Home Economics Goldman Explains How Traders Made 4,364% Since 2009 With This ‘Simple’ Strategy

Goldman Explains How Traders Made 4,364% Since 2009 With This ‘Simple’ Strategy

31
SHARE


This is too easy. Forget NFLX Calls. As Goldman explains, the road to real riches over the past 8 years (off the ‘666’ lows in the S&P) is simple – Sell Vol!

The S&P 500 VIX Short-Term Futures Daily Inverse Index which tracks the return of being short a one-month VIX future was up 4364% from March 9, 2009 through 1Q 2017.

Policy uncertainty is elevated around the globe and yet the VIX has posted one of its lowest starts to a calendar year on record. Many investors want to get long volatility via VIX ETPs. But buyer beware. Not a single VIX ETP actually tracks the VIX. They track VIX futures and the performance differential can be large. In our view it is important to understand how VIX ETPs are constructed, key return drivers and historical performance across bull, bear, and boring markets before trading.

VIX ETPs have almost $4 billion in assets under management.

And Open Interest has grown dramatically.

Trading: VIX ETPs are benchmarked to VIX futures; not the VIX

In our client discussions the number one misunderstanding regarding VIX ETPs is the simple fact that not a single VIX ETP is actually benchmarked to the VIX. What actually trades? Most VIX ETPs are benchmarked to indices constructed by S&P Dow Jones Indices and track the daily return on a constant maturity one-month or five-month VIX future. In practice, this typically involves the daily rebalancing of two-to-four VIX futures contracts depending upon the target maturity.

VIX ETP performance

Long vol performance: Strong gains during market shocks; difficult to buy and hold. Many VIX ETPs track variations of the S&P 500 VIX Short-Term Futures Index which measures the performance of a constant maturity one-month VIX future. That benchmark index is down 99.9% since December 2005, equating to an annualized return of -46.1% over the period. Its double-levered counterpart has had a median loss of -1.2% per day; -14.5% per month and an annualized return of -80.7% back to 2005 (Double-levered long: From 100,000 to under a penny since 2005).

Timely hedges have performed very well: While long vol strategies have been tough to hold for long periods of time; timely hedges can perform very well. We estimate that the S&P 500 has dropped by an average of 9% across the top ten calendar-month declines back to December 2005. One-month VIX strategies were up in each of the ten months by an average of +37.3%.

Short volatility performance: It stands to reason that if being long volatility has done poorly over time then the short has done well. The XIV and SVXY should have daily returns similar to the S&P 500 VIX Short-Term Futures Daily Inverse Index (SPVXSPI) which tracks the returns from selling a constant maturity 1m VIX future. The SPVXSPI was up 4364% from March 9, 2009 – March 31, 2017. The S&P 500 was up 249% over the same window. The inverse index has had an annualized return of 21.5% since December 20, 2005. The S&P 500 VIX Short-Term Futures Daily Inverse Index, which tracks the return of being short a one-month VIX future, was up 4364% from the market bottom on March 9, 2009 through the end of the first quarter of 2017. The index has had a median daily return of 0.6% and an annualized return of 21.5% since December 2005.

The best of times for short volatility: After vol spikes and slow markets with low realized volatility. Being short volatility often performs well after market shocks as volatility mean reverts and over periods of persistently low volatility and stable market returns. Exhibit 6 shows that the top 10 calendar month gains in the S&P 500 immediately followed many of the largest down moves. The S&P 500 gained 7.9% on average versus 21.6% on the 1m -1x index.

The worst of times for short volatility: Market shocks. Being short volatility can generate high returns but it can also suffer severe drawdowns. The index has shown a positive correlation of 0.82 and a beta of 4.3 to daily S&P 500 returns since 2013. The high beta means that drawdowns can be large when the market is down and the 1m -1x index was down -71.4% in 2008 and -45.5% in 2011. The worst calendar month for the strategy occurred in October 2008 when the S&P was down -16.9% but the 1m -1x index dropped -58.7%. The worst calendar day for the 1m -1x index occurred on June 24, 2016 (Brexit): The index was down 32.7% versus the S&P 500 down 3.6%.

Contango: Why long VIX ETP performance has been so poor

The VIX curve is said to be in contango when VIX futures are trading at higher levels than the VIX. The VIX curve has been in contango 70% of the time back to December 2005. ETPs which track one-month VIX futures strategies have lost an average of 83 basis points per day when the VIX curve has been in contango. That is not a fluke. We provide a quick mathematical illustration that contango leads to negative performance for long vol VIX ETPs all else equal.

VIX ETP scenario analysis: VXX returns if the VIX stays where it is We outline the methodology used to create many VIX ETPs. Once you understand the methodology, some simple, yet powerful, questions can be answered. For example, what are the potential returns on a VIX ETP over the next month if the VIX stays where it is and futures roll down the curve? The answer depends upon both the level and the term structure of implied volatility.

Compounding: A short is not the mirror image of a long

Most VIX ETPs are rebalanced daily, which means they track the one-day return of an index. By construction, that means that a long and a short can track their respective benchmarks perfectly each day but they may not be mirror images of each other over longer holding periods due to the compounding of returns. August of 2015 was an interesting example in a tough market. The benchmark for a one-month VIX futures strategy was up 71%, the inverse was down 48% and the double-levered was up 170%.

*  *  *

So that’s it America. Easy! Instead of buy-and-hold, it’s sell-vol-and-roll – what could possibly go wrong?



Source link