Gender voting gap in the dawn of urbanization: evidence from a quasi-experiment with Greek special elections

The electoral law of 31 May 1952 extended the voting rights to all adult women in Greece. This paper examines the impact of womenÕs enfranchisement on party vote shares by employing a unique dataset of 385 communities located in seven prefectures in Greece where by-elections took place in 1953 and 1954 (for strictly exogenous reasons). To estimate causal effects, we exploit the observed heterogeneity in the proportion of women in the electorate across communities as the identifying source of variation, and employ a difference-in-differences design that holds unobserved local characteristics fixed. Our results provide strong evidence in favour of the Òtraditional gender voting gapÓ (women voting more conservatively compared to men) in the urban prefecture of Thessaloniki, and no evidence of gender voting differences in the remaining (six) predominantly rural prefectures of our sample. Our results also reveal that the existence of a gender voting gap is highly conditional upon the proportion of economically inactive women; that is, women tend to vote for right parties when they are outside of the labour force. Interestingly, when we account for this conditionality, a suffrage-induced pro-right shift can also be observed in communities outside Thessaloniki. Building on the economic bargaining models of the family, we argue that, in an economic environment characterized by limited demand for female labour force participation, women support more vigorously the sanctity and the strength of family values and tend to vote more conservatively compared to men.