This paper studies how spousal bargaining power affects consumption patterns of married households in the US, using a detailed barcode-level dataset. We use two distribution factors as proxies for spousal bargaining power: (1) spouses' relative education and (2) spouses' relative potential wage, which is our preferred distribution factor. As an arguably exogenous measure of bargaining power, our relative potential wage is constructed as a Bartik style measure of female-to-male wage ratio, exploiting county-level variations in heterogeneous exposure to different industries and state-wide wage growth. We find that the expenditure shares on women's beauty goods increase and the expenditure shares on alcohol decrease significantly both when relative education of wife increases and when relative potential wages of wives increase. These results are consistent with household bargaining explanations. For couples with children, improved women's household bargaining position is associated with higher budget share on books, stationary, and school supplies, which are potentially related to investment in children. For singles, we do not find statistically meaningful effects of relative potential wage on any of their consumption outcomes, which strengthens the interpretation that the relative wage only affects couples' consumption decisions.