Wenting Liao, Jun Ma, Chengsi Zhang

Read Full Paper

This study explores the time-varying effects of uncertainty shocks, identified using an external instrument method, on the broad-based movement of commodity returns since the early 1990s. It employs a vector autoregression-augmented dynamic factor model with time-varying parameters and stochastic volatility to extract a common factor from 43 commodity returns. The results indicate that uncertainty shocks reduce commodity returns through this common factor, and their effects vary significantly over time, becoming much stronger during recessions. Additionally, uncertainty shocks often lead to dollar appreciation, which may explain the negative correlation between commodity returns and the strength of the US dollar.

By jointly estimating a VAR and DFM, the study enhances the identification of factors by incorporating additional macroeconomic information. The external instrumental variable (IV) method is used in the factor-augmented VAR (FAVAR) framework to identify uncertainty shocks. The model allows for the study of not only the time-varying effects of uncertainty shocks on commodity returns but also the interactions between these shocks and macroeconomic fundamentals.

This study contributes to the literature by directly linking uncertainty shocks to commodity returns through a common factor, highlighting the increasing influence of aggregate economic activity on commodity returns. Unlike traditional two-step estimation methods, the study uses a joint estimation approach that simultaneously extracts the common factor and studies its interactions with other shocks, avoiding the generated regressor issue. The main findings show that uncertainty shocks immediately reduce the common factor of commodity returns, which recovers after about six months. The US dollar strengthens with a lag following the shock, and then gradually weakens. The study also finds that the US short-term interest rate decreases gradually following the shock, while global demand drops initially and then recovers over time.

In conclusion, this study provides empirical evidence on the effects of uncertainty shocks on commodity markets and reveals how these shocks vary across different economic cycles, particularly during recessions when their impact is most pronounced. The study's innovation lies in using a time-varying factor model to accurately capture the dynamic changes of uncertainty shocks and their interactions with macroeconomic factors.