![Sample our Tourism, Hospitality and Events journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5950/TOC-Subject-Sample-2021-Tourism-Hospitality-Events.jpg"})
ABSTRACT
If, when and where to travel is normally planned in advance. Hence, the exchange rate at the time of booking accommodation is likely to play an important role in determining inbound tourist demand. That fluctuations in the exchange rate influence inbound hotel demand is well established in the literature, but the lag with which it influences demand has received limited attention. This paper sets out to remedy this, by estimating the elasticity of hotel demand and its lag structure with respect to exchange rate fluctuations. Our data consists of a panel of hotel beds sold per month in Norway to tourists from ten different source countries. Models with source country fixed effects were used to obtain the estimates. We find evidence that demand is approximately unit elastic and that it operates with a lag of two to three quarters with respect to the exchange rate. The latter result is key since it enables improved medium-term demand forecasts in the tourist industry.
Acknowledgements
We are grateful for the helpful comments and assistance from our colleagues Gjermund Grimsby, Jens Fredrik Skogstrøm, Torbjørn Bull Jenssen and Kristoffer Midttømme. We would also like to thank André Kallak Anundsen at Norges Bank for his valuable comments. Data and codefiles are available on request from the authors.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 Using the same dataset and methodology, we have also estimated if elasticities facing hotels which primarily serve business travellers as opposed to holiday travellers differ. We find that the price elasticity facing the former group of hotels (0.824) is about twice as high as that facing the latter (1.641), while the lag structure is identical to the remaining estimates in the paper. As this merely confirms a result that is well known in the existing literature (see e.g. Peng et al., Citation2015), we do not report the results in detail.
2 Several papers have employed the relative price level or tourism-specific relative price levels. We reference several of these in the data chapter where we argue that exchange rates are a more appropriate price measure in our context.
3 While producer-country pricing is found to be more prevalent in many types of goods in the long run, in this study of 23 OECD countries.
4 The NOK/DKK exchange rate is not included in as it was pegged to the Euro during the entire sample period.
5 Levin, Lin, and Chu recommend using their procedure for moderate-sized panels, with an N between 10 and 250 and a T between 25 and 250. Our panel has N = 10 and T = 130.
6 The number of observations falls by ten when the real household consumption variable is included. The most recent observation for overnight stays is May 2016. Since the latest available observation of real household consumption is Q1 2016, we are forced to not include May 2016 in specifications where real household consumption is controlled for.
7 All are highly significantly different from zero, apart from that of the mid-priced hotels, which is significant only at the 11 percent level.
8 The test is implemented in STATA by use of seemingly unrelated regression (command -suest-), followed by a Wald test. It tests the joint fit of the two models with the restriction of equality between the coefficient for against the joint fit of the models with no such restrictions. As -suest- does not allow models estimated with the standard fixed effects command, all variables were manually demeaned. An OLS-model on these transformed variables is equivalent to running the standard fixed effects regression command on the non-transformed variables, but by doing the transformation manually, it is possible to use seemingly unrelated regression tests on the models.