A1307
Title: Safe haven properties of gold: An application of a dynamic score-driven rotation model
Authors: Ruven Zapf - Georg-August-Universität Göttingen (Germany) [presenting]
Helmut Herwartz - Georg-August-University Goettingen (Germany)
Abstract: Gold's safe-haven properties are evaluated through a tail-risk perspective rooted in higher-order moment dynamics. A framework is developed that integrates a multivariate GARCH specification with a dynamic score-driven rotation, enabling the extraction of independent components and time-varying conditional skewness and kurtosis. These higher-order dynamics are translated into one-step-ahead value at risk and expected shortfall forecasts via the Cornish-Fisher expansion, thereby linking distributional asymmetries directly to risk measures. Using daily data for global equities and gold from 2007 to 2025, three main findings emerge. Dynamic rotations substantially improve the calibration of value at risk and expected shortfall forecasts relative to Gaussian, Student-t, and unrotated benchmarks, as verified by standard backtests. Optimal portfolio weights derived from higher-moment risk criteria tilt sharply toward gold during episodes of financial turmoil, including the global financial crisis, COVID-19, and the Russian invasion of Ukraine, while reverting toward equities in tranquil periods. By jointly modeling volatility, skewness, and kurtosis, the analysis reveals a contingent rather than universal safe-haven role for gold, emphasizing that its protective capacity is crisis-specific and state-dependent. The results highlight the economic relevance of higher-order moment dynamics for risk management, portfolio allocation, and the assessment of safe-haven assets.
