Abstract: In this study, we collect annual data for 30 regions in China using tourism receipts and gross domestic product over the period of 1980-2020. Empirical results from Bootstrap Toda-Yamamoto Granger Causality test with a Fourier function, which consider both cross-sectional dependence and multiple smooth breaks, support the evidence of feedback hypothesis in most of the regions except for Jilin that we find the validity of tourism-led-growth hypothesis, Shanxi, and Qinghai where we find the supportive evidence of growth-led-tourism hypothesis, and for Gansu, Ningxia, and Xinjiang where we find the strong evidence of neutrality hypothesis. These results further show the importance of incorporating cross-sectional dependence and smooth breaks in testing the Granger causality. Our empirical results have important policy implications for the regional governments in China conducting tourism policy to sustain its regional economic growth.

Keywords: Regional Analysis, Panel Data, Sustainability, Cross-Sectional Dependence, Smooth Breaks

JEL codes: C33, Z3

DOI: ...