Masking turnout inequality. Invalid voting and class bias when compulsory voting is reinstated

Compulsory voting increases turnout. However, this does not imply that CV reduces the socioeconomic gap among voters. While conventional accounts argue that CV corrects turnout inequality, some literature challenges such effect because CV also increases invalid voting. We evaluate the empirical merits of these theories turning to the case of Chile. We compare the results of the Constitutional referendums in 2020 –with voluntary voting– and 2022 –under mandatory rules. We report the transition to CV (1) boosted turnout; (2) reduced the class-biased turnout; (3) increased invalid voting, yet wasn't socioeconomically biased in all municipalities; (4) invalid voting inequality emerged when dividing municipalities into those within and outside the Metropolitan Region. These findings hold for the 2021 –with VV– and 2023 constitutional representative elections –under CV. Our results show that CV raised participation and reduced class-biased turnout, but it doesn't automatically correct inequalities, as it produces heterogeneous effects over invalid voting.