Professional gamblers understand the odds for every bet they make. When playing blackjack, they keep a running count of cards played in order to assess whether the remaining deck favours them or the dealer.
Similarly, investors would improve their performance if they knew when their strategies were more and less likely to pay off. In the January issue of Advisor’s Edge, we found that the worst stock market losses were greater than the best stock market gains. By eliminating both the best and worst days, the return for the remaining periods would be higher.
Chart 1 shows daily returns for the S&P/TSX Composite Index from June 30, 1992 to Oct. 31, 2017. Removing the 10, 50 and 100 best and worst daily returns would have resulted in higher average compounded returns compared with the index return of 9.48%: by 0.50% annually for the 10 best and worst days, by 1.51% for the 50 best and worst days, and by 2.48% for the 100 best and worst. Eliminating the 550 best and worst days would have yielded a return 5.55% per annum higher than the benchmark.
The distribution appears normal, but it is actually negatively skewed, meaning there are more extremely negative returns than extremely positive returns. The difference in average returns for the top and bottom 10% was -0.69% (see Table 1 farther down).
SOURCE: PÜR Investing Inc.
How to increase returns
Like card counting in blackjack, avoiding more best and worst return days leaves the stock market stacked in the investor’s favour. But how can these tails be reliably identified and monetized?
The tails represent periods of higher volatility. Sixteen of the worst 20 days and 16 of the best 20 days occurred between September 2008 and June 2009, a period that experienced sustained volatility more than two times the average. Clustering of highest and lowest returns is a characteristic of volatile markets. Eliminating both accomplishes two useful things:
- portfolios demonstrate lower volatility, thus making compounding returns easier (volatility drag is reduced); and
- remaining average portfolio returns are higher.
Maintaining a constant level of risk (standard deviation) reduces exposure to volatile tails. De-risking a portfolio when market volatility rises reduces exposure to extreme declines. The most important lesson from this exercise is to avoid large declines. Maintaining consistent risk levels is one way to do it.
Eliminating only the negative returns and keeping the positive ones would be ideal, but trading costs would be extreme because large gains and losses are clustered at the tails.
Table 1: The negative skew
|S&P/TSX Composite||June 1992 to October 2017 average return|
Source: TMX Group
Sensitivity to market changes determines how successfully additional returns can be harvested. Because volatility (standard deviation) tends to mirror market prices, casual observers dispute its usefulness as an indicator (see “Is it time to nix the VIX?”). But volatility is persistent and therefore somewhat predictable. Like card counting, a simple idea needs discipline, dedication and a strong statistical foundation.
Volatility clusters, outliers and data noise benefit from statistical adjustment. A moving average of market fluctuations benefits from exponential weighting or application of a generalized autoregressive conditional heteroskedastistic (GARCH) model, whose definition is beyond the scope of this article. We prefer the GARCH model.
The method we have described takes market changes and actively reorders the portfolio’s return distribution to favour the investor, rather than assume or hope that the market will reward strategies like growth or value over time.
Professional gamblers know that, over time, the casino wins. Controlling the odds, whenever possible, is always the better bet.