Switching between states and the COVID-19 turbulence
Abstract
In Aarab (2020), I examine U.S. stock return predictability across economic regimes and document evidence of time-varying expected returns across market states in the long run. The analysis introduces a state-switching specification in which the market state is proxied by the slope of the yield curve, and proposes an Aligned Economic Index built from the popular predictors of Welch and Goyal (2008) (augmented with bond and equity premium measures). The Aligned Economic Index under the state-switching model exhibits statistically and economically meaningful in-sample ($R^2 = 5.9\%$) and out-of-sample ($R^2_{\text{oos}} = 4.12\%$) predictive power across both recessions and expansions, while outperforming a range of widely used predictors. In this work, I examine the added value for professional practitioners by computing the economic gains for a mean-variance investor and find substantial added benefit of using the new index under the state switching model across all market states. The Aligned Economic Index can thus be implemented on a consistent real-time basis. These findings are crucial for both academics and practitioners as expansions are much longer-lived than recessions. Finally, I extend the empirical exercises by incorporating data through September 2020 and document sizable gains from using the Aligned Economic Index, relative to more traditional approaches, during the COVID-19 market turbulence.
Summary
This paper investigates the economic value of using an Aligned Economic Index within a state-switching framework to predict U.S. stock returns. The research builds upon Aarab (2020), which introduced the Aligned Economic Index, constructed from Welch and Goyal's (2008) predictors and augmented with bond and equity premium measures, using partial least squares (PLS). This paper extends the analysis by evaluating the economic gains for a mean-variance investor using the Aligned Economic Index under both a one-state regression model and a state-switching model, where the market state is proxied by the slope of the yield curve. The study uses monthly data from January 1960 to September 2020, focusing on the post-war period and including the COVID-19 market turbulence. The core methodology involves comparing the certainty-equivalent returns (CERs) of portfolios managed using different forecasting models: (1) historical average, (2) Aligned Economic Index (E_PLS), PCA-based index (E_PCA), and forecast combination (E_FC) under a one-state regression, and (3) the same indices under a state-switching regression. The paper also considers a buy-and-hold strategy as an additional benchmark. Optimal portfolio weights are computed monthly, considering transaction costs and leverage limits. The key finding is that the state-switching model, particularly when using the E_PLS index, delivers statistically and economically significant CER gains compared to the historical average and other models across various market states (expansions, recessions, up states, and down states). The state-switching model demonstrates superior performance, especially during the COVID-19 market turbulence, because it allows for more timely and smoother adjustments to equity exposure around turning points. This research contributes to the field by demonstrating the practical benefits of incorporating state dependence and a comprehensive economic index for real-time portfolio allocation, especially during periods of economic uncertainty.
Key Insights
- •The Aligned Economic Index (E_PLS) under the state-switching model yields the highest certainty-equivalent return (CER) gains, with a statistically significant ∆CER of 6.09% before transaction costs and 5.54% after transaction costs.
- •The state-switching models generate substantially higher monthly Sharpe ratios (0.22-0.25) compared to the historical-average benchmark (0.13), indicating improved risk-adjusted performance.
- •During recessions, all three models (E_PLS, E_PCA, E_FC) produce statistically significant CER gains under the one-state model, with E_PCA yielding the highest gains of 5.82%. However, these gains are often negated by transaction costs.
- •One-state models exhibit highly volatile equity weights and fail to reduce equity exposure sufficiently in advance of recessions, leading to significant drawdowns. In contrast, state-switching models adjust more smoothly and proactively.
- •The buy-and-hold strategy outperforms the one-state forecasting strategies but is significantly outperformed by the state-switching models, highlighting the value of incorporating state dependence in portfolio allocation.
- •The study imposes realistic portfolio constraints, including leverage limits (up to 50%), no short selling, and restrictions on month-to-month portfolio adjustments, enhancing the practical relevance of the findings.
- •The state-switching models show consistent outperformance across all market states (expansions, recessions, up states, and down states), while the one-state models and the buy-and-hold strategy exhibit state-dependent performance.
Practical Implications
- •The research provides a valuable framework for professional practitioners to improve real-time portfolio allocation by incorporating state-dependent information and a comprehensive economic index.
- •The findings suggest that asset managers can significantly enhance portfolio performance by using the Aligned Economic Index (E_PLS) within a state-switching framework, potentially justifying an annual management fee of up to 5.54%.
- •Investment strategies can be tailored to different market states by monitoring the yield curve slope, which serves as a proxy for economic regimes, allowing for proactive adjustments to equity exposure.
- •Future research could explore alternative state variables, refine the construction of the Aligned Economic Index, or investigate the model's performance in different asset classes or international markets.
- •The study's findings can be used to develop more robust risk management strategies, particularly during periods of economic uncertainty or market turbulence, by incorporating the state-switching model to dynamically adjust portfolio risk exposure.