Several models posit a positive cross-sectional correlation between markups and firm size, which, among others, characterizes misallocation, factor shares, and gains from trade. Yet, taking labor market power into account in markup estimation, we show that larger firms have lower markups. This correlation turns positive only after conditioning on wage markdowns, suggesting interactions between product and labor market power. Our findings are robust to common criticism (e.g., price bias) and hold across 19 European countries. We discuss the resulting implications and highlight studying input and output market power within an integrated framework as an important next step for future research.