Asymmetric pass-through of oil prices to gasoline prices with interval time series modelling

Modelling and analyzing the pass-through of crude oil prices to retail gasoline prices has received an increasing attention in existing literature. However, most of the ongoing efforts focus on point-valued data, which may suffer from the informational loss. This paper first employs the threshold autoregressive interval-valued (TARI) models developed recently by Sun et al. (2018a) to investigate this transmission, and proposes a novel consistent interval-based test to detect threshold effects, and derives its asymptotic properties. These interval-based TARI model and test, superior to traditional point-based methods, can produce more efficient parameter estimation and more powerful inference due to the informational gain of interval data. That is, interval data simultaneously contain more information (e.g., trend and volatility) than point-valued data during the same period. Empirical results suggest that retail gasoline prices are adjusted upwards faster than downwards, and this short-run asymmetry is statistically significant during the whole sample. Our findings reveal that both the level and volatility of oil prices have a positive impact on the price of gasoline, which contributes to the asymmetries in the transmission of oil price shocks. These results are robust before and after financial crisis. Furthermore, it is interesting to find that after financial crisis, both the error correction adjustment speeds and the asymmetries decrease.