SOEPpapers on Multidisciplinary Panel Data Research 767
The aim of this study is to estimate the causal effect of cultural participation on health status. Cultural activities may directly impact upon health through palliative coping or substituting health-compromising behaviors. Cultural engagement may also facilitate the development of social networks, which can improve health via social support and the dissemination of social health norms. Previous estimates on the arts-health relationship are potentially biased due to reverse causality and unobserved heterogeneity. Using individual-level data from Germany, we employ propensity-score matching methods. The treatment group is confined to individuals that visit cultural events at least once a month. The participation equation includes a rich set of personal characteristics that cover the respondents' demographic and social background, social capital and leisure-time activities, health-related lifestyle, personality and childhood environment. We explicitly consider reverse causality by including the pre-treatment trends in health outcomes among the covariates. To deal with time-fixed unobserved heterogeneity, we combine the matching model with a difference-indifference approach. We find that frequent cultural-event visits are unrelated to health once we account for unobserved persistent differences across individuals. However, examining the dose-response relationship we find positive mental-health effects of low levels of cultural participation compared to non-attendance. Our results may thus yield important insights on the effectiveness of arts participation as a means to reduce social inequalities in health.
cultural participation mental health physical health propensity-score matching multivalued treatment