describes the empirical strategy, and section 6 presents the results. Section 7 concludes. 2. Literature This article is closely related to the literature on media bias and persuasion. In particular, it contributes to the empirical literature on the effects of social media by providing new causal estimates. This article also contributes to the literature on the impact of communication technology on political outcomes.2 The literature distinguishes roughly two broad channels of how internet influences political outcomes. First, providing information may persuade receivers to change their behavior or beliefs. Second, it can enhance communication and coordination between agents. Studies on democratic regimes have focused on how access to internet affects voter participation. The effect varies depending on the relative importance of information and entertainment content (cf. DellaVigna and Gentzkow 2010; Strömberg 2015). The role of internet as a low cost channel of information may be especially important in an environment where traditional media is under state control. Miner (2015) shows that in Malaysia, where the government held strict control over mass media, expanding uncensored internet penetration led to a decrease in the ruling party’s support. Similarly, Enikolopov, Petrova, and Zhuravskaya (2011) find that in Russia availability of an independent TV channel decreased vote for the government party and increased vote for the opposition. Guriev, Melnikov, and Zhuravskaya (2019) suggest that wider access to internet may reduce government approval, particularly when traditional media is censored, by exposing corruption. However, if internet is also censored, increased access does not affect opinion on the government. Mass media may be captured by an authoritarian regime and used to spread propaganda. Studying cross-border media, DellaVigna et al. (2014) find that exposure to nationalistic Serbian radio in Croatia increased ethnic animosity towards Serbs and increased vote for Croatian nationalist parties, whereas Peisakhin and Rozenas (2018) find that access to Russian state owned TV in Ukraine increased support for the pro-Russian candidates in 2014. Adena et al. (2015) examine the effect of radio before and after Hitler became the chancellor in Germany. During the democratic period pro-government radio had a mitigating influence on Nazi support, but After Hitler’s rise to power the radio content also shifted to reflect the views of the regime, which increased Nazi popularity. Yanagizawa-Drott (2014) shows that radio propaganda had an important role in inflaming the Rwandan genocide, and significantly increased killings and participation in violence. In an opposite setting in Uganda, Armand, Atwell, and Gomes (2020) show that defection messaging was effective in mitigating conflict. Research on enhanced communication in conflict situations exhibits mixed results. 2. For a more thorough literature review, see Zhuravskaya, Petrova, and Enikolopov (2020). 4