¹ÏÌïºÚÁÏ

We test the effect of a change in the sex ratio on marital assortative matching by social class using a large negative exogenous shock to the French male population due to WWI casualties. We analyze a novel data set that links marriage-level data to both French censuses of population and regional data on military mortality. We instrument the sex ratio in a region with military mortality, which exhibits exogenous geographic variation. We find that men married women of higher social class than themselves (married up) more in regions that experienced larger decreases in the sex ratio. A decrease in the sex ratio from one man for every woman to 0.90 men for every woman increased the probability that men married up by 8 percentage points. These findings shed light on individuals' preferences for spouses. Rather than preferring to marry spouses from the same social class, men seem to prefer to marry higher-class spouses, but cannot do so when the sex ratio is balanced.