The share of gasoline consumption in household expenditures decreases with income, gasoline demand being least elastic for low-income households. Based on this empirical evidence for non-homotheticities, we develop a quantitative heterogeneous-agent general equilibrium model to quantify the distributional consequences of oil price shocks. Although oil price shocks have small aggregate effects, they hurt low-income households considerably with costs to lifetime utility two to three times larger for those in the bottom decile of income relative to those in the top decile. Additionally, the 2014/15 oil glut depressed gasoline prices, which delivered comparable welfare benefits to the 2018 tax cuts.