If the impact of electronic voting on turnout is rather well documented, its impact on invalid votes is often disregarded by the literature. The implementation of e-voting has a direct technical consequence: voters can no longer express a null vote. For those voters that want (or have) to participate in elections but want to express discontent, casting a blank vote is therefore a relevant option. This paper investigates the share of blank votes in all local elections in Belgium (1994-2018) and tests whether differences across municipalities can be explained by e-voting. Belgium is the ideal country to analyse the impact of e-voting on blank votes, as voting is compulsory (so there are less variation of turnout) and as e-voting has been implemented on about half of the municipalities. Taking into account various aspects of the electoral competition as well as socio-demographic factors, models demonstrate that e-voting has a significant impact on voting behaviour as increases the amount of blank votes in these municipalities. The paper finally discusses the measurement of turnout in settings where voting is compulsory and where e-voting is implemented.
