Research

Zero-sum thinking and oppositional identities pose essential threats to democracies and their ability to conduct mutually beneficial reforms. Zero-sum narratives contribute to this problem, but can aspirational narratives also be part of the solution? We test this conjecture in the context of a surprising and highly contentious political decision that forced society and politics to decide within a limited time frame on a particularly polarized issue: the future of artisanal gold mining in the Peruvian Amazon. We conducted a competition among professional filmmakers to create an aspirational narrative video treatment portraying a possible mutually beneficial future with cleaner and more formalized mining. We then use this video as a treatment, and run a pre-registered experiment with 400 participants in two major Peruvian cities. Compared to a passive control group, the narrative video significantly shifts beliefs about miners from being clear villains to possible heroes. We find that this not only shifts beliefs and expectations, but also real-stakes outcomes: Treatment participants are about 15% more likely to sign a letter supporting mutually beneficial reforms. We highlight a decrease in zero-sum thinking and a change in in- vs. out-group identity as two main mechanisms.

Narratives significantly shape economic and political decision-making processes, yet understanding their causal impact remains a critical challenge. This paper investigates the influence of narratives on economic choices, specifically within the context of environmentally sustainable mining in the Peruvian Amazon. We focus on how weaving information on clean technology into narratives affects artisanal miners' decisions. The study contrasts groups of miners receiving information embedded in an aspirational narrative against a control group that receives information in a factual manner. Our study aims to provide valuable insights into the effectiveness of narratives in policy settings.

Does populist rhetoric merely exploit existing social divisions, or does it actively reshape political discourse? We study this question by exploiting the sudden rise of Donald Trump during the 2016 US presidential campaign as a natural experiment. Using a comprehensive dataset of over 1 million Twitter posts from 2010–2021 and a difference-in-differences design comparing the US to other English-speaking democracies, we show that Trump's nomination caused a significant increase in villain narratives — a defining feature of populist communication that casts political opponents, institutions, and media as enemies rather than legitimate adversaries. Focusing on climate change discourse to isolate narrative effects from real-world events, we find that villain narratives increased by 6 percentage points following Trump's May 2016 nomination. We document heterogeneous effects across Republicans, Democrats, and Independents, suggesting that the effect operates through social learning and strategic adaptation, not just mobilization of existing grievances.

The effectiveness of political narratives as a communication technology depends on their virality and on the persuasiveness of single narrative exposure. To analyze narratives empirically, we introduce the political narrative framework and a pipeline for its measurement using large language models (LLMs). The framework captures the essence of a narrative by its characters, who are either neutral or cast in one of three drama triangle roles: hero, villain, or victim. Using 1.15 million U.S. climate policy tweets from 2010–2021, we find that political narratives are consistently more viral than comparable neutral tweets. This result is robust to conditioning on a rich set of fixed effects, author characteristics, language metrics and emotionality. Hero roles increase virality by 56%, but the biggest virality boost stems from using villain roles (152%) and from combining other roles with villain characters. To examine the persuasiveness of single exposure to some of the most frequent and viral character-role combinations, we use three pre-registered online experiments with 3000 participants. The results show that narrative exposure influences beliefs and revealed preferences about a character, while a single exposure is not sufficient to move support for specific policies. Political narratives also lead to consistently higher memory of the narrative characters and their roles, while memory of objective facts is not improved. Taken together, the political narrative framework provides a measure that moves beyond emotions and linguistic features, helps to explain virality, and is linked to shifts in beliefs, revealed preferences, and memory.

The spread of propaganda and misinformation from autocratic regimes is a growing concern in democracies. We study the European Union's ban on Russian state-backed news outlets after the 2022 invasion of Ukraine, analyzing 677,780 tweets from 146,633 users with a difference-in-differences design in a daily panel. The ban reduced pro-Russian tweets by 10.9% per active user day, with strongest effects among users directly connected to the banned outlets. We find no evidence of substitution to secondary suppliers. Evidence on mechanisms indicates that the ban curtailed pro-Russian content by removing key agenda-setters. Finally, we examine the costs of censorship in a democratic context: A pre-registered experiment finds reduced satisfaction with free speech, particularly among political centrists.