Are shocks to tourist arrivals permanent or transitory? A comprehensive analysis on the top 20 most-visited countries

ABSTRACT

Tourism sector is highly vulnerable to shocks due to its characteristics. The purpose of this paper is to investigate whether the effects of the shocks on the tourism sector would be permanent or transitory. To this end, I analysed the number of tourist arrivals in the top 20 most-visited countries for the 1995–2018 period. I employed a battery of cross-sectional dependency tests, second-generation panel unit root tests without and with structural breaks. The findings significantly change from non-stationary to stationary when structural breaks are incorporated. The results suggest that tourist arrivals to top 20 most-visited countries are stationary under sharp structural breaks. This finding is important in that shocks hitting tourism sector are transitory and the effects of shocks will eventually die out, but the new equilibrium will not converge to its pre-shock level.

KEYWORDS:

• Tourism
• tourist arrivals
• shocks
• panel unit root test
• cross-sectional dependency
• structural breaks

Acknowledgements

I would like to thank the editors Dr. C. Michael Hall and Dr. Chris Cooper and anonymous reviewers for their insightful comments, which significantly improved the manuscript.I would also like to express my gratitude to Dr. Şaban Nazlıoğlu and Çağın Karul of Pamukkale University for their help in the estimations. The codes for all the tests applied in this study can be accessed through https://github.com/aptech/tspdlib. Last but not least, I am grateful to Halil Yücel of Erciyes University for carefully proofreading the manuscript. Needless to say, all the remaining errors are solely my responsibility.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 By tourist arrivals, we employ arrivals of non-resident overnight visitors at national borders. By this definition, both the arrivals of the same day visitors spending only a few hours and foreign residents in the host country are excluded.

2 The United Arab Emirates is the only country which was dropped in the top 20 due to missing data between 2006–2013. Instead of the United Arab Emirates, I included India.

3 It is worth noting that each cross section test proposed in the literature is applicable only in certain cases. If the results contradict each other, the researcher should examine the power and size properties of these tests. Interested reader may refer to Pesaran et al. (2008, pp. 112–118) for a comparative table of power and size properties of LM, LM_adj and CD tests.

4 Interested reader may refer to Bai and Ng (2004) and Breitung and Pesaran (2008) for a comprehensive survey of the first generation and second generation panel unit root tests.

5 It might seem as too restrictive to assume that there are only one or two structural shifts. To overcome this problem, there are some studies which performed tests with multiple sharp structural breaks. We restrict our analysis to consider only up to two breaks for two reasons. Firstly, our time dimension in each section is not long enough to allow for multiple breaks. Secondly, as Lee and Tieslau (2019a, p. 7) argues, a series with more than two breaks might best be modelled as a non-linear process.