Backlash or norm clash: understanding rising gender violence in equal times

Across many democracies, gender-based violence against women (GBVAW) persists—and even appears to rise—despite notable progress in gender equality. This apparent paradox has become central to contemporary scholarship, prompting competing explanations. Three hypotheses dominate the debate: (1) disclosure effects, whereby violence is not objectively increasing but seems to rise in high-equality contexts because it is perceived and reported more; (2) backlash, in which rising violence reflects resistance to feminist progress and is a response to anti-gender mobilisation; and (3) norms clash, whereby broader gender equality norms in society, fostered by advances in public and economic spheres coexist with persistent inequalities in private and intimate domains. Distinguishing between these mechanisms requires assessing whether increases in GBVAW reflect genuine escalation or heightened visibility produced by institutional and cultural change. Given the difficulty of distinguishing between these dimensions of violence, no study to date has empirically tested these mechanisms jointly to explain the evolution of GBVAW, nor systematically compared objective and reported levels of violence over time. This paper contributes new evidence through a comprehensive longitudinal analysis that, for the first time, integrates all available indicators of GBVAW into a single dataset. Existing studies often examine individual forms of violence in isolation, risking biased conclusions because different forms of GBVAW may respond to different social, political, or cultural drivers. By combining survey-detected and administrative data, our integrated dataset enables systematic differentiation between objective prevalence and reporting effects, offering a clearer picture of long-term trends. We empirically test the mechanisms proposed in the literature. Using a newly constructed dataset covering all forms of GBVAW in Spain from the 1990s to 2025, we apply time-series analysis, difference-in-differences, and regression discontinuity approaches to disentangle these mechanisms. To our knowledge, this is the first dataset worldwide to integrate all forms of GBVAW over time and across regions, offering a unique empirical foundation for understanding the evolution and drivers of GBVAW.