The credit crisis has rekindled interest in collars and protective strategies in general, and for good reason. In 2008-09, the PowerShares QQQ exchange-traded fund (ETF) experienced a drawdown of roughly 50% from peak to trough. Even diversification, which many rely on to be a strong guard against portfolio-wide losses, didn’t protect as expected.
Many asset classes that are generally considered effective equity diversifiers also faced significant losses. Correlations of most asset classes with broad equity indexes tended to be significantly higher in 2007 and 2008 than in previous years, negating much of the expected benefits of diversification. This type of contagion across asset classes suggests that in times of major systemic stress, direct hedges through protective options strategies provide equity portfolios with more benefits than standard diversification programs.
A new study has found that a long protective collar strategy using six-month put purchases and consecutive one-month call writes earned far superior returns compared to a simple buy-and-hold strategy, while reducing risk by almost 65%. The study (“Loosening Your Collar: Alternative Implementations of QQQ Collars”) was conducted by Edward Szado and Thomas Schneeweis at the Isenberg School of Management’s Center for International Securities and Derivatives Markets at the University of Massachusetts.
Szado and Schneeweis evaluated more than 10 years of data on the PowerShares QQQ exchange-traded fund (Ticker: QQQQ) and its associated options. They also extended the analysis to a more active implementation of the strategy. While the passive collar used a constant set of fixed rules, the active collar used rules that adapt the collar to changing macroeconomic variables and market conditions. The active collar implementation generated higher returns than the passive implementation, while volatility was only slightly higher. Over the 122-month study period, the passive collar returned almost 150%, while the QQQ lost one-third of its value. The active collar outperformed both strategies and returned more than 200% (see “Active or passive”).
KNOWING ALL OPTIONS
A variety of options strategies provide capital protection for equity-based portfolios. The most obvious choice is the use of protective puts. However, protective puts tend to be relatively expensive, especially in periods of high volatility.
Another options-based approach is the buy-write or covered call strategy. The covered call strategy entails the writing of call options against a long underlying position, typically at a one-to-one ratio. Several empirical studies have suggested that covered call writing can enhance returns as well as mitigate losses from market downturns. However, covered call writing still leaves an investor exposed to large down moves.
The collar strategy essentially adds a long protective put to a covered call strategy. This addition provides significant downside protection, which the covered call lacks. The purchase of the long put is financed by the sale of the call. In essence, the collar trades upside participation for downside protection. A tight collar provides less upside participation and more downside protection than a loose collar. At one extreme, the tightest collar using at-the-money (ATM) puts and calls effectively immunizes the portfolio from market movements. At the other extreme, a loose collar uses far out-of-the-money (OTM) puts and calls. Between these far OTM strike prices, the collar is essentially equivalent to a long underlying position.
The research assessed the effectiveness of the passive and active variations of the collar strategy from March 1999 to May 2009. The analysis considered a number of implementations of long collar strategies with varied moneyness of the puts and calls, as well as times to expiration. In addition, the collars’ performances were analyzed with the time period segmented into three sub-periods. These sub-periods feature different market environments reflecting conditions generally favorable and unfavorable to a collar strategy.
The protective collars significantly outperformed the QQQ in the overall period, as well as in the periods covering the technology bubble and the credit crisis. While the collar variations underperformed the QQQ in the interim period between 2003 and 2007, in all of the implementations in all time periods, both collars significantly reduced risk compared to the buy-and-hold strategy. The study further indicated that the collar variations using six-month put purchases outperformed the one-month and three-month put strategies in almost all measures. The slower time decay of the longer maturity six-month puts was a significant benefit to this collar implementation.
Rolling standard deviations clearly showed the risk-reduction benefits of the collar strategy (“Rolling your risk” ). The collar strategies exhibited lower standard deviations throughout the entire period, with the differences ranging from five to 45 percentage points. It is also worth noting that the risk-reduction benefits of the active collar strategy over the passive collar tended to be relatively subtle, particularly when compared to the difference between the collars and the QQQ.
The passive strategy using the 2% OTM six-month put and 2% OTM one-month call generally exhibited the best performance and represented a middle ground between ATM and far OTM. This strategy uses fixed rules and does not allow for flexibility in different market conditions.
During the April 1999 to September 2002 period, the QQQ was extremely volatile and lost more than three-quarters of its value from peak to trough. Specifically, the QQQ had an annualized loss of 23.3% with a staggering 42% volatility. In contrast, the passive collar strategy generated an annualized positive return of 21% at a volatility of only 13.7%.
The collar was able to turn a sizeable loss into a significant gain while at the same time reducing risk by more than two-thirds. The capital protection ability of the collar strategy truly shines in this case. The collar could have earned investors a very impressive 21.2% per year over the sub-period with a maximum loss of capital of 7.5%, regardless of how poorly investors timed their entry into the strategy. The collar was an effective way of capturing a significant return from the bubble run-up without facing the tremendous losses that came with the collapse.
In the period between October 2002 and September 2007, steady positive returns, low volatility and few sharp down moves of the index explain why the collar strategy was expected to perform relatively poorly. The annualized return of the QQQ over this period was an impressive 20.4% at a relatively moderate volatility of 17.5%. The collar only provided a 5.2% annualized return over this period. It did, however, do so at a far lower volatility. Nevertheless, this under-performance was not nearly as significant as the QQQ’s under-performance in the earlier period.
While most asset classes became more correlated and collapsed during the credit crisis period from October 2007 to May 2009, the collar again provided significant capital protection. The annualized loss of 19.8% in the QQQ was reduced to a loss of only 1.4%, while the standard deviation was cut from 29.2% to 11.6%. Again, the strategy provided capital protection where the implementation was not a cost but rather additive to the returns of the underlying position.
The active implementation of the collar strategy used three different sets of signals: momentum, volatility and a macroeconomic indicator (see www.optionseducation.org/institutional for a complete discussion of the active market signals). Various time horizons were reviewed in the full study. Because results based on the shorter-term signals were superior to alternative combinations, they are summarized below.
Changes in the signals are incorporated into the strategies only on roll dates. Because puts tend to be more expensive than calls for a given level of moneyness, the active strategies begin with puts further OTM than calls. This allows for the initial construction of the option component of the strategy to be close to zero cost. The initial moneyness of the puts and calls is set to 3% OTM and 2% OTM, respectively.
From this initial point, the momentum signal will widen or tighten the collar by increasing or decreasing the OTM amount. The macroeconomic signal will shift the collar up by increasing the OTM amount of the calls and decreasing the OTM amount of the puts, or shift the collar down by moving the strike prices in the opposite direction. The volatility signal will determine whether the ratio of the call write will be neutral, overwritten or underwritten.
The momentum signal is a simple moving average crossover (SMACO) of the Nasdaq 100 index. A SMACO compares a short-term moving average (SMA) and a long-term moving average (LMA) to determine whether an upward or downward trend exists. This summary highlights the shorter 1/50 day moving average crossover on each roll date (vs. a 5/150 day moving average for medium-term and 1/200 day moving average for long-term) to determine whether to widen the collar (increase the upside participations with a corresponding reduction in downside protection) or tighten the collar in response to a sell signal (increasing downside protection, while reducing upside participation).
The volatility signal uses 50-day moving average of the daily Vix close of as an indicator of implied volatility levels (vs. 150-day SMA for medium term and 250-day moving SMA for longer-term variations). The CBOE Volatility Index (Vix) is a market estimate of expected volatility of the S&P 500 index calculated by using bid/ask quotes of near-term and next-term OTM SPX options with at least eight days left to expiration, weighted to yield a constant, 30-day measure of the expected volatility.
The strategy writes 0.75 calls to each long index position when the markets short-term anxiety level is high (as indicated by a situation in which the Vix is more than one standard deviation Bollinger band above its current 50-day moving average level), and writes 1.25 calls per index position when the anxiety level is low (when Vix is more than one standard deviation Bollinger band below the 50-day average level). When the one-month implied volatility level is within the one standard deviation, the strategy follows a standard 1:1 ratio buy-write. The goal in varying the quantity of written calls is to have a longer exposure to the market in times of high anxiety and shorter exposure in times of complacency.
The macroeconomic signal is based on the trend of initial unemployment claims and the state of the economy with respect to the business cycle as pronounced by the National Bureau of Economic Research. These announcements are generally considered the authority on the current state of the business cycle. Because there is often a significant delay in announcement dates, the authors base the signals on announcement dates to avoid hindsight biases. During expansionary periods stocks rise, counter-intuitively, on bad unemployment news, while the opposite relationship holds in contractionary periods.
Most typically expect rising unemployment to negatively affect stock prices regardless of the business cycle, but the authors relied on the existing literature that suggested that rising unemployment in expansionary economies causes expected future interest rates to decline, increasing the value of equities, while rising unemployment in contractions indicates slower future earnings growth rates, reducing the value of equities.
This summary highlights the results of the shorter SMACO unemployment signal using 1/10 weeks (vs. 1/30 week for medium-term and 1/40 week for long-term). Because a rising unemployment claims number in an expansionary economy is a bullish stock market price and volatility signal, and if the SMA is greater than the LMA, the authors shift the collar toward the ATM put and OTM call (increasing both strike prices). In contractionary economies, rising unemployment claims would cause a shift of the strike prices in the opposite direction.
IT PAYS TO BE ACTIVE
The active collar outperformed both the QQQ and the passive collar. While the volatility was slightly higher for the active collar than for the passive collar, annual returns of the active exceeded those of the passive by more than two percentage points. On the other hand, maximum drawdown and minimum monthly returns were slightly higher for the active collar.
In the first period covering the technology bubble, the active collar significantly outperformed the passive collar, generating almost a one-third higher annualized return, while standard deviations were essentially identical. In the second period, the active collar mitigated the under-performance of the passive strategy. Annualized returns improved from 5.2% to 6.7%, while volatility was slightly reduced.
While the improvements of the active strategy were certainly appealing in the favorable period, they were even more appealing in this period in which both collars underperformed the QQQ. The credit crisis period was the only period in which the active strategy underperformed the passive strategy (albeit only slightly) but still significantly outperformed the QQQ. Annualized losses of active vs. passive collars were increased from 1.4% to 3% and standard deviations were increased slightly from 11.6% to 13.7%. These results suggest that even a conditionally implemented strategy had difficulty reacting optimally to the extraordinary events of late 2008. A truly active approach with minimal time lags in its inputs may have been able to overcome the small performance deficits between the passive and active strategies. Nevertheless, the active collar still provided significant protection from the annualized 20% QQQ loss.
The 10-year time horizon since the introduction of the QQQ options has provided a variety of market conditions for testing the performance characteristics of collar strategies. The collar underperformed the QQQ in the strong market climb of October 2002 to September 2007. However, over the entire 122-month period and in the sub-periods around the technology bubble and credit crisis, the collar strategies significantly outperformed a buy-and-hold strategy and provided much needed capital protection.
Phil Gocke is managing director of The Options Industry Council. The Options Industry Council (OIC), which provides education and research to institutional investors, helped sponsor the CISDM study of the performance of a collar strategy on the PowerShares QQQ exchange-traded fund (ETF).