Spillovers Between Lithium and Electric Vehicle Markets: GAS vs GARCH Approach

This thesis models the stock returns of electric vehicle (EV) manufacturers to investigate the dynamics between the EV market and the lithium market. We employ the Generalized Autoregressive Score (GAS) framework alongside the Exponential GARCH (EGARCH) model to test three specific research hypotheses. First, we test the hypothesis that there is a significant return spillover effect between the lithium and EV markets, while the volatility spillover effect remains insignificant. Using jump intensity estimates from an EGARCH-EARJI model as external regressors, our results support this hypothesis. They identify return spillovers from lithium producers to be significant, while volatility spillovers are found to be insignificant across all models. Second, we examine the hypothesis that the GAS model outperforms GARCH-type models in modeling EV returns. We find that while the GAS model does not outperform the EGARCH model in terms of AIC, it offers a more flexible framework with superior theoretical properties for handling time-varying parameters in financial time series. Third, we test whether the Fama-French 5-factor model improves the model fit compared to the 3-factor model. The additional two factors slightly improve the fit of both EGARCH and GAS models while showing the same results regarding the spillover effects. This further validates the results of the 3-factor models, as well as those of previously conducted studies.