The Aligned Economic Index & the State-Switching Model

Battle Count: 110

Highly relevant to quantitative trading as it directly addresses equity premium predictability with practical out-of-sample forecasting. The state-switching model provides actionable regime information (yield curve inversion as a leading indicator) that can inform tactical asset allocation and market timing strategies. The Aligned Economic Index offers a parsimonious predictor that outperforms benchmarks across both expansionary and recessionary states. However, the monthly frequency and one-month horizon limit direct application to high-frequency trading. The economic significance (R²_oos of 4.12% overall, up to 11.21% in recessions) is meaningful for systematic equity strategies.

Moderate complexity. The state-switching regression is a straightforward extension of OLS with interaction terms. The PLS construction involves two-step regressions (time-series then cross-sectional) following Kelly and Pruitt (2015). The recursive out-of-sample forecasting procedure requires careful implementation to avoid look-ahead bias. The main complexity lies in correctly implementing the PLS algorithm with recursive weight updates and the state indicator construction. Standard econometric software (R, Python statsmodels, MATLAB) can handle all components.

Data is publicly available from Amit Goyal's webpage (www.hec.unil.ch/agoyal/). The methodology is well-described with explicit equations for the state-switching model, PLS construction, and recursive out-of-sample forecasting. The τ parameter is fixed at 9 months with robustness checks for τ between 3 and 12. However, no code repository is provided, and some implementation details of the PLS algorithm (e.g., number of components, convergence criteria) are referenced to Appendix A.