Precious Metals

Over our main 6-12 month investment horizon, valuation has very little ability to predict asset returns. Over longer periods, however, valuation is highly relevant and should be an important part of strategic asset allocation decisions. Our review of the valuation condition of major asset classes suggests that ex-US stocks and ex-US developed market currencies are the primary assets that could reasonably be considered “cheap” today. The value of long-maturity US Treasurys has improved considerably compared with the average of the past decade, when they were significantly overvalued. 10-year US Treasurys are fairly valued today, but not cheap. The US dollar is extremely overvalued to a degree that has only been associated with subsequent structural declines. The yen and European currencies are the most undervalued relative to the dollar, with commodity currencies (outside of NOK) still cheap but less so. We find that no major commodity groups are cheap, although oil is fairly valued. Gold’s overvaluation is extreme. Based on our expectation that the US will enter a recession at some point over the coming year, we are cyclically positioned against the message from our valuation indicators in favoring US versus global ex-US equities, the US dollar, and growth versus value stocks. We are likely to recommend reducing exposure to these assets earlier than would otherwise be the case once recessionary dynamics take hold in financial markets. Chart II-1Valuation Does A Good Job Of Predicting Long-Term Returns August 2023 August 2023 BCA Research’s primary objective is to provide investors with top-down investment strategy recommendations aimed at improving portfolio performance on a 6-12 month time horizon. Over that time horizon, valuation has been shown to have nearly zero ability to predict asset returns. However, ten-year regressions relating current equity valuations to future ten-year compound returns tend to be highly predictive (Chart II-1), even though these relationships tend to be influenced by the fact that the global financial crisis occurred roughly 10-years after the equity market bubble of the late-1990s. Thus, over our primary investment horizon, valuation is not an especially important investment tool. Over a structural time horizon, however, valuation is important, and should strongly inform investors’ strategic asset allocation decisions. In this Special Report, we discuss some simple approaches that we use when valuing the major asset classes that we cover, and provide some conclusions about where investors can find value today in global financial markets. We conclude that global ex-US equities and ex-US developed market currencies are the main assets that can be considered “cheap” today. US stocks are meaningfully overvalued, and this overvaluation goes beyond the region’s heavy overweight toward broadly-defined technology stocks. Some of the US equity premium is warranted due to the US’ large ROE advantage over other countries, but the pricing of US stocks is extreme even after accounting for this effect. Chart II-2Valuation Today Implies Poor Future Returns From US Equities Valuation Today Implies Poor Future Returns From US Equities Valuation Today Implies Poor Future Returns From US Equities If global ex-US outperformance does occur over the coming decade due to equity multiple mean-reversion, it is an open question whether it will occur due to active outperformance of global ex-US stocks or simply due to the stagnation of the US equity market. Chart II-2 highlights that today’s 12-month forward P/E ratio would imply a 10-year future total nominal return from US stocks of just 4.2%. That is roughly in line with the total return of global ex-US stocks over the past five years, and slightly below the average of the past ten years. We previously had more confidence in poor US equity performance due to disappointing earnings growth for broadly-defined technology stocks, but the potential impact of AI models – whether they can be effectively commercialized and whether they can escape regulatory shackles – has significantly raised the uncertainty over this question. In contrast to what has been a long period of overvaluation for US long-maturity government bonds, we find that 10-year Treasurys are fairly valued today (but not cheap). Still, in contrast to expensive global stocks (driven entirely by US overvaluation), investors should clearly be overweight US government bonds. Within the currency and commodity space, we find the US dollar to be extremely overvalued to a degree that has only been associated with subsequent structural declines. The yen and European currencies are the most undervalued relative to the dollar, with commodity currencies (outside of NOK) still cheap but less so. We do not find any commodities to be cheap, although oil is fairly valued. Industrial metals and gold both show up as expensive according to our approach. Gold’s overvaluation is extreme. One wildcard for industrial metals will be the net impact of the interplay between a likely decline in China’s base metals demand, a structural rise in developed market demand due to decarbonization policies, and seemingly tight supply. It is possible that industrial metals are richly priced due to expectations of a price squeeze, making the timing and magnitude of any structural decline in Chinese base metals demand an extremely important factor for investors to monitor. Cyclically, we are positioned consistent with the message from valuation on our global asset allocation recommendation that investors should be underweight stocks, our recommendation that investors should be long fixed-income portfolio duration, and our view that investors should avoid cyclically-sensitive commodity exposure. Based on our expectation that the US will enter a recession at some point over the coming year, we are positioned against the message from our valuation indicators in favoring US versus global ex-US equities, the US dollar, and growth versus value stocks. In each of these cases, we agree that reversion to more favorable valuation is possible or likely following the next US recession, and we are likely to recommend reducing exposure to these positions earlier than would otherwise be the case once recessionary dynamics take hold in financial markets. Jonathan LaBerge, CFA Vice President The Bank Credit Analyst Equities One of the first ways that we examine the valuation of global equities is by assessing the valuation of the US equity market. The US market accounts for over 60% of global equity market capitalization, and has been an important driver of the global market over the past decade. We publish our suite of US equity indicators in Chart III-1 of each month’s report, which includes our BCA US Valuation Indicator. There is no single accepted way to value the equity market. The most common measure is the price-earnings ratio (PER), but even that has complications because there are many different definitions of earnings (e.g. reported, operating, trailing, forward, cyclically-adjusted). Meanwhile, the market can also be valued in terms of balance sheet ratios or yields relative to those from competing assets. The BCA approach has been to combine a variety of different valuation measures into one composite index. In building our valuation indicator, we started by sorting valuation measures into three broad groups: those related to earnings (or earnings proxies), balance sheets, and relative yields. Valuation indicators for each of these sub-groups are then combined to produce the overall BCA Valuation Index. The three sub-groups are given equal weight in the overall index, which has been constructed using data back to 1955. Below we list each of the components included in the index and their justification, in order to understand what is driving the message from the index: Chart II-3Earnings-Based Measures Point To US Equities Being Expensive Earnings-Based Measures Point To US Equities Being Expensive Earnings-Based Measures Point To US Equities Being Expensive The Earnings Group There are five components to the earnings group (Chart II-3): The S&P PER based on trailing operating earnings: This is a straightforward calculation of the PER using a four-quarter total of operating earnings as published by Standard & Poor’s. The S&P PER based on a 10-year moving average of trailing reported earnings: The idea here is to use a long-run moving average of earnings in order to smooth out cyclical swings. Moreover, we use reported earnings because persistent write-offs should not be ignored. The S&P 500 Growth PER: Investors are paying for both income (dividends) and growth when they invest in the market. The price paid for income can be estimated by dividing dividends per share by the corporate bond yield. Subtracting this value from the S&P 500 index leaves the price paid for growth. As growth must come from retained earnings, the growth PER is calculated by dividing the imputed price paid for growth by retained earnings. A three-year smoothing of dividends and earnings is used in the calculation of this ratio. The S&P 500 price-to-sales ratio: Measures of the PER can be influenced by temporary shifts in margins. A price-to-sales ratio can thus give a different perspective. Stock market capitalization as a percent of GDP: Unless the corporate sector was taking an ever-rising share of the economy, there is no reason why stock market capitalization should increase its share of the economy over time. Thus, major over (or under) shoots in this ratio can highlight shifts in valuation. The Yield Group There are four components in this group (Chart II-4): The real corporate bond yield minus the S&P earnings yield: the earnings yield is a “real” concept and thus should be compared to a real bond yield. We take the average of the AAA and BAA corporate yield and deflate it with our measure of long run inflation expectations. The earnings yield is simply the inverse of the PER based on trailing operating earnings. The real corporate bond yield minus the smoothed S&P earnings yield: The calculation is the same as described above, except that the earnings yield is based on a 10- year average of reported earnings (i.e., the inverse of the second PER listed above in the Earnings Group). The 10-year Treasury yield minus the S&P earnings yield: This is the so-called “Fed Model” and is the inverse of our equity risk premium (except that bond yields here are measured in nominal terms for simplicity). The 3-month Treasury Bill yield minus the dividend yield: This captures competition that stocks face from cash. The Balance Sheet Group There are two components to this group (Chart II-5): Chart II-4Interest Rate Measures Point To Fairly Valued US Equities August 2023 August 2023 Chart II-5Balance Sheet Measures Point To US Equities Being Expensive Balance Sheet Measures Point To US Equities Being Expensive Balance Sheet Measures Point To US Equities Being Expensive Chart II-6Our Composite Valuation Index Shows US Stocks Are Significantly Overvalued Our Composite Valuation Index Shows US Stocks Are Significantly Overvalued Our Composite Valuation Index Shows US Stocks Are Significantly Overvalued The market value of corporate equities relative to corporate net worth at market value: This data comes from the Financial Accounts of the United States (Table B.103) and relates to the non-financial corporate sector. The market value of corporate equities relative to corporate net worth at historical cost: This measure is the same as above, except that the sector’s tangible assets are valued at book cost as opposed to replacement value. Chart II-6 shows the three groups together with the overall Valuation Indicator. The chart makes it clear that the US equity market is significantly overvalued, just not as extremely overvalued as it was during the late-1990s and during the early phase of the pandemic. While the interest rates group is signaling neutral valuation levels and the balance sheet group appears somewhat distorted by the value of equities relative to net worth at historical cost, the overvaluation of US stocks is clearly reflected in the earnings group. Is the apparent overvaluation of US equities justified by superior fundamentals? Chart II-7 shows that the US does deserve to be priced at a premium relative to global ex-US stocks, given that the return on equity of the US market is almost seven percentage points higher than is the case for global ex-US. However, Chart II-7 suggests that US stocks are seemingly 30% overvalued versus global ex-US stocks even considering the relative ROE advantage. Clearly, a good portion of the US equity market’s overvaluation versus its global peers relates to broadly-defined technology stocks. Chart II-8 highlights that global growth stocks are very expensive compared to value stocks, even after accounting for the premium typically paid for the former. The US market is significantly overweight growth / broadly-defined technology stocks, which accounts for an important part of US / global ex-US valuation discrepancy. Chart II-7Some Premium Is Justified For US Stocks, But Better Fundamentals Do Not Explain US Valuation Levels Some Premium Is Justified For US Stocks, But Better Fundamentals Do Not Explain US Valuation Levels Some Premium Is Justified For US Stocks, But Better Fundamentals Do Not Explain US Valuation Levels Chart II-8The US Is Overweight Growth Stocks, And Growth Stocks Are Especially Expensive The US Is Overweight Growth Stocks, And Growth Stocks Are Especially Expensive The US Is Overweight Growth Stocks, And Growth Stocks Are Especially Expensive Chart II-9Almost Every US Equity Sector Is More Expensive Than Its Global Ex-US Counterpart August 2023 August 2023 But Chart II-9 shows that it is not just broadly-defined technology stocks. The chart shows the current 12-month forward P/E ratio for all GICS level 1 sectors for the US and global ex-US MSCI indexes, as well as the current P/B ratio. Broadly-defined technology encompasses stocks that are included in the consumer discretionary, communication services, and information technology sectors, and the chart shows that US stocks in these sectors are comparatively expensive (in addition to their higher weight in the index). US equity sectors are, however, also comparatively expensive in almost every case on both measures, with the exception of US health care stocks based on the 12-month forward P/E ratio. So it is clear that the US market carries a valuation premium, that is augmented by a heavy index skew toward richly-priced tech stocks. For now, it is not clear whether a catalyst exists to cause global ex-US stocks to outperform US stocks. Cyclically-speaking, we expect US stocks to outperform global ex-US because the US dollar is a reliably countercyclical currency, and because growth/tech stocks should disproportionately benefit from falling interest rates. We would expect a period of active global ex-US outperformance following the next US recession, but whether that is likely to be sustained is an open question. As we concluded in our October 2021 report,1 global ex-US outperformance, or the end of its structural underperformance, may instead occur passively via the end of US outperformance. While this could come in response to a structural decline in the dollar (which is also overvalued, see below), it is more likely that a period of poor US equity performance would only occur in response to a structural slowdown in broadly-defined tech sector earnings. We previously had fairly high confidence that this would occur at some point over the coming few years, although today the structural trend in tech sector earnings is likely very linked to whether AI-related software products can be effectively commercialized and whether they will face significant regulatory headwinds. This remains an open question, and should be revisited after the period of cyclical outperformance that we expect from global ex-US stocks following the next US recession. US Government Bonds While there are a variety of methods that investors can use to value equities, there is considerably more agreement about how to (conceptually) value government bonds (especially US Treasurys). All fixed income investors have a choice when buying US government bonds: they can purchase a 3-month Treasury bill and perpetually roll over the position as it matures, or they can purchase a Treasury bond of a longer maturity. This means that yields on longer maturity Treasury bonds simply reflect investor expectations for the average 3-month T-bill rate over the life of the bond, plus some positive risk premium (the “term premium”) to compensate for the inherent uncertainty of the path and tendency of short-term yields. Thus, government bonds are fairly valued when they reflect the right expectations for future short-term interest rates and when the term premium is reasonable. All major estimates of the term premium treat it as the difference between prevailing 10-year yields and an estimate of the future path of short-term rates. Of course, correctly projecting the future path of short-term interest rates is not an easy task, nor is estimating the term premium. It is for this reason that bonds are often valued empirically, or using an equilibrium framework. Chart II-10 presents one empirical approach: our US 10-Year Bond Valuation index. The index is constructed by regressing the nominal 10-year Treasury yield against an average of real short-term interest rates and inflation, and has done a fairly good job in the past of signaling extreme over/undervaluation. One flaw in the model is that it assumes the recent history of the real fed funds rate is what should be expected under normal conditions, which clearly was not true during the last economic cycle. As such, the model suggested that 10-year Treasurys were fairly valued or cheap from 2013-2018 (we disagree). But abstracting from the global financial crisis and its aftermath, the model has done well. Chart II-11 presents our preferred equilibrium approach. In theory, short-term interest rates should be in line with neutral or equilibrium policy rates over long periods of time. And as we discussed at length in a previous report,2 trend GDP growth is by far the dominant driver of the neutral rate. Faster (slower) growth will incentivize firms to expand (reduce) capacity in anticipation of rising (falling) demand, and faster (slower) growth in aggregate income will raise (lower) the sustainable level of interest on borrowed funds for both investment and consumption. This will elevate (reduce) the neutral rate of interest. Chart II-10US 10-Year Treasurys Are Fairly Valued According To Our Bond Valuation Index US 10-Year Treasurys Are Fairly Valued According To Our Bond Valuation Index US 10-Year Treasurys Are Fairly Valued According To Our Bond Valuation Index Chart II-11An Equilibrium Approach Says US Government Bonds Are Not Cheap, But Are Fairly Valued An Equilibrium Approach Says US Government Bonds Are Not Cheap, But Are Fairly Valued An Equilibrium Approach Says US Government Bonds Are Not Cheap, But Are Fairly Valued As such, our equilibrium approach to US government bond valuation focuses on comparing nominal 10-year yields to available estimates of trend growth. Chart II-11 illustrates this approach, and highlights that while US government bonds have offered better value in the past, they are essentially fairly valued today after a long period of overvaluation. What about the term premium? On this front, the perspective is more complicated. Chart II-12 shows that model-based estimates still point to a low term premium, whereas our best available survey-based estimate is pointing to a significantly higher premium. Our approach has typically been to average these estimates, and based on that the term premium is low but nowhere near as low as it was during the depths of the pandemic. Chart II-13 highlights that changes in the expected average policy rate have been far more responsible for the structural decline in US 10-year Treasury yields than the term premium has over the past fifteen years. As such, despite a low reading on our estimate of the term premium, we are more inclined to focus on the neutral valuation readings from our Bond Valuation index and our preferred equilibrium approach. Thus, US long-maturity government bonds do not offer good value, but they are no longer overvalued like they were for many years and thus offer a good alternative to global equities (especially US stocks). Chart II-12The US 10-Year Term Premium Is Not High, But Not As Low As Before The US 10-Year Term Premium Is Not High, But Not As Low As Before The US 10-Year Term Premium Is Not High, But Not As Low As Before Chart II-13The Expected Future Fed Funds Rate Has Been A More Important Driver Of Yields Than The Term Premium The Expected Future Fed Funds Rate Has Been A More Important Driver Of Yields Than The Term Premium The Expected Future Fed Funds Rate Has Been A More Important Driver Of Yields Than The Term Premium       Currencies Chart II-14The US Dollar Is Extremely Expensive According To Our PPP Models The US Dollar Is Extremely Expensive According To Our PPP Models The US Dollar Is Extremely Expensive According To Our PPP Models When valuing currencies, BCA’s Foreign Exchange Strategy service relies heavily on a PPP valuation approach.3 Chart II-14 presents two separate PPP models for the US dollar: our original PPP model and a new version with refined methodology. Our original PPP models were rooted in comparing national CPI baskets with equal weights assigned to the constituents. First, we divided the national CPI indices into 5 subgroups. These included Group A (food, restaurants, and hotels), Group B (shelter), Group C (apparel, health, culture, and recreation), Group D (energy and transportation) and Group E (household goods). Our new model now includes 10 groups, in order to get closer to an apples-to-apples comparison across countries. Second, we ran two regressions with the exchange rate as the dependent variable. The first regression shown in Chart II-14 used the relative price ratios of each subgroup as independent variables. The second regression took a weighted average of the subgroups to form a single “synthetic price ratio.” The weights were determined as the average of the two countries. So, for example, if housing is 38% of US CPI and 18% of European CPI, the synthetic ratio for EUR/USD will assign a 28% weight to housing. This gives as a more accurate cross-sectional comparison of the two national CPIs. Chart II-15 presents another method of valuing the US dollar, which involves regressing the US real effective exchange rate against relative productivity trends and real bond yield differentials. The core point of Charts II-14 and II-15 is that the US dollar is extremely expensive based on either of these approaches, and that similar periods of overvaluation in history have been associated with subsequent structural downtrends in the currency. Chart II-16 presents over/undervaluation estimates for major developed market currencies based on the PPP approach noted above. The chart highlights that US dollar overvaluation is primarily mirrored by undervaluation in the yen and European currencies excluding Sterling. Commodity currencies such as the CAD, AUD, and NZD are the least undervalued against the dollar, although NOK trades at close to a 20% discount to fair value according to our models. Chart II-15The Dollar Is Also Expensive Based On Relative Productivity And Real Yields The Dollar Is Also Expensive Based On Relative Productivity And Real Yields The Dollar Is Also Expensive Based On Relative Productivity And Real Yields Chart II-16The Yen And European Currencies Are Very Cheap According To Our PPP Models August 2023 August 2023 The bottom line for investors is that the US dollar is extremely overvalued and thus vulnerable to a structural decline. We currently recommend that investors position favorably toward USD, but for cyclical reasons that are strongly linked to our expectation of a US recession emerging over the coming 6-12 months. Given how richly priced the dollar is relative to other major currencies, we are highly likely to recommend reducing USD exposure before the end of the next US recession. Commodities It is very difficult for investors to value commodities using traditional approaches because direct commodity holdings provide no cash flow for investors to discount. Many of the commodity “valuation” models offered up to investors actually track deviations in prices from cyclical drivers, which is not truly what investors want to know when they ask whether an asset is fairly priced. For industrial/cyclically-sensitive commodities, one method of measuring fair value is to gauge whether the cost of consuming the commodity is elevated relative to economic activity. For oil, Chart II-17 illustrates our calculation of the “global oil bill”, i.e., global expenditure on oil as a percent of global GDP. Based on this measure, the chart shows that oil is fairly valued. Chart II-18 presents a similar approach for global copper consumption, albeit over a shorter time horizon. Chart II-18 suggests that copper prices are likely overvalued, in the sense that the “global copper bill” is much closer to its 2005-2015 average than is the case for oil. Chart II-17Oil Is Fairly Valued Oil Is Fairly Valued Oil Is Fairly Valued Chart II-18Copper Seems Overvalued Copper Seems Overvalued Copper Seems Overvalued Chart II-19Real Raw Industrial Prices Are Too High Real Raw Industrial Prices Are Too High Real Raw Industrial Prices Are Too High Another potential method for valuing commodity prices is to examine the long-term trend in real commodity prices and whether current real prices are elevated or depressed relative to that trend. In the case of industrial commodities, we would expect that the trend in real prices is either flat or down over time due to technological improvements in commodity extraction and the relative abundance of most industrial metals in the earth’s crust. Chart II-19 highlights that the very long-term trend in real raw industrials prices is indeed down, and that prices today are overvalued in a way that is consistent with Chart II-18. Under normal circumstances, the combination of seemingly overvalued industrial metals prices and a clear shift in China’s attitude toward its metals-intensive property sector would lead to a very negative structural outlook for base metals prices. However, one complication is that decarbonization efforts globally will clearly require a significant increase in industrial metals consumption from developed economies, and there have been several warnings from major industry players that key metals supply is tight and could threaten the transition to green energy. Thus, it is possible that elevated real industrial metals prices today reflect expected future demand / scarcity from the coming shift away from fossil fuels. Finally, turning to gold, a similar analysis of the long-term trend in the real price of gold is the only reasonable valuation method that we feel comfortable using given that gold provides no cash flow and has very little industrial demand as a share of total demand. Unlike real industrial metals, we agree that gold should act as a hedge against inflation over the long term, so we would expect the real price of gold to at least trend sideways over time. Chart II-20Gold Is Extremely Overvalued Gold Is Extremely Overvalued Gold Is Extremely Overvalued Chart II-20 highlights that this has indeed been the case since the 1970s, and that gold is extremely expensive based on this valuation approach. Somewhat similar to the case of industrial metals, one could potentially point to future sources of demand for gold as a basis for currently expensive gold prices. First, while widespread concern about the US fiscal position has existed for many years, it is possible that the US government’s interest burden will break above its early-1990s levels before the end of this decade. Given that a fiscal crisis in the US would very likely be dollar-negative, this could potentially justify elevated gold prices today. Second, following Russia’s invasion of Ukraine and the sanctions levied in response, China has significantly increased its gold holdings in an attempt to “de-dollarize”. We are skeptical that China can continue to do this over the longer term, but it is a possibility – especially considering China’s ambitions toward Taiwan. The bottom line for investors is that none of the major commodities classes are cheap in absolute terms, but our valuation indicators point to the order of least expensive to most expensive being: oil, then industrial metals, and then precious metals. Footnotes 1 Please see The Bank Credit Analyst "The “Invincible” US Equity Market: The Longer-Term Outlook For US Stocks In Relative And Absolute Terms," dated September 30, 2021, available at bca.bcaresearch.com 2  Please see The Bank Credit Analyst "The Fed’s Low Neutral Rate View: Fact Or Folly?" dated February 23, 2023, available at bca.bcaresearch.com 3    Please see Foreign Exchange Strategy "Currency Valuation And Long-Term Returns: Part I," dated January 20, 2023, available at fes.bcaresearch.com

In this short weekly report, we review some of the most common questions clients asked us in the last few weeks.

Markets continue to be tossed to and fro by central-bank policy, and risks of higher commodity prices. These are due to fiscal stimulus and exogenous weather and war-related risk, which could send food and energy prices higher this winter. We remain long gold outright, energy and metals producers via the XOP, XME and PICK ETFs, direct commodity exposure via the COMT ETF, and futures exposure to backwardation in copper (long 4Q23 copper futures vs. short 4Q24 copper futures).

Recession is on track to start around year-end. Stocks usually peak shortly before recession begins. So, position defensively but be prepared for a few more months of the rally.

We build a four-stage business cycle framework based on economic growth and capacity utilization, and then analyze historical returns for most major asset allocation decisions for each stage. Given that we are in the early recession stage (negative growth coupled and an overheated economy), our framework recommends a defensive positioning across all asset classes.

What’s going on? The market-weighted stock market is up. But the equally-weighted stock market is not up. Neither is credit. Neither are industrial metal prices. Neither is the oil price, despite two waves of OPEC output cuts. We explain the dichotomy. Plus: European basic resources stocks can rebound, but Netherlands is likely to reverse.

The US stock price / bond yield (SBY) correlation shifted into negative territory over the past year, significantly departing from the positive correlation regime of the past two decades. The inflation regime is the primary macro driver of the SBY correlation. History suggests that the inflationary threshold for the SBY correlation is nontrivially above the Fed’s target, suggesting that a shift back to a positive correlation is likely in the lead up to and during the next US recession. Whether the SBY correlation remains positive during the subsequent economic recovery will depend – importantly – on the magnitude of the recessionary effect on inflation, and whether investors will have reason to believe that the responsiveness of inflation to the output or jobs gap has permanently increased. A severe US recession does not appear necessary to bring inflation below the positive/negative SBY correlation threshold, but only so long as long-term inflation expectations remain anchored. Generally-speaking, cross-asset performance has not been unduly impacted by the shift in the US SBY correlation, with the exception of the performance of tech versus ex-tech or growth versus value, precious metals, and the yen. Our conclusions, alongside our economic expectations detailed in Section I of our report, support defensive portfolio positioning over the coming 6-12 months, with long-maturity government bond yields as the diversifying asset. They also imply the outperformance of growth versus value or broadly-defined tech versus ex-tech stocks within a global equity portfolio, and the outperformance of the yen within the currency space. The fact that inflation was far more elevated last year than many investors expected led to negative performance from both stocks and bonds. That was a major departure from the better part of the past two decades, when government bonds provided portfolio protection during risky asset selloffs. Quantitatively, this shift manifested itself as a decline in the US stock price / 10-year government bond yield (SBY) correlation from positive to negative territory. While the SBY correlation has become less negative over the past several months, it has not yet durably reverted into positive territory. In this report, we present our best estimate of the inflationary threshold that results in a positive or negative SBY correlation, and whether inflation is likely to approach this level over the coming one-to-two years. We also examine the performance of major financial assets since the US SBY correlation turned negative last year, to understand the investment implications of a potential shift back to a positive SBY correlation. We conclude that core US inflation does not likely need to return to the Fed’s target in order for the SBY correlation to return to positive territory, which is good news for multi-asset investors. Additionally, some simple estimations of the modern-day Phillips curve suggest that a severe recession is not required to bring inflation back to its target. That that may be increasingly less true, however, the longer it takes for a recession to occur, as it risks the emergence of backward-looking inflation expectations. Cross-asset performance has not been unduly impacted by the shift in the US SBY correlation, with the exception of the performance of tech versus ex-tech or growth versus value, precious metals, and the yen. From an investment strategy perspective, our conclusions support defensive portfolio positioning over the coming 6-12 months, with long-maturity government bond yields as the diversifying asset. They also imply the outperformance of growth versus value or broadly-defined tech versus ex-tech stocks within a global equity portfolio, and the outperformance of the yen within the currency space. Precious metals remain a wildcard, but we are sticking with a low-conviction recommendation to be overweight/bullish previous metals. The Inflationary Threshold Of The SBY Correlation Chart II-1Visualizing The Stock Price / Bond Yield Correlation Visualizing The Stock Price / Bond Yield Correlation Visualizing The Stock Price / Bond Yield Correlation Chart II-1 presents the rolling 1-year correlation between the S&P 500 and US 10-year Treasury yields. The chart illustrates that the SBY correlation was persistently negative from the mid-1960s until the late-1990s, when it shifted durably into positive territory. The SBY correlation was also modestly positive during the mid to late 1960s. Over the past year, the SBY correlation has been at its most negative since it moved durably into positive territory in the late-1990s and early-2000s. It has been a long-held view at BCA that the correlation between US stock prices and long-maturity US government bond yields turned positive in the late-1990s due to a shift in the inflation regime, and, it is thus not surprising to us that well above-target inflation has caused the SBY correlation to move back below the zero line. For example, our US Investment Strategy service presented a detailed analysis of the macro drivers of the SBY correlation in a 2012 special report,1 and more recently our Emerging Market Strategy service warned of a likely shift in the SBY correlation back to negative territory.2 Both reports underscored the importance of the inflation regime as the major driver of the SBY correlation with high (low) inflation associated with a negative (positive) correlation. It is not difficult to rationalize why the inflation regime is the primary driver of the SBY correlation. A negative correlation means that rising bond yields are bad for stocks, which was clearly the case from the late-1960s until the mid-1990s – because rising bond yields during this time reflected the Fed’s attempt to rein in inflation. In the parlance of fixed-income investors, this was the era when the inflation component of bond yields rather than the real component was the dominant driver of yields. This also explains why the correlation was modestly positive in the early-to-mid 1960s, given that core inflation averaged just 1.4% percent during that time. Chart II-2Inflation Expectations Did Not Stabilize Until The Late-1990s Inflation Expectations Did Not Stabilize Until The Late-1990s Inflation Expectations Did Not Stabilize Until The Late-1990s What changed in the late-1990s? We believe two factors were at play. First, while the idea of the “Greenspan put” had been around since the 1987 “Black Monday” stock market crash, the notion that the Fed would be quick to ease monetary policy in response to slowing economic growth was strongly reinforced by its reaction to the LTCM crisis – shown as the vertical line in Chart II-1 – as well as the 2001 recession. Second, while actual core PCE inflation essentially fell back to target levels by the early-to-mid 1990s, measures of long-term inflation expectations, such as the University of Michigan’s median 5-10 year household inflation expectation series, as well as our adaptive expectations model did not truly stabilize until the late-1990s (Chart II-2). That implies that investors were not truly convinced that inflation had reached low and stable levels until that point. Looking forward, the core question for investors is not whether the inflation regime will be the main determinant of the SBY correlation, but rather whether inflation needs to fall all the way back to the Fed’s 2% target in order for the correlation to turn durably positive again. This question is relevant to address not just whether investors can expect portfolio protection from falling long-maturity bond yields during the next recession, but also whether the SBY correlation is likely to be positive or negative on average over the coming few years. Looking at the empirical record, we address this question in two ways. First, we use a logistic regression approach to test for the inflationary threshold, using either headline or core CPI or the PCE deflator, that has the best power to predict when the rolling 1-year SBY correlation has been negative (Chart II-3). And second, using a quarterly SBY correlation (calculated using daily data), we simply calculate how often a negative correlation occurs alongside core inflation above a given value (Chart II-4). The results of Charts II-3 and II-4 are clear, and paint a somewhat positive picture for multi-asset investors over the coming year or two. Chart II-3 highlights that the core inflation regime is a more reliable predictor of a negative SBY correlation than headline, and that core CPI is a more important predictor than core PCE inflation. This likely reflects the fact that, typically, investors more closely follow the CPI data, given that it is released earlier than the PCE data. Chart II-3The Inflationary Threshold For A Negative SBY Correlation Is Higher Than The Fed’s Target… June 2023 June 2023 Chart II-4…And Is Possibly As High As 3.5% June 2023 June 2023 But, more importantly, Chart II-3 shows that the core inflation rate consistent with a negative SBY correlation has historically been higher than 2%, roughly between 2.5-3%. Chart II-4 echoes this point, by showing that negative quarterly SBY correlations have been more often associated with annualized core inflation rates of up to 3.5%. To us, this underscores that a negative SBY correlation does not reflect investor expectations of target inflation, but rather whether inflation is at a sufficiently high level that it becomes the dominant driver of monetary policy decisions. Charts II-3 and II-4 highlight that investors appear to believe that the inflationary threshold for the SBY correlation is nontrivially above the Fed’s target, suggesting that a shift back to a positive correlation is likely in the lead up to and during the next US recession. Whether the SBY correlation remains positive during the subsequent economic recovery will depend – importantly – on the magnitude of the recessionary effect on inflation, and whether investors will have reason to believe that the responsiveness of inflation to the output or jobs gap has permanently increased. Will The SBY Correlation Durably Shift Back Into Positive Territory? Determining the likely impact of the next US recession on inflation and the likely inflationary impulse of the subsequent recovery are analytically challenging exercises. The degree to which core inflation surged over the past two years underscores the inherent difficulty in estimating the magnitude of inflation over a cyclical time horizon, especially when unique circumstances – such as the supply chain impacts of the COVID-19 pandemic and its aftermath – are present. We noted in our January 2021 Special Report that the modern-day version of the Phillips curve expresses actual inflation as a function of expected inflation, economic or labor market slack, and other atypical shocks to prices. For headline inflation, these atypical shocks very often occur from sharp changes in food and energy prices. In core space, price shocks typically stem from changes in imported goods prices and the US dollar, as well as core feed-through effects from large or long-lasting shocks to food and energy prices. Despite the fact that pandemic-related effects are clearly still exerting an impact on core PCE inflation in the US, it is still worthwhile to estimate core inflation using the first two terms of the modern-day Phillips curve to gauge what kind of labor market adjustment may be required to bring inflation back to the Fed’s target. Chart II-5 presents a variety of estimates based on a linear regression of core inflation on different measures of inflation expectations, as well as the unemployment rate gap as defined by the CBO's estimate of NAIRU. We present the results of the regression calculated over two different estimation periods; in all cases, the inflation expectations component is the most significant driver, but the unemployment rate gap shows up as highly significant in all six models. Chart II-5As Long As Inflation Expectations Remain Forward-Looking, Only A Mild Or Average US Recession Will Be Needed To Bring Inflation Back To Target June 2023 June 2023 Chart II-5 indicates that the range of unemployment rate estimates that would be required to return core PCE inflation to the Fed’s 2% target is very large. Still, it highlights a very important point, which is that ostensibly forward-looking measures of inflation expectations (such as those reported by the University of Michigan’s Surveys of Consumers) imply a considerably lower unemployment rate needed to return core inflation back to the Fed's target. This is strongly consistent with academic research showing that the monetary policy response to a rise in inflation must be significantly greater the more that inflation expectations are adaptive (i.e., backward-looking). For investors hoping that long-maturity government bonds will once again provide some form of protection to a balanced portfolio during periods of falling equity prices, Chart II-5 is encouraging. It suggests that because households recognize a large part of the surge in inflation over the past two years was uniquely driven by pandemic related effects, a large rise in the unemployment rate will not be needed to bring inflation back to the Fed’s target. However, Chart II-5 does show that a recession will likely be needed for inflation to return to target given that even the smallest unemployment rates shown imply a recessionary shock to the labor market, unless disinflation or outright deflation from housing and goods inflation is more pronounced over the coming year than we currently expect. It also reinforces a crucial point about why we think a recession is likely over the coming year: the greater weight on inflation expectations from the models shown in the chart underscores that the Fed must prevent household inflation expectations from becoming backward-looking, or else it would require a much more severe recession in order to return to low and stable inflation. Chart II-6Some Concerning Signs Of A Renewed Rise In Household Inflation Expectations Some Concerning Signs Of A Renewed Rise In Household Inflation Expectations Some Concerning Signs Of A Renewed Rise In Household Inflation Expectations Disinflation from energy prices and global supply chain factors have bought the Fed some time in this regard, but Chart II-6 highlights that we have recently seen a concerning rise in both one-year and five-to-ten year expected inflation. To the extent that a recession would simply accelerate the housing and goods-related disinflation that is either already occurring or will soon, the Fed is likely to prefer a recession sooner rather than later – were they to conclude that an economic contraction is unavoidable in order to bring inflation back to 2%. So, while we agree that the SBY correlation is likely to turn positive over the coming 6 to 12 months, this will very likely occur in the context of falling stock prices. To us, the greater weight of inflation expectations in the models shown in Chart II-5 also supports the idea that a shift in the SBY correlation back into positive territory will likely be a durable one if a recession occurs over the coming year, as we expect. We acknowledge that demographic effects are likely to impact the US labor market structurally for years to come, and agree that US inflation may be above target on average over the course of the next economic recovery. However, unless long-term inflation expectations become unmoored, the implication is that US core inflation will probably grow at a 2-3% annual rate during the next economic recovery – which is below the negative SBY correlation threshold that we identified in Charts II-3 and II-4. This reflects our belief that the greatest threat to very elevated structural inflation in the US does not come from the demographic outlook, but rather from a tepid monetary policy response from the Fed to the currently elevated rate of inflation. This tepid response would likely be inadvertent, as we believe the Fed is truly serious about bringing inflation back down to target levels. Rather, it would stem from the Fed’s misguided views about the neutral rate of interest, which we have noted in past reports could cause the Fed to cut interest rates before the job of wringing out excess inflation from the system is complete. We have noted in past reports that monetary policy rules that are commonly-cited by Fed officials could justify a fed funds rate below our estimate of neutral if core PCE inflation falls below 3%, given that the Fed’s view of the neutral rate of interest is meaningfully below ours. For now, this is a plausible but not probable scenario. Were we to see mounting evidence of this scenario, we may shift our 6-12 month investment recommendations more in favor of a pro-risk stance, as interest rate cuts into easy territory would stimulate economic activity and meaningfully push out the onset of a recession. But as noted above, this would also significantly raise the odds of elevated structural inflation, suggesting that the SBY correlation would only briefly turn positive during the next US recession in such a scenario. Asset Performance And A Negative SBY Correlation Given the very large change in the SBY correlation over the past two years, one natural question for investors to ask is whether we have seen highly atypical cross-asset performance owing to this correlation shift. However, the surprising reality is that most of the examples of atypical asset performance since the beginning of 2022 have occurred for reasons other than the decline in the SBY correlation – besides some well-known performance dynamics like the failure of US stocks to outperform in the face of falling global stock prices. When analyzing this question, we use an admittedly quantitative approach, but one that we feel is still accessible and relatively easy for investors to understand. Essentially, we examine the beta of a variety of assets relative to global stocks since the beginning of 2022, and also look at beta-adjusted performance. Assets whose performance has been significantly driven by the shift in the SBY correlation into negative territory should have experienced a sharp change in their beta versus global stocks and beta-adjusted out/underperformance that is logically consistent with the asset’s relationship to inflation. Charts II-7 and II-8 present the results of our approach, and highlight several points. Chart II-7Only A Few Major Asset Classes… June 2023 June 2023 Chart II-8…Have Had Their Performance Materially Affected By The Shift In The US SBY Correlation June 2023 June 2023 First, while it is true that several assets experienced a significant change in their beta to global stock prices since 2022 versus the five years prior to the onset of the COVID-19 pandemic, several of these changes were related specifically to Russia's invasion of Ukraine or China's zero-COVID policy. This is especially true when examining the performance of euro area, UK, and emerging market equities, emerging market sovereign bonds, and oil and base metals prices. Chart II-9Core Services Ex-Shelter Inflation Drove The SBY Correlation Into Negative Territory, And Its Rise Preceded The War In Ukraine Core Services Ex-Shelter Inflation Drove The SBY Correlation Into Negative Territory, And Its Rise Preceded The War In Ukraine Core Services Ex-Shelter Inflation Drove The SBY Correlation Into Negative Territory, And Its Rise Preceded The War In Ukraine China’s decision to pursue strict lockdown and control measures for most of 2022 had nothing to do with above-target inflation in the developed world, so it is clear that the underperformance of China-related assets since the beginning of last year has not been connected to a shifting US SBY correlation. And while it is clearly the case that Russia's invasion of Ukraine and its impact on the global energy market made the global inflation situation worse, it was the rise in core services ex-housing inflation that was the root cause of the shift in the US SBY correlation into negative territory, which was already well underway prior to the war (Chart II-9). To us, Charts II-7 and II-8 highlight just three major asset classes or investing styles whose performance appears to have been materially affected by the shift in the US SBY correlation from positive to negative territory: the performance of tech versus ex-tech or growth versus value, precious metals, and the yen. In the case of tech/growth stocks and the yen, we have seen beta-adjusted underperformance since the beginning of 2022, whereas in the case of precious metals it has been the opposite. The abnormal performance of these three assets does appear to have been driven by the same factors that pushed the US SBY correlation into negative territory: The fact that bond yields were rising in conjunction with falling stock prices had an outsized impact on tech or growth stocks because of the perception by investors that these are comparatively longer duration assets. As we highlighted in Section I in last month’s report, this perception is still in effect given that the relative performance trend of tech versus ex-tech stocks continues to be strongly negatively correlated with long-maturity government bond yields. Given the US equity market’s heavy weight toward growth/tech stocks, it is thus unsurprising that US equities have failed to outperform global stocks since the beginning of 2022. Chart II-10The Yen's Decline Has Been Related To A Shift In The US SBY Correlation The Yen's Decline Has Been Related To A Shift In The US SBY Correlation The Yen's Decline Has Been Related To A Shift In The US SBY Correlation As the most prominent major country that has historically struggled with below target core inflation, Japanese interest rate differentials collapsed versus other DM economies last year as investors priced in a much more aggressive monetary policy response in the US and euro area than in Japan (Chart II-10). This resulted in significant underperformance of the yen on a broad trade-weighted basis (panel 2), despite the fact that the yen is typically a risk-off currency. Over the past year, gold has massively outperformed what its historical relationship with real government bond yields and the dollar would have implied (Chart II-11). Part of this outperformance may have been caused by a negative SBY correlation, in the sense that some investors may have flocked to gold as an alternative to stocks and bonds while both were falling in value. It is also likely that gold has benefited from a significant increase in central bank gold reserves (in lieu of US dollars) in the aftermath of Russia’s invasion of Ukraine and the US’ freeze on Russia foreign currency reserves. At the same time, it is also possible that gold is starting to benefit from fears of an eventual fiscal crisis in the US, given that US federal government interest payments are exploding higher as a share of GDP. This explosion in payments is the result of the sharp rise in the fed funds rate over the past year (Chart II-12). While it is true that the government’s interest expense will fall during the next US recession as the Fed cuts the policy rate, the key point is that investors have become more aware of the US government’s eventual interest burden based on normalized interest rates and may have increased their structural allocations to gold over the past year in response. Chart II-11Gold Has Significantly Outperformed What Real Interest Rates And The Dollar Would Have Implied Gold Has Significantly Outperformed What Real Interest Rates And The Dollar Would Have Implied Gold Has Significantly Outperformed What Real Interest Rates And The Dollar Would Have Implied Chart II-12Gold May Be Benefitting From Fears Of An Eventual US Fiscal Crisis Gold May Be Benefitting From Fears Of An Eventual US Fiscal Crisis Gold May Be Benefitting From Fears Of An Eventual US Fiscal Crisis Investment Strategy Conclusions Our analysis highlights several conclusions about the US SBY correlation: The historical evidence suggests that the US SBY correlation will shift back into positive territory once US core inflation falls back or below 3%. A full return to target inflation is not likely required. We expect that US core inflation will fall below 3% at some point over the coming year or two, but it is only likely to occur in the context of a recession. A severe US recession does not appear necessary to bring inflation below the positive/negative SBY correlation threshold, but only so long as long-term inflation expectations remain anchored. Generally-speaking, cross-asset performance has not been unduly impacted by the shift in the US SBY correlation, with the exception of the performance of tech versus ex-tech or growth versus value, precious metals, and the yen. From an investment strategy perspective, our conclusions support defensive portfolio positioning over the coming 6-12 months, with long-maturity government bond yields as the diversifying asset. They also imply the outperformance of growth versus value or broadly-defined tech versus ex-tech stocks within a global equity portfolio, and the outperformance of the yen within the currency space. Precious metals remain somewhat of a wildcard, and we are conflicted about the outlook for gold and silver prices over the coming year. On the one hand, precious metals have significantly outperformed over the past year as the US SBY correlation has fallen into negative territory, implying a potential reversal of performance if the correlation becomes positive again. In addition, the real price of gold is extremely elevated relative to history, suggesting that precious metals are quite expensive. On the other hand, we expect real interest rates to fall meaningfully at some point over the coming 6-12 months as the US economy slips into recession, which has historically been bullish for precious metals. On balance, we are sticking with our recommendation to be overweight/bullish previous metals, but investors should note that this is a relatively low-conviction view. Investors who are heavily overweight precious metals should respond to increasing evidence of 1) an impending US recession and 2) a major technical breakdown in precious metals prices as a sign to reduce their exposure significantly. Finally, even though structurally elevated inflation has recently fallen off investors’ radars as a major source of concern, investors should continue to monitor long-term inflation expectations closely for further signs of a renewed breakout. While we strongly believe that the SBY correlation will turn positive during the next US recession regardless of its severity, a strong breakout in long term inflation expectations would increase the probability of a negative SBY correlation during the next economic recovery. It would also significantly raise the odds of a more severe recession than we expect. That would still justify conservative portfolio positioning, but it would imply meaningfully higher long-maturity government bond yields over the nearer term, and thus would delay the point at which long duration positions would be warranted. As noted in Section I of our report, fixed-income investors would lose money on 10-year Treasury positions over the coming year if yields rise above 4.2%, something that could occur if the odds of the “no landing” scenario increase further over the near-term. We are sticking with our view that investors should wait for meaningful labor market weakness or a rise in 10-year yields above 4% to shift to a long duration stance, but investors should be prepared to extend duration quickly and significantly in response to either of these events. Jonathan LaBerge, CFA Vice President The Bank Credit Analyst   Footnotes 1  Please see “Shifting Sands: Why Are Stocks And Bond Yields Positively Correlated, And When Will This Change?”, US Investment Strategy, March 12, 2012, available at usis.bcaresearch.com 2  Please see “A Paradigm Shift In The Stock-Bond Relationship”, Emerging Markets Strategy, February 25, 2021, available at ems.bcaresearch.com