Typical measures of wages, such as average hourly earnings, fail to capture cyclicality in the effective cost of labor in the presence of (i) cyclical fluctuations in the quality of worker-firm matches, or (ii) wages being smoothed within employment matches. To address both concerns, we estimate cyclicality in labor's user cost exploiting the long-run wage in a match to control for its quality. Using NLSY data for 1980 to 2019, we identify three channels by which hiring in a recession affects user cost: It lowers the new-hire wage; it lowers wages going forward in the match; but it also results in higher subsequent separations. All totaled, we find that labor's user cost is highly procyclical, increasing by about 3\% for a 1 pp decline in the unemployment rate, with three-quarters of this reflecting a procyclical quality-adjusted new-hire wage.