- Volatility products can be used to hedge downside risk in equity portfolios
- RVXSM futures provides a U.S. small cap equity focused alternative to CBOE Volatility Index® (VIX® Index) futures
- RVX futures hedges can potentially be more effective when the risk is specific to small cap stocks
RVX volatility futures (CBOE Russell 2000® Volatility Index Futures Contracts) track the implied volatility expectations of the U.S. small cap equity market and can be a particularly effective hedge for the risk inherent in the small cap asset class. The potential usefulness of RVX futures in hedging small cap-specific risk is illustrated in this paper with results of a simulation case study that includes two hedging examples based on market movements in spring and summer 2014. One example shows that a simulated small cap portfolio could have been better hedged with RVX futures than with VIX futures during the period examined.
This paper starts with a brief introduction to volatility products and the concept of small cap-specific risk. It then presents the case study results, starting with a review of market conditions leading up to the small cap market correction in 2014 and continuing through to a comparative estimation of the results of a particular hedging strategy implemented with both RVX and VIX volatility futures. RVX trades are conservatively modeled. A methodology appendix provides details of how the hedges are constructed and valued. It also includes an examination of the differences between volatility indexes and the prices of volatility futures based on those indexes, basic details of the construction of the futures hedges, and the approach used here to estimate the prices used in implementing and rolling RVX and VIX hedges.
VIX-based volatility products have expanded over the last 10 years - from futures contracts, to volatility options, and to increasingly sophisticated ETFs (e.g., VXX and XIV). VIX, the CBOE Volatility Index, is a measure of the implied volatility of the U.S. large cap asset class as represented by the S&P 500®.1 The implied volatility of a security is determined by the prices of options on the security and is an estimate of the market-implied expected volatility over the remaining life of the options.2
RVX is the implied volatility index for the U.S. small cap asset class as represented by the Russell 2000 Index.3 Small cap-specific risk, which is driven by factors associated with small cap stocks, is a type of asset class-specific risk. Many risk factors affect all asset classes. The asset class-specific risk examined in this case study is the market perception that small cap stocks might be overpriced relative to the market as a whole. The greater liquidity of the VIX contract has made it a frequent choice for managers seeking to hedge small cap stocks against market-wide risks. Yet the examples developed here show that hedging with VIX may not be the best strategy for small cap-specific risks.
Our baseline assumption is that an investor wants to use volatility futures to hedge a small cap portfolio. (Options offer another tool for hedging portfolio risk, but will not be studied here.4)
VIX positions are modeled with entry and exit trading at the contract daily settlement price. RVX entry/exit trades are modeled more conservatively by taking the observed bid/ask spread into account, with purchases at the daily settlement price plus one-half the daily quote spread and sales at the daily settlement price minus one-half the daily quote spread. Thus simulated RVX purchases are at the ask price, and simulated sales are at the bid price. The primary cost of maintaining a volatility futures hedge is typically the cost of rolling the contracts, which is when contracts nearing expiration are sold and replaced with longer-dated contracts. VIX contracts are rolled at average time-matched trade prices for both contracts. RVX contracts are rolled conservatively, based on the average time-matched bid price for the expiring contract and ask price for the next-month contract. The methodology appendix provides further details.
Hedging the 2014 small cap correction
The spring of 2014 was a time of increasing investor uncertainty about U.S. small cap equity prices. Many believed small cap was overvalued. On February 6, 2014, for example, a Bloomberg article noted that implied volatility on iShares Russell 2000 ETF options had spiked and that strategists were seeing a downturn coming for small cap prices.5 This uncertainly was reflected in RVX values. Figure 1 shows the RVX Index from July 1, 2013, to February 6, 2014. A spike in small cap implied volatility can plainly be seen to have started on January 24, 2014. A peak can be seen on February 3, followed by a small decline to February 6, as highlighted in red.
Figure 2 displays the RVX premium relative to VIX. The RVX premium is defined as RVX Index values as a percentage relative to VIX Index values. A zero premium corresponds to RVX and VIX indexes having the same value. The RVX premium is a representation of the market perception of the magnitude of small cap risk relative to large cap risk. An increase in expected small cap-specific risk should be reflected in an increase in the RVX premium.
Looking at the period January 1, 2014 through August 31, 2014, it can be seen in Figure 2 that there is also a spike in the RVX premium starting on January 24. The RVX premium value of 40.9% on February 6 is highlighted in red, both to provide orientation for the different time periods shown in Figures 1 and 2 and to show that the RVX premium continues to spike on this date, even though the RVX values in Figure 1 decline. This indicates that VIX values declined faster than RVX values from February 3 to February 6. Figure 2 displays the RVX premium until August 31, 2014. After February 6, the premium first reverts back into the range in which it had been trending and then trends unevenly upward through at least the beginning of June. One interpretation of this upward trend is that full risk of an expected small cap correction was not immediately reflected in options prices and volatility indexes as of February 6. It can also be seen that the RVX premium declines sharply at the end of July, suggesting a potential end to the period of small cap-specific risk.
The simulation case study is based on the time period January 1, 2014 to August 31, 2014, depicted in Figure 2. This is obviously a retrospective study. The selection of the times to hedge was based on the benefit of a hindsight-based timing ability that is not available to investors. The dates that define the first hedge were selected according to the idea that they might have been chosen by an investor with moderate timing ability. So, for example, the small cap volatility spike starting on January 24, which has already been discussed, is not used as a hedging example. It seems unlikely that an investor with moderate timing ability would have been able to predict this volatility spike with sufficient accuracy.6
The first hedge is implemented on February 24, about two weeks after the end of the volatility spike and publication of the Bloomberg article. This gives our hypothetical investor time after the February 6 spike to do an assessment of the small cap market and to design and implement a hedge.
The second hedge is started on June 30 and ends on August 6. The primary purpose of this second hedge is to compare RVX and VIX hedge performance in a changing market environment, wherein small cap-specific risk is not as prominent.
The dates for ending the hedges are not crucial. A change of a few days does not affect the qualitative results.
Figure 3 shows the effects of hedging Russell 2000 Index performance over the two periods just described. The cumulative total return of the Russell 2000 is shown, along with cumulative performance when the Russell 2000 is hedged using RVX futures or VIX futures. The shaded areas indicate the time periods of the hedge examples. The RVX-based hedge outperformed the Russell 2000 and the VIX-based hedge strategy. The VIX-based hedge was initially successful, in the early period of the first hedge, but quickly had increasing periods of underperformance relative to the Russell 2000. It is only toward the end of the second hedging window that the VIX-based hedging strategy outperformed the Russell 2000 on a sustained basis. As of August 31, 2014, the RVX-based hedge outperformed the Russell 2000 by 12.95% and outperformed the VIX-based hedge by 7.60%.
Figure 3. Russell 2000 Index cumulative total return performance and the performance of two Russell 2000 50%-hedged strategies, one using RVX and one using VIX futures. First hedge period is February 24, 2014 to May 8, 2014. Second hedge period is June 30, 2014 to August 6, 2014. Data is from Russell Indexes and CBOE. Details of hedge construction are provided in this paper’s appendix. Complete time period is January 1, 2014 to August 31, 2014.
Table 1 presents performance statistics for the complete study history and for the time periods of the two hedges. The results over the complete study history show that the RVX hedge protected investor value with minimal increase in portfolio volatility. The maximum drawdown statistics for the complete history time period show that the RVX hedge reduced maximum drawdown from 9.06% to 7.40% while the VIX hedge had a 17.00% maximum drawdown. The cumulative return of the RVX hedged portfolio was also considerably better. Note that the complete history statistics in Table 1 involve volatility hedges only over the two time periods of the hedging examples.
Table 1. Summary performance statistics for the Russell 2000 and RVX-hedged and VIX-hedged Russell 2000 strategies as defined here
Results presented for the entire study history and for the two hedge periods: February 24, 2014 to May 8, 2014 and June 30, 2014 to August 6, 2014.
Figure 3 shows the RVX functioned successfully as a hedge in the first period, almost completely neutralizing the Russell 2000 Index loss in value. This result was achieved with lower levels of volatility than experienced by the Russell 2000. VIX failed as a hedge in the first time period. This failure is visible both in the low returns and the large drawdown reported in Table 1.
The results for the second hedge reported in Table 1 show that the RVX hedge performed effectively and provided positive net returns, but was outperformed by the VIX hedge. The performance of the RVX hedge relative to the Russell 2000 index is 9.53%, and that of the VIX hedge is 12.17%. Note, however, that the VIX hedge outperformance is accompanied by considerably higher volatility. As a result of this increase in volatility, the information ratio of the VIX hedge is only marginally higher than that of the RVX hedge.
The likely reason for the strong performance of the second VIX-based hedge can be seen in Figure 4. This chart shows a long period of steady declines in VIX values that starts in April and reverses over the month of July, before spiking on July 31 (highlighted in red). This spike occurred on the same day that the Dow Jones industrial average fell 317 points, as a result of investor fears relating to domestic and global concerns.7 The RVX premium declines sharply at the time of the VIX spike and stabilizes for the first time since the February 6 spike, in a range mostly below that spike. The second hedging period can thus be seen to embody a transition from small cap stock risk being driven primarily by asset class-specific issues to being driven by market-wide factors.
A volatility hedge is utilized to protect against the marketplace volatility under the assessment that the likely benefits outweigh the more certain costs entailed. Because a hedge is costly to maintain, some ability to time implementation and liquidation is important.
The effect of the lower liquidity of RVX is modeled in this study by taking positions at ask prices and liquidating positions at bid prices. VIX contracts, in contrast, are modeled to trade at average and settlement prices. These adjustments provide for a more conservative estimate of the performance of RVX hedges relative to VIX hedges.
One result of the conservative pricing approach is that the RVX continuous hedge underperformed the VIX continuous hedge over the total study period. In spite of this handicap, the RVX hedge significantly outperformed the VIX hedge in the first hedging example, which is most directly related to small cap-specific risk. In the second example, where perceived risk is macro in nature, the VIX hedge did better than the RVX hedge, but the RVX hedge still enabled the hypothetical RVX-hedged Russell 2000 portfolio to deliver net positive returns for the time period studied.
In conclusion, this case study suggests that when the risk to be hedged is specific to small cap stocks, RVX futures-based hedges may provide for results superior to those of VIX-based hedges.
In the appendix below, the method of determining hedged performance is reviewed.
This appendix explains how the examples in the simulation case study were constructed and provides some relevant background information. The methodology used to construct, value and liquidate hedges was designed to enable realistic comparison of RVX-based and VIX-based hedge strategies.
A1. Data and futures prices used in this study
This study covers the period January 1 to August 31, 2014. RVX and VIX index values, contract prices, trade data and RVX quotes are obtained from CBOE, the Chicago Board Options Exchange.8 Russell 2000 index total return values come from Russell Indexes.
The daily value of both RVX and VIX positions is determined by daily settlement prices. VIX entry and exit trades are also based on settlement prices. Prices for VIX trades may be obtained through the “trade at settlement” (TAS) transaction system available on the CBOE Futures Exchange (CFE®), the trading venue for CBOE volatility futures. The prices assumed for RVX trades and contract rolls are more conservatively estimated, because of the relative thinness of trading. RVX positions are established based on the RVX settlement price but with payment of a premium of one-half the average bid/ask spread quoted that day. This is an estimated ask price. RVX positions are liquidated based on payment of the daily settlement price minus one-half the average bid/ask price quoted that day, representing an estimated bid price.
Contract rolls are assumed to take place at the average price ratio between the front-month and second-month contract on the day of the roll. The roll ratio is defined as the price paid for a second-month contract divided by the price at which the first-month contract is sold. Average prices of VIX trades and average RVX bid and ask quotes are determined for each 15-minute block over the study period. VIX roll ratios are based on time-block-matched average front-month and second-month contract trade prices. RVX roll ratios are based on time-block-matched average ask prices for the second-month contract and bid prices for the front-month contract. The average roll ratio on the day of the roll is the average of the ratios defined for each 15-minute block.
— Download the complete paper —