“Buy the rumor, sell the fact”: political narratives and stock prices in an emerging equity market

Purpose This paper asks whether political narratives can generate large, temporary price movements in an emerging equity market, not fully explained by observable fundamentals and benchmark movements, using Taurus Armas S.A. (TASA4), Brazil's only listed firearms manufacturer, around the 2018 election, in which gun policy was a defining issue. Design/methodology/approach Daily B3 prices for TASA4 (2000–2026), the Ibovespa and 15 comparison firms support mean-adjusted and market-model event studies across four Brazilian election cycles (2010–2022), with FDR correction and placebo tests, triangulated with a pooled CAR regression, an EPU check, Wikipedia data, a DCF valuation and a 2022 electoral-cycle check. Findings TASA4 rose 386.7% and then fell 62.6% within three sessions around the 2018 election, with abnormal returns up to +52.4% and −68.6%, robust to the market model, FDR correction and placebo tests, and volume up to 165 times baseline. Neither the comparison-firm nor cross-sector benchmarks show comparable movement; a pooled regression indicates TASA4 remains elevated after controlling for liquidity and uncertainty, though attenuating with policy salience. Wikipedia attention builds ahead of the peak. Research limitations/implications This is a single-firm case study, not a definitive causal test; order-flow and social-media data remain unavailable, and the 2022 electoral-cycle check should be read as exploratory. Practical implications Politically exposed emerging-market stocks can move far more than diversified benchmarks around elections, relevant for risk managers, retail investors and investor-protection policy. Social implications The paper underscores the growing social importance of narrative economics in understanding how political discourse affects economic decisions, investor psychology and public perceptions of risk and opportunity. Originality/value The paper combines a comparison-firm benchmark, cross-sector panel, pooled CAR regression and Wikipedia data in one study, evaluating three competing mechanisms – rational updating, behavioral overreaction, narrative speculation – rather than assuming mispricing by default.