Assessing the impact of visa regime liberalization on inbound tourism: a combination of the gravity model and machine learning

Abstract

The article assesses the determinants of inbound tourist flows to Uzbekiston and the impact of the 2018–2019 visa liberalization reform. A “Tourist Attractiveness Factors Model” is proposed that combines a gravity model, an event study and machine learning algorithms. The gravity model is estimated by Poisson pseudo-maximum likelihood using official country-level data for 2014–2025. Tourist flows are driven by the population (elasticity 0.83), income (0.89) and distance (−1.59) of origin countries; flows from neighbouring countries are 7.5 times higher and from CIS countries 11.4 times higher. The visa effect built up over several years and in 2022–2024 corresponded to a 102–178% increase in flows. XGBoost achieved the highest out-of-sample accuracy (R² = 0.850). The potential of the Indian, Pakistani and Bangladeshi markets is found to be largely untapped.