Periodic Reporting for period 2 - POLITICALMIND (Explaining Politicians' and Voters' Behavior)
Reporting period: 2018-02-01 to 2019-07-31
Both research questions have been tackled empirically, combining the collection of original datasets (referred to Italy) with the implementation of state-of-the-art econometric strategies, namely, randomized controlled trials in the field, laboratory experiments, and regression discontinuity designs. In particular, the causal effects of campaign messages on voters’ beliefs and behaviors have been estimated by convincing real-world politicians to randomize a part of their campaign in Italian local elections.
In a set of studies on the impact of traumatic event on voters’ behavior, we estimate the long-lasting effect of political violence on electoral outcomes. In particular, we exploit a geographic discontinuity in the intensity and duration of the Nazi occupation of Italy and civil war in 1943-45. We find that the Communist party gained votes in the postwar elections where the Nazi occupation and the civil war lasted longer. This effect persists until the early 1990s. The empirical evidence also suggests that this is due to an effect on political attitudes rather than party organizations. Thus, the foreign occupation and the civil war left a lasting legacy on the newborn Italian democracy.
In a set of studies on the impact of campaign messages by politicians, we document three original results thanks to both randomized political campaigns in the field and online survey experiments. First, there are stark gender differences in the response to negative campaigning, which appears to be effective with male voters but to be counterproductive with females. Second, there is causal evidence of cognitive dissonance in voting behavior. Third, negative campaigning produces spillover effects in multi-candidate environments, as both the receiver and the sender of the campaign attack lose votes in favor of third-party candidates.