On June 16th, our bear market indicator signaled that we had reached a point that historically, since 1943, has always allowed us to conclude that the bear market was behind us. This demarcation for the strategy is important because it completely changes the behavior of it, reverting it back to its real mode: a data-driven algorithm. I've previously discussed the bear signal; the purpose of this post is to provide instead a recap of how the strategy performed in bear mode since the beginning of the year, and more importantly, to discuss what to expect from the strategy going forward, now that we're back in bull mode.
2023 Recap: End of the Bear Market
The S&P 500 hedging signal has been live since mid-August 2022 although we only launched it in April. Apart from some bug fixes, it underwent a fundamental update in early January of this year. This update was designed to make the conditions for buying back into a bear bounce more restrictive to avoid re-entering the market ahead of a highly anticipated event. While the previous strategy required the VIX to be under a certain level and to have cooled down significantly, the new strategy also considers the structure of the VIX. When the market becomes scared of a given near event such as a CPI report, or Powell's speech, the VIX9D usually reacts strongly as close range bets in the option market become affected. This is an effective method to quantitatively measure subjective market sentiment and the January update will prevent to re-enter the market if the VIX9D is considerably higher than the VIX1M and VIX. This improvement had no impact in 2023, as it only affected three trades that occurred in the past.
So far this year, the results of the S&P 500 strategy have exceeded my expectations by capturing any significant moves. After re-entering the market on January 11th, it sent us two bear sell signals. The first one occurred during the sideways movement at the peak of the January move. This hedge signal protected us just in time for the regional bank crisis, capturing an impressive short. It then rapidly unhedged on March 16th for another upward move. We were then stopped out of that move on April 11th. While the NASDAQ consolidated for about a month before beginning an upward trend, the S&P 500 moved sideways for an additional two weeks. The strategy just re-entered the market in time to capture this new uptrend that brought us out of the bear market on June 16th. Here's an image of the moves as captured by the signal.
From January 11th to June 16th, someone riding the strategy on SPY by flipping from long to short would have yielded a return of 12.12%, or 10.92% if changing from long to neutral, compared to a buy-and-hold return of 9.43%. Playing the same signal long/short on UPRO would have resulted in an impressive 53.64% (or 42.43% for long/neutral), compared to a buy-and-hold return of 30.29%. Naturally, since the overall trend in 2023 was up with only one real dip, we couldn't expect to outperform buy-and-hold by an incredible margin, but the interesting thing is that we still managed to outperform it while maintaining peace of mind anytime the market experienced some turbulence. Indeed, this is where I think using a hedging strategy shines the most. If it can generate alpha, that's a very good thing, but the main benefit is that it can give us the confidence to ride the market on high beta without the fear of being caught in a pullback that could wipe out our account. The great thing is, if this allows us to be invested in leveraged ETFs while always being positioned with the trend, this can transform what people often refer to as the infamous decay of leveraged ETFs into an additional compounding effect. Indeed, leveraged ETFs promise to track a given index with a certain beta on a daily basis, within a certain margin of error. The fact that this return is tracked on a daily basis can create significant decay over a certain period when the market begins a downtrend or goes sideway. You can read the section of our ebook dedicated to leverage ETF if you want to learn more, but to keep it simple, the signal allowed us to achieve 5.68 times the SPY buy-and-hold value over a six-month duration. This is an outstanding performance that surpasses the 3X daily leverage of UPRO. The strategy had a hit rate of 80% for long/short (and 100% for long/neutral), which aligns with our expectations even though it's higher than the backtested success rate of around 75%. The reason I say it aligns with our expectations is that the success rate is not uniform over time, and I expect a higher success rate in the bear market bounce mode which is built to be restrictive. More importantly, the only failed short resulted in a 1.2% loss on SPY and a 1.35% loss on UPRO, within the range we could expect from the backtested highly asymmetric win-to-loss ratio. One note: it may seem strange that using SPY or UPRO resulted in a very similar loss. The reason for this is linked to the path dependency of inverse ETFs, as this "hedged" signal did see some red candles that initially created more profit on the leveraged short. Here is a table containing more stats for the period between January 11th and June 16th (end of the bear market per our criteria).
Note that these data relate to our signal, not our own results. We recently inaugurated our new WealthUmbrella Fund (WUF), and transferred our funds, which we hope to have audited in the future for the next fiscal year. This makes the exact computation of our return somewhat challenging at this moment. We were in UPRO for all these events except when we took the May 25th signal on a beta equivalent to SPY. A quick calculation tells me that we are at around a 30% return on our UPRO/SPXU play. I will aim to obtain the exact figure for the launch of our new method of tracking our portfolio.
What to expect now ?
Now that the strategy is out of the bear market, it will revert to its true nature: a data-driven algorithm. The bear market mode was constructed around the premise that the risk lay in being in the market. This meant that in this mode, the strategy was very restrictive regarding any market bounces, generally opting to stay out of the market. Most conventional data sources, prone to misinterpreting the timing of these bounces, were deactivated in this environment. Instead, the strategy was steered mainly by price action-driven metrics (Phase angle, VIX).
However, in bull mode, the strategy takes the exact opposite viewpoint. In this scenario, the risk is in being out of the market. As a result, you should generally expect the strategy to be much more forgiving of minor dips. Indeed, it's essential to remember that the strategy was designed to overlook most of the minor dips encountered during a bull market, preferring to stay in the market and only exit when a major correction is imminent. This is because most dips in a bull market are usually very shallow, rarely exceeding 6-8%, and often fall within the 4% range. Since it usually takes 1-3 days to distinguish a downturn from a red candle, it's often too late to exit and then reenter at a profit. I have worked on versions of the strategy that flag mid-sized dips. The Phase Angle (Threshold at -0.17 and 0.01 on QQQ) was an excellent tool for this, but I found that the extra alpha delivered by this strategy was minimal compared to the increase in the number of signal it generated.
That being said, it's not because I would like to ignore all small dips that the strategy will never get caught flagging one. In fact, I've tried everything possible so far, and nothing I've tested has allowed me to completely ignore all of them. A prime example of this is the recent Hedge signal we had on January 16th. This signal was triggered by a very high peak of short selling on the DarkPool that turned out to be a very shallow move. However, this doesn't mean that these shallow drops, which we will capture against our will, will result in losing trades, as evidenced by this recent successful short. But some of them will, though they should remain close to breakeven.
In summary, although the recent hedge signal provided by the bull part of the strategy flagged a very minor dip, most of the time expect these small dips along the bull market road to be simply ignored. To give you an idea of what I mean, here is an image of the period from 2020 to 2022. Although the strategy quickly and effectively captured the Covid drop and the onset of the 2021-2023 bear market, it only triggered during the pullback in September 2020.
While it may seem easy to hold one's position for the duration shown in this chart, I remember the emotional impact of some of the red candles during that period. It can feel quite alarming, especially if you're riding the market on leverage. The main difference between these small dips that we ignore and the larger ones are actually hidden in the undercurrents of the market. Indeed, while the Covid drop triggered a signal on February 20th, 2020 and three others on February 24th, and four signals between November 2021 and January 2022 triggered the hedge strategy, some of these smaller dips were very healthy according to all datasets. This image, which I've shared before, shows one of the most important indicators of our strategy (that I must keep confidential). For me, this is a good example of how data can differentiate between healthy pullbacks and more threatening ones.
Over time, I will keep you informed about what our data indicates as we remain invested through regular pullbacks, to help you understand why we aren't selling.
At the moment, every one of these metrics is showing an incredibly healthy reading (including the one above, as you can see). The only one that currently concerns me is the skewed option distribution in the market. This suggests that we should expect a healthy pullback in the coming weeks, if not months, which would bring down this reading to around 125-135 (it's currently at 150). Such a pullback would be normal after the good start we've had this year. This could turn out to be an excellent buy-the-dip opportunity and might finally lead to a broader rally.