Welcome to the second issue of the fourth volume of The Journal of Investment Strategies. We are happy to present four papers devoted to the modeling of asset return trends and volatilities, bond yield curve strategies and a discussion of how to construct an investment strategy that navigates risk cycles.
In the first paper of the issue, "Does Google Trends data contain more predictability than price returns?", Damien Challet and Ahmed Bel Hadj Ayed ask a question that many in the industry have speculated about: namely, might a careful analysis of social media, such as Google Trends, help us predict stock returns? To answer this question accurately, the authors carefully examine various types of biases that such measurements could be subject to, and come up with a clean metric for predictions based on Google Trends. They find that, while these trends can be used to generate positive returns, the scale of these returns and their predictability is not any better than if they simply used past price data. While it might be somewhat discouraging that the authors did not uncover a "diamond in the rough", I must say that their findings chime with my own perception, which is that Google Trends are, by definition, a result of mass population interest and that such interest is more naturally coincidental with (or follows) real news and insight rather than preceding it.
In our second paper, "A supply-and-demand based price model for financial assets", Takashi Kanamura presents a model of financial asset prices, returns and volatilities using the fundamental supply-demand modeling approach. The nontrivial angle- shying away from either purely empirical or purely stochastic-analytic approaches- is interesting as it allows the author to tackle head on the causes of certain wellknown stylistic facts about asset returns and volatilities: primarily the asymmetry of the volatility-return relationship. Moreover, the author is able to distinguish between the pure leverage effect and the feedback effect, which are indeed known to have different signatures empirically. I would recommend this paper to all those who study volatility dynamics across asset classes, especially if they keep an open mind regarding interpretation of the model parameters.
In the issue's third paper, "Bootstrapping the relative performance of yield curve strategies" Razvan Pascalau and Ryan Poirier investigate the oft-used yield curve factor modeling approach in application to building effective bond trading strategies. They demonstrate how a choice of parameterization affects the error bounds in the statistical estimation of the factor dynamics. Selecting the "best" such parameterization, they also study the effect of parameter comovement, and conclude that it indeed results in the tightest in-sample and out-of-sample fits. Finally, they study the effects of stock market shocks on the performance of the yield curve fits and corresponding bond bullet and barbell strategies.
Joakim Agerback and Tor Gudmundsen Sinclair have written a very interesting discussion paper, "Navigating risk cycles", for our investment strategy forum, showing the importance of risk cycles on the performance of conventional investment strategies. Rather than building a complex model of risks and returns, the authors follow a pragmatic empirical approach, demonstrating the existence of highly visible cyclical patterns in volatility (specifically, in equity market returns) and the coincidental changes in the prospective market returns. Having established the pattern, they then also study it in other markets, before putting it all together in a diversified strategy that they argue shows a better profile of responses to risk cycles than the simpler long-only or conventional trend-following strategies.
I would like to thank our readers for their continued support and interest, and hope that they will find something useful in this issue of The Journal of Investment Strategies, just as in the ones before and the ones that will come after.
Arthur M. Berd
General Quantitative LLC