Carlos Molina

Carlos Molina

Job Market Candidate

Research Fields

Political Economy, Development Economics, Behavioral Economics

Contact Information

Phone (857) 253-1749
Email Address
Office E52-301

I am a PhD Candidate in the MIT Department of Economics. My research agenda explores several of the most salient political events in recent decades, including the rise in affective polarization and protests, the perceived increase in voters’ dissatisfaction with the democratic system, and the high levels of corruption in developing countries.  These events can all severely affect the success of democratic institutions, which are vital for economic development.

Research Statement


Job Market Paper 1

Social Influence and News Consumption (with Alex Moehring)

Abstract: Populations in several countries have become decidedly more polarized in recent decades. Many believe that social media, which facilitates interactions within echo chambers, is partly to blame. These interactions can trigger two distinct effects on the demand for biased news. First, individuals can be influenced by their peers' news consumption, for example, because they value keeping a news diet that is ideologically congruent with that of their peers. Second, individuals might purposefully skew their news consumption in anticipation that their peers will observe these choices. We design a field experiment on Twitter (renamed X in 2023) to separately identify the importance of both mechanisms. Our main result documents that, through these two mechanisms, online interactions with like-minded peers are not a major contributor to the demand for polarized news content. Our experiment induces variation in an individual's perceptions of the political leanings of their peers' news consumption and the visibility of their own news consumption to their social media followers. We track participants' sharing behavior and news consumption, proxied by the news outlets they follow. We find no evidence to support the first channel: our experimental variation influences respondents' beliefs about the news diets of their peers, but they do not respond by changing their own news diets. In contrast, we find that participants alter their news diet considerably when they believe their peers will observe these choices, as in the second channel. Interestingly, individuals primarily wish to present themselves as following a balanced set of news. Therefore, our paper uncovers one mechanism through which social media can attenuate the demand for polarizing content: as these platforms amplify the visibility of user interactions, which increases the importance of social image concerns, users adjust their news consumption to be more balanced.


Job Market Paper 2

Electoral Transhumance: Voter Registration Shifting and Corruption Engagement Before Election Day

Abstract: The literature on corruption distinguishes between two forms of vote buying: corrupt candidates can either bribe voters to change their vote (voter-buying) or bribe non-voters to vote (turnout-buying). This reasoning relies on the assumption that the electoral roll is fixed, overlooking a third method explored in this paper: electoral transhumance. In its most common form, corrupt candidates pay citizens from neighboring areas to change their registration location, so that they can vote in the locations where the candidates plan to run. I study electoral transhumance in Colombia by looking at failed registration attempts – i.e., registrations of people who illegally attempted to vote in a municipality that was not their residence – and studying whether these respond systematically to politicians' incentives to increase votes. Overall, I find that `failed' transhumance – i.e. cases where voters were caught switching their localities -- represents 5.03% of the electoral roll. I then study a reform that made public, more than one year prior to election day, information about drastic changes in the future allocation of resources to the municipalities that increased the presumed returns from office. I exploit a discontinuity in the (future) allocation of transfers from central to local governments to show that electoral transhumance is more common in districts with expected larger budgets. Based on a back-of-the-envelope calculation using intra-municipal migration rates, I estimate that `successful' cases of transhumance represent at least 2.23% of the electoral roll, which is a meaningful fraction because 15% of the municipal elections were won by a margin less than that value. Consistent with a model in which open economies can ``trade'' voters, a positive shock in a municipality's revenue makes electoral transhumance more likely in three types of municipalities: (i) those with small populations, (ii) those with poorly functioning institutions, and (iii) those with neighboring municipalities that have a large electorate. I show that candidates are forward-looking: they engage in this behavior as early as one year before the election.


Research Papers

Facebook Causes Protests (with Leopoldo Fergusson)

Abstract: Using Facebook’s release in a given language as an exogenous source of variation in access to social media where the language is spoken, we show that Facebook has had a significant and sizable positive impact on citizen protests. By exploiting variation in a large sample of countries during close to 15 years and combining both aggregate and individual-level data, we confirm the external validity of previous research documenting this effect for specific contexts along a number of dimensions: geographically, by regime type, temporally, and by the socioeconomic characteristics of both countries and social media users. We find that “coordination” effects that rest on the “social” nature of social media play an important role beyond one-way information transmission, including a “liberation effect” produced by having a direct outlet to voice opinions and share them with others. Finally, we explore the broader political consequences of increased Facebook access, helping assess the welfare consequences of the increase in protests. On the negative side, we find no effects on regime change, democratization or governance. To explain this result, we show there are no effects on other political engagements, especially during critical periods, and that social media access also helps mobilize citizens against opposition groups, especially in less democratic areas. On the positive side, we find that Facebook access decreases internal conflict, with evidence that this reflects increased visibility deterring violence and that social media and the resulting protests help voice discontents that might otherwise turn more violent. 


Political Incentives and Corruption: Evidence from Ghost Students (With Leopoldo Fergusson, Arturo Harker and Juan Camilo Yamin)

Abstract: We study the effect of links between politicians on corruption under prevailing clientelism. Connections between politicians increase fabricated “ghost” students to obtain more national transfers, without raising the quality or quantity of education. Bureaucratic turnover, temporary and discretionary hiring, electoral fraud, and complaints against functionaries also increase. Effects on ghosts are larger in municipalities with more clientelism, discretion over resource spending, and weaker oversight. The findings favor a venal view of corruption, where politicians divert resources for personal gain rather than to favor their constituencies. Nonetheless, they have better future career prospects, reflecting a failure of electoral control.



(Successful) Democracies Breed Their Own Support (with Daron Acemoglu, Nicolás Ajzenman, Cevat Giray Aksoy and Martin Fiszbein), Review of Economic Studies (forthcoming).

Abstract: Using large-scale survey data covering more than 110 countries and exploiting within-country variation across cohorts and surveys, we show that individuals with longer exposure to democracy display stronger support for democratic institutions, and this effect is largely driven by individuals who have been exposed to democracies that have performed well in terms of fostering economic growth, avoiding political instability, and providing public goods. We bolster these baseline findings using an instrumental-variables strategy exploiting regional democratization waves and focusing on immigrants’ exposure to democracy before migration. In all cases, the timing and nature of the effects are consistent with a causal interpretation. We also show that support for democratic institutions matters: when there is greater support for democratic institutions, democratic political instability, and coups are less likely, and democracies perform better in the face of negative shocks.


Converging to Converge: A Comment (with Daron Acemoglu), NBER Macroeconomics Annual 2022, v.36. University of Chicago Press.

Abstract: We revisit recent evidence by Kremer, Willis, and You (2021) suggesting that the lack of economic convergence in early years has now been replaced by modest convergence. We show theoretically and empirically that failure to include country fixed effects will create a bias in convergence coefficients toward zero and this bias can be time varying, even when the underlying country-level parameters are stable. Our reanalysis finds no evidence of major changes in patterns of convergence and, more importantly, no flattening of the relationship between institutional variables and economic growth.


The Weak State Trap (with Leopoldo Fergusson and James Robinson), Economica 2022, v. 89 Issue 354.

Abstract: Development outcomes come in ‘clusters’ that seem difficult to exit. Typically juxtapositioned are states that are both weak in the sense of lacking fiscal resources, but also patrimonial or clientelistic in the way in which they operate. We document the individual behaviour underlying such a cluster using original data from Colombia. We show that tax evasion, as a measure of state weakness, and vote buying, as a measure of clientelism, are highly correlated at the individual level. We argue that while state weakness creates the right environment for clientelism to flourish, clientelism sets in place a structure of incentives for politicians and citizens that is detrimental to building state capacity. We also document that both practices are widely accepted in society, a result consistent with a deeply entrenched relationship of mutually reinforcing influences. Finally, we present evidence of a vector of other types of behaviour and beliefs that are highly correlated with both clientelism and tax evasion, which suggest the presence of multiple feedback loops that we argue justifies calling this situation a trap.


Research in Progress

Voter Learning and Measuring Support for Democratic Institutions (With Daron Acemoglu, Ceren Baysan and Cevat Giray Aksoy)

Abstract: There is widespread concern about the reported global democratic decline and its implications for economic growth. A critical factor that may determine the future of democracy is the level of voter understanding and appreciation for the quality of democratic institutions. However, it remains unclear how voters assess these institutions and if they have a shared understanding of which political parties and leaders uphold them. This study aims to evaluate the effect of credible information about democratic institutions on voter behaviour and beliefs in Turkey using a large- scale experimental information campaign, electoral data, and survey data. The information source is experimentally varied as non-partisan or aligned with either right or left-wing ideologies, allowing us to control for voter beliefs of the credibility of the information or reputation of the information source. Our research design also allows us to isolate the impact of information on voter beliefs and behaviour from any persuasive effects.


The Return of Pachamama (With James Robinson and Pablo Selaya)

Abstract: We study the political and economic consequences of the violation of the "moral economy" of rural Bolivia, based on coca, caused by the escalation of coca eradication in the 1990s. We show that this policy is associated with the rise of the MAS political party - their vote share is significantly higher both in coca-suitable places and in the presence of traditional socio-political institutions notably the Aymara ayllu. We then study the consequences of controlling the state after 2005. We present evidence of an "empowerment effect" in the form of an increased use by indigenous people of indigenous first names for their children. Using survey data we show that indigenous people significantly increased their expectations of social mobility after 2005.    


Voting Behavior and Female Representation: Experimental Evidence from Turkey (With Daron Acemoglu, Ceren Baysan, Antonia Paredes-Haz and Gamze Zeki)

Abstract: This project explores whether policies promoting gender equality remain underrepresented in conservative countries because political elites misperceive voter demand for such policies. To estimate voter demand, we use experimental variation to evaluate voter response to campaign promises on "gender issues" compared to voters in a control group who are not exposed to any campaign. To disentangle party or canvasser persuasion effects (supply) on voter behavior from campaign content (demand), we estimate the differential effect of the gender-related campaign to a placebo, a second campaign treatment arm on "general issues." The outcomes for each campaign treatment arm include local election vote shares and civic engagement, as measured by participation in town hall meetings. The meetings are cross-randomized as “deliberative” or "top-down." This design also allows us to observe whether supply-side factors, like elite persuasion, affect specific policy preferences from demand-side factors like voter deliberation and democratic processes.