The local press as an external public governance power

ABSTRACT

Do national and local newspapers both impact voter turnout? This paper tackles this question by employing data from a unique dataset that collects detailed national and local newspapers, politico-institutional and socio-economic information for Italy during 1980–2007. We use a Dynamic Panel Data with Instrumental Variables methodology that allows for the consideration of potential sources of endogeneity that may affect the relation between the press and voting behavior. We strengthen this methodology by considering the industry’s entry and exit of newspapers – i.e. turbulence – as an external shock. Our analysis provides robust evidence that local newspaper readership affects both national and local turnout and improves political accountability.

KEYWORDS:

• Division of powers
• External versus internal public governance institutions
• Knowledge dispersion
• Local press
• Political accountability
• Voter turnout

JEL CLASSIFICATION:

• D72
• H00
• L82

Acknowledgments

We benefited from feedback from the following audiences: DiSSE Sapienza – Università di Roma; ERSA, Johannesburg; European Public Choice Society (EPCS); PEARL Workshop; Società Italiana di Diritto ed Economia (SIDE-ISLE); Società Italiana di Economia Pubblica (SIEP). Conversations with Lyndal Keeton, Uma Kollamparambil, Yasmina Rim Limam, Nicola Pontarollo and Roberto Ricciuti were fruitful as well. We are extremely grateful to Aldo A. Sitoe for his very valuable research assistance. Last but not the least, we thank the referees.

Disclosure statement

The authors declare that they have no relevant or material financial interests that relate to the research presented in this work.

Notes

¹ On the neglected importance of the international newsreel industry, which can be thought of as a rudimentary version of satellite channels before the television era, see Althaus et al. (2018).

² On the relation between the Internet and political participation see, among others, Czernich (2012).

³ Indeed, Tocqueville ultimately perceived the role of the local press to be so important to elevate it to the primary power – after the consumer-voters themselves – for political control: ‘each newspaper individually has little power; but the periodical press, after the people, is still the first of powers’ (de Tocqueville 2012[1835–1840], p. 298).

⁴ World Bank data indicate that Internet use in Italy started to pick up momentum around 1998, reaching almost 41% of the population in 2007. See https://data.worldbank.org/indicator/IT.NET.USER.ZS?locations=IT (last accessed November 9, 2022). More generally, local television coverage seems to have little to no effect on political knowledge (Snyder and Strömberg 2010); scant evidence on radio so far suggests that there is reinforcement of already held local political beliefs (Adena et al. 2015). Clearly, one must be mindful of external validity as context matters. Media specialists moreover point out that digital media does not pose a genuine challenge to local newspapers because (i) it creates more information awareness but less genuine civic participation (Howard 2005); (ii) it does not cover local government as much (Fico et al. 2013); and (iii) it is increasingly centralized (Zittrain 2008). See Bunting (2017) for a recent elaboration.

⁵ Citizens who are 18 and older are the voting base of the Chamber of Deputies; while citizens who are 25 and older are the voting base of the Senate.

⁶ Note that during our sample period there is never coincidence between national and regional election day.

⁷ Visit https://rousseau.movimento5stelle.it/main.php (last accessed February 8, 2020).

⁸ An exception is the region of Molise. Following a decision of the Council of State (Italy’s highest administrative judging organ), Molise’s 2000 elections were cancelled due to several irregularities, and held in 2001 and in 2006.

⁹ It is useful to point out here that the Italian electoral system cannot genuinely be considered compulsory. Although during 1945–1993 possible social sanctions were provided, they were never effective (so-called sanzioni innocue). Therefore, we do not consider the sanctions to have an impact on our dependent variable.

¹⁰ http://www.fieg.it/ (last accessed November 3, 2022).

¹¹ https://www.adnkronos.com/ (last accessed November 3, 2022).

¹² With the exceptions of $ElectoralDummy$, $Localnewspapers$, $Nationalnewspapers$, $School$, $Senate$, and $Turbulence$, all other variables are expressed as logarithms for reasons of scale.

Additional information

Funding

The work was supported by Sapienza-University of Rome, Rome, Italy (RM11916B8902F7D7).