Table of Contents
Fetching ...

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

Demetrio Lacava

Abstract

This paper examines how trade policy uncertainty influences the correlation between U.S. stock indices and short-term government bonds. The objective is to assess whether policy-related shocks, especially those linked to trade tensions, alter the traditional stock-T bill relationship and its implications for investors. We extend the Dynamic Conditional Correlation (DCC) framework by incorporating exogenous variables to account for external shocks. Three specifications are analyzed: one using the Trade Policy Uncertainty (TPU) index, one including a dummy variable reflecting presidential-cycle effects, and one combining both through an interaction term. The analysis is based on daily data for major U.S. stock indices and the 3-month Treasury bill. Results indicate that trade policy uncertainty exerts a significant effect on stock-T bill correlations. Moreover, its influence becomes stronger under specific political conditions, suggesting that political agendas can amplify the impact of trade-related shocks on financial markets. Crucially, augmenting the DCC framework with trade-policy-related variables improves also the economic relevance of correlation forecasts. Therefore, this study contributes to the literature by explicitly integrating policy-related uncertainty into correlation modeling through an augmented DCC framework. The findings provide new insights for portfolio allocation and risk management in environments characterized by heightened trade tensions.

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

Abstract

This paper examines how trade policy uncertainty influences the correlation between U.S. stock indices and short-term government bonds. The objective is to assess whether policy-related shocks, especially those linked to trade tensions, alter the traditional stock-T bill relationship and its implications for investors. We extend the Dynamic Conditional Correlation (DCC) framework by incorporating exogenous variables to account for external shocks. Three specifications are analyzed: one using the Trade Policy Uncertainty (TPU) index, one including a dummy variable reflecting presidential-cycle effects, and one combining both through an interaction term. The analysis is based on daily data for major U.S. stock indices and the 3-month Treasury bill. Results indicate that trade policy uncertainty exerts a significant effect on stock-T bill correlations. Moreover, its influence becomes stronger under specific political conditions, suggesting that political agendas can amplify the impact of trade-related shocks on financial markets. Crucially, augmenting the DCC framework with trade-policy-related variables improves also the economic relevance of correlation forecasts. Therefore, this study contributes to the literature by explicitly integrating policy-related uncertainty into correlation modeling through an augmented DCC framework. The findings provide new insights for portfolio allocation and risk management in environments characterized by heightened trade tensions.

Paper Structure

This paper contains 10 sections, 13 equations, 7 figures, 5 tables.

Figures (7)

  • Figure 1: Heatmap representing the sample correlations of the de-GARCHed returns as defined in Eq. \ref{['eq:stand_res']}. Darker colors indicate stronger positive correlations, while lighter colors indicate weaker or negative correlations.
  • Figure 2: 60-day rolling correlation between stock and T bill returns (black line), with periods exceeding one standard deviation highlighted in red. The dashed red line represents the threshold of one standard deviation.
  • Figure 3: Conditional correlations from the DCC model (black line, left axis), Trade Policy Uncertainty (TPU) index (red line, right axis), and key trade-related events (vertical blue-dashed lines)
  • Figure 4: Impulse Response Function (IRF) of the stock-T bill correlation to a one-standard-deviation shock in trade policy uncertainty (black line, left axis) and VIX (red line, right axis), expressed in percentage points
  • Figure 5: $\hat{\theta}_3$ rolling window estimated coefficient (black line) from the $DCC-X_{TPU}$ model and break dates (dashed blue line, details in the text).
  • ...and 2 more figures