Indexology^® Blog

Risk is once again part of investors’ vocabulary. Through yesterday’s close, the S&P 500 lost a total of 6%, made all the more jarring by the practically straight line rise in most of 2018 prior to the losses. Volatility has, of course, ticked up, but in the context of the broader 27 year history, not dramatically so.

ROLLING 21-DAY VOLATILITY FOR S&P 500

While the stock market’s recent declines seem quite traumatic (in point terms, the Dow Jones Industrial Average’s decline was a record), insight into the factors that contribute to volatility may alleviate some worries.  The dispersion-correlation map below gives us a way to put the recent jump in market volatility into context.  A rolling 21-day look at the dispersion and correlation levels so far this year shows a general trend of increasing dispersion.  As of the market’s close on February 5^th, dispersion had increased to slightly higher than median levels, and correlation made a significant jump.

ROLLING 21-DAY DISPERSION-CORRELATION YEAR-TO-DATE

The longer-term dispersion-correlation map below shows that 2017 was among the sleepiest of years and, in the context of stock dispersion and correlation, was similar to the boom years of the mid 1990s when volatility was also subdued.  Current conditions are at approximately the same levels as in 2011 (a year that experienced significant volatility mid-year but ended more or less flat).

Years of extreme market distress such as 2000 and 2008 witnessed dispersion levels that were much higher than the current environment.  If the past is any gauge, therefore, what has happened in the last week hasn’t propelled us to crisis levels.

DISPERSION-CORRELATION MAP

In recent years, strategies selling volatility (and VIX® futures in particular) garnered substantial attention due to the low levels of VIX and the eye-watering returns achieved by associated benchmarks such as the S&P 500® VIX Short-Term Futures Inverse Daily Index (we’ll call it the “short VIX index” here for convenience).   At the end of last week, the short VIX index could boast of a 10-year total return of 1,518%.

We’ve previously warned that short VIX strategies should be exclusive to the courageous, especially in times of low volatility.  The past 24 hours remind us why this must be so: by the end of Monday’s trading, as both VIX and VIX futures recorded their largest ever one-day percentage increases, the 10-year total return of the short VIX index had fallen from a 15-multiple gain, to a loss of 3.6%. 

So, what just happened?

One candidate explanation for the remarkable swings in VIX futures yesterday relates to a phenomenon similar in spirit to a combination of a classic “short squeeze” with the types of trading patterns generated by portfolio insurance strategies.  It is best illustrated by the hypothetical example of an investor following a strategy selling VIX futures in equal proportion to their invested capital.

Let’s suppose that – at the market’s open on Monday morning – an investor had a collateralized position in futures (or other index-linked products) tracking the short VIX index, with a notional position size of $100.

By the end of the Monday’s trading, as VIX futures rose by a total 96%, the value of the futures which our investor was short rose from $100 to $196.  The $96 increase in the VIX futures position necessitates an equal and opposite decline in the value of our investor’s position: from $100 down to $4.

In order to maintain his short position at a fixed proportion of capital, the investor would have to reduce the size of his exposure from $196 to $4; in other words to purchase $192 of VIX futures.  This adjustment would naturally have had the effect of putting further upward pressure on VIX levels.

It is not clear how many investors were following such a strategy overall, but a lower bound might be approximated  by the $3bn of assets in ETPs (exchange-traded notes and exchange-traded funds) that were tracking the short VIX index as of Friday’s close.  From that approximation, we would have anticipated a little under $6bn of “short covering” in VIX futures into yesterday’s close – an amount that may have indeed exacerbated the market’s moves.

And is it over yet?

There are plenty of other candidate explanations for the market’s movements yesterday.  However, if the primary driver was indeed the rebalancing flows (and short-covering) from investors who were short volatility, the good news is that it’s quite possible the market will return to normal.  The same $3bn of assets tracking short volatility at Friday’s close would be worth only $0.12bn after yesterday’s losses, which is far less likely to move the market.  Of course, if the real reason for the market moves lay elsewhere, higher volatility may well continue.  At any rate, we suspect that after such a lesson in the possibility of sharp spikes in the volatility market, investors will approach short VIX strategies with renewed caution.

The results are in for the latest S&P Persistence Scorecard. Based on data as of Sept. 30, 2017, the results again highlight the lack of performance persistence among actively managed equity funds. Produced semiannually, the S&P Persistence Scorecard highlights the degree of difficulty faced by active managers to stay at the top of their peer group consistently.

Of the 222 large-cap managers that were in the top quartile as of September 2013, zero could retain this mark for the subsequent four periods. Similarly, for mid- and small-cap managers, within their respective starting universes, no manager was able to maintain their performance. Furthermore, it only took mid- and small-cap managers three years to reduce the universe to zero in the top quartile (see Exhibit 1).

The S&P Persistence Scorecard also provides transition matrices for managers in various quartiles. The transition matrices track the path of managers across different quartiles over time, and therefore present the probability of where a fund may end up. Exhibit 2 shows a direct relationship between the starting group and the probability of being liquidated in the next five-year period. We observe that funds starting in the top quartile, based on the past five-year period, have a 10.24%, 8.96%, 11.11%, and 10.23% change of being merged for All Domestic, All Large, All Mid, and All Small-Cap mutual fund categories, respectively. Furthermore, when the fund was in the bottom quartile to start, these statistics rise to 31.81%, 31.34%, 33.33%, and 37.5%, respectively. This means that a fund is roughly three times as likely to disappear if it started in the bottom quartile versus in the top quartile.

We can also note that with the exception of mid-cap managers, the relationship is strictly increasing. The implication for market participants is such that they should be aware of (and consider) the increased probability that a manager might not be around given their relative performance ranking over a certain period. While past performance does not have an impact on future performance, past performance has some correlation with the likelihood of being liquidated or not.

Past performance is often viewed as a way to select an investment going forward, even though the body of research suggests it to be a poor indicator of future results. The S&P Persistence Scorecard has repeatedly shown that past performance is not a robust metric to use when selecting active managers. Analysis such as that in Exhibit 1 shows that if a market participant were to select a manager based only on performance, there would be less than a random chance that the manager would stay in the first quartile.

The S&P 500 just posted its best January since 1997, and also had its highest measured optimism, a 6.6% risk premium,  since October 2015.  Whether history repeats itself is yet to be seen, but just a few days after that high risk premium, the stock market topped on Nov. 3, 2015.  By December investors saw the glass as half empty, as the exuberant premium turned into a discount, and in the following period through Feb. 11, 2016, the market dropped 12.7%.

Now the stock market just logged its worst week, down 3.9%, since that last drop in 2016.  There are worries of rising inflation, interest rates and uncertainty over the dollar.  So, perhaps it is helpful to review some of the U.S. equity market sensitivities to the rising interest rates, inflation and the dollar, but in this post the focus will be solely on the impact of the falling dollar.

Since the U.S. Dollar peaked on Dec. 20, 2016, it has fallen 13.8% while the S&P 500 (TR) gained 24.3%.  While the relationship is opposite overall, when correlation is measured on a daily basis the correlation is nearly zero at 0.16.  This is not surprising since many other factors may drive the stock market on any given day.

However, when looking further back since 1988, there is positive correlation between the U.S. dollar and the S&P 500 (TR) until the Aug. 2000 stock market top (using monthly data) and then negative correlation since the Sep. 2002 bottom.  In the earlier period, the U.S. dollar and S&P 500 (TR) gained a respective 27% and 486%, and in the latter period the U.S. dollar fell -17% while the S&P 500 (TR) gained 246%.  The correlation since the global financial crisis has been consistently negative though with the U.S. dollar’s strength from 2014, the correlation rose to be less negative.

What changed?  Many things may have changed but the decline from the tech bubble burst certainly changed the makeup of that sector.  Before the decline in 2000, the information technology sector held the biggest weight at 33% of the S&P 500, but with the stock market drop, fell almost 2/3 to just 13%, dropping in size behind financials, health care and consumer discretionary.  Today, the sector is the biggest again comprising 24% of the S&P 500.  While the market cap weight is proportionally not as large as it used to be, the foreign revenue exposure is much bigger.  The increased globalization led to more international revenue that is reflected in the S&P 500 Foreign Revenue Exposure Index where the information technology sector is the biggest by far, making up 41% of foreign revenue in the S&P 500.  This may be driving much of the opposite relationship between the S&P 500 and the U.S. dollar in at least the last decade.

While all sectors across styles and sizes are negatively correlated with the U.S. dollar, most are only moderately negatively correlated.  Though of course energy and materials are much more negatively correlated since the commodity prices underlying the businesses in their sectors are priced in U.S. dollars.  Interestingly, the heavy weight of 40% information technology in the S&P 500 Growth (TR) makes it slightly more negatively correlated to the U.S. dollar than the S&P 500 Value (TR) that has energy as its second biggest sector, weighted at 12%.

Though correlation matters, especially for diversification purposes, the sensitivity of the U.S. equities to the U.S. dollar is meaningful too.  In particular, only energy and materials have negative beta less than -1 despite size, meaning it has more volatility and in the opposite direction of the U.S. dollar.  The small-cap energy sector has the biggest magnitude of negative beta of -3.14, meaning if the U.S. dollar were to fall by 10%, then small-cap energy stocks may be expected to rise 31% from that drop.  That is more than the -2.76 and -1.69 respective beta measures from energy mid-cap and large-cap.  Though large-caps may do more international business than the smaller companies, the smaller companies may have fewer resources to hedge against oil price volatility, leaving them more susceptible to the U.S. dollar moves.   Similarly the small-cap materials sector with a 1.62 beta to the U.S. dollar is slightly more sensitive than its mid and large size sector counterparts that have respective beta of 1.57 and 1.50.

Lastly, the falling U.S. dollar helps U.S. equities but mid-caps benefit most. For every 1% drop in the U.S. dollar, the S&P MidCap 400 (TR) rises 3.20% on average. This is more than the 2.63% on average from the S&P 500 (TR) and 2.96% from the S&P SmallCap 600 (TR).  The major exception to this, happens in energy where a falling dollar helps more than a rising dollar hurts, boosting the less hedged small-cap energy sector by 4.64% on average for every 1% dollar drop. The mid-caps are well positioned to grow international revenues with a falling dollar if they are on the cusp of expanding globally and need a catalyst.  While the large caps can benefit from the falling dollar, it is likely the mid-caps can capture more new business with the global growth opportunity.  Mid-caps are also positioned generally better than small-caps for international business due to their greater size and resources.

Knowing which sectors are exposed to which countries can help determine how to benefit most from the falling U.S. dollar.  Also, the distinct market cap split that the S&P 400, 500 and 600 offer is important for intentionally getting exposures and improving risk management as macro economic factors move.

Over the past five years, the more worrisome government-issued debt in Europe has made significant progress in managing the normal mechanism of higher-perceived risk equaling higher yields. The Wall Street Journal focused on Portugal’s debt in their article, “Decade of Easy Cash Turns Bond Market Upside Down”. Quantitative easing (QE) by the European Central Bank (ECB) not only lowered rates to spur economic development in the region, but it also seems to have eased concerns about the risks of buying from some of the heaviest debt-ridden European countries. The negative yield environment created by the ECB’s QE program pushed investors to find higher yields in countries with higher risk profiles. Portugal, Italy, and Spain have all seen yields tighten significantly over the past five years, as measured by the S&P Portugal Sovereign Bond Index, the S&P Italy Sovereign Bond Index, and the S&P Spain Sovereign Bond Index, respectively. As of Jan. 31, 2018, yields for the overall indices saw both Portugal and Spain below 1%, with Italy hovering above 1% at 1.12%. Yields for all three indices have tightened from a range of 3.28%-5% back in the beginning of 2013 to a range of 0.79%-1.12%. When compared to the yield of U.S. Treasuries (2.36%), as measured by the S&P U.S. Treasury Bond Index, one may wonder what the reaction will be when the ECB reverses QE. While all four countries have different durations between 5.24 and 6.47, the U.S. has the shortest duration of them all, with the lowest risk profile and the highest yield.