Market & Economic Commentary

Is the rebound in earnings estimate revisions peaking?

Our measures of aggregated earnings estimate revisions trends have shown some of their most dramatic movements on record this year, and now may be looking extended.

After reaching historically extreme negative readings in April/May amid the initial COVID-19 lockdowns, earnings estimate revisions activity has now lurched back up to extremely positive readings. Better-than-expected Q2 earnings reports and the effects of massive monetary and fiscal stimulus are now finally reflected in analyst earnings forecasts. However, with fiscal stimulus weakening (and little imminent sign of movement toward new stimulus) and no meaningful further scope for interest rate cuts, the “snap-back” in earnings estimate activity could soon drop off.

The first chart below shows our measure of aggregated analyst earnings estimate revision activity in the US, for our broad universe of over 2000 stocks (equally-weighted) on a longer-term time frame. The data are month-end values except for the latest data point.  The red line represents the “breadth” of estimate revisions, meaning the aggregate net proportion of positive versus negative revisions (changes) to forward 12-month earnings estimates over the prior three months (i.e., number of analysts who have raised earnings forecasts minus the number who have reduced forecasts, as a percentage of the total number of analysts for each stock, scale right). The blue bars represent the “magnitude” of the month-on-month changes in forward 12-month forecasts, i.e., the average percentage change in earnings forecasts from a month ago (scale left).

United States_AbsERS

We can see that the low point in April matched (or exceeded) the extremes seen in the 2008 Great Financial Crisis (GFC) period, which is not surprising given that the drop-off in economic activity this year was greater than in the GFC. However, the combined fiscal and monetary stimulus recently produced in response was also greater than any previous post-WWII period, and so revisions metrics have shown a faster and more extreme rebound than at any previous point in our data. Stock prices appear to have moved ahead of aggregate estimate revisions, raising the question of whether this apparent good news for earnings is already priced in.

Perhaps more concerning is the risk that revisions (i.e., analyst sentiment) have reached highly optimistic readings now and may already be starting to revert. The chart below is calculated identically to the one above, but plots the daily figures (rather than monthly) over the last three years. Here we can see that the blue bars are already coming down from their latest peak, suggesting that the upward momentum of earnings estimate revisions may be fading now that Q2 earnings reports are over. The breadth series (red line) is based on revisions over the last three months, so it encompasses a full calendar quarter and is thus more stable. If revisions breadth starts to turn down (as it did after the tax-cut surge in early 2018) alongside current elevated valuations for equities, then the recent signs of higher stock market volatility could persist into Q4.

United States_AbsERS_Daily

Unemployment is still a huge issue

The latest weekly report on unemployment claims was released yesterday and provoked mixed responses depending on how the data was (or was not) analyzed. While unemployment claims data is not a perfect measure of the national job market, it is one of the most timely measures and gives a good picture of what is going on (though long-term comparisons can be difficult). Recent methodology and seasonal adjustment changes along with reporting of state-only (not federal) claims data have caused some confusion among investors.

The chart below shows the current preferred measure of ALL continuing (not initial) unemployment claims over the last 12 months, i.e., including the total of all the newly created or expanded federal unemployment assistance in addition to the regular state-level unemployment benefits that are normally reported. The data are not seasonally adjusted, and are shown on a log scale to better capture percentage changes. The picture is pretty clear: the trend in unemployment has not improved materially since May despite the signs of a rebound in economic activity. The positive side is that more unemployed people than ever before are getting at least some benefits while out of work thanks to the expanded federal support programs started in April (far more so than in the 2008-09 Great Recession). This highlights the economy’s current reliance on federal fiscal support.

Total Unemployment Claims

We can see that before February there were around 2 million people getting unemployment benefits (and fewer than that most of last year), which were all from regular state-level programs, and since May there have been steadily between 27 and 30 million people getting benefits from state and federal programs combined. The latest data (for Aug. 15th) shows the number back to the upper end at 29.2 million.  The US labor force (people over 16 deemed working or looking for work) as of the end of 2019 was about 159 million people. So over 18% of the pre-pandemic labor force is now collecting some kind of unemployment benefits, with no real trend of improvement visible yet.

Since part of the confusion about the weekly claims reports is related to the presence of the new federal programs in addition to the standard state programs, we break down the two categories in the chart below. Many state unemployment benefits normally only last 26 weeks, though there have been extensions in some cases. The major federal programs (Pandemic Unemployment Assistance, PUA, is the largest, with Pandemic Emergency Unemployment Compensation, PEUC, the other major one) are designed to both expand who can receive benefits (self-employed, gig workers, people working to support family at home, etc.) and extend the duration of benefits when state programs run out. As the months have passed, some people originally on state benefits (or who would normally have been ineligible) have switched over to the federal programs. We can see this in the chart, where the state-level data (blue line) has been declining from its peak, but the federal programs (red line) have been rising to largely offset the decline in the state data. Indeed, the total in federal programs has now exceeded those in state programs for the first time. This is why the total for all programs together has shown no material improvement (as in the first chart above), but those watching only the standard state-level data argue that employment conditions are improving.

US Unemployment Claims State vs Federal

While the supplemental benefit of $600/week for many of those getting benefits expired at the end of July, the benefits from federal support programs (PUA) are due to run out at the end of this year unless they are extended. Thus there will continue to be much attention focused on Congress as they debate if and how to provide additional economic support, with much uncertainty especially as the November elections approach.

The S&P 500 is top-heavy, but so are fundamentals

There has been much discussion about the increasing concentration of the market cap weighted indices in the US, with the S&P 500 now showing some of the highest levels of concentration among the largest constituents in history. The top 20 S&P 500 stocks (4% of the constituents) currently comprise 38.6% of the index weight, while the top five companies alone make up 23.8% of the weight.

But what about the underlying fundamentals? Are they as concentrated as the capitalization weightings? Broadly speaking, the answer is yes.

Following are a chart and two tables based on the top 20 stocks in the S&P 500. The chart below shows the proportion of total estimated operating income and sales made up by the top 20 S&P 500 stocks each month over the last 15 years. The figures are based on consensus analyst forecasts for net income and sales for the current fiscal year (currently FY2020), rather than trailing reported financials.

Top 20 SP500 Earn Sales

The first message from the chart is that the top 20 stocks do in fact make up an increasing proportion of income and sales, but the proportion is not historically outlandish. The largest stocks have tended to produce 30-40% of total index earnings and 31-35% of total sales, and are near the upper end of those ranges now.

And possibly more surprising to some, the proportion of earnings expected to be generated this year by the top 20 stocks is very similar to the proportion of market cap they hold (39.1% of earnings as of the end of July, vs 38.5% of market cap). The proportion of sales from the top 20 is similar but slightly lower at 35%. So the most heavily weighted stocks in the index are in fact producing a comparable proportion of earnings and sales in aggregate (not for every stock of course). The S&P 500 has long been “top-heavy” to varying degrees due to the presence of a few dominant mega-cap stocks, but right now that broadly captures the similarly concentrated nature of the underlying fundamentals.

What are the top 20 stocks by earnings and sales? The tables below show the latest monthly lists. While some names naturally appear on both lists, there are some notable differences in the lists of earnings versus sales, as profitability varies significantly in some cases. Amazon.com, for instance, has the second largest sales but is not among the top 20 for earnings. The same is true for a number of other well-known names whose earnings are either depressed right now or whose profitability is structurally lower.

List of Top 20 SP500 Net Inc List of Top 20 SP500 Sales

And finally, for some interesting historical perspective, the tables below show the top 20 S&P 500 stocks by earnings and sales 10 years ago (as of July 2010). Notably, the range of estimated profits among the top 20 was narrower 10 years ago, and a number of the names have changed over the last decade.

List of Top 20 SP500 Net Inc 10 yrs Ago List of Top 20 SP500 Sales 10 yrs Ago

Negative sentiment on the dollar becoming extreme

One of the most notable trends in markets recently has been the weakness in the US dollar. After a choppy strengthening trend since early 2018 (as the Fed was tightening policy), the US Dollar Index surged in the immediate aftermath of COVID-19’s arrival in March. Since then, however, as monetary and fiscal stimulus engulfed markets and investor risk appetite returned, the dollar has been weakening versus a number of other currencies. Indeed, the dollar’s weakness has accelerated recently, and the Dollar Index has reached its lowest levels since May 2018.

The latest movement is in large part attributable to strength in the euro, which has a big (58%) weight in the Dollar Index among the six major developed market currencies in the index. However, the recent decline can also be seen to a lesser degree in the Fed’s Broad Dollar index that encompasses a wider range of currencies. Certainly the prospect of US interest rates across most of the yield curve being near zero for the foreseeable future makes the dollar less attractive now versus other currencies.

However, we are now seeing signs that bearishness towards the US dollar has become extreme, and such sentiment is likely correlated with broader trends in investor risk appetite.

When reviewing the data on sentiment toward the dollar, one key metric that has jumped out at us is shown in the chart below. It plots an average of the proportion of market commentators (strategists, newsletter writers, etc.) who are bullish on the US dollar as reported by Consensus Inc. and Market Vane (services which track investor sentiment). We can see that bullishness on the dollar has now plunged to its lowest levels in nearly a decade. And while sentiment can remain bullish or bearish for some time, we can clearly see where the heavy consensus view on the dollar is right now. It is thus more likely that the dollar’s decline is in its later stages rather than a newly developing trend.

Dollar Indices and Sentiment

Since the US dollar behaves as a “risk-off” currency (along with the Japanese yen and Swiss franc), selling the dollar is often a sign of risk-on behavior, and consistent with recent risk-on activity in equity and debt markets. Bullish sentiment toward stocks and bonds is also back to pre-COVID levels, so the dollar sentiment readings are consistent with those trends as well. With monetary and fiscal stimulus in the US fading now, the surge in the supply of US dollars is also likely to ease, and the US economy and inflation are likely to slow without additional stimulus. A reversal of the current lopsided sentiment trends would potentially be bullish for the US dollar and bearish for equities, though the precise timing of any such change is always a challenge.

Going over the fiscal cliff

As we discussed in an earlier blog post (and in our client research), much depends on the path of monetary and fiscal policy, particularly the fiscal stimulus programs put in place in response to the COVID-19 crisis in March and April.

Back in March, hopes were high that by the end of July, the trends in COVID-19 would look better and there would be less need for such aggressive fiscal support. Sadly, that has not been the case, and the markets have clearly come to expect a new fiscal package to maintain support for the still shaky US economy. The concern is that a “fiscal cliff”, i.e., a sudden drop-off in government support before the economy has regained self-sustaining momentum, will lead to a renewed bout of heavy economic weakness.

Unfortunately for the millions of unemployed people who have been relying on a combination of state and federal benefits to get by, the fiscal cliff has already begun. The US Congress has allowed much of the fiscal support provided in the initial rounds of stimulus legislation to expire (or has been used up) and has not yet agreed on what (if anything) will replace it.

US Treasury Unemployment Benefit Expenditures

One of the key supports for the unemployed has been the additional federal unemployment benefit of $600/week (Federal Pandemic Unemployment Compensation, FPUC) that was being added to the regular state unemployment benefits (which average less than $400/week). The Pandemic Unemployment Assistance (PUA) program also made benefits available to those who would not normally be able to claim unemployment (self-employed, gig workers, those caring for a family member, etc.). There were also provisions that provided federal funding for extended unemployment benefits to those who would have exhausted the normal duration of benefits (often 26 weeks) to a total of 52 weeks.

The provision of the extra $600 expired on July 25th and has so far not been replaced (Trump’s recent memoranda on the topic notwithstanding), while other provisions will expire at the end of this year. Congress and the White House continue to debate whether to extend the federal unemployment support and at what level, among many other things. Until they come to an agreement and pass new legislation, the additional federal unemployment benefits will remain unavailable. And even after a bill is signed, state unemployment offices will need time to make changes to their (mostly old and creaky) systems before claimants can receive the benefits, potentially leaving an extended gap in benefits.

The chart above plots data released daily by the US Treasury, which shows the amount disbursed for all federal unemployment benefits. Prior to March of this year, the numbers were normally much smaller but surged to levels of about $5 billion per day, or $25 billion per week, from April to July. Since July 27th, the payments have begun to drop off sharply, with the weekly average already down below $3 billion/day and falling. A drop from the $5 billion/day rate to $2 billion or less (as it looks to be heading toward) equates to removing something like $15 billion per week (~$750 billion annualized), or more, of income support for unemployed people, who have a high average propensity to spend.

If the financial markets, earnings, and investor sentiment have been supported heavily by fiscal stimulus, and that stimulus (and political support for it) is now clearly fading even while COVID-19 continues to spread across the country, it raises questions as to whether the markets will have sufficient tailwinds to continue higher in the coming months.

Risk on? Not really since early June

In financial markets, it seems like “everything” is going up recently. Stocks, bonds, precious metals, even Bitcoin. Perhaps that should not be surprising given the huge amount of liquidity being produced by global central banks in addition to the fiscal stimulus earlier this year. That tends to have the effect of pushing asset prices up generally.

But when we look at relative returns of some key assets, it looks more like the “risk on” trend has not really gone anywhere since early June. That is, owning the riskier option within various asset classes has not generated excess returns to compensate for that extra risk since the recent peak in risk about June 8th.

As shown in the chart below, buyers of risk in many areas have not been rewarded for about two months now (shaded area since June 8th). Brief comments on each section of the chart are below it.

This could be a natural consolidation after a surge in return to risk after the late-March lows, or a sign that the impact of stimulus, and stimulus itself, is fading.  Many prices/valuations are back near pre-COVID levels even while the economy and earnings are still far weaker than they were in January. Uncertainty about additional fiscal stimulus, now that much of it has expired or been spent, and worries about the continued aggressive spread of COVID-19 in the US are potentially countering the positive hopes for vaccine developments and ongoing central bank support.

We will be watching these returns closely for indications of whether investors are getting properly “paid” for taking on additional risk.

High vs Low Risk Relative Return Measures

  • Top section: Global stocks, measured by the MSCI All-Country World Index (ACWI) have performed no better than long-term US Treasury bonds since early June, and are still far behind bonds on a year-to-date basis.
  • 2nd section: Similarly, within the fixed income market, long-term Baa-rated US corporate bonds have done no better than long-term US Treasuries since early June (despite the ongoing support from the Federal Reserve), and remain well behind Treasuries for the year-to-date.
  • 3rd section: High-beta stocks in the S&P 500 have lagged low-volatility stocks in the index since early June, even as the S&P 500 itself has moved somewhat higher.
  • 4th section: US small-caps have lagged large-caps since early June, and small-caps remain significantly more volatile.
  • Bottom section: Among commodities, industrial metals (copper, aluminum, zinc, etc.) that are used in manufacturing are typically a measure of global growth that rise when the economy is improving (“risk on”). Precious metals are more often preferred as an inflation or currency hedge (“risk off”). So while both industrial and precious metals prices have individually risen significantly recently, industrial metals prices have lagged those of precious metals since June.

Global Technology: still outperforming, but is it expensive?

Much of the attention in equity markets has been focused on the Technology sector, many of whose constituents are reporting Q2 earnings now. The Technology sector has outperformed dramatically both in the US and globally in recent years as well as for the year-to-date. This has raised questions about whether the sector is “overowned” and overvalued, particularly given its unusually high weighting in the S&P 500 index now.

While it may seem almost a foregone conclusion that after extraordinary outperformance the sector must be overvalued, our data suggest this is not necessarily the case.  We review a couple of interesting charts below, focusing on relative returns and valuations (not absolute valuations).

Global Technology Relative Return

  • Tech stocks have indeed outperformed globally, extending a long-term trend in place for seven years and counting now. The chart above shows the relative return of global Technology stocks in our stock universe relative to the return of the entire global universe (~6000 stocks currently). The returns are calculated on an equal-weighted basis (to avoid the potential distortions of a few mega-cap stocks), though the trends using cap-weighted returns look similar.

Global Technology Relative Valuation

  • Based on the relative forward earnings yields1 in Technology versus the global average, the typical Tech stock is not currently very far out of line with historical norms on a relative basis in our work. And Tech stocks are actually cheaper on a relative basis now than they were at the start of the year.
  • The chart above shows the relative forward earnings yield for the median2 stock in the global Tech sector relative to the median stock in the global universe. The solid horizontal line is the long-run average, and the dashed lines are +/- 1 standard deviation from the average.
  • The current reading of -1.4% reflects the Tech sector’s current median forward earnings yield of 3.5% (equal to a forward P/E of 28.5) being 1.4% lower than the global median forward earnings yield of 4.9% (equal to a forward P/E of 20.4).
  • Technology is a growth sector, so it almost always trades at a lower forward earnings yield (higher P/E) than the overall market. Historically, the earnings yield differential has averaged about -1.2%, so the current valuation spread (-1.4%) is actually quite close to the long-run average.
  • Even though Technology has outperformed this year, the sector’s relative valuation has actually improved since the start of the year (from -1.6% to -1.4%). This is because expected earnings for Technology have also outperformed the global average by a substantial margin this year.
  • Confirming the global figures, the chart below shows the same calculation for the US Technology sector, where it may come as a surprise to some to see that the median US Technology stock is actually slightly cheaper than average on a relative basis now, and significantly cheaper than at the start of this year.

US Technology Relative Valuation

Using valuation alone as a timing tool can be quite challenging, but our data suggest that Tech stocks are not especially overvalued on a relative basis (i.e., putting aside whether the entire equity market is fairly valued or not on an absolute basis). So long as earnings in Technology continue to outpace the average, valuation does not appear to be a major headwind to the sector’s relative performance at this point, either globally or in the US.

1 We use earnings yields (earnings/price ratios) instead of price/earnings ratios in aggregate calculations to properly account for the presence of companies with negative earnings. Consequently, higher numbers are more favorable. The forward earnings are based on rolling 12-month forward consensus estimates.

2 Using the median stock avoids potential distortions that can occur in cap-weighted calculations, where unusual movements in a few mega-cap components can skew the sector’s valuation figures at times.

Relative volatility risk in US small-caps remains high

Among the various asset allocation decisions for which we provide guidance to clients is whether to favor small-caps or large-caps (i.e., the “size” factor) within the US equity market. In our view, small-caps do not reliably outperform large-caps consistently over time (as some models and studies might suggest), and instead view the “size premium” (outperformance of small-caps) more as a cyclical phenomenon that tends to show up under certain macroeconomic and market conditions.

While there are many potential conditions that might affect small-cap/large-cap relative performance, much of our work is oriented around the idea that small-caps give the best “bang for the buck” in the periods just before and through the early stages of a new economic or market cycle. That is, recessionary troughs in the economy and equity market set up the conditions for future small-cap outperformance, as smaller companies tend to benefit most from the re-acceleration of economic growth that typically occurs after recessions. This is also when monetary and fiscal stimulus tend to be strongest. In these scenarios, small-caps have typically underperformed before and during the preceding recession/bear market and become out of favor and potentially undervalued. Conversely, the later stages of an economic cycle and the early stages of a recession or bear market tend to be unfavorable for the riskier and more economically sensitive small-caps.

Looking at conditions now, there is certainly evidence that a recessionary trough has occurred or is in process, and both monetary and fiscal stimulus have been very aggressive. This would potentially argue for favoring small-caps over large-caps, and indeed small-cap relative returns have stabilized after a sustained period of significant outperformance by large-caps since mid-2018 (and arguably longer). However, it may be worth an extra dose of caution before making heavy overweight allocations to small-caps on an intermediate-term (6-12 month) basis. This is not only due to the unusual nature of the current cycle, but also more specifically to the extremely elevated relative risk still apparent in the volatility of small-caps versus large-caps.

Using the Russell 2000 Index to measure returns for US small-caps and the Russell 1000 Index for large-caps, the chart below shows the rolling three-month annualized volatility of daily returns for both indices over the last 20 years (small-caps in blue, large-caps in red) in the top section, and the difference between them in the bottom section.

US Small Large Cap Relative Volatility

A few things jump out: first, volatility for all equities surged to extreme levels earlier this year, matched only by the Great Financial Crisis in the last 20 years, and has been declining rapidly thanks to the Federal Reserve’s extraordinary interventions. And second, the difference between small-cap and large-cap volatility has remained near historic extremes even as volatility has declined for both size categories.

As shown in the bottom section of the chart, small-caps are almost always more volatile than large-caps (i.e., the volatility spread is usually above zero), with a historical average of about 4.5%. However, the latest readings on the volatility spread of over 14% are the highest in the last 20 years. That is, while large-cap volatility has dropped to a (still elevated) level of just under 24% (equal to average daily index movements of about 1.5%), the small-cap index volatility has only managed to decline to 38% so far (equal to average daily movements of about 2.4%). This is still far above the normal level of volatility for small-caps historically of about 22%.

The bottom line is quite straightforward: even after sharp rallies in equities and lower market volatility in general, owning small-caps remains much riskier on a price volatility basis than owning large-caps. So if expected returns must be commensurate with expected risk, then a decision to allocate heavily to small-caps requires either 1) an unusually high excess return expectation, or 2) an expectation of drastically lower small-cap volatility soon.

Given the economic backdrop and the reliance on government stimulus as well as the relative fundamentals of small-caps versus large-caps (a possible topic for another post), it may take a little while longer to have confidence that small-cap excess returns will be sufficient to compensate for their unusually high extra risk relative to large-caps.

In case you need a refresher course, it’s all stimulus these days

*with apologies to Irwin M. Fletcher

Much has been made about the divergence between the path of the US (and global) economy and that of the stock and corporate bond markets. Even while economic and earnings growth is historically weak and remains under severe pressure from a rapidly spreading virus, major stock market indices have rallied and are at or near all-time highs. Market valuations based on forecasted earnings over the next 12 months have clearly risen sharply.

While the divergence between financial asset prices and the “real economy” as experienced by many Americans is striking, the reason is not hard to see: enormous levels of stimulus by both fiscal and monetary authorities.

Essentially, the US Congress has voted (in four separate pieces of legislation so far) to drastically increase deficit spending to support individuals and businesses as well as the costs of health care and virus abatement. The trailing 12-month US federal deficit has jumped to about $3 trillion, easily a new record in nominal terms and the largest deficit as a percentage of US GDP since World War II.

As shown in the bottom section of the chart below, the current 12-month deficit is now nearly 14% of nominal GDP, significantly more than the peak following the Great Financial Crisis, and up from the deficit levels of 4-5% of GDP in 2018-19, i.e,. adding roughly 10% of GDP to spending. This has helped fill a huge gap in personal incomes caused by lockdowns associated with COVID-19.

US Monetary and Fiscal Stimulus

And crucially, at the same time, the Federal Reserve has also engaged in an unprecedented level of money creation (“quantitative easing”, or QE) to essentially monetize the increase in Treasury debt. After immediately cutting its fed funds policy target back to zero (0 – 0.25%), the Fed has expanded its balance sheet from around 20% of GDP before COVID-19 hit to over 30% of GDP currently, i.e., adding roughly 10% of GDP to the money supply. This is also far bigger than its balance sheet was after the three previous rounds of QE between 2008 and 2014.

While the Fed has a number of different programs in place to support financial markets and the economy, most of what it is doing is buying up Treasury bonds, mortgage backed securities, and more recently corporate bonds and bond ETFs. It does this with newly created money that can then circulate in the economy. A key effect is to essentially offset the absorption of money caused by the Treasury issuing huge amounts of new bonds, and has drastically expanded the US money supply. And by using newly-granted powers (under the recent CARES Act) to buy corporate debt (including some high-yield or “junk” debt), it has also compressed the credit spreads in the corporate bond market and made it cheaper for companies to borrow than it otherwise would be.

So an extreme oversimplification would be that the Treasury engaged in new spending of something on the order of 10% of GDP, and the Fed printed roughly a similar amount of new money to “pay” for it (a liquidity conversion of the Treasury bonds to cash), with some of that Treasury money being used to allow the Fed to take on corporate credit risk for the first time. That extra ~10% of the economy that has been conjured up by Congress and the Fed was meant to fill some of the hole left by the meteor that hit the global economy in the form of COVID-19.

The key question now is to what degree is additional stimulus needed, and will policy makers provide it to the appropriate degree? Current virus trends, structural economic damage COVID-19 has already caused, and rising debt levels suggest significant further stimulus will be required to bridge the gap until the economic effects of the virus have passed, though that remains a point of debate.

Investors who take the view that the policy response will be sufficient (or possibly even excessive) are likely those who are content to pay the same (or higher) prices for stocks and some bonds than they did in 2019, and assume that earnings will fully recover and stocks will be worth more in the future due to the lack of any competition from near-zero interest rates on safe debt (Fed policy rates are expected to remain near zero for several years at least).

Investors who worry that the policy response will falter may be more cautious, given the potential further weakness in corporate earnings that would also result. And one outcome in particular bears consideration: what if the fiscal response falters (due to Congress being unable to agree on sufficient stimulus legislation), but the Fed continues to intervene heavily in financial markets with quantitative easing and corporate credit spread compression (i.e., printing more money than can readily be spent)? Then financial assets, and unprofitable companies, might continue to hold up and inflation could eventually rise, even while the underlying economy weakens (with implications for income and wealth inequality).

That is arguably not far from what is happening now, as the Fed has adopted a “whatever it takes” approach and has few limits to the amount of monetary support it can provide. The question of further fiscal stimulus is the key question right now, as there reportedly remains a wide gap between Democrats and Republicans in Congress (and the White House) about the size and scope of any additional stimulus. With much of the original stimulus plans now fully utilized or expiring soon, there is a clear danger of a de facto sharp contraction in fiscal policy relative to the current extreme levels of support (the so-called “fiscal cliff”). Congress is scheduled to be in session for only a couple of weeks before going on August recess, and the key $600/week of extra unemployment benefits are due to expire on July 25th. The path of the economy and financial markets may depend on decisions made by Congress in the next few weeks.

Global risk appetite measures slipping recently

July 14, 2020

Stock prices globally have remained unusually buoyant in the face of well-known health and financial risks. Thanks largely to aggressive global monetary and fiscal stimulus starting in March and still going on (though arguably fading), risk appetite jumped dramatically following the severe but relatively brief sell-off from late February to late March.

Most recently, however, several metrics of global risk appetite that we track have either plateaued or weakened. This coincides with a reduction in the pace of central bank activity and growing uncertainty about further fiscal stimulus programs, especially in the US. It also coincides with the recent turn higher in the growth of COVID-19 cases in the US and globally.

The complete explanation of the chart below follows and makes this post somewhat longer, but the key point is that multiple metrics show at least a pause if not an incipient downturn in risk appetite that is worth watching, particularly given that it occurs with the Fed and other central banks still actively pursuing asset purchase programs, even if at a less frantic pace recently.

 Global Risk Appetite Measures

The chart above plots four price-based measures of investor risk appetite. Three are based on equity prices, and the bottom section shows US high yield credit spreads from the bond market. In all four, the preference for higher-risk assets has either lost its earlier momentum or is weakening.

The top section plots the relative return of the MSCI All-Country World Index (ACWI), one of the broadest measures of global equity returns. It includes both developed and emerging markets and has approximately 3000 constituents. Like most equity indices, it is capitalization-weighted, so the largest global companies get the most weight in the index. The US currently makes up about 57% of the weight in the index. We plot its total return relative to that of the ICE/BofA Merrill Lynch 10+ Year Treasury total return index. As its name indicates, it tracks the returns of all US Treasury bonds with 10 or more years to maturity, and thus better aligns with the long-term nature of equities as an asset class but is still considered a “risk-free” asset from the standpoint of return of principal. US Treasury bonds are a widely used global benchmark and the largest and most liquid bond market.

The relative returns of global stocks versus bonds captures the severe drop in stock prices starting in late February of this year, and the subsequent rebound. It bears noting, however, that even with the sharp rebound in stock prices, long-term US Treasury bonds have still outperformed global equities since the pandemic began (i.e., the stock/bond relative return series is lower than it was at the start of this year). The relative return series is also still below its own one-year average, which is itself declining. We also see that global stocks had struggled to outperform long-term Treasuries over a longer period even before COVID-19 became a factor earlier this year, reflecting the general weakness in the global economy developing in 2019. High total returns in bonds are potentially harder to generate now that interest rates are so low (though long-term bond prices can be quite volatile), so if stocks begin to lag bonds again on a sustained basis, it would reflect a clear shift in preference for “return of capital” rather than “return on capital”.

The second section of the chart plots our own index of high-volatility stocks globally. On a quarterly basis, we screen the global stock universe for all stocks with at least USD$200 million market cap (and at least $1 million/day average trading value) and identify those ranking in the top 10% (top decile) based on their trailing two-year price volatility (price risk). We track the daily returns of those top-decile volatility stocks as a proxy for global equity risk appetite. The index thus shows risk appetite in absolute terms, and confirms the stock/bond relative return pattern, though more dramatically. From its initial peak on January 16th 2020, the index plunged -49% to its trough on March 18th. From there it posted an extraordinary rise up to a new recovery high, gaining 123% from the low to its peak on June 8th, putting it higher for the year-to-date at that point. After a short-term sell-off from the peak, the index has been moving sideways recently, off about 10% from the June peak. It remains above its one-year average, and in line with levels seen in late 2019. The positive trend is thus arguably still intact, but has lost momentum.

The third section is somewhat similar but plots the relative returns of the S&P 500 High Beta index versus the S&P 500 Low Volatility index. These indices are constructed based on the 100 stocks within the S&P 500 with the highest market beta (sensitivity to market movements) and the 100 index stocks with the lowest historical volatility (rebalanced quarterly). The relative return series thus shows relative preferences for risk within the US equity market. It has followed a similar pattern to the global volatility index above it, but we can see a notable recent difference. The relative returns of high beta stocks versus low volatility stocks has turned down more distinctly recently, and is now back to about its one-year average. A drop below the average would potentially mark a reversal of investor risk preferences to at least neutral if not outright risk aversion, at least temporarily. We will be watching this series closely.

The final section of the chart plots the average credit spread (the option-adjusted spread, which accounts for call features on corporate bonds) for bonds in the Bloomberg Barclays US High Yield Index. Thus it measures the yield premium on high yield (“junk”) debt (rated BB/Ba or lower by Moody’s or S&P) over US Treasuries of the same maturity. A higher spread (rising line) indicates greater risk aversion and expectations of higher default risk on high yield bonds, while a lower spread indicates lower perceived credit risk. Within the fixed income market, high yield debt is closest to equity in risk, making this another useful proxy for investor risk perceptions.

This series is arguably slightly more difficult to interpret right now than it would have been historically, due to the Fed’s new direct influence on corporate debt, including high yield debt1. Nonetheless, the message is similar. After surging as COVID-19 and the associated lockdowns halted much economic activity and the corresponding ability for companies to service debt, yield spreads have come back down sharply thanks to Fed support. But most recently, the narrowing of credit spreads has stopped and may be starting to reverse. It also still remains above (worse than) its own one-year average (which would be expected given the severity of the economic weakness). The underlying fundamentals for many junk-rated companies have not improved substantially and face continued risk from the virus, but monetary and fiscal support (along with the historically low yields on “safe” debt) have increased investors’ willingness to bear credit risk recently. If that risk tolerance recedes (due to policy changes, or worsening fundamentals), both the high yield debt market and the stock market would potentially come under renewed pressure.

Overall, investor sentiment is arguably quite positive (high equity valuations, low recent demand for put options, high proportions of bullish advisors represented in surveys, etc.), and this has been reflected in returns. But the most recent price action has shown some early warning signs of stalling or reversing that risk appetite, which could lead to renewed volatility in equities.

1 While historically the Fed has been prevented from owning anything other than government-guaranteed debt, the recent legislation (CARES Act) passed in the wake of the economic and market distress allowed the Fed to act in concert with the Treasury to buy corporate and municipal debt, as well as backstopping direct loans to companies. This has directly impacted the credit spreads seen in debt markets, as intended, though at the potential cost of reducing the ability to see investors’ true risk preferences.