Modelling tourism receipts and associated risks, using long-range dependence models

Tourism receipts have important policy implications for destination countries in terms of government revenues and the management of tourism-related policies. This article uses time series models to analyse the risk exposure reflected in the growth rates of tourism revenues. To do so, we apply risk management measures based on value-at-risk (VaR) and the expected shortfall (ES), analysing monthly data for six Spanish regions from January 2004 to March 2017. Two main results were obtained. Firstly, tourism receipt growth rates present negative long-range dependence. In other words, they have intermediate memory or anti-persistence and therefore show signs of dependence between widely separated observations. Moreover, we detected the existence of long-range dependence in these volatilities in one of the six regions considered. Secondly, we show that VaR based on Generalized Autoregressive Conditional Heteroscedasticity (GARCH)-type models is a valid means of analysing the risk exposure of tourism receipt growth rates, doing so by evaluating various in-sample and out-of-sample VaR thresholds and the ES.