This paper introduces an innovative nonparametric panel data approach to model the long-run relationship between the monthly oil price index and stock market price indices of ten large net oil importing countries; namely, the United States, Japan, China, South Korea, India, Germany, France, Singapore, Italy and Spain. In the proposed model, we allow the coefficient on the oil price index to be a time-varying function which evolves over time in a way that is assumed to be unknown. We also allow the common trend function to evolve over time, as well as extending the model further to incorporate country-specific trend functions. We employ a data-driven local linear method to estimate these time-varying trend and coefficient functions. The results show that, despite being largely positive, there are several downward trends, reflecting the aftermath of the Iraq war and the recent unprecedented drop in the oil price. Overall, we find that the nonparametric panel data model better captures the way in which the underlying stock-oil price relationship has evolved over time in comparison to the point estimates of the parametric counterpart. Moreover, we find that stock market fundamentals play a significant role in determining the oil-stock price relationship. Our findings have important implication for policymakers and financial speculators.
- Local linear estimator
- Panel data analysis
- Time-varying coefficient functions
- Time-varying trend functions