<p>This paper investigates the behavior of the changes in gasoline prices and crude oil prices using a Threshold Vector Error Correction Model (TVECM). Daily average gasoline prices and daily crude oil prices are modeled together to predict the next day’s change in the average gasoline price. A TVECM is also used to impute the missing values in daily crude oil prices. A model that includes both prediction and imputation using TVECM is compared with other multivariate time series models and other imputation methods over a wide range of lags. Based on root mean squared errors, TVECM clearly outperforms when predicting the changes in the next day’s average gasoline prices while imputing missing crude oil prices using the last known value. Additionally, TVECMs seem to recognize the well–known 7–day gasoline price cycle in gasoline prices. The models are applied to gasoline prices from suburbs of Western Australia and crude oil prices from the region. Software for implementing TVECM with enhancements allowing imputation and inclusion of additional covariates are made freely available. Given that fluctuations in gasoline prices have a significant impact on consumers’ lives and that consumers are increasingly eager to track daily price changes, the proposed model offers a novel approach for predicting next day’s gasoline price behavior.</p>