Tourism, real output and real effective exchange rate in Malaysia: a view from rolling sub-samples

The objective of this study is to examine the tourism-growth nexus for Malaysia with the cointegration and Granger causality tests. This study covers the monthly data from January 1989 to May 2010. The Johansen’s cointegration and the residuals-based test for cointegration with regime shift consistently suggest that tourist arrivals, real output, and real effective exchange rate in Malaysia are cointegrated. In terms of Granger causality, this study finds different sources of causality. In the short run, real output and real effective exchange rate Granger-cause tourist arrivals, while tourists arrivals also Granger-cause real output and real effective exchange rate. In the long run, this study shows that all the variables are bi-directional causality. Moreover, we also extend the study to analyse the stability of causality between tourism and real output by using rolling regression procedure into the Granger causality test. Interestingly, the rolling Granger causality test demonstrates that the growth-led tourism hypothesis is valid and stable, while tourism-led growth hypothesis is valid and but unstable in particular after 2005. Although tourism contributes to economic growth, it is not a persistence source for long-term economic growth in Malaysia.