Multiscale spillovers and portfolio management in major assets: insights from a Jump, Oscillation and Trend decomposition

This study develops an integrated multi-scale framework for structurally decomposing major financial asset prices, including stocks, bonds, currencies, commodities, and precious metals, into trend, oscillation, and jump components. This decomposition provides the basis for examining cross-market spillovers and evaluating dynamic portfolio allocation strategies. Empirical results reveal regime-dependent spillover mechanisms: (1) Trend-based spillovers demonstrate high persistence and structural stability, closely corresponding to the Merrill Lynch Investment Clock, and support the macro-fundamental transmission channel. (2) Oscillation components capture medium-term behavioral sentiment and liquidity cycles, providing potential early-warning signals during market transitions. (3) Jump components dominate during systemic shocks, with equities and precious metals acting as primary transmitters of instantaneous contagion. Portfolios that use these component-specific spillover dynamics exhibit stronger benchmark performance than traditional strategies in the pre-transaction-cost setting, particularly in trend- or jump-dominated regimes. By integrating structural decomposition with dynamic asset allocation analysis, this study contributes to the understanding of heterogeneous spillover mechanisms and provides a structured reference for adaptive, multi-layered portfolio strategies.