The Dynamic Hedge Rule During Confused Market Environments by Mazin Jadallah, Alpha Clone
We’ve written extensively about our dynamic hedge, which is one of the features of our methodology that we often get asked about. Lately that discussion has been timely. The hedge has been triggered twice in just the past seven months, making it a good time to review the rule and our analysis of historical results.
New in our analysis {this time around} is a focus on periods when the hedge rule has yielded multiple hedge events in short succession – there have been seven since 1950. We isolate these “clustered hedge” events and look at whether their effect on performance is any better/worse than discreet hedge events on the one hand and a buy-and-hold S&P 500 strategy on the other.
For the uninitiated, the mechanics of our dynamic hedge rule are very simple. When the S&P 500 closes below its 200 day moving average at any month end, strategies that employ the hedge will vary their exposure from long only to “market neutral”. Subsequently, the hedge is removed completely when the S&P 500 closes above its 200 day moving average at any month end.
Exhibit 1 summarizes the basic statistics for the rule. There have been 46 hedge events since 1950 with an average hedge duration of 5 months. Most of the events (78%) are False Signal events which means that the hedge results in a losing trade. The average loss on False Signal events is 4.22 percentage points in total return over the duration of the hedge. Conversely 22% of the time the hedge rule yields a “Real Deal” event which means that the hedge results in a winning trade. The average gain from winning trades is 19.1 percentage points in total return and the average duration of a winning hedge is 14.1 months (about 5x longer than the average duration of a losing hedge).
Our dynamic hedge rule has historically paid off for long-term investors (benefits out weight costs) and looking at the historical statistics is certainly helpful but we know from experience that some hedge events are harder to live through than others. Specifically, when the hedge rule is triggered multiple-times in short succession it can feel like the rule is arbitrary and confused. Most important perhaps to investors, if the average losing hedge event costs 4.22 percentage points, how much does 2, or 3 or even 5 (yikes!) successive events cost in aggregate?
Exhibit 2 summarizes the seven historical periods where the hedge rule has triggered multiple times in short succession, each hedge instance less than six months from each other. Of the seven events, five, including most recently, are “double dip” events where the hedge was triggered twice in succession, one event was a “triple” and amazingly the period 1978-80 yielded 5 consecutive hedge instances, three of which lasted only one month! Interestingly, the average loss for each False Signal within clusters is 3.89 ppt., less than the 4.22 ppt. average loss for all False Signal events.
Dynamic Hedge Rule
But what about the cumulative effect of holding a hedged strategy through these clustered periods? To get at the answer, we compare a strategy that holds the S&P 500 from the date of first entry in each cluster to the date of last exit with a strategy that is long the S&P 500 but relies on the dynamic hedge rule to adjust exposure. We also beta adjust the hedged strategy by a factor 1.13 which equals the beta of the longs in the AlphaClone Hedge Fund Downside Hedged Index since the index went live on October 11, 2011. An example might help. Let’s assume that on a given day the daily return for the S&P 500 is 0.5%. We derive daily returns for our portfolios as follows:
- Buy and hold daily return = 0.50%
- Hedged portfolio daily return if hedge is off = 0.50% x 1.13 = 0.57%
- Hedged portfolio daily return if hedge is on = (0.50% x 1.13) – 0.50% = 0.07%
Exhibit 3 summarizes the total returns for both portfolios over the period that spans each cluster. The opportunity cost of utilizing the hedge on the beta adjusted portfolio is 1.78 total return ppts., significantly less than the average cost across all False Signal events of 4.22 ppts. This makes intuitive sense since the hedged strategy runs long only in between hedged instances where the probability is higher that the strategy’s beta will benefit, not harm, returns. And as we’ve stated above, the average False Signal cost in this sample is already lower.
Bottom Line
The cost to investors during False Signal hedge events, whether single or clustered, is outweighed by the benefit the hedge yields during Real Deal events. Clustered hedge events may feel more confused than the one-and-done variety, but they simply are reflective of the reality of the market at the time.
