Large fluctuations of China's commodity prices: Main sources and heterogeneous effects

Assuming that the economic variables obey the normal distribution, existing studies often use traditional mean models to explore commodity prices. In fact, the distribution of economic data is often skewed. When the economic data is left‐biased or right‐biased, the quantile regression can more fully characterize the distribution of economic variables, and thus obtain a comprehensive analysis result. Moreover, the coefficient estimates of the quantile regression are more robust than those in the ordinary least squares regression. Therefore, this paper uses the quantile regression method to investigate commodity price fluctuations. The results show that the money supply has a greater impact on commodity prices in the lower 10th and 10th–25th quantile groups. Among all quantile groups, the impact of international oil prices on commodity prices in the upper 90th quantile group is the highest. The effect of the exchange rate in the 25th–50th quantile group is greater than those in other quantile groups. However, the real economy has a minimal impact on the 25th–50th quantile group. Thus, the heterogeneous effects of these influencing factors should be taken into consideration when stabilizing commodity prices.