Abstract:
Based on representative data for 1614 citizens in Germany, this paper empirically examines the relationship between different types of environmental protection activities and subjective well-being (SWB) in terms of life satisfaction by specifically considering the role of economic preferences for this relationship. With respect to pro-environmental behavior, we differentiate between stated non-climate environmental and climate protection activities as well as revealed climate protection activities, which are measured in an incentivized donation experiment and thus are more meaningful than stated climate protection activities. Our empirical analysis reveals that climate protection activities are more robustly and more strongly positively correlated with life satisfaction than non-climate environmental protection activities. Furthermore, not only stated climate protection activities, but also revealed climate protection activities are significantly positively correlated with life satisfaction. These results suggest that climate protection activities lead to stronger warm glow feelings and reputation gains than non-climate environmental protection activities. Our empirical analysis additionally shows that economic preferences play an important role since especially patience and trust, but also risk-taking preferences and (less robust) altruism are significantly positively correlated with life satisfaction. In particular, economic preferences are also relevant for the relationship between pro-environmental behavior and life satisfaction. When economic preferences are included in the econometric analysis, the estimated correlations between climate protection activities and life satisfaction become weaker and the estimated correlation between non-climate environmental protection activities and life satisfaction even becomes insignificant. These results strongly suggest omitted variable biases in cross-sectional econometric analyses of the relationship between pro-environmental behavior and SWB when economic preferences are not included as control variables.