Evaluating the impacts of imported cultural goods on carbon emission: Cross-country evidence

Abstract

Low attention has been paid to the environmental effects of imported cultural commodities, which serve as a medium for cultural engagement and information sharing around the globe. Using the generalized method of moment (GMM) approach, this article empirically assesses the effects of imported cultural products on carbon emissions using cross-country panel data from 2004 to 2017 for 121 countries. The empirical findings reveal that the environmental consequences of imported cultural commodities vary among income level country groups. Overall, a 1% increase in the imports of total cultural goods imports would decrease carbon emissions significantly by 0.002%. After dissecting the sample into two groups, we find that a 1% increase in total cultural goods imports would lead to a significant decrease in carbon emissions of higher-income countries by 0.008%, but a significant increase in the emissions of carbon dioxide of lower-income countries by 0.002%, thus showing the importance of worldwide cultural interaction to carbon emissions. In addition, we find that performance & celebration goods and audiovisual & interactive media goods are primarily responsible for the inhibiting effects of imported cultural goods on the carbon emissions of higher-income countries. Imports of the majority of cultural products (with the exception of heritage items) would result in a significant rise in carbon emissions for low-income nations, a conclusion that is consistent with that of overall cultural goods imports. Gross domestic product per capita and economic complexity index have consistently negative effects on carbon emissions across several models for both nation groups, although energy consumption is the most significant contributor to carbon emissions. In addition, using Lasso Poisson regressions, we undertake robustness tests to assess the dependability of our empirical results based on the GMM methodology. The Lasso Poisson estimation results are consistent with our earlier findings.