Popularity Shocks and Political Selection

Abstract

A B S T R A C TWe observe that popularity shocks are crucial for electoral accountability beyond their effects on voters' behaviors. By focusing on Brazilian politics, we show that the disclosure of audit reports on the (mis)use of federal funds by local administrators affects the type of candidates who stand for election. When the audit finds low levels of corruption, the parties supporting the incumbent select less-educated candidates. On the contrary, parties pick more-educated candidates when the audit reveals a high level of corruption. These effects are stronger in municipalities that have easier access to local media.

Corruption and Firms

“…43 For example, experimental evidence by Arias et al (2018) in Mexico lends support to this explanation in the context of politicians' malfeasance and voter learning. Cavalcanti et al (2018) study our same CGU program and show that detecting little corruption justifies political parties to present a pool of politicians that is of lower quality. 44 For instance, audits may lead politicians to under-spend or they may discourage them to use audit-sensitive but potentially efficient procurement procedures, as shown by Lichand and Fernandes (2019) and Gerardino et al (2017), respectively.…”

“…A leading explanation for why the direct experience of an audit might have an effect is the presence of political incentives, as outlined by several standard political agency models and a large literature on the importance of political institutions to limit rentseeking (Persson and Tabellini, 2002;Besley, 2007). Audits may have an impact on political selection because they lead to an overall better pool of candidates (Ferraz and Finan, 2008;Cavalcanti et al, 2018). Moreover, audits may have a disciplining measure of GD-Corrupt, but without scaling by the number of firms (GD-Corrupt-NS).…”

Who’s getting the office? Party dominance and elected executives’ career path

“…In this section, I derive a testable hypothesis that organizes and motivates the empirical analysis. Following Cavalcanti et al (2018) and Galasso and Nannicini (2011) the analysis is based on the idea that parties select and allocate candidates based on their likelihood of winning elections. This analysis is based on four main assumptions:…”

“…Following the literature on political parties’ strategic allocation of candidates (see Cavalcanti et al, 2018; Galasso & Nannicini, 2011), the paper argues that, when selecting candidates, a dominant party will respond differently to a loss of popularity than non‐dominant parties. Parties will select candidates with less political experience (and more technical or administrative experience), and the decrease in political experience will be larger in dominant states than in competitive states.…”

Election by Community Consensus: Effects on Political Selection and Governance

“…More broadly, this paper adds to the literature on the impact of electoral institutions on political selection and governance (Diermeier et al 2005, Mattozzi & Merlo 2008, Keane & Merlo 2010, Banerjee et al 2011, Kotakorpi & Poutvaara 2011, Gagliarducci & Nannicini 2013, Baltrunaite et al 2014, Daniele & Geys 2015, Banerjee et al 2017 and how political selection affects governance characteristics. 6 Research also shows that political selection responds to factors like politician murders (Daniele, 2017), audits (Brollo et al 2013, Cavalcanti et al 2018, electoral competition (Galasso & Nannicini, 2011), the threat of bribes and punishment (Dal Bó et al , 2006) and even political predecessors (Caselli & Morelli, 2004). 5 00970 outcomes both in India and other countries (Pande 2003, Chattopadhyay & Duflo 2004, Powley 2007, Washington 2008, Rajaraman & Gupta 2012, Afridi et al 2013.…”