Political Uncertainty-Managed Portfolios

Forward-looking metrics of uncertainty based on options-implied information should be highly predictive of equity market returns in accordance with asset pricing theory. Empirically, however, the ability of the VIX, for example, to predict returns is statistically weak. In contrast to other studies that typically analyze a short time-series of option prices, I make use of a ‘VIX-type’ but a text-based measure of uncertainty starting in 1890, which is constructed using the titles and abstracts of front-page articles of the Wall Street Journal. I hypothesize that uncertainty timing might increase Sharpe ratios because changes in uncertainty are not necessarily correlated with changes in equity risk and, therefore, not offset by proportional changes in expected returns. Using a major US equity portfolio, I propose a dynamic trading strategy and show that lagged news-based uncertainty explains future excess returns on the market portfolio at the short horizon. While policy- and war-related concerns mainly drive these predictability results, stock market-related news has no predictive power. A managed equity portfolio that takes more risk when news-based uncertainty is high generates an annualized equity risk-adjusted alpha of 5.33% with an appraisal ratio of 0.46. Managing news-based uncertainty contrasts with conventional investment knowledge because the strategy takes relatively less risks in recessions, which rules out typical risk-based explanations. Interestingly, I find that the uncertainty around governmental policy is lower and, by taking less risk, it performs better during periods when the Republicans control the senate. I conclude that my text-based measure is a plausible proxy for investor policy uncertainty and performs better in terms of predictability compared to other options-based measures.