Key points are not available for this paper at this time.
This study evaluated the impact of Spain's 2010 Anti-Smoking Law on cigarette sales using a hybrid counterfactual and forecasting framework that combines econometric and machine learning models. Monthly provincial data observed from January 2005 to August 2025 are used to estimate the historical effects of the law (2011-2013) and to generate prospective projections under alternative scenarios for the period 2025-2027. Six time-series models-ETS, Harmonic, NNAR, SARIMA, STLAR, and STLM-were compared using a multi-metric validation scheme to ensure robustness. Results indicate that the 2010 law generated an immediate and persistent reduction in cigarette sales throughout Spain. Across models, the cumulative national decline over 2011-2013 ranged between 0. 37 and 2. 45 billion cigarette packs, equivalent to a 15%-30% contraction in the tobacco market. Projections from September 2025 to December 2027 suggest continued stabilization or slight decreases in cigarette sales, indicating the long-term persistence of the law's impact. Methodologically, the study demonstrates that hybrid time-series and machine learning ensembles outperform single-model approaches in capturing structural breaks, nonlinearities, and seasonality shifts associated with behavioral regulation. The results confirm the effectiveness of comprehensive tobacco control policies in achieving sustained public health gains. Overall, Spain's experience exemplifies how evidence-based regulation can produce lasting declines in harmful consumption and provides a replicable framework for evaluating future public health interventions.
Álvarez et al. (Thu,) studied this question.