A PDF of this commentary can be downloaded here.
- Dalbar studies tell us that investors often sell after losses and wait for markets to reclaim high water marks before re-entering – behavior that is guaranteed to lead to underperformance
- Other, more dynamic, approaches may help investors achieve their dual goals of participating with market growth while managing capital loss
- While nobody has a crystal ball, pretending to have had a crystal ball over the last decade can tell us something about how we should behave
- Losses happen; we should remain de-sensitized to losses in bull markets in effort to participate with growth
- Whipsaw can be worse than market losses; being overly sensitive to losses can compound drawdowns
- Each bear market is unique: being highly sensitive to losses may be the optimal strategy in one bear market and the worst strategy in another
Q2’s end-of-quarter Greek drama served as a great reminder that the market can, and often will, move violently. And, when it does, often the best action is inaction.
Unfortunately, as investors, inaction during losses seems to go against every emotional fiber we have in our bodies.
This preference for action over inaction can be highly detrimental to long-term performance results. According to Dalbar’s “Quantitative Analysis of Investor Behavior,” investors often sell after they have big paper losses and then sit on the sidelines until the market has recovered.1 This approach mathematically guarantees lagging returns because investors are realizing their losses during declines and missing out on gains during market growth.
For example, if we sell after every 5% dip and wait for the market to recover, the market may return to parity, but we’ll only be sitting on 95% of our wealth. In other words, we’ve paid a 5% premium for protection against any losses greater than 5%.
Below I’ve (poorly) drawn the basic idea behind this strategy with a hypothetical market scenario to highlight how while it may help avoid losses, and even create outperformance during those periods (green shaded areaa) it eventually locks in underperformance (red shaded area).
Just how detrimental can this behavior be? Below I’ve created the equity curves for this strategy, assuming different thresholds of losses that an investor will sell after. The results, not surprisingly, are catastrophic:
Any cursory analysis of such a strategy would lead to its quick dismissal of it being a valid method for protecting capital: the asymmetric nature of participation in losses versus participation in growth guarantees long-term underperformance.
While investors care deeply about capital preservation, participating in growth is also mandatory for outpacing inflation.
Frankly, based on this replication of the method outlined in Dalbar’s study, I find the behavior to be so illogical that I struggle to believe that investors actually behave this way.
Perhaps a bit more realistic scenario would be investors selling after a decline and waiting for the market to recover back above the point they sold at.
Below I’ve drawn how we can expect such a strategy to behave. We can see that ideally we are able to avoid significant losses, but can begin participating again without a significant performance sacrifice.
Unfortunately, the simulated equity curves don’t quite live up to this ideal behavior. Whipsaw and entry/exit timing end up playing a fairly significant role.
The results did not improve much over the original strategy of reentering at the previous high water mark. Those strategies with increased sensitivity (quicker triggers to sell) spend more time on the sidelines during market growth; those strategies with less sensitivity are able to keep pace with the market during growth, but fail to protect against any significant losses.
Part of the problem with such an approach is that we’re not able to take advantage of being right in our bear market calls. We may improve our risk-adjusted performance by avoiding the downside, but by waiting for the market to rally back to the point at which we sold out means we mathematically cannot outperform the market on a total return basis.
Perhaps a better concept may be to sell after the market has fallen X% from its high water mark and re-purchase after the market has rallied X% from its low water mark. For example, if the market is at 10,000 and drops to 9,000 (10% loss), it only needs to recover to 9,900 (10% gain) before we get back in. This may allow us to avoid significant market losses and re-purchase near lows.
We’ll call this approach the “symmetric strategy” and it shares similarities with a traditional momentum approach, reacting to market weakness by selling and waiting for confirmed strength before re-entering.
Unfortunately, reality fails to live up to the platonic ideal of how the strategy should work.
We can see that buy & hold still largely out-performs, with the exception of a single outlier strategy, which is more likely to be an artifact of data-mining than a robust and reusable result.
The problem is that while the more sensitive strategies offer protection, they have little market participation, and while the less sensitive strategies participate, they fail to meaningfully protect. Ultimately, the threshold level that works in one market environment eventually struggles in another.
Changing market dynamics means that there is no magic threshold level that will allow us to always avoid losses and significantly capture growth. The optimal threshold value will be variable both over time and for each asset class we evaluate.
What can be illuminating, however, is looking at how these optimal threshold values change over time.
To do this, we will assume we have a crystal ball that lets us peer into the future. At the beginning of each month, we’ll look into our crystal ball and invest in either the Symmetric 5%, 10%, 15%, or 20% strategy that does best over that month. Because we are allocating flexibly among the strategies, we are indirectly controlling the X and Y values over time.
But why is assuming to know the future helpful? Understanding our optimal behavior when we do know the future can educate us on both the behavior we should adopt when we don’t as well as enlighten our understanding of performance limits we can expect from such an approach.
Below are the allocations for our optimal strategy. Those allocations above the zero line mean the strategy we selected for the month was invested in the market on that day. Below the zero line means the strategy was out of the market.
As is generally the case whenever we use a crystal ball, the strategy performance dwarfs buy & hold.
But what lessons can we learn from this exercise of using our knowledge of the future to identify the optimal threshold value?
- In bull markets, don’t be so sensitive
In the raging bull markets from 1993-2000, 2003-2007 and 2009 through present, the reigning strategy was that with the biggest buffer to losses: the 20% threshold. In other words, increasing sensitivity in a bull market will only lead to whipsaw and missed opportunities.If we care at all about returns, participating in the market is critical.
- Each bear market is unique
The rapidly changing allocations we see in the tech crash and the credit crisis are our crystal ball taking advantage of high volatility mean-reversion opportunities. This occurs because we built our strategy to optimize for total return, always selecting the strategy that gives us the best return over the next month.What we can look at, however, is what occurs in the underlying strategies during each of the bear markets. What we can see is that the strategy that worked in the tech crash – namely, being highly sensitive to changes – was the worst strategy to take during the credit crisis.
So while “be robust” – or relatively insensitive – can be broadly said to be the right strategy in bull markets, a dynamic approach is required to successfully navigate a bear market.
- Even with perfect foresight, drawdowns still happen.
Even though we used our crystal ball to look into the future and pick the strategy that maximized our returns each month, the strategy still ended up suffering a 17.31% max drawdown.It cannot be overstated that this large a drawdown still occurred despite the fact that our process involved knowing the future.Tapping into market growth requires market participation. This, unfortunately, often means stomaching losses in the short-term to access opportunity over the long run.
- Sometimes you have to ride out losses to avoid whipsaw
The max drawdown in the strategy occurred during the August 2011 market sell-off.So why was the model not invested in the Symmetric 5% strategy, which would have gotten out of the market faster? In a word, “whipsaw.”Investing in the Symmetric 05% strategy subjected the investor to quite a bit of mistimed buying and selling, resulting in a greater loss than if the investor had just stayed invested over the period.
Similar to our take away with lesson three, August 2011 teaches us that losses happen, even in bull markets. Worse than losses, however, is an overly sensitive reaction that can only serve to compound losses further.
Newfound was founded on the simple, but powerful premise that investors care deeply about capital preservation. Nevertheless, we also understand that capital growth is only possible with market participation. Emotional responses to market losses – proxied in these tests by overly sensitive buy and sell triggers – only serve to compound losses.
As we have seen, static entry and exit rules may outperform during one period but severely lag during the next market regime. In bull markets, it pays to be robust to losses, willing to suffer through short-term sell-offs to access long-term growth. In bear markets, our response must be dynamic: sensitive in some markets and robust in others.
Constantly evolving markets require us to adapt to the new environments as they arise, but regardless of what environment we are currently in, an objective, unemotional process is critical to long-term success.