Welcome to the fourth issue of the sixth volume of The Journal of Investment Strategies. This issue is truly diverse, featuring a paper on life-cycle investing in wealth management, an investigation of backtesting techniques, a study of using options in an enterprise management context, and, finally, a study of the role of leverage and uncertainty in defining optimal capital allocation.
In the first paper of this issue, “Life-cycle investing with the profitable dividend yield strategy: simulations and nonparametric analysis”, Wai Mun Fong presents a rigorous study of and a novel solution to the long-standing life-cycle investment problem in wealth management. The problem is to define an intertemporal capital allocation rule for a person who starts investing while young, and then reaches retirement age, eventually consuming the invested assets in retirement. An often-posited heuristic solution is to skew the investments heavily toward the equities in earlier years, then gradually move the allocation toward balanced as people reach retirement, and finally toward more bond-oriented portfolios during the post-retirement period. While it is a commonly accepted practice in wealth management as well as a foundation for a large target date fund industry, this solution style has not been demonstrated to be optimal, and there have in fact been many studies showing that one would do better by following other approaches, including passive buy-and-hold pure equity investment.
Wai Mun Fong correctly identifies profitable dividend-yielding stocks as suitable candidates for a life-cycle investment portfolio because of their combination of growth-and income-generating capabilities. The paper demonstrates that, given the history from which we could draw statistical inferences, such a dividend-and profitability-oriented portfolio strategy as part of life-cycle investing would produce superior results with high statistical confidence. I tend to agree with the conclusions of the paper inasmuch as the characterization of the relative performance of the passive and dividend- and profitability-oriented portfolios is very much in line with known market anomalies including quality, low-beta and high-dividend excess returns. All of these anomalies are well established, and at least some of them are believed to have structural underpinnings that would make them persistent. The main lesson I draw from this paper is that commonsensical wealth management approaches such as life-cycle investing do have their place, but that it is worthwhile going deeper and not just accepting them blindly; one should ask what specific investment approach might be best suited to such a context.
"An uncertainty quantification framework for the achievability of backtesting results of trading strategies” by Raymond Hon-Fu Chan, Alfred Ka-Chun Ma and Lanston Lane-Chun Yeung focuses on a core problem for most quantitative investment strategists: how to estimate the implementation achievability of backtesting results. This is indeed a critical question. On the one hand, almost all quants rely in one way or another on backtests; there is not much choice in that, as we lack a reliable model of true market behavior, so relying in some form on the actually observed past is often the closest approximation one gets to the truth. On the other hand, it is a well-known problem that many strategies that look good in backtests fail to deliver equally good results in practice. There are two distinct reasons why this can happen. The first reason is that the backtest was not constructed in an appropriate manner, ie, some sort of data snooping, overfitting or other such quant heresy has been committed. This is usually a methodological problem, and one should deal with it by following best practice in order to avoid such pitfalls. However, even if one has been extremely careful to avoid overfitting or data snooping, this still does not guarantee that the backtested strategy characteristics will actually materialize in the future. The second reason is a reliance on the assumptions on implementation of the strategy, such as when and at what prices one could actually invest and divest when trading it.
The authors focus on this second source of potential backtest failure in their paper, and specifically on the impact of possible delays and execution price uncertainty. They introduce the concept of return-at-risk and demonstrate how this expanded view of backtesting reveals that certain technical trading strategies might indeed have a disconnect between the backtest and expected future performance. Their approach is less detailed than the extensive literature that focuses on market impact and trading slippage as a result of market microstructure, but it nonetheless presents an interesting and practical way to estimate the viability of backtests, which remain, without a doubt, a useful tool in the quant arsenal.
In our third paper, “Enhancing enterprise value by trading options”, Dilip B. Madan and Yazid M. Sharaiha consider the impact of systematic options trading/positioning on the enterprise value of the holder. The setting of the problem is quite different from the more-often-asked question about the attractiveness of similar strategies in stand-alone comparison with cash or passive benchmarks. The enterprise value is sensitive not only to the added value but also to added risk, and this is what drives the results of the paper. After a rigorous definition of the model, the authors demonstrate that it is indeed possible to formulate a systematic options overlay strategy that would add value, based on the definition of risk charges from Madan’s two-price economy expanded framework.
In “Leverage and uncertainty”, the fourth paper in this issue, Mihail Turlakov undertakes a conceptual study of issues of uncertainty and leverage as they relate to optimal capital allocation rules. Specifically, the author extends the Kelly criterion approach to a case where an additional fat-tail risk is driven by uncertainty, ie, a fundamentally unknowable risk rather than the potentially knowable risk of market dynamics. By starting from the Kelly optimal portfolios, the author demonstrates that one can span the range from conventional Markowitz portfolio rules to risk-parity portfolio rules by simply varying the degree of uncertainty embedded in the dynamics. Further, a notion of Kelly parity is introduced as a potential alternative definition of the balanced portfolio allocation principle. Those readers who prefer the Kelly approach and intuition to the more conventional mean–variance approaches, including Markowitz and risk parity, will be interested to find that there is a commonality between all of them.
In conclusion, I hope that this issue will please readers with its diversity of topics and approaches. One of our guiding principles in The Journal of Investment Strategies is to help spread novel ideas across various industry and academic areas of expertise, and this issue is a good example of that.
Lifecycle investing with the profitable dividend yield strategy: simulations and nonparametric analysis
Using simulations, the author shows that life-cycle investing implemented on highly profitable and high dividend yield stocks (the profitable dividend yield strategy) provides a compelling solution to the suboptimality problem by leveraging on the…
An uncertainty quantification framework for the achievability of backtesting results of trading strategies
In this paper, the authors propose a framework for implementing and backtesting trading strategies.
This paper considers the problem of enhancing an investment activity by regularly adding an option trade to the portfolio mix and presented results for the single underlier of the S&P 500 index, with the underlying activity being either long the index or…
By extending the Kelly criterion to a simple probabilistic model with an additional tail risk outcome associated with uncertainty, this paper looks beyond risk and evaluates how uncertainty constrains optimal leverage.