year return for a wide range of different asset classes. Our methodology is basically unchanged from our last Return Assumptions report published in 2019, though we have refined our analysis and use of data in some areas. Returns over the next decade will be very low compared to history. We project that a standard global portfolio (50% equities, 30% bonds, and 20% alternatives) will return only 3.0% a year in nominal terms. That compares to a historic return of 6.3%. There are still some assets that will produce better returns, most notably small caps (4.9% a year in the US) and alternatives (6.2% for private equity, for example). But they also carry higher risk. Spreadsheets are available with detailed data. Introduction This is the third edition of our work on return assumptions. Since publishing the previous reports in November 2017 and June 2019, we have had many opportunities to discuss our methodologies with clients and in the Global Asset Allocation course at the BCA Academy. This has allowed us to test and, in many cases, refine our approach. We believe the methodologies we use have stood the test of time. We have always emphasized that this sort of capital markets assumptions (CMA) analysis is an art, not a precise science. We continue to prefer to project returns over a somewhat undefined 10-15 year period, since this allows us to think about the underlying trend of likely returns. Many other CMA papers use five (or even three) year time horizons which, in our view, are problematical since they rely heavily on a forecast of the timing, length, and severity of the next recession. Our approach is based on the concept that the return on the risk-free long-term government bond is the cornerstone to projecting asset returns, and that this return is rather predictable: It is approximately the current yield. Most other asset returns can be built up from that – the return on high-yield bonds, for example, by assuming that their historic spread over government bonds, and default and recovery rates will continue in the future. For equities, we continue to use six different methodologies, which are based on a mixture of valuation and projected earnings growth. This approach – that assumed returns can be built up from a combination of current yield plus forecast future growth in capital values – also works for most alternative asset classes, for example real estate. We have made a few minor changes to our methodology in this edition. We have, for example, made our use of historical data (for spreads, profit margins, growth relative to GDP, etc.) more consistent, using the 20-year average where possible. The biggest change this time is that clients can download here a spreadsheet with all the data in this report in order, for example, to use the data as inputs into their own optimizers. In addition, we have set up our detailed spreadsheet to allow clients to see the underlying inputs, the formulae behind our methodologies, and to input their own assumptions. This will also allow us to update the results of our analysis as often as needed. Please let us know here if you would like more details about this additional service. This Special Report is structured as follows. First, we analyze the overall results: What is the probable return from each asset class over the next 10-15 years, and how do these differ from historical returns. Next, we describe in detail the methodologies we use, for (1) economic growth, (2) fixed-income instruments, (3) equities, and (4) 12 different alternative asset classes. Then, we describe our way of forecasting currency returns, and show the return assumptions in different base currencies. Finally, we update the numbers for volatility and correlations, which many investors need as inputs into optimization programs. The summary of our results is shown in Table 1. The results are all average annual nominal total returns, in local currency terms (except for global indexes, which are in US dollars). The data is updated to end-April 2021 (except for some alternative asset classes where only quarterly data is available). Table 1BCA Assumed Returns Overall Results Returns over the coming decade are likely to be very disappointing compared to history. Our assumptions suggest a typical global portfolio, consisting of 50% large-cap equities, 30% bonds, and 20% alternatives, will produce an annual nominal return of only 3.0%, compared to an average of 6.3% over the past 20 years. A US-only portfolio with a similar composition is likely to produce only a 3.1% return, compared to 7% in history. The reason is simple: Valuations currently are very stretched in almost every asset class. The risk-free rate (the 10-year government bond yield) in the US is 1.6% (compared to a 20-year average of 3.1%). It is negative in the euro area (in nominal terms) and zero in Japan. These rates are the anchor for the returns of all other asset classes, which are theoretically priced off the risk-free rate plus a risk premium. We have long argued that valuations are not a good timing tool for investors. An asset can remain very expensive or very cheap for a considerable period. But all the evidence shows that the valuation at the starting point is a very powerful indicator of long-run returns. The yield on government bonds, for example, has a strong correlation with their 10-year return (Chart 1). In the equity market, the Shiller PE has historically had little correlation with the return over one or two years, but has a 90% correlation with the return over the subsequent 10 years (Chart 2). Chart 1Starting Yield Determines Bond Returns Chart 2Valuation Drive Long-Run Equtiy Returns     With valuations in equity markets now expensive relative to history (for example, forward PE for US stocks of 22x compared to a 20-year average of 16x, and 18x in the euro zone compared to 13x), investors should expect that equity market returns will be low relative to history. Our assumptions point to a 2.6% annual return from US stocks, 2.3% from the euro zone, and 1.6% from Japan (compared to 8.5%, 3.9%, and 3.5% over the past 20 years). Our assumptions are significantly lower than when we last published our analysis in 2019; then we projected 5.6% for US stocks, 4.7% for the euro zone, and 6.2% for Japan. The difference is that equity multiples have risen and risk-free rates have fallen significantly since then. So what should investors do? They have only two choices: Lower their return assumptions, or increase their weightings in riskier asset classes. Chart 3Hard To See How US Pension Funds Will Achieve Their Targets The average US public pension fund (Chart 3) still assumes a return of 7% a year, and private pension funds’ assumption is not much lower. And yet corporate pension funds have been pushed by their consultants in recent years to increase their weighting in bonds, to more closely match their liabilities (Chart 4). It is almost mathematically impossible to achieve their targets with that sort of portfolio. In other countries, such as Australia or Canada, pension funds’ return targets are typically inflation or cash plus 3-4 percentage points. But even those targets are challenging.   Chart 4...Especially With Over 50% In Bonds There are asset classes which will produce higher returns. For example, we project a return of 4.9% from US small-cap stocks – and 9.7% from UK small caps. US high-yield bonds should produce a return of 3.2% a year (even after defaults) and Emerging Markets local currency sovereign debt 2.7% (in USD terms) – not exactly exciting, but at least a pick-up over other fixed-income securities. The projected returns from illiquid alternative assets continue to look relatively attractive. An equal-weighted portfolio of the 12 alternatives we cover is projected to return 5.7% a year, not much lower than the forecast of 6.1% from our 2019 report (and compared to an average of 7.1% of the past 20 years). There are some alt assets where returns have started to trend down: Private equity, for instance, is projected to return 6.2% a year, compared to 11.1% in history, and hedge funds 4.5%, compared to 5.9%. But the illiquidity premium should not disappear completely, even if the move of alternative investments to become more mainstream has reduced it to a degree. So adding more risky assets to a portfolio is an answer, at least for those investors with a long enough time-horizon that allows them to bear the inevitable big drawdowns that come with having a more volatile portfolio. And, unfortunately, lower returns mean that the incremental return gained for each unit of risk taken has declined compared to the past 10 or 20 years (Chart 5) – the efficient frontier has flattened significantly. Chart 5You Need To Take More Risk To Produce Return How We Came Up With The Assumptions GDP Growth Several of our methodologies use assumptions (for example, in equity methods (2) and (3), based on projections of earnings growth, real-estate capital-value growth, and commodities prices) which require estimates of nominal GDP growth in each country and region. To make these forecasts, we assume that nominal GDP growth can be decomposed into: (1) growth of the working-age population, (2) productivity growth, and (3) inflation. This ignores capital intensity, but it has been relatively stable over history and is difficult to forecast. Table 2 shows the assumptions we use, and our forecasts for real and nominal GDP in each country and region. Table 2Calculations Of Trend GDP Growth For population growth we use the United Nations’ median forecast of annual growth in the population aged 25-54 between 2020 and 2040. This ranges from -1% in Japan to +1% in Emerging Markets – although note that the range of forecast population growth in EM varies widely from 1.2% in India to -1.1% in Korea (and in China, too, is negative at -0.7%). This estimate is reasonably reliable, although it does miss some possible factors, such as changes in the female participation rate, hours worked, and changing openness to immigration. Productivity is much harder to forecast. Over the past 10 to 20 years, productivity growth has trended down in most countries (Charts 6A & B). We take a slightly more optimistic view, assuming that productivity growth over the next 10-15 years will equal the 20-year average. We base this on the belief that part of the decline in productivity since the Global Financial Crisis is due to cyclical reasons which are now dissipating, and also to expectations that new technologies coming through (artificial intelligence, big data, automation, robotics etc) will boost productivity in the coming years. Others take a more pessimistic view. The Congressional Budget Office’s forecast of trend real US GDP growth in 2022-2031 of 1.8%, for example, is lower than our estimate of 2.2% mainly because of its more cautious estimate of productivity growth. Chart 6AProductivity Growth (I) Chart 6BProductivity Growth (II)   To derive nominal GDP growth, we assume that inflation over the next 10 years will be on average the same as over the past 20 years, for example 2% in the US, 1.6% in the euro area, 0.1% in Japan, and 3.9% in Emerging Markets (using a weighted average of EM by equity market cap). This estimate, too, has a high degree of uncertainty. One could imagine a scenario whereby inflation picks up significantly over the next decade due to excessively easy monetary policy, overly generous fiscal spending, growth in protectionism, rising labor pressure for wage increases, and the effects of a rising dependency ratio (the ratio of non-working people, especially retirees, to total population).1 But another scenario of continued “secular stagnation” and disinflation, caused by automation-driven job losses and a chronic lack of aggregate demand, is also conceivable. We think our middle-path forecast is the most sensible one to use in projecting likely asset returns, but investors might also want to plan based on these alternative scenarios too. Note that for Emerging Markets, we continue to show two different scenarios, which vary according to different projections of productivity growth. EM productivity growth has been declining steadily since around 2010, and in all major emerging economies, not just China. Our first scenario assumes that this decline ends and that, as in our assumption for developed economies, productivity growth reverts to the 20-year average. The more pessimistic (and, in our view, more likely) scenario assumes that the deterioration in productivity continues and that in 10 years’ time, EM productivity is the same as the average of developed economies. Which scenario will be correct depends on whether emerging economies, not least China, are able to implement structural reforms over the next decade, for example liberalizing the labor market, allowing a greater role for the private sector, improving corporate governance, and institutionalizing more orthodox fiscal and particularly monetary policy. Fixed Income Our anchor for calculating assumed returns is the return on long-term risk-free assets, specifically the 10-year government bond in the strongest countries. It is a reasonable assumption that an investor who buys, for example, a 10-year Treasury bond today and holds it for 10 years will make 1.6% a year in nominal US dollar terms. While this is not perfectly mathematically correct (since it ignores reinvested interest payments, for instance), empirically the return on government bonds has been very closely linked to the yield at the start-point in history (see Chart 1). From this starting-point in each country, we can easily build up the return for other fixed-income assets. These assumptions and the results are shown in Table 3. Table 3Fixed-Income Return Calculations Government bonds in most countries have an average duration of less than 10 years. Over the past five years, in the US it has averaged 6.4 years, and in the euro area 8.0 years. Only in the UK is the average over 10 years: 12.4 years to be precise. To calculate the return from the government bond index for each country we therefore assume that the shape of the yield curve (using the spread between 7-year and 10-year bonds) in future will be the same as the historic 20-year average. Cash. We assume that over the next 10 years the yield on cash will gradually revert to an equilibrium level. We calculate a market-implied real long-term neutral rate from the 10-year historical average of 5-year/5-year OIS implied forwards deflated by the 5-year/5-year implied CPI swap rate. This is a change from the methodology we used in 2019, when we based this off the neutral rate, r*, as calculated by the Holston Laubach-Williams model. But the New York Fed has temporarily stopped updating its calculation of this due to pandemic-induced volatility in the data, and anyway it was not available for every country. We turn the real cash rate into a total nominal return using our assumption for inflation described in detail in the GDP section above, the 20-year historical average of CPI. For inflation-linked securities, such as TIPS, we take the average yield over the past 10 years (a 20-year average was not available in many markets) and add the assumption for inflation described above. Corporate credit. We assume that spreads, and default and recovery rates, while highly volatile over the cycle, remain stable in the long run (Chart 7). We use 20-year averages for these, except that data for investment-grade default rates in Japan, the UK, Canada, and Australia are not available and so we use the average of the US and the euro zone. High-yield default rates are not available for the UK either, and so we do the same. Other bonds. For government-related debt (which is a big part of some bond indexes, 28% in the US for example) we assume that the 20-year historical average of the option-adjusted spread over government bonds will apply in the future too. We use the same methodology for securitized debt (for example, mortgage- and asset-based bonds): The 20-year average spread over the return on government bonds. Emerging Market debt. The assumptions and results for the three categories of EM debt (US dollar sovereign debt, US dollar corporate debt, and local currency sovereign debt) are shown in Table 4. We here assume that the 20-year average historical spread will continue in future. Default and recovery rates are a little harder to calculate, due to a lack of data. For USD sovereign debt (where defaults are rare and so hard to project), we use the rating-based default rate, calculated by Aswath Damodaran of NYU Stern School of Business.2 For USD-denominated EM corporate debt, we use the historical average, calculated by Moody's 2.5%.3 For local-currency debt, we use the same rating-based default rate as for USD sovereign debt. To translate the return into hard currency, we assume that currencies will move in line with the inflation differential between Emerging Markets and the US. For EM inflation we use an average of the IMF’s inflation forecasts for the nine largest emerging markets weighted by their weights in the J.P. Morgan GBI-EM Global Diversified local government bond index, and compare this to our US inflation forecast. This produces an EM inflation forecast of 2.9% a year, compared to 2.2% for the US, thus lowering the USD-based return from local EM debt by 0.7 percentage point. (See a more detailed discussion of forecasting long-term EM currency changes in the Currency section below). Index returns. Table 3 also shows the assumed return for the Bloomberg Barclays bond index for each country and for the global bond index, based on a weighted average of our assumption for each fixed-income asset class and country. Chart 7ACredit Spreads & Default Rates (I) Chart 7BCredit Spreads & Default Rates (II)   Table 4Emerging Market Debt   Equities The assumptions and detailed results for seven different equity markets are shown in Table 5. We have not made any substantial changes to our methodology for equities. We continue to use the average of six different methods to calculate the probable equity returns over the next 10-15 years. These are: Equity Risk Premium (ERP). The return from equities equals the yield on government bonds (we use 10-year bonds) plus an equity risk premium. For the US, we use an equity risk premium of 3.5%. This is based on work by Dimson, Marsh and Staunton4 showing that this is approximately the average excess return of equities over bonds in developed economies since 1900. We scale the equity risk premium for other countries using their average beta to the US market over the past 10 years. This varies from 0.66 for Japan (giving an ERP of 2.3%) and 1.2 in the euro area (ERP is 4.2%). Growth model. Here we assume that the return from equities equals the current dividend yield plus dividend growth. We need to adjust the dividend yield, however, to take into account that in some countries, particularly the US, it is more tax efficient for companies to do buybacks than to pay out dividends. We do this by adding equity withdrawals to the dividend yield. But this needs to be done on a net basis (taking into account equity issuance). We calculate this using the average annual change in the index divisor over the past 10 years. For the US, this is -0.8%, meaning there are more buybacks than new share issues. But in all other regions, the number is positive, and as high as 5.9% a year for Emerging Markets. This dilution is something that many calculations of assumed equity returns miss. For dividend growth, we assume that the dividend payout ratio remains stable, and that earnings growth is correlated with nominal GDP growth. However, history shows that earnings grow more slowly than GDP (logically so, when you consider that companies usually grow fastest before they list on a stock exchange). So we deduct 1% from nominal GDP growth to derive our earnings growth assumption. Note that for Emerging Markets, we use two different measures of dividend growth, depending on future productivity growth, as detailed above in our explanation of the GDP projections. Growth model (with reversion to mean). To take into account that valuations and profit margins typically revert to mean over the long run, we adjust the standard growth model (No. 2 above) by assuming that the current 12-month forward PE ratio and forward net profit margin for each country gradually revert over the next 10 years to their 20-year average. In the US, for example, that would mean that the current 12-month forward PE of 22.5x falls back to 16.0x, and profit margin of 12.5% falls to 10.7%. In every country and region, the profit margin is currently above the long-run average, and in all except the UK the PE is too. Note that we have changed from using the trailing PE and margin, because to use these now would be misleading given the big pandemic-driven decline in profits in 2020. Earnings yield. An intrinsically intuitive (and empirically demonstrable) way of estimating future returns is to use the earnings yield. This is based on the idea that an investor’s return from owning a stock comes either from the company paying a dividend, or from it investing retained earnings and paying a dividend in future. In the US, for example, a forward PE of 22.5x translates into an earnings yield of 4.4%. Again, here we switched this time to using 12-month forward forecast earnings yield, rather the trailing. Shiller PE. There is a strong correlation between valuation at the starting-point and the subsequent return from equities, at least over the long-run, although not over a period of less than 3-5 years (Chart 2). We regressed the Shiller PE (current price divided by average real earnings over the past 10 years) against the return from equities over the subsequent 10 years for each country and region. Composite valuation metric. The Shiller PE has its detractors. Using a fixed 10-year period does not reflect the different lengths of recessions and bull markets. It may say more about the mean-reverting nature of earnings than about whether the current price level is too high. So we also use the BCA Compositive Valuation Metric, which comprises eight indicators including, besides standard valuation measures such as price/sales and price/book, more esoteric ones such as market cap/GDP and Tobin’s Q. Again, we regress the metric against the subsequent 10-year return. Table 5Equity Return Calculations Alternative Assets Real Estate & REITs. We use the same basic methodology for both: The current yield (cap rate or dividend yield) plus projected capital value appreciation (linked to GDP growth). For US direct real estate, for example, we use the simple average cap rate of the five categories of commercial real estate (CRE), apartments, office, retail, industrial, and hotels in major cities: 6.1%. We also use the simple average of available city and category data for other countries. Cap rates are notoriously hard to estimate precisely; our data include a range of real estate, not just prime locations. We assume that capital values will grow in line with nominal GDP growth (using the same assumptions for this as we used for equities, 4.2%). We then deduct 0.5% for maintenance. This produces an expected return of 9.8% for the US. The only difference for REITs is that we do not deduct maintenance since this should already be reflected in the dividend yield. US REITs have a dividend yield currently of 3.5%, which produces an assumed return of 7.7% (Table 6). One risk with this methodology is that in the post-pandemic world, work and life practices might change. This will hurt office and residential real estate in major cities (which are overrepresented in investible CRE), though smaller cities and rural areas might benefit. As a result, capital values might fall. Table 6Alternatives Return Calculations Farmland & Timberland. Our methodology is similar to that for real estate: Current yield plus projected growth in capital values. For farmland, we use the farmland renter yield, sourced from the US Department of Agriculture. To estimate future land values, we take the gap between land value growth over the past 40 years (3.7%) and nominal growth of world GDP over that time (5.2%), assume that gap will continue and so deduct it from our estimate of global nominal GDP growth going forward (3.6%). This gives a result of 6.5%. For timberland, we assume that annualized returns in the future are the same as over the past 20 years. This produces a return assumption of 5.7%, which is (logically) moderately lower than our assumed return for farmland. Private Equity & Venture Capital. We project the return for private equity (PE) using the 30-year time-weighted average of the three-year rolling annualized return of PE over US large-cap equities, 3.6% (Chart 8). This produces an assumed return of 6.2%. For venture capital (VC), we use the same historical average for VC over PE (0.4%) to arrive at an assumed return of 6.6%. Hedge Funds. We use the 20-year time-weighted return of the Hedge Fund Composite Index over cash, 3.5% (Chart 9). This projects a future annual nominal return of 4.5%. Commodities. We previously used a methodology based on the idea that commodities’ bear markets in history have been rather fairly consistent, lasting on average 17 years, with an average decline of 50%, and that the current bear market began in 2012 (Chart 10). However, there are arguments that a new “commodities super-cycle” may be starting, driven by government infrastructure spending, and investment in alternative energy.5 We are agnostic for now on whether that will be the case, but it makes sense to switch to a neutral methodology, more in line with what we use for other assets classes: The return from commodities relative to GDP over the long run. Specifically, the CRB Raw Industrials Index has risen by an annualized 1.6% since 1951, during which time US nominal GDP growth averaged 6% (Chart 11). We assume that the differential will continue in future (although we calculate growth using global, not US, GDP), giving an annual return from commodities over the next 10-15 years of -0.9%. Gold. We calculate this using a regression of the gold price against nominal GDP growth and the annual change in the real 10-year yield over the past 40 years. For the forward-looking return assumption, we use a forecast of real rates (based on the equilibrium cash rate plus the average historical spread between the 10-year yield and cash) and a forecast of global nominal GDP growth. This produces an assumed return of 3.8%. Structured products. This asset class consists mainly of mortgage-backed and other asset-backed securitized instruments. In the US, these have historically returned 0.6% over US Treasurys. We assume that this premium continues, producing a total future return of 1.1% a year. Chart 8Private Equity Premium Chart 9Hedge Fund Return Over Cash     Chart 10Commodity Prices In History Chart 11Commodity Prices Vs. GDP Growth     Currencies Chart 12Currencies Tend To Revert To PPP To translate our local currency returns into an investor’s base currency, we need to arrive at some projections for FX movements over the next decade. Fortunately, for developed market currencies at least, it is relatively straightforward to use purchasing power parities (PPP) to do this since, over the long run, all the major currencies have tended to revert to PPP (Chart 12). We assume that in 10 years’ time all currencies will trade at PPP. We use the IMF’s estimate of today’s PPP for each currency to calculate the current under- or over-valuation. We assume that PPP will change in future years according to the relative inflation between each country and the US. The IMF provides five-year inflation forecasts and we assume that inflation will continue at this rate until 2031. For the euro zone, we calculate the PPP of the euro using the GDP-weighted PPPs of the five largest economies. The results (Table 7) suggest that the US dollar is currently overvalued and, given the forecast of higher inflation in the US than elsewhere in the future, will depreciate significantly against all major currencies except the Australian dollar. The USD is projected to depreciate by 1.7% a year against the euro and 1.1% against the yen over the next 10 years. It is likely to appreciate by 1.3% a year against the AUD, however. Table 7Currency Return Calculations Emerging Markets (Table 8) are more complicated. There is no evidence that EM currencies move towards PPP over time. All the major EM currencies are currently very cheap versus PPP (varying from 34% undervalued for the Chinese yuan to 67% for the Indonesian rupiah) but they were 10 years ago, too, and have not significantly moved towards PPP over that time. Table 8EM Currencies To calculate likely EM currency moves against the USD, therefore, we carry out a regression of the nine largest EM currencies against their relative CPI inflation rate to US inflation in history. We assume an intercept of zero. The regression coefficients vary from +0.5 for China to -1.7 for Malaysia. Apart from China, Malaysia, Poland and South Africa, the coefficients were negative, meaning that historically the USD has strengthened against the EM currency at least partly in line with relative inflation. To calculate likely future currency movements, we use the IMF’s five-year inflation forecasts and assume that the same rate of inflation will continue for our whole projection period. This methodology points to moderate annual depreciation of most EM currencies against the USD, varying from 0.8% a year for the Russian ruble to 0.1% for the Indonesia rupiah. The Chinese yuan and Taiwanese dollar are projected to appreciate moderately. We calculate the average EM currency movement using the weights of these nine large economies in the EM J.P. Morgan GBI-EM Global Diversified local-currency sovereign bond index. This produces a small (0.1%) a year appreciation. However, the IMF’s EM inflation forecasts may be too optimistic. It forecasts, for example, that Brazilian inflation will be only 3.3% a year in future, compared to an average of 6.1% over the past 20 years, and Russian inflation 4.0% versus a historical average of 9.3%. This suggests that EM currency performance could be worse than our projections. Table 9 shows the returns for the major asset classes expressed in local currency terms for six base currencies, based on the calculations explained above. Table 9Returns In Different Base Currencies Correlation And Volatility Below, in Table 10, we provide correlations for clients who need these inputs for their optimization calculations. Table 10Long-Run Correlation Matrix Returns can be calculated using the sort of forward-looking methodologies we have described above. For volatility, we think it is reasonable to use historical average data (Table 1, far right column), since volatility does not tend to trend over the long run (Chart 13). But correlation is a different matter. Correlations have varied significantly in history due to structural changes or regime shifts. The correlation of equities to bonds, it is well known, has moved from positive in the 1980s and 1990s, to negative since 2000 – probably because inflation disappeared as a factor moving bond prices (Chart 14). The correlation between equity market has risen as a result of the globalization of investment flows, though note that it fell back in 2010-2019. Chart 13Volatility Is Fairly Stable In The Long Run Chart 14Correlations Are Not Stable   So what correlations should investors use in an optimizer? Our recommendation would be to use the longest period of history available. A US investor, for example, might take the average correlation between Treasury bonds and large-cap US equities since 1945, 0.1%. Table 10 shows the correlation since 1973 of all the major asset classes for which data is available. Unfortunately, this misses some important asset classes such as high-yield bonds and Emerging Market equities, whose history does not go back that far. The results are intuitive – and prudent. From these numbers, it would seem sensible to use an assumption of a small positive correlation between US Treasurys and US equities, for example. US investment-grade debt has a correlation of 0.4 against equities. Global equity markets are all fairly highly correlated to each other, ranging mostly from 0.4 to 0.7. The most non-correlated asset class is commodities, especially gold.   Garry Evans, Senior Vice President Global Asset Allocation garry@bcaresearch.com   Amr Hanafy, Senior Analyst Global Asset Allocation amrh@bcaresearch.com   Footnotes 1 These are themes that BCA Research has been writing about for several years. See, for example, please see Global Investment Strategy, "1970s-Style Inflation: Could It Happen Again? (Part 1)," dated August 10, 2018; and " 1970s-Style Inflation: Could It Happen Again? (Part 2)," dated August 24, 2018. 2 Please see http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/ctryprem.html 3 Annual Emerging Markets Default Study: Coronavirus Will Push Up Default Rates https://www.moodys.com/researchdocumentcontentpage.aspx?docid=PBC_1214906 4 Please see, for example, https://www.credit-suisse.com/media/assets/corporate/docs/about-us/research/publications/credit-suisse-global-investment-returns-yearbook-2021-summary-edition.pdf. 5 Please see Commodity & Energy Strategy, "Industrial Commodities Super-Cycle Or Bull Market?", dated March 4, 2021.

Special Report

Highlights The selloff in crypto-currencies on May 19 may be overblown but the risk of government intervention is a rising headwind for this asset class. While environmental concerns are a threat to Bitcoin, the entire crypto-currency complex faces a looming confrontation over governance. Digital currencies are a natural evolution of money following coinage and paper. Moreover a sizable body of consumers is skeptical of governments and traditional banking. Loose monetary conditions are fueling a speculative mania. However, governments fought for centuries to gain a monopoly over money. As crypto-currencies become more popular, governments will step in to regulate and restrict them. Central bank digital currencies (CBDCs) threaten to remove the speed and transactional advantage of crypto-currencies, leaving privacy/anonymity as their main use-case. Feature The prefix “crypto” derives from the Greek kruptos or “hidden.” This etymology highlights one of the biggest problems confronting the crypto-currency craze in financial markets today. Speed and anonymity are the greatest assets of the digital tokens. But the former advantage is being eroded by competitors while the latter is becoming a political liability. In the 2020s, governments are growing stronger and more interventionist, not weaker and more laissez faire. Chart 1Loose Money Fuels Crypto Mania Bitcoin and rival crypto-currency Ethereum fell by 29.5% and 43.2% in intra-day trading on May 19, only to finish the day down by 13.8% and 27.2%, respectively. The market panicked on news that China’s central bank had banned firms from handling transactions in crypto-currencies. What really happened was that China’s National Internet Finance Association, China Banking Association, and Payment and Clearing Association issued a statement merely reiterating a 2013 and 2017 policy that already banned firms from handling transactions in crypto-currencies. These three institutions also warned about financial speculation regarding crypto-currencies.1 The crypto market suffered a spike in volatility because it is in the midst of a speculative mania. In the last five years, total market capitalization of crypto-currencies has risen from around $7 billion to $2.3 trillion,2 recording a 34,000% gain. Some crypto-currencies have even recorded returns in excess of that number over a shorter horizon. Price gains have been driven by retail buyers who may or may not know much about this new asset class (Chart 1). Prior to the May 19 selloff, prices had grown overextended and recent concerns over the environment, sustainability, and governance (ESG) had shaken confidence in Bitcoin and its peers. Chinese authorities have already banned financial firms from providing crypto services in a bid to deter ownership of crypto-currencies. And China is not alone. The latest market jitters are a warning sign that government interference in the crypto-currency market is a real threat. Regulation and sovereign-issued digital currencies are starting to enter the fray. While ultra-dovish central bank policies are not changing soon, and therefore crypto-currency price bubbles can continue to grow, crypto-currencies will remain subject to extreme volatility and precipitous crashes. In this report we argue that the fundamental problem with crypto-currencies is that they threaten the economic sovereignty of nation-states. Environmental degradation, financial instability, and black market crime, and other concerns about cryptos have varying degrees of merit. But they provide governments with ample motivation to pursue a much deeper interest in regulating a technological innovation that has the power to undermine state influence over the economy and society. Government scrutiny is a legitimate reason for crypto buyers to turn sellers. Does The World Need Crypto-Currencies? Broadly speaking, there are two primary justifications for crypto-currencies, centered on a transactional basis: speed and privacy/anonymity. The crux of crypto-currency creation rests on these two use cases.3 The speed of crypto-currencies comes from their ability to increase efficiency in local and global payment systems by facilitating financial transactions without the need of a third party (e.g. a financial institution). Cross-border settlement of traditional (fiat) currency transactions processed through the standard SWIFT communications system takes up to two business days. Most transactions involving crypto-currencies over a blockchain network are realized in less than an hour, cross-border or not.4 The fees involved with third-party payments are often more expensive than transacting with crypto-currencies. Simply put, excluding the “middleman” can save money. This is a selling point in a global market that expects to see retail cross-border transactions reach $3.5 trillion by the end of 2021, of which up to 5% are associated with transaction-based fees.5 But this breakthrough in payment system technology can be overstated and is not the main reason for using crypto-currency. Speculation drives current use, especially given that there is speculative behavior even among those who believe that cryptos are safe-haven assets or promising long-term investments (Chart 2). Chart 2Crypto-Currency Use Driven By Speculation Chart 3Consumers Growing Skeptical Of Banking Regulation If a person wants to buy an item from a company in a distant country, that person could use a crypto-currency just as he or she could use a credit card. Both parties would have a secure medium of exchange but, unlike with a credit card, both would avoid using fiat currencies. Neither party could conduct the same transaction using gold or silver. The crucial premise is the existence of an online community of individuals and firms who for one reason or another want to avoid fiat currencies. From a descriptive point of view, the crypto-currency phenomenon implies a lack of trust in modern governments, or at least their monetary systems, and an assertion of individual property rights. The list of crypto-currencies continues to grow. To date, there are approximately 9,800 of them. Some are trying to prove their economic value or use, while others have been created with no intended purpose or problem to solve. Even so, there has yet to be a crypto-currency that overwhelms the use of slower fiat money. In a recent Special Report, BCA Research’s Foreign Exchange Strategist Chester Ntonifor showed that crypto-currencies still have a long way to go to have a chance at replacing fiat monies. While crypto-currencies are showing signs of significant improvement as mediums of exchange, they still fall short as stores of value and units of account. The other primary case for crypto-currencies is privacy or anonymity. The bypassing of intermediaries implies a greater control of funds by the two parties of a transaction. Crypto-currencies are said to be more “private” compared to fiat money. Fiat money is controlled by governments and banks while crypto-currencies have only “owners.” Crypto-currencies are anonymous because they are stored in digital wallets with alphanumeric sequences – there is a limited personal data trail that follows crypto-currency compared to those of electronic fiat currency transactions. In a post-9/11, post-GFC, post-COVID world where a sizable body of consumers is growing more skeptical of government surveillance and regulation and banking industry practices (Chart 3), crypto-currencies give users more than just a means to transact with. However, privacy is not the same as security. Hacking and fraud can affect cryptos as well as other forms of money and attacks will increase with the value of the currencies. Bitcoin At The Helm Of Crypto-Currency Market Chart 4Bitcoin Slows Bitcoin has cemented its status as the number one currency in the crypto-verse.6 It is considered to be the first crypto-currency created, it is the most widely accepted, it is touted as a store of value or “digital gold,” and it is the most featured in quoting alternative crypto-currency pairs across crypto exchanges. As it stands, Bitcoin accounts for around 42% of total crypto-currency market capitalization.7 This share has declined from around 65% at the start of 2021 on the back of the frenzied rise of several alternative coins.8 But rising risks to Bitcoin’s standing will cause the entire crypto-market to retreat. In a Special Report penned in February, BCA Research’s Chief Global Strategist Peter Berezin argued that Bitcoin is more of a trend than a solution and that its usefulness is diminishing. Bitcoin’s transaction speed is slowing and its transaction cost is rising (Chart 4). Slowing speed and rising cost on the Bitcoin network are linked to a scalability problem. The crypto-currency’s network has a limited rate at which it can process transactions related to the fact that records (or “blocks”) in the Bitcoin blockchain are limited in size and frequency. This means that one of its fundamental justifications, transactional speed, will become less attractive over time, should the network not address these issues. Bitcoin also consumes a significant amount of energy, a controversy that is gaining traction in the crypto-currency market after Elon Musk, the “techno-king” of Tesla, cited environmental concerns in reversing his decision to accept Bitcoin payment for his company’s electric vehicles. Energy consumption rises as more coins are mined, since mining each new Bitcoin becomes more computer-power intensive. The need for computing power and energy will continue to increase until all 21 million Bitcoins (total supply) are mined, which is currently estimated to occur by the year 2140. Strikingly, the energy needed to mine Bitcoin over a year are comparable to a small country’s annual power consumption, such as Sweden or Argentina (Chart 5). Chart 5Bitcoin Consumes More Energy Than A Small Country … Bitcoin also generates significant quantities of electronic waste (Chart 6). Chart 6… And Generates A Lot Of Electronic Waste Bitcoin mining is heavily domiciled in China, which accounts for 65% of global mining activity (Figure 1). China’s energy mix is dominated by coal power, which makes up approximately 65% of the country’s total energy mix even after a decade of aggressive state-led efforts to reduce coal reliance. Of this, coal powered energy makes up approximately 60% of Bitcoin’s energy mix in China.9 With several countries aiming to minimize carbon emissions, and with approximately 60% of Bitcoin mining powered by coal-fired energy globally,10 Bitcoin imposes a major negative environmental impact. Figure 1Bitcoin Mining Well Anchored In Asia Bitcoin does not shape up well when compared to gold’s energy intensity either. Bitcoin mining now consumes more energy than gold mining over a single year. While the energy difference is not large, the economic value is. Gold’s energy consumption to economic value trade-off is lower than that of Bitcoin. The production value of gold in 2020 was close to $200 billion, while Bitcoin was measured at less than $25 billion (Chart 7A). On a one-to-one basis, gold even has a lower carbon footprint than Bitcoin (Chart 7B). Chart 7AGold Outshines Bitcoin On Production Value And Carbon Footprint Chart 7BGold Outshines Bitcoin On Production Value And Carbon Footprint Crypto-currency energy consumption and carbon footprint will attract the attention of government regulators. Of course, not all crypto-currencies are heavy polluters. But if the supply of cryptos is constrained by mining difficulties then they will require a lot of energy. If the supply is not constrained then the price will be low. Government Regulation Is Coming Environmental concerns point to the single greatest threat to crypto-currencies – the Leviathan, i.e. the state. In this sense the crypto market’s wild fluctuations on May 19, at the mere whiff of tougher Chinese regulation, are a sign of what is to come. Governments around the world have so far left crypto-currencies largely unregulated but this laissez-faire attitude is already changing. Environmental regulation has already been mentioned. Governments will also be eager to expand their regulatory powers to “protect” consumers, businesses, and banks from extreme volatility in crypto markets. But investors will underrate the regulatory threat if they focus on these issues. At the most basic level, governments around the world will not sit idly by and lose what could become significant control of their monetary systems. The ability to establish and control legal tender is a critical part of economic sovereignty. Governments won control of the printing press over centuries and will not cede that control lightly. If crypto-currencies are adopted widely, then finance ministries and central banks will lose their ability to manipulate the money supply and the general level of prices effectively. Politicians will lose the ability to stimulate the economy or keep inflation in check. Most importantly, while one may view such threats as overblown, it is governments, not other organizations, that will make the critical judgment on whether crypto-currencies threaten their sovereignty. Throughout the world, most crypto-currency exchanges are regulated to prevent money laundering. Crypto-currencies are not legal tender and, aside from Bitcoin, their use is mostly banned in China (Table 1). However, more specialized regulation that targets energy and economic use has yet to be brought into law across the world. Table 1World Governments Will Not Relinquish Hard-Fought Monopolies Over Money Supply In China, initial coin offerings (ICOs – the equivalent of an initial public offering on the stock market) and trading platforms are banned from engaging in exchanges between the yuan and crypto-currencies or tokens. In fact, China recognizes crypto-currencies only as virtual commodities or virtual property. India is another country where exchanges and ICOs are banned. While crypto-currencies are not banned, they are not legal tender. Indian policymakers have recently proposed banning crypto-currencies, however. The proposed legislation is one of the world’s strictest policies against crypto-currencies. It would criminalize possession, issuance, mining, trading, and transferring crypto-assets. If the ban becomes law, India would be the first major economy to make holding crypto-currency illegal. Even China, which has banned mining and trading, does not penalize possession. In the US, Secretary of the Treasury Janet Yellen has already expressed concerns regarding the illicit use of cryptos for supposed criminal gain.11 She is in alignment with European Central Bank President Christine Lagarde. Because of the anonymity of crypto-currencies, identifying users behind illicit transactions is difficult. This means regulators face headwinds in identifying transactions that are made for criminal gain, as compared to fiat transactions. Governments have long dealt with the anonymity of cash but they have ways of monitoring bank accounts and paper bills. Crypto-currencies are beyond their immediate sight of control and therefore will attract growing scrutiny and legislative action in this regard. The Colonial Pipeline ransomware attack on May 7, which temporarily shuttered about 45% of the fuel supply line for the eastern United States, illustrates the point. The DarkSide group of hackers who orchestrated the attack demanded a ransom payment of $4.4 million worth of Bitcoin, which Colonial Pipeline paid them on May 7. Shortly thereafter, unspecified “law enforcement agencies” clawed back the $4.4 million from the hackers’ account (transferring it to an unknown address) and DarkSide lost access to its payment server, DOS servers, and blog. This episode should not be underrated. It was a successful, large-scale cyber-attack on critical infrastructure in the world’s most powerful country. It highlighted the illicit uses to which crypto-currencies can be put. True, criminals demand ransoms in fiat money as well – and many crypto-currency operators will distance themselves from the criminal underworld. Nevertheless governments will give little slack to an emerging technology that presents big new law enforcement challenges and is not widely used by the general public. Ultimately governments will pursue their sovereign interests in controlling money, the economy, and trade, listening to their banking lobby, expanding their remit to “protect” consumers, and cracking down on illicit activity. Governments are not capable of abolishing crypto-currencies altogether, or the underlying technology of blockchain. But they will play a large and growing role in regulating them. Central Banks Advancing On Digital Currencies Central bank digital currencies (CBDCs) will leave crypto-currencies in the realm of speculative assets. CBDCs are a form of digital money denominated in a country’s national unit of account and represent a liability on a central bank’s balance sheet. This is different from current e-money that represents a claim on a private financial institution’s balance sheet. It is also different from crypto-currencies, because there is a central authority behind a CBDC, unlike with crypto-currencies due to their decentralized nature. In China, the People’s Bank of China (PBoC) has suggested its rollout of a digital yuan is “ready” despite no release to date. Beta testing is ongoing in several provinces. The PBoC’s justification for a digital yuan comes from China’s growing cashless economy. The transition away from cash is largely thanks to mobile payment platforms like Alibaba’s Alipay and Tencent’s WeChat Pay, which, between the two of them, control almost the entire mobile payments market of some 850 million users. There is a significant amount of systemic risk in this system – one reason why Chinese authorities have recently subjected these companies to new scrutiny and regulation. Should Alibaba or Tencent go bankrupt, the local payment system will crash. The PBoC’s efforts will increase competition in the local payments space and reduce this systemic risk. Policymakers are also concerned that as Chinese citizens choose to hold their money in digital wallets provided by Alibaba and Tencent instead of bank accounts, liquidity is being drained from the traditional banking system, putting deposit levels at banks under strain, and posing risks to liability matching. The digital yuan will still involve a third party, unlike crypto-currencies which do not. Doing away with commercial banks is not a reality – indeed the Chinese Communist Party seeks to buttress the state-owned commercial banks in order to maintain control of the economy. What the digital yuan does, and other CBDCs will do too, is utilize blockchain technology, which is faster and more secure than traditional payment networks. In the US, the Fed has been studying the viability of a CBDC US dollar. The Fed has stated that it is carefully exploring whether a CBDC will lead to “safer, less expensive, faster, or otherwise more efficient payments.” While the Fed has yet to find a single standout case for a CBDC US Dollar, Fed Chair Jerome Powell said last year that the US has a “competitive payments market” with “fast and cheap services, particularly in comparison to other nations exploring a CBDC.” To date, the Fed’s observation is that many of the challenges that CBDCs hope to address do not apply to the US, including disuse of physical cash, narrow reach or high concentration of banking, and weak infrastructure for payment systems. Rather, the Fed is more focused on developing the FedNow real-time payment system for private banks. This is much the same as in Europe, where physical cash still plays a major role in day-to-day economic activity and where local payment systems are fast and secure. But central banks around the world continue to engage in work centered around CBDCs (Charts 8A and 8B) – and China’s progress will encourage others to move faster. Advanced economies are mostly interested in creating a safer and more efficient payment system, while emerging and developed economies have interest across several areas such as financial stability, monetary policy setting, and inclusiveness of banking, as well as efficiency and safety (Chart 9). CBDCs are especially attractive to emerging market policy makers at targeting those who lack access to traditional banking. Chart 8ACentral Banks Advancing On CBDC Work Chart 8BCentral Banks Advancing On CBDC Work Chart 9Central Banks CBDC Interest Areas In remote areas, access to banking is scarce and expensive. CBDCs can help solve this problem. Individuals would have CBDC accounts directly on a central bank ledger. They could then access their money and transact through a digital wallet application that is linked to the CBDC account. Giving people access to digital currency would allow them to transact quickly, in remote settings, without the need of hard currency. Monetary policy transmission is also better in advanced economies. In emerging markets, there are bottlenecks in local financial markets. Looser central monetary policy does not always translate into cheaper financing across the economy. In remote and poverty stricken areas, monetary policy transmission is sticky, meaning high costs of borrowing can persist even through accommodative policy cycles. This is a smaller issue in advanced economies. Payment systems in advanced economies are due an overhaul in security and efficiency, and CBDCs and blockchain technology will provide this. CBDCs will prove to be just as efficient to transact with as any crypto-currencies available today. CBDCs will also be legal tender and accepted by all vendors. The anonymity factor will be lost but this will not be a problem for most users (whereas legal issues will become a problem for crypto-currencies). The probability of central banks issuing CBDCs in both the short and medium term, both in the retail and wholesale space, is rising. If advanced economies like those of the G7 issue CBDCs soon, policy makers will undoubtedly ensure the use of it over the currently circulating and partially accepted crypto-currencies. The endgame will leave crypto-currencies in the highly speculative asset class, perhaps even in the black market where anonymity is valued for transactions that wish not to be tracked. Investment Takeaways Prices of crypto-currencies may continue to rise given sky-high fiat money creation amid the COVID pandemic and ultra-low interest rates. Digitalization is the natural next step in the evolution of money from precious metals to paper banknotes to electronic coin. But the market leader, Bitcoin, is encountering more headwinds. The primary case for the use of Bitcoin is challenged due to slowing transaction speeds and rising transaction costs. The virtual currency is primarily mined using coal-powered energy, resulting in growing scrutiny from governments and consumers. Government regulation is entering the ring and policymakers will take an increasingly heavy-handed role in trying to ensure that cryptos do not undermine economic sovereignty, financial stability, and law and order. When central banks begin to rollout digital currencies, especially those domiciled in advanced economies, crypto-currencies as medium of exchange will lose much of their allure. Crypto-currencies will remain as anti-fiat currencies and speculative assets. Risks To The View Given the controversy surrounding crypto-currencies, it is only fair to state outright the risks to our view. We would also recommend clients read our colleague Dhaval Joshi’s latest bullish take on Bitcoin. First, scaling up Bitcoin’s network and processing transactions in batches instead of single transactions will resolve transaction time and cost risks, restoring efficiency. This is a clear solution to efficiency concerns. However, scaling and batching transactions are not on the immediate horizon of Bitcoin developers. Bitcoin’s network will still need to undergo another “halving” in order for this risk to subside and for the network to scale. A halving of the network will only occur again in 2024.12 Second, on the environment: Bitcoin mining is not solely dependent on fossil fuel energy that gives it a “dirty” footprint. Renewables already make up some 25% of Bitcoin mining. Increasing the use of renewables in Bitcoin’s energy mix will help lower its environmental impact. However, this is easier said than done. Global renewable energy has yet to scale up to a point where it can consistently out-supply existing fossil-fuel energy. Mining hardware also has its associated carbon footprint that would need to be addressed. And location matters too. Crypto-currency mining farms are large-scale projects. Simply uprooting operations to a country that could lower the carbon footprint of a mining farm or two is not viable due to the costs involved. Hence crypto-currency mining will probably continue to be a “dirty” operation but a rapid shift to renewables would challenge our thesis. Bitcoin’s network is also based off a “proof of work” protocol. Miners must prove that a certain amount of computational effort has been expended for confirming blocks on the network, allowing transactions to be processed. Proof of work is energy intensive. Other crypto-currencies, like Ethereum, will adopt a “proof of stake” protocol. Simply put, transactions are confirmed by users and their stake in the associated crypto-currency. Proof of stake is less energy intensive compared to proof of work. Third, as to government regulation, the longer policymakers take to enact legislation targeting crypto-currencies, the larger their market will grow. Regulation in China and India may set a benchmark for major economies but not all will follow in the Asian giants’ footsteps. Some governments have been slow to study crypto-currencies, meaning legislation aimed at governing or regulating them may still be long in coming. Innovation is a good thing and free economies will not wish to restrain crypto-currencies or blockchain technology unduly, for fear of missing out. Fourth, on CBDCs, some central banks may only adopt them based on their respective economic needs. However, rising crypto-currency populism drives associated economic risks which can force the hands of central banks to adopt CBDCs in lieu of said needs. Each country faces unique challenges. Some central banks may not want to be left behind even if they believe their policy framework is facilitating economic activity efficiently. While the Fed has stated that it will not adopt a CBDC for the primary reason of ensuring payment security since it believes it already has a safe system in place, this view will change. The Fed could justify a move to a CBDC US dollar on the single basis of transitioning to a more sophisticated technology for the future. The Fed will not want to be caught behind the curve considering the PBoC is priming its digital yuan for release soon. Technological leadership is a strategic imperative of the United States and that imperative applies to financial technology as well as other areas.   Guy Russell Research Analyst GuyR@bcaresearch.com Matt Gertken Vice President Geopolitical Strategy mattg@bcaresearch.com Footnotes 1 Muyao Shen, “China Reiterates Crypto Bans From 2013 and 2017”, coindesk, May 18, 2021, coindesk.com. 2 As of May 11, 2021. 3 There are several other reasons or “problems” that crypto-currencies are created for or to solve, but speed and privacy form the basis of crypto-currencies first coming into existence. 4 Not all crypto-currencies transact in less than an hour. But there are many that transact in several minutes and in some cases, mere seconds. As the leading crypto-currency, Bitcoin takes approximately one hour for a transaction to be fully verified over its network. 5 “McKinsey’s Global Banking Annual Review”, McKinsey, Dec. 9, 2020, mckinsey.com. 6 We use Bitcoin as an example to understand the risk and impact of forthcoming government regulation and competition. Because of Bitcoin’s status, any significant risks that threaten the crypto-currency’s standing as the number one currency will threaten the entire market. 7 As of May 20, 2021. Figure varies daily. See www.coinmarketcap.com for more information. 8 Alternative currencies such as Ethereum, Ripple, Binance Coin, Dogecoin, and Cardano have chipped away at Bitcoin’s crypto-market dominance through 2021. 9 According to The Center For Alternative Finance, The University Of Cambridge. 10 According to The Center For Alternative Finance, The University Of Cambridge. 11 Data on the use of crypto-currencies for illicit activitiessays otherwise. Of all crypto-currency transactions, it is estimated that only 2.1% are used for illicit activities. See “2021 Crypto Crime Report”, Chainalysis, chainalysis.com. 12 A Bitcoin halving is when the reward for mining Bitcoin transactions is cut in half. This event also cuts Bitcoin's inflation rate and the rate at which new Bitcoins enter circulation, in half. Bitcoin last halved on May 11, 2020.

Highlights The drubbing that cryptocurrencies have received over the past two weeks is just a taste of things to come. Crypto markets will continue to face tighter regulation, as this week’s announcements from China and the US Treasury underscore. The hope that cryptocurrencies can ever truly “go green” is wishful thinking. Given their decentralized nature, cryptocurrencies require real resources to be expended to permit secure transactions to take place. In addition to their technical limitations, cryptocurrencies face a fundamental constraint, which we dub the “Crypto Impossibility Theorem.” The Crypto Impossibility Theorem states that cryptocurrencies will be viable only if they offer a higher return than equities. The assumption that cryptos can generate a return in excess of equities is almost certain to be false since it would require that cryptocurrency holdings rise more quickly than income in perpetuity. In the near term, the pain in crypto markets could drag down other speculative assets such as tech stocks. In the long term, diminished investor interest in cryptos will benefit the stock market, as investor attention focuses back on equities. Cryptos: Can’t Have It All Investors who track the cryptocurrency market might be aware of the “blockchain trilemma.” It posits that cryptocurrencies can possess only two of the following three attributes: decentralization, security, and scalability. Bitcoin is both highly decentralized and reasonably secure. However, because control of the Bitcoin blockchain is distributed across thousands of individual computer nodes, it is also very slow. The Bitcoin network can barely process five transactions per second, compared to over 20,000 for the Visa network (Chart 1). The average fee for a Bitcoin transaction is around $30, a number that has risen over the past few years (Chart 2). Chart 1Speed Of Transactions, Or Lack Thereof Chart 2Rising Cost Per Transaction   The elaborate puzzles that the Bitcoin algorithm must solve to verify transactions are extremely energy intensive. Bitcoin mining consumes more energy than entire countries such as Sweden, Argentina, and Pakistan (Chart 3). About two-thirds of Bitcoin mining currently takes place in China, often using electricity generated by burning coal. Chart 3Bitcoin And Ethereum: How Dare You! Some claim that Bitcoin and other cryptocurrencies are shifting to renewable energy sources, a trend that will continue in the years ahead. However, this argument misses the point, which is that the “proof of work” mechanism that underpins Bitcoin requires that real resources be expended. Suppose that all Bitcoin mining could be performed entirely for free using solar energy. This would reduce the cost of running a “mining rig,” incentivizing more mining. The Bitcoin algorithm operates in such a way that the difficulty of mining coins increases as the total computational power of all miners grows. In this computational rat race, miners would need to purchase more servers with ever more powerful specifications to keep up with their competitors. And semiconductors do not grow on trees. It takes real resources to produce them. As this recent Bloomberg article pointed out, Taiwan Semiconductor generates almost 50% more greenhouse emissions than General Motors. Like Bitcoin, Ethereum uses the “proof of work” mechanism to verify transactions. There have been active discussions to shift Ethereum to a “proof of stake” mechanism, which would greatly expedite transactions.1 However, some have argued that a proof of stake system would degrade security, allowing for “double-spend attacks” where someone transfers coins to someone else but then spends the coins before the transaction is completed. The Crypto Impossibility Theorem We will not delve any further into the technical nature of the blockchain trilemma other than to note that it poses a serious challenge to the entire cryptocurrency project. Instead, let us highlight another obstacle that has received less attention – one that could be even more damaging for the prospects of cryptocurrencies in the long run. Let us hyperbolically call it the “Crypto Impossibility Theorem.” The Crypto Impossibility Theorem states that a cryptocurrency will be viable only if it offers a higher return than equities. As we discuss below, the assumption that cryptos can generate a return in excess of equities is almost certain to be false since it would require that cryptocurrency holdings rise more quickly than income in perpetuity. This implies that the value that investors currently attach to cryptos will turn out to be illusory. To see the theorem in action, recall that money serves three functions: As a unit of account, as a medium of exchange, and as a store of value. It is doubtful that anyone seriously thinks that the price tag on a box of cereal will ever be displayed in units of Bitcoin, ether, or any of the various dog coins currently in vogue. Thus, we can scratch “unit of account” off the list of possible crypto uses. What about medium of exchange? One can imagine a scenario where the prices of goods and services are still listed in dollars, but one may transfer the equivalent in cryptocurrencies to purchase them. However, this raises an obvious question: Why would anyone choose to hold a cryptocurrency if wages and prices are denominated in fiat currencies such as US dollars or euros? The only possible answer is that people must see cryptocurrencies as fulfilling the third function of money, namely being a store of value. Would people be willing to hold cryptocurrencies if their prices generally moved sideways? It is doubtful. Cryptocurrencies are risky. Cryptocurrency accounts are not subject to deposit insurance. Crypto prices are also extremely volatile. During the pandemic, the S&P 500 fell by 34%, but the price of Bitcoin sank by an even greater 53%. Other cryptocurrencies fared even worse. In contrast, the trade-weighted US dollar strengthened by about 4% while gold prices only fell marginally (Chart 4). Thus, to incentivize people to hold cryptos, the prospective capital gain has to be large enough to offset the inherent volatility in owning these currencies. Chart 4Cryptocurrencies Fared Badly During Last Year’s Equity Sell-Off This is where the Crypto Impossibility Theorem comes in. Unlike dividend-paying stocks, cryptocurrencies do not provide any income to their holders. Thus, even if cryptos were just as risky as stocks, the price of cryptos would still need to rise more than the price of stocks in order to ensure that investors remain indifferent between the two asset classes. In practice, as the experience of the pandemic demonstrates, cryptos are even riskier than stocks. Thus, the expected return on cryptos has to exceed the expected increase in stock prices by more than the dividend yield. The problem for crypto holders is that this is not mathematically possible. Even if one controls for the rise in price-earnings multiples over time, equity returns have generally exceeded nominal GDP growth (Table 1). Hence, if cryptos need to offer superior returns to equities, and if the return on equities is at least equal to nominal GDP growth, then the market capitalization of cryptocurrencies will not only end up rising faster than for stocks, it will rise faster than aggregate national income. In a digital world where people need ever-less money to facilitate transactions, there is no good reason to expect this to happen. Table 1Equity Returns And GDP Growth A Fashion Choice Crypto-optimists might argue that the required rate of return to holding cryptos will decline as the market matures. This is wishful thinking. Equities derive their value from the fundamentals of a company’s business. In contrast, cryptocurrencies have no intrinsic value. Their value is whatever others are willing to pay for them. Not only does this make cryptocurrencies inherently more risky than equities, it also makes them highly susceptible to fashion trends. It is not surprising that many upstart cryptocurrencies have crafted ties with celebrities and other “influencers.” The whole point is to get enough people interested in a cryptocurrency to generate a feedback loop of wider adoption, thus allowing the currency’s early backers to cash out. The drubbing that cryptocurrencies have received over the past two weeks is just a taste of things to come. In this sense, cryptocurrencies are even more vulnerable to affinity scams than other assets such as precious metals. While apocalyptic warnings of “currency debasement” have long been used to sell bullion, at least with gold and silver, you truly do get something that is in short supply. In the case of cryptocurrencies, while the supply of any individual cryptocurrency may be limited, the overall supply is unbounded. This means that the average price of each currency is likely to rise much less than the aggregate value of all cryptocurrencies, making the entire asset class even less viable over time.   Cryptogeddon The drubbing that cryptocurrencies have received over the past two weeks is just a taste of things to come. As Matt Gertken and Guy Russell discuss in this week’s Geopolitical Strategy report, crypto markets will continue to face tighter regulation (Table 2). Just this week, China reiterated its ban on financial companies offering cryptocurrency services. As part of its broader effort to crack down on tax evasion, the US Treasury Department also announced that it will require any cryptocurrency transfer worth $10,000 or more to be reported to the IRS. Table 2Regulation Of Cryptos: What Can And Cannot Be Done The blockchain trilemma will make it impossible for cryptos to overcome ESG concerns, while the Crypto Impossibility Theorem will prevent cryptocurrencies from ever being stable stores of value. In the meantime, an ebbing of input price inflation will take some of the wind out of the sails from the argument that cryptos are an indispensable hedge against the “inevitable” debasement of fiat monies. Chart 5 shows that DRAM prices have rolled over. Lumber prices have dropped 11% so far this week. Corn, soybean, and steel prices have also backed off their highs. Cryptos are like sharks; they need to move forward or they will sink. Back when they were unknown to most investors, a speculative case could have been made for buying cryptos. However, that case vanished earlier this year when the aggregate value of cryptocurrencies briefly surpassed the entire stock of US dollars in circulation (Chart 6). Even with the recent correction, there are 17 cryptocurrencies with market capitalizations above $10 billion (Table 3). Chart 5To The Moon And Back? Chart 6Aggregate Value Of Cryptos Briefly Surpassed The Entire Stock Of US Dollars In Circulation Table 3Close To 20 Cryptos Have A Market Cap In Excess Of US$10bn What will the ongoing crypto collapse mean for the broader investment landscape? In the near term, the pain in crypto markets could drag down other speculative assets such as tech stocks. In the long term, diminished investor interest in cryptos will benefit the stock market, as investor attention focuses back on equities. For the broader economy, the impact of a crypto bear market will be limited. The banking system has very little exposure to cryptos. There will be a modest adverse wealth effect from falling crypto prices. However, the inability of a few laser-eyed crypto traders to buy their Lambos is hardly going to matter much against the backdrop of strong stimulus-fueled consumption growth in the US and a number of other economies. Investors should continue to overweight stocks in a global asset portfolio, favoring value over growth, cyclicals over defensives, and non-US stocks over their US peers. Peter Berezin Chief Global Strategist pberezin@bcaresearch.com Footnotes 1 Proof of Work (PoW) and Proof of Stake (PoS) are two methods used to ensure the integrity of a coin’s ledger or record of transactions. PoW achieves this by requiring miners (those who add transactions to the ledger) to solve a time-consuming mathematical puzzle. PoS achieves this through a different mechanism, where anyone who stakes their own coin can be randomly selected to add new transactions to the ledger. Those holding or “staking” more coin have a higher probability of being selected. Global Investment Strategy View Matrix Special Trade Recommendations Current MacroQuant Model Scores

Highlights Global oil markets will remain balanced this year with OPEC 2.0's production-management strategy geared toward maintaining the level of supply just below demand.  This will keep inventories on a downward trajectory, despite short-term upticks due to COVID-19-induced demand hits in EM economies and marginal supply additions from Iran and Libya over the near term. Our 2021 oil demand growth is lower – ~ 5.3mm b/d y/y, down ~ 800k from last month's estimate – given persistent weakness in realized consumption.  We have lifted our demand expectation for 2022 and 2023, however, expecting wider global vaccine distribution and increased travel toward year-end. The next few months are critical for OPEC 2.0: The trajectory for EM demand recovery will remain uncertain until vaccines are more widely distributed, and supply from Iran and Libya likely will increase this year.  This will lead to a slight bump in inventories this year, incentivizing KSA and Russia to maintain the status quo on the supply side. We are raising our 2021 Brent forecast back to $63/bbl from $60/bbl, and lifting our 2022 and 2023 forecasts to $75 and $78/bbl, respectively, given our expectation for a wider global recovery (Chart of the Week). Feature A number of evolving fundamental factors on both sides of the oil market – i.e., lingering uncertainty over the return of Iranian and Libyan exports and the strength of the global demand recovery – will test what we believe to be OPEC 2.0's production-management strategy in the next few months. Briefly, our maintained hypothesis views OPEC 2.0 as the dominant supplier in the global oil market. This is due to the low-cost production of its core members (i.e., those states able to attract capital and grow production), and its overwhelming advantage in spare capacity, which we reckon will average in excess of 7mm b/d this year, owing to the massive production cuts undertaken to drain inventories during the COVID-19 pandemic. Formidable storage assets globally – positioned in or near refining centers – and well-developed transportation infrastructures also support this position. We estimate core OPEC 2.0 production will average 26.58mm b/d this year and 29.43mm b/d in 2022 (Chart 2). Chart of the WeekBrent Prices Likely Correct Then Move Higher in 2022-23 Chart 2OPEC 2.0 Will Maintain Status Quo The putative leaders of the OPEC 2.0 coalition – the Kingdom of Saudi Arabia (KSA) and Russia – have distinctly different goals. KSA's preference is for higher prices – ~ $70-$75/bbl (basis Brent) to the end of 2022. Higher prices are needed to fund the Kingdom's diversification away from oil. Russia's goal is to keep prices closer to the marginal cost of the US shale-oil producers, who we characterize as the exemplar of the price-taking cohort outside OPEC 2.0, which produces whatever the market allows. This range is ~ $50-$55/bbl. The sweet spot that accommodates these divergent goals is on either side of $65/bbl for this year. OPEC 2.0 June 1 Meeting Will Maintain Status Quo With Brent trading close to $70/bbl, discussions in the run-up to OPEC 2.0's June 1 meeting likely are focused on the necessity to increase the 2.1mm b/d being returned to the market over the May-July period. At present, we do not believe this will be necessary: Iran likely will be returning to the market beginning in 3Q21, and will top up its production from ~ 2.4mm b/d in April to ~ 3.85mm b/d by year-end, in our estimation. Any volumes returned to the market by core OPEC 2.0 in excess of what's already been agreed going into the June 1 meeting likely will come out of storage on an as-needed basis. Libya will likely lift its current production of ~ 1.3mm b/d close to 1.5mm b/d by year end as well. We are expecting the price-taking cohort ex-OPEC 2.0 to increase production from 53.78mm b/d in April to 53.86mm b/d in December, led by a 860k b/d increase in US output, which will take average Lower 48 output in the US (ex-GOM) to 9.15mm b/d by the end of this year (Chart 3). When we model shale output, our expectation is driven by the level of prompt WTI prices and the shape of the forward curve. The backwardation in the WTI forward curve will limit hedged revenues at the margin, which will limit the volume growth of the marginal producer. We expect global production to slowly increase next year, and the year after that, with supply averaging 101.07mm b/d in 2022 and 103mm b/d in 2023.  Chart 3US Crude Output Recovers, Then Tapers in 2023 Demand Should Lift, But Uncertainties Persist We expect the slowdown in realized DM demand to reverse in 2H21, and for oil demand to continue to recover in 2H21 as the US and EU re-open and travel picks up. This can be seen in our expectation for DM demand, which we proxy with OECD oil consumption (Chart 4). EM demand – proxied by non-OECD oil consumption – is expected to revive over 2022-23 as vaccine distribution globally picks up. As a result, demand growth shifts to EM, while DM levels off. China's refinery throughput in April came within 100k b/d of the record 14.2mm b/d posted in November 2020 (Chart 5). The marginal draw in April stockpiles could also signify that as crude prices have risen higher, the world’s largest oil importer may have hit the brakes on bringing oil in. In the chart, oil stored or drawn is calculated as the difference between what is imported and produced with what is processed in refineries. With refinery maintenance in high gear until the end of this month, we expect product-stock draws to remain strong on the back of domestic and export demand. This will draw inventories while maintenance continues. Chart 4EM Demand Will Recovery Accelerates in 2022-23 Chart 8China Refinery Runs Remain Strong COVID-19-induced demand destruction remains a persistent risk, particularly in India, Brazil and Japan. This is visible in the continued shortfall in realized demand vs our expectation so far this year. We lowered our 2021 oil demand growth estimate to ~ 5.3mm b/d y/y, which is down ~ 800k from last month's estimate, given persistent weakness in realized consumption. Our demand forecast for 2022 and 2023 is higher, however, based on our expectation for stronger GDP growth in EM economies, following the DM's outperformance this year, on the back of wider global vaccine distribution year-end (Table 1). Table 1BCA Global Oil Supply - Demand Balances (MMb/d, Base Case Balances) Our supply-demand estimates continue to point to a balanced market this year and into 2022-23 (Chart 6). Given our expectation OPEC 2.0's production-management strategy will remain effective, we expect inventories to continue to draw (Chart 7). Chart 6Markets Remained Balanced Chart 7Inventories Continue To Draw CAPEX Cuts Bite In 2023 In 2023, we are expecting Brent to end the year closer to $80/bbl than not, which will put prices outside the current range we believe OPEC 2.0 is managing its production around (Chart 8). We have noted in the past continued weakness in capex over the 2015-2022 period threatens to leave the global market exposed to higher prices (Chart 9). Over time, a reluctance to invest in oil and gas exploration and production prices in 2024 and beyond could begin to take off as demand – which does not have to grow more than 1% p.a. – continues to expand and supply remains flat or declines. Chart 8By 2023 Brent Trades to /bbl Chart 9Low Capex Likely Results In Higher Prices After 2023 Bottom Line: We are raising our 2021 forecast back to an average of $63/bbl, and our forecasts for 2022 and 2023 to $75 and $78/bbl. We expect DM demand to lead the recovery this year, and for EM to take over next year, and resume its role as the growth engine for oil demand. Longer term, parsimonious capex allocations likely result in tighter supply meeting slowly growing demand. At present, markets appear to be placing a large bet on the buildout of renewable electricity generation and electric vehicles (EVs). If this does not occur along the trajectory of rapid expansion apparently being priced by markets – i.e., the demand for oil continues to expand, however slowly – oil prices likely would push through $80/bbl in 2024 and beyond.   Robert P. Ryan Chief Commodity & Energy Strategist rryan@bcaresearch.com Ashwin Shyam Research Associate Commodity & Energy Strategy ashwin.shyam@bcaresearch.com   Commodities Round-Up Energy: Bullish The Colonial Pipeline outage pushed average retail gasoline prices in the US to $3.03/gal earlier this week, according to the EIA. This was the highest level for regular-grade gasoline in the US since 27 October 2014. According to reuters.com, the cyberattack that shut down the 5,500-mile pipeline was the most disruptive on record, shutting down thousands of retail service stations in the US southeast. Millions of barrels of refined products – gasoline, diesel and jet fuel – were unable to flow between the US Gulf and the NY Harbor because of the attack, which was launched 7 May 2021 (Chart 10). While most of the system is up and running, problems with the pipeline's scheduling system earlier this week prevented a return to full operation. Base Metals: Bullish Spot copper prices remained on either side of $4.55/lb (~ $10,000/MT) by mid-week following a dip from the $4.80/lb level (Chart 11). We remain bullish copper, particularly as political risk in Chile rises going into a constitutional convention. According to press reports, the country's constitution will be re-written, a process that likely will pave the way for higher taxes and royalties on copper producers.1 In addition, unions in BHP mines rejected a proposed labor agreement, with close to 100% of members voting to strike. In Peru, a socialist presidential candidate is campaigning on a platform to raise taxes and royalties. Precious Metals: Bullish According to the World Platinum Investment Council, platinum is expected to run a deficit for the third consecutive year in 2021, which will amount to 158k oz, on the back of strong demand. Refined production is projected to increase this year, with South Africa driving this growth as mines return to full operational capacity after COVID-19 related shutdowns. Automotive demand is leading the charge in higher metal consumption, as car makers switch out more expensive palladium for platinum to make autocatalysts in internal-combustion vehicles. Ags/Softs: Neutral Corn prices continued to be better-offered following last week's WASDE report, which contained the department's first look at the 2021-22 crop year. Corn production is expected to be up close to 6% over the 2020-21 crop year, at just under 15 billion bushels. On the week, corn prices are down ~ 15.3%. Chart 10 Chart 11     Footnotes 1     Please see Copper price rises as Chile fuels long-term supply concerns published 18 May 2021 by mining.com. Investment Views and Themes Strategic Recommendations Tactical Trades Commodity Prices and Plays Reference Table Trades Closed in 2021 Summary of Closed Trades

Highlights US growth has likely peaked. Economic momentum will slow over the coming quarters as the tailwind from stimulus fades and the vaccination campaign winds down. Historically, a slowdown in US growth, as proxied by a decline in the ISM manufacturing index, has been associated with lower overall equity returns, the outperformance of defensive stocks over cyclicals, large caps over small caps, and US equities over their overseas peers. A falling ISM has also been associated with a strengthening dollar, lower Treasury yields, wider credit spreads, a decline in the US Treasury/German bund spreads, falling oil prices, and an increase in the gold-to-copper price ratio. Compared to past episodes, there are three reasons to expect the coming US slowdown to be relatively benign: First, growth is slowing from exceptionally strong levels; second, growth in many other parts of the world is still speeding up; and third, monetary policy will remain highly accommodative in the face of what is likely to be a transitory increase in inflation. We continue to maintain a positive 12-month view on global equities. Nevertheless, with global growth momentum likely to slow later this year, investors who are maximally overweight risk should pare back cyclical exposure. Crypto update: We warned that “Bitcoin is on a collision course with ESG” two weeks ago. Elon Musk’s flip-flop on allowing customers to pay for Teslas in Bitcoin is yet another piece of evidence that ESG concerns will win out. With that in mind, we are going short Bitcoin. Beware The Second Derivative US growth has likely peaked. Economic momentum will slow over the coming quarters as the tailwind from fiscal stimulus fades and the vaccination campaign winds down. According to the Brookings Institution, fiscal easing contributed nearly seven percentage points to US growth in the first quarter (Chart 1). However, fiscal policy is set to detract from growth in the remainder of the year, reflecting the one-off nature of some of the stimulus measures. Chart 1After A Strong Boost, Fiscal Thrust Is Turning Negative On the pandemic front, the number of new cases continues to trend lower in the US, thanks mainly to a successful vaccination campaign. A falling infection rate has allowed states to dismantle lockdown measures. Conceptually, it is the change in social distancing measures that correlates with economic growth. While some restrictions remain in place (especially in the educational sector), we are now well past the point of maximum loosening. How have financial markets performed during episodes of slowing US economic growth? To answer this question, we looked at the performance of various assets during periods when the ISM manufacturing index was falling and when it was rising. To add a bit more granularity to the analysis, we also looked at cases when the ISM was trending up and above 50, trending down and above 50, trending down and below 50, and trending up and below 50. As summarized in Table 1 and the Appendix Charts, the key results are as follows: Stocks tend to do best when the ISM is rising. Since 1950, the S&P 500 has risen on average by 1.51% during months when the ISM was trending higher, compared to 0.49% during months when the ISM was trending lower. The results were virtually the same if one restricts the sample to the post-1995 period. While the change in the ISM generally matters more for the S&P 500, absolute levels matter too. Since 1995, the best period for the S&P 500 was when the ISM was below 50 but trending higher (S&P 500 up 2.07%), while the worst period was when the ISM was below 50 and trending lower (S&P 500 up 0.03%). This suggests that swings in the ISM have a bigger effect on the stock market during periods of economic contraction. During periods where the ISM was falling but still above 50, the S&P 500 has delivered a positive – though far from stellar – monthly return of 0.69%. US defensively-geared equities outperformed cyclicals when the ISM was trending lower. During periods when the ISM was falling but still above 50, defensives beat cyclicals by 0.45%. Defensives outperformed cyclicals by 0.84% during periods when the ISM was below 50 and trending lower. US small caps underperformed large caps during periods when the ISM was falling. Non-US stocks also underperformed their US counterparts in a falling ISM environment. The relationship between the ISM and value/growth performance is more ambiguous. To the extent that there is one, value generally outperforms growth when the ISM is below 50. Treasury yields tend to increase, while the yield curve tends to steepen, when the ISM is trending higher. Reflecting the higher beta that Treasuries have to the global business cycle, Treasury yields generally rise more than Germany bund yields when the ISM is on the upswing. Corporate credit spreads tend to widen when the ISM is falling. Spreads narrow the most when the ISM is below 50 but rising. As a countercyclical currency, the US dollar tends to weaken when the ISM is rising and strengthen when the ISM is falling. The prices of cyclically-sensitive commodities such as oil and copper normally decline when the ISM is trending lower, although in general, the bulk of the decline in commodity prices usually occurs only when the ISM has dipped below 50. There is not much of a relationship between gold prices and the ISM. Table 1The Economic Cycle And Financial Assets Implications For Today Assuming that the ISM has peaked but remains above 50, the analysis above suggests that the S&P 500 will rise modestly over the coming months; US stocks will edge out non-US stocks; defensives will outperform cyclicals; and large caps will perform slightly better than small caps. The analysis also suggests that Treasury yields will move lower; the Treasury-bund spread will narrow; corporate credit spreads will be flat-to-wider; the dollar will strengthen modestly; and commodities will move broadly sideways. Our own 12-month view is more pro-risk than implied by the ISM analysis. There are three reasons for this: First, US growth is slowing from exceptionally strong levels; second, growth in many other parts of the world is still accelerating; and third, monetary policy remains highly accommodative. Let’s examine each assumption in turn. Reason #1: US growth is slowing from exceptionally strong levels While payroll growth surprised sharply on the downside in April, we suspect this was mainly due to pandemic-induced distortions to the seasonal adjustment mechanism used by the Bureau of Labor Statistics. Seasonally unadjusted payrolls rose by 1.1 million in April, which is broadly consistent with the strong pace of GDP growth tracking estimates. The Atlanta Fed GDPNow model points to growth of 11% in Q2. Bloomberg consensus estimates have US real GDP rising by 8.1% in the second quarter. Growth will decline to 7% in Q3 and 4.7% in Q4, but still average 4% in 2022 (Table 2). Table 2Growth Is Peaking, But At A Very High Level Chart 2Firms Will Need To Rebuild Inventories US households were sitting on $2.2 trillion in excess savings as of the end of April. This is money they would not have had in absence of the pandemic. Slightly less than half of that stockpile can be attributed to transfer payments, mainly in the form of stimulus checks and unemployment benefits. The rest stems from decreased spending during the pandemic. Not all of this money will be spent immediately. However, given the large sums involved – $2.2 trillion is equivalent to 15% of annual personal consumption – even a partial depletion of these excess savings will be enough to power consumption for the foreseeable future. Meanwhile, firms will have to boost production in order to restore depleted inventories. The inventory-to-sales ratio stands at record low levels (Chart 2). The decline in inventories pushed up the ISM new orders-to-inventory ratio in April, even as the overall ISM index slid from 64.7 in March to 60.7. The new orders-to-inventory ratio tends to lead the ISM index, which suggests that any decline in the ISM index over the coming months will be gradual.    An easing of supply-side constraints should also support growth. Even though overall employment was still 5.2% below pre-pandemic levels in April, a record share of small firms surveyed by the NFIB reported difficulty in filling vacant positions (Chart 3). Enhanced unemployment benefits have eroded the incentive to find work. In addition, many schools remain partially shuttered. Chart 4 shows that mothers with young children have seen a much larger decline in labor force participation than other groups. Chart 3Firms Are Struggling To Find Workers Chart 4Mothers With Children Had To Leave The Labor Force Enhanced unemployment benefits will expire in September. As schools resume normal operations, more workers will flow back into the labor market. At the same time, some of the bottlenecks currently gripping the global supply chain should abate, allowing for increased output.   Reason #2: Growth in many other parts of the world is still accelerating Chart 5Over 40% Of S&P 500 Revenues Come From Abroad Chart 6Euro Area Data Has Surprised On The Upside S&P 500 constituent firms derive 43% of their revenues from abroad (Chart 5). While Bloomberg estimates suggest that US growth will peak in the second quarter, growth in the euro area is not expected to peak until the third quarter. Mathieu Savary, who heads BCA’s European Investment Strategy service, sees upside risks to European growth estimates for the second half of this year. Consistent with Mathieu’s observations, recent economic data has been surprising to the upside in the euro area (Chart 6). Just this week, economic expectations for both Germany and the wider euro area leaped to the highest level in more than 20 years, according to the ZEW economic research institute. Growth in Japan should also pick up in the remainder of the year. Japan’s vaccination campaign has gotten off to a very slow start, with less than 3% of the population being inoculated to date. The government imposed its third state of emergency on April 25 in response to rising viral case counts. It subsequently extended those restrictions on May 11. The authorities intend to vaccinate the country’s 36 million elderly people by July, when the Olympics are set to begin. This should permit some easing in lockdown measures. Investors are worried that the Chinese economy will slow this year. The Chinese PMIs peaked in November 2020, about the same time as the combined credit/fiscal impulse reached an apex (Chart 7). Jing Sima, BCA’s chief China strategist, expects the general government budget deficit to remain at a still-ample 8% of GDP this year, similar to where it was last year. She expects credit growth to slow by 2%-to-3%, converging towards the pace of nominal GDP growth. Keep in mind that China’s credit-to-GDP ratio stands at 270%. Thus, if credit grows in line with nominal GDP growth of about 10%, this would still leave the stock of credit roughly 27% of GDP higher at the end of 2021 compared to the end of 2020. This hardly constitutes “deleveraging”. A resilient Chinese economy should buoy other emerging markets. Progress on the pandemic front should also help. The UN estimates that as many as 15 billion vaccine doses could be produced by the second half of 2021, enough to inoculate most of the world’s population (Chart 8). The shortages of vaccines in emerging markets could turn into a surfeit by the end of this year, something that market participants do not seem to fully appreciate. Chart 7China: Peak Stimulus And Peak Growth The rotation in growth momentum from the US to the rest of the world should put downward pressure on the US dollar. A weaker dollar, in turn, has usually coincided with the outperformance of non-US stock markets (Chart 9). Chart 8Vaccine Production Set To Ramp Up Further Chart 9A Weaker Dollar Has Coincided With The Outperformance Of Non-US Stock Markets   Reason #3: Monetary policy remains highly accommodative The slowdown in US growth is coming at a time when inflation is rising. The core CPI increased by 0.9% month-over-month in April. This was the biggest monthly jump since August 1981. The year-over-year rate climbed to 3.0%, the highest in 25 years. The “whiff of stagflation” helped push the S&P 500 down this week. As we discussed last week, we are very much in the camp that expects inflation to rise significantly over the long haul. Over the next one or two years, however, we would fade inflationary fears. As the example of the 1960s illustrates, a long period of overheating is often necessary to push up inflation in a sustained manner. The US unemployment rate reached its full employment level in 1962. However, it was not until 1966 – when the unemployment rate was two full percentage points below equilibrium – that inflation finally took off (Chart 10). The official core CPI likely overstates underlying inflationary pressures. The pandemic threw all sorts of prices out of whack. Stripping out volatile food and energy prices from inflation is not enough. One needs more refined measures of inflation. Luckily, they exist. Chart 11 shows that median CPI, trimmed-mean CPI, and sticky price CPI all remain well contained. Similarly, relatively clean measures of wage growth, such as the Atlanta Fed Wage Tracker, do not point to an imminent wage-price spiral (Chart 12). Chart 10Inflation Started Accelerating Quickly Only When Unemployment Reached Very Low Levels In The 1960s Chart 11Cleaner Measures Of Inflation Are Telling A Different Story Chart 12Wage Growth Is Still Lackluster All this means that the Fed can afford to sustain exceptionally easy monetary policy. That should keep growth at an above-trend pace and continue to support to equity valuations.   Investment Conclusions My “golden rule” for investing is to stay bullish on stocks unless one thinks there is a recession around the corner (Chart 13). Seeing around the corner is not easy, of course, but it is not impossible either. Chart 13Recessions And Bear Markets Tend To Overlap Last year’s recession was caused by a true exogenous shock – the pandemic. Most recessions are endogenous in nature, however. They result from growing imbalances that are usually laid bare by tighter monetary policy. One can debate the extent to which the global economy is plagued by imbalances of one form or another. But one thing is clear, monetary policy is unlikely to turn contractionary any time soon. In this environment, one should remain positive on equities and other risk assets over a 12-month horizon. Nevertheless, with global growth momentum likely to slow later this year, investors who are maximally overweight risk should pare back cyclical exposure. Go Short Bitcoin We warned that “Bitcoin is on a collision course with ESG” two weeks ago in a report entitled “How To Short Bitcoin, Or Anything Else, Without Losing Your Shorts.” Elon Musk’s flip-flop on allowing Tesla customers to pay for Teslas in Bitcoin is yet another piece of evidence that ESG concerns will win out. News that Colonial Pipeline paid hackers 75 bitcoin (nearly $5 million) in ransom further cements Bitcoin’s status as the currency of choice for criminals around the world. With all that in mind, we are going short Bitcoin as of midnight Eastern Daylight Time (EDT) using the shorting technique described in that report. The technique flips the usual risk-reward from shorting on its head. Normally, when you short a stock, your gain is capped at 100% of the initial position whereas your potential loss is unlimited. With our shorting technique, your potential loss is capped at 100% while your potential gain is unlimited. This makes shorting as an investment strategy a lot safer. APPENDIX The Economic Cycle And Financial Assets APPENDIX CHART 1A APPENDIX CHART 1B Appendix Chart 1C Appendix Chart 1D Peter Berezin Chief Global Strategist pberezin@bcaresearch.com Global Investment Strategy View Matrix Special Trade Recommendations Current MacroQuant Model Scores