Shifting Correlations: How Trade Policy Uncertainty Alters stock-T bill Relationships

Battle Count: 66

Highly relevant for quantitative trading and portfolio management. The paper provides actionable insights on how trade policy uncertainty affects stock-bond correlations, which is critical for: (1) dynamic asset allocation strategies, (2) risk parity and mean-variance portfolio construction, (3) hedging strategies using T-bills, (4) regime-aware trading systems that adjust to political cycles, (5) stress-testing portfolios under trade war scenarios. The finding that TPU weakens the safe-haven role of T-bills during high-uncertainty periods directly impacts diversification strategies. The out-of-sample forecast results show economic improvements in portfolio variance prediction when incorporating TPU and political regime variables.

Moderate complexity. The two-step estimation procedure (GJR-GARCH followed by DCC-X) is computationally tractable and well-established in financial econometrics. The augmentation with exogenous variables is straightforward (adding regressors to the Q_t equation). However, ensuring positive definiteness of the correlation matrix, handling the interaction terms, and implementing the MCS forecasting evaluation require careful attention. Standard econometric packages (R with rugarch/dccGarch, Python with arch/statsmodels, or MATLAB) can implement the core model. The structural break analysis and rolling window estimation add additional complexity.

The TPU index is publicly available (Caldara et al., 2020, via Matteo Iacoviello's website). Stock and T-bill data are sourced from Yahoo Finance. However, the paper states data will be made available on request rather than providing a direct download link. The GJR-DCC-X estimation procedure is well-documented with explicit equations, and standard econometric software (e.g., R, Python, EViews) can implement it. Robust standard errors (White, 1980) are used. No code repository is provided.