Indexology^® Blog

Equity markets have fared reasonably well aided by liquidity in ETFs, as my colleague Craig Lazzara highlights. Steep discounts to net asset values (NAVs) on popular fixed income ETFs are bringing an onslaught of doomsday projections. But while the signs of stress are evident, it’s important to decouple the dysfunction of the bond market from the investment product as well as the manager’s skill.

Let us first consider the market context. The rise of interest rate, credit, and volatility risk was previously discussed as the S&P U.S. Treasury Current 10-Year Index yield fell below 1%. Yields then halved and are now bouncing between 0.5% and 1%. Treasury bid-ask spreads were reported to widen to beyond 1 point and NAV discounts were seen in treasury and credit ETFs. Despite that illiquidity, fixed income ETFs experienced their largest daily volume in treasury and credit sectors. Liquidity is a premium. The Cboe/CBOT 10-year U.S. Treasury Note Volatility Index hit its highest point since the end of Lehman Brothers. This measure, along with its credit and swap variants, highlights the stress experienced across the fixed income markets.[1]

In terms of measuring the impact of these shocks on ETFs, Andrew Upward at Jane Street produced a great historical summary that was covered in a recent S&P Global webinar. In times of credit stress, the discount as a percentage of NAV reflects the price buyers are willing to pay.  In recent days, the largest investment-grade and high-yield ETFs traded at over a 5% discount. This could be misinterpreted as a sign a credit ETF isn’t functioning well. There are a few critical points to counter that view. First, prices in the over-the-counter bond market are typically shown as “request for quote”—as soon as one wishes to sell at an advertised price, the trader showing the bid can remove it. The fact that an ETF has an executable price goes well beyond the unwilling participants in the OTC market. The second critical point is the latency in NAV calculation. As an index provider, we pride ourselves on using independent transparent pricing.  These price providers play a heavy role in NAV calculation and are often referred to as price “evaluators,” since they are providing their evaluation on what the price of a bond should be on a given day. Having an independent resource is critical for the index and fund administration community. But they are not employing exacting measures to determine the price of a bond that may or may not have traded that day. The last point is a timing issue; the official index closes at 3PM, while the ETF and its NAV close at 4:00PM.  This also causes a mismatch to NAV during times of volatility.

In the webinar, Bill Ahmuty, Head of SPDR Fixed Income Group at State Street Global Advisors, spoke about how fixed income volumes tend to grow during times of stress, while the underlying cash bond market volume tends to shrink. It appears that ETFs fulfill a critical need of liquidity when liquidity is needed most. ETF structure lends well to this, as investors can trade ETF shares without having to source the individual bonds. This only works to the extent that buyers and sellers can match their trades. Once they are matched, the liquidity must be met by the underlying bond market. As the fixed income ETF market grows, it has a better opportunity to meet or improve liquidity, similar to the equity market.

Finally, as investors look for those who successfully navigated these markets, the active versus passive debate will return. While that argument may be over for equity, new index-based strategies are proving their worth in fixed income. We will cover passive strategies and their performance in upcoming posts, but now, we want to highlight how the S&P U.S. High Yield Low Volatility Corporate Bond Index has outperformed its benchmark by 2.4% YTD.

While fixed income ETFs have largely performed in line with this market, the growth of secondary market trading will continue to help face future liquidity needs.

[1] For more information on this topic, please see the S&P Global webinar Measuring Fixed Income Volatility.

The fall in equity market values since February’s peak has been sudden and dramatic.  During this period, the equity markets have functioned well at their primary task of facilitating price discovery at a time when values were changing rapidly.  Equity investors who wanted to trade have been able to trade.  (Whether they were wise to do so is a separate question.)

This is in no small part due to the liquidity provided by index-linked vehicles.  Index funds have been part of the investment landscape for nearly 50 years, during which time superior performance and low cost have earned them a growing share of investor assets.  What’s less well-appreciated has been the growth of a trading ecosystem linked to indices, most importantly to the S&P 500.  The vast majority of trading in index-linked vehicles is not done by index funds, but rather by active investors and traders.

The coronavirus-induced bear market which we’re now enduring illustrates the logic of index-linked trading well.  A pandemic of still-unknown duration and severity can be expected to affect every business adversely.  The degree of adversity will obviously vary across industries, with some (e.g., airlines and hotels) suffering more than others.  But investors who want to adjust equity exposures quickly will be far better served by trading an entire index than by trying to sort out relative winners and losers in the traditional way.

In this environment, it’s unsurprising to see higher volumes and higher dealer spreads, as we did during the last two weeks’ market carnage.  It’s estimated that in the first week of March, “exchange-traded products powered more than a third of equities trading on U.S. exchanges.”  We can see this effect in Exhibit 1, which measures the average value traded in the largest ETF tracking the S&P 500.  Normally the most heavily-traded issue in the U.S. market, daily volume skyrocketed at the end of February.

Exhibit 1. Trading Volume Spiked During the Market Decline

At the same time trading costs rose, as measured by the width of the bid-offer spread pictured in Exhibit 2.

Exhibit 2.  Spreads Widened

As recently as January 2020, spreads were as low as 0.003%; more recently they doubled.  Spreads are not at all-time highs, however, and trading the ETF is still much cheaper than trading individual securities.

It’s important to understand that the widening of dealer spreads is a natural and appropriate reaction to the extraordinary levels of market volatility we’ve witnessed since mid-February.  Exhibit 3 describes the relationship between the VIX index and the ETF’s bid-asked spread.  When VIX is at or below its median level, for example, spreads average just over $0.01/share.  The relationship slopes gently upward over most values of VIX.  Extreme levels of VIX, however, indicate heightened dealer risk, and thus increased bid-asked spreads.

Exhibit 3.  Higher VIX Leads to Wider Spreads

At time when it was badly needed, index-based products helped to ensure equity liquidity was available at a reasonable price.

By this point, the infamous COVID-19 has made its way around the world and wreaked havoc through equity markets across the globe. We’ve already highlighted the value of protection in the U.S. and Canada.  Pandemics show no respect for international boundaries, so year to date charts for equity indices everywhere show very similar patterns and conclude with the same trajectory. Everyone started 2020 on a good note until the reality of the virus settled in.

We often say that the low volatility anomaly is universal, but it is rare that we see the strategy at work simultaneously in multiple markets.  The charts on the left below track year to date performance for the parent index in each region and the respective low volatility index. On the right shows the performance from the beginning of the year to the peak for each region and then the performance from the peak through March 16. In every market, the low volatility strategy has allowed for participation on the way up while offering protection on the way down.

Six minutes after trading began on the New York Stock Exchange on March 9, 2020, the S&P 500^® plummeted 7% and market-wide circuit breakers kicked in for the first time since the stock market crash of Oct. 27, 1997. However, the indices fell again later in the week and triggered another circuit breaker on Thursday, March 12. As I was writing this blog, on Monday, March 16, the third circuit breaker rocked the market soon after the open.

In such a tumultuous market, the best way to win is to not lose. The S&P 500 Dynamic VEQTOR Index was among the few strategies that posted positive returns in the week of March 9, 2020; it gained 5.3% in one of the worst weeks in S&P 500 history (-8.7%).

Although the VEQTOR index lagged the broad market most of the time in 2020 before last week, it finally pulled ahead with its recent success. Exhibit 2 shows how dramatic the reversal of fortune has been. As of March 13, 2020, the S&P 500 lost 15.7% YTD with a volatility of 45.5%, while the S&P 500 Dynamic VEQTOR Index gained 4.2% YTD with a volatility of 16.2%.

The outperformance of the VEQTOR strategy mainly comes from its dynamic allocation to the S&P 500 VIX^® Short Term Futures Index, a liquid, tradable proxy of VIX spot. This defensive strategy index monitors the implied and realized volatility in the market and allocates to equity and VIX futures accordingly. Its allocation to VIX futures usually increases in a volatile market and decreases otherwise. In the calm market of January 2020 when VIX was below 15 on most days, VEQTOR maintained a low allocation of 2.5% to VIX futures. However, as VIX shot up more than 300% in March 2020, VEQTOR quickly increased its allocation to VIX futures to 40%. We also observed that VEQTOR took precaution and went to all cash briefly when the index dropped more than 2% in the course of one week (see Exhibit 3).

Historical data shows that both VIX spot and VIX futures tend to have a negative correlation with the equity market, and this correlation is generally strengthened during market crisis, when traditional geographical hedging and cross-asset-class hedging breaks down. Furthermore, the negative correlation between VIX spot and futures and the equity market is convex. This means that when the S&P 500 drops, VIX spot and futures tend to rise even higher. The combination of negative correlation to the equity market and convexity of VIX futures contribute to the recent success of the VEQTOR strategy with an allocation to VIX of less than 50%.

That said, investors must keep in mind that the VEQTOR strategy is designed to hedge tail risk in the broad equities market. Similar to holding a continuously rolling put option, hedging costs over time tend to create a significant performance drag in the long term. As shown in Exhibit 4, the S&P 500 Dynamic VEQTOR Index lagged the S&P 500 in the past decade’s bull market. In general, the index tends to outperform when VIX rises and stays high for a prolonged period.

The famous football saying, “the best offense is a good defense” is true because a good defense means that you are always in the game. The S&P 500 Dynamic VEQTOR Index has lagged the broad equity market in the long stretch of the 10-year bull market. In the recent market shakeup, however, investors who opted for its defensive feature are getting what they paid for.

If you are interested in the S&P 500 Dynamic VEQTOR Index allocation process, please refer to its methodology and our Research paper: Dynamically Hedging Equity Market Tail Risk Using VEQTOR.

Backwardation is incredibly uncommon in the VIX^® futures curve. While the reason behind this term structure is not perfectly understood, the conclusion is clear: long and hold does not work for VIX futures, as the roll cost burns.

There are different ways to measure VIX futures backwardation: by using the relationship between the VIX level and the front-month futures, between the first and second month futures, or between points further out on the curve.

One way I think highly insightful is to calculate the roll yield by taking the return of the S&P 500^® VIX Short-Term Futures ER MCAP Index (ER measures the price return plus the roll return) less the returns of the S&P 500 VIX Short-Term Futures Index (which measures the price return only). Backwardation was implied by a positive result, whereas contango was implied by a negative result. This approach also allows us to decompose the return of the S&P 500 VIX Short-Term Futures ER MCAP Index into the price change of VIX futures (at constant one-month maturity) and roll yield/cost.

Backwardation is incredibly uncommon in the VIX futures curve. Since 2005, there have only been four periods where the roll yield was wider than 1%—during the financial crisis, when the U.S. lost its ‘AAA’ credit rating in 2011, in February 2018, and now.

The implication of this is that when VIX futures are backwardated, exchange-traded products that track the S&P 500 VIX Short-Term Futures ER MCAP Index may earn a positive return from rolling into a cheaper contract before expiry, independently from the futures price change. Thus if the VIX level is unchanged, the index can still provide positive returns through the roll yield. For example, this roll yield averaged 1.2% per day last week (March 9-13, 2020).

We know backwardation is an uncommon occurrence, and Exhibit 3 provides some historical context of how long backwardation has lasted in prior periods.

Note the longest streaks in Exhibit 3 were 76 and 63 days and they occurred in 2011 and 2008, respectively. During both periods, roll yield contributed 115% and 45% to the S&P 500 VIX Short-Term Futures ER MCAP Index, respectively.

We have been in backwardation for three weeks (as of March 13, 2020), and VIX is approaching an all-time high; if the markets continue to be volatile, we could be in this situation for some time.