Switching between states and the COVID-19 turbulence

Battle Count: 135

Highly relevant to quantitative trading. The paper directly addresses real-time portfolio allocation using predictive signals, with explicit transaction cost modeling, leverage constraints, and turnover limits. The state-switching framework provides actionable regime-based allocation rules. The finding that EPLS under state switching delivers 5.54% annualized CER gains net of costs is economically significant for practitioners. The yield-curve-based state indicator is implementable in real time without ex-post NBER dating. However, the monthly frequency and one-month horizon limit applicability to higher-frequency strategies.

Moderate complexity. The PLS construction (following Kelly and Pruitt 2013, 2015) requires careful implementation of the three-pass regression filter. The state-switching specification adds a binary regime indicator based on yield-curve slope. The mean-variance optimization with constraints (leverage limit, no short selling, turnover bounds) is standard but requires careful numerical implementation. The recursive out-of-sample estimation procedure is straightforward but computationally intensive over 40+ years of monthly data. Overall, a skilled quantitative analyst could implement this in 2-4 weeks using standard statistical software (R, Python, MATLAB).

The paper provides detailed methodology including the PLS construction, state-switching specification, mean-variance optimization, and transaction cost assumptions. Data is publicly available from Amit Goyal's webpage (Welch and Goyal 2008 dataset). The recursive estimation procedure, risk aversion parameters (γ=3), leverage limits (50%), and turnover constraints are clearly specified. However, no code repository is provided, and some implementation details (e.g., exact PLS algorithm parameters) rely on references to Kelly and Pruitt (2013, 2015).