Title: A predictive model of revenue per available room in the Eastern Mediterranean using bilateral exchange rates
Authors: Christopher M. Hannum; Kerem Yavuz Arslanli
Addresses: Istanbul Technical University, Istanbul, Turkey ' Istanbul Technical University, Istanbul, Turkey
Abstract: In an era of exchange rate volatility, performance of hospitality industry assets may be highly sensitive to currency shocks which affect international demand, such as the fall in the value of the Russian Ruble in late 2014. We aim to create a systematic and meaningful predictive model to give insight for managers and investors in the hospitality industry in this region regarding which bilateral exchange rates to watch most closely and particularly the timing of anticipated effects. This study will examine the relationship between Revenue per Available Room (RevPAR), the key measure of operational performance in the hospitality industry, and bilateral exchange rates in eight countries of the Balkans and Eastern Mediterranean. The study methodology follows a LASSO penalised regression model for simultaneous and systematic selection of variables and lag lengths and estimation of model parameters.
Keywords: hotels; hospitality industry; tourism; hotel demand; demand forecasting; penalised regression; LASSO; elastic net regression; cross-validation; prediction accuracy; predictive modelling; revenue per available room; Balkans; Eastern Mediterranean; bilateral exchange rates; operational performance.
International Journal of the Built Environment and Asset Management, 2015 Vol.1 No.4, pp.350 - 367
Received: 16 Nov 2015
Accepted: 31 Dec 2015
Published online: 01 Mar 2016 *