This study analyzes the heterogeneous treatment effects of the COVID-19 pandemic on regional tourism demand in South Korea, focusing on the role of geographic distance from the metropolitan area to tourist destinations and the spatial characteristics of tourist destinations. Since a substantial portion of the population resides in the capital region, it can be expected that regional tourism demand is largely driven by residents of the capital region. In addition, the pandemic has particularly discouraged visits to indoor and densely populated areas due to increased perception of infection risk. To estimate these effects, we use a causal machine learning approach using double machine learning, analyzing monthly visitor data from 994 major tourist sites between the years 2019 and 2020. Tourist destinations are classified by spatial characteristics, including indoor, outdoor, and mixed settings as well as by tourism type. The analysis reveals that the impact of COVID-19 was more pronounced for indoor destinations located closer to the metropolitan center, whereas outdoor and mixed destinations showed little variation in treatment effects by distance. These findings highlight the importance of adopting distance-sensitive and space-specific policy measures in tourism planning during pandemics. Our study also demonstrates the practical utility of causal machine learning in tourism analytics, suggesting its potential for enhancing policy precision and resilience against future public health crises.

1. 서 론
2. 선행연구
    2.1 인과머신러닝과 이질적 처치효과
    2.2 코로나19와 관광수요
3. 연구 설계
4. 분석 결과
5. 결 론
Acknowledgement
References

• Won Seok Yang(Department of Business Administration, Hannam University) | 양원석 (한남대학교 경영학과) Corresponding author