Predicting Safe Haven Assets through Implied Treasury Yield Skewness: A Time-Varying Nonparametric Quantile Causality Analysis

Using both static and time-varying quantile Granger causality frameworks, we relate the implied Treasury yield skewness (ISK), a newly proposed measure of the market's expectation of asymmetric risks in US yields, to the returns and volatility of eight major safe-havens (gold, German bunds, silver, Swiss franc, the Japanese yen, the US Dollar index, platinum and palladium). The analysis covers daily data from January 1988 to April 2025. The static causality analysis shows that the ISK significantly predicts the entire conditional distribution of returns and volatility for all safe havens. It is notable for the returns of precious metals and very pronounced for the volatility of the eight safe-havens, especially during moderate volatility states. The time-varying causality analysis reveals that return predictability is unstable and highly episodic, often clustering around crisis periods. In contrast, volatility predictability is more persistent. These findings suggest that information about the market's expectation of asymmetric risks in US treasury yields embeds signals that affect the price and volatility dynamics of various safe-havens, which have important implications for market participants, underscoring the need for dynamic portfolio and risk management of safe-haven assets.