||An Empirical Investigation of Tourism Demand Variability:The Gini Index and Entropy Measure Approach
NAKAHIRA KazuhikoYABUTA Masahiro
Focusing on the tourism in Japan, this paper examined the seasonality of tourism demand in Japan by utilizing Gini coefficient and some kinds of entropy measures. Regarding the estimated Gini coefficient, it is large in Hokkaido, Hiroshima, and Okinawa, and small in Tokyo and Osaka. In addition, the east Japan great earthquake happened in 2011 might affect seasonal fluctuations of many prefectures. After the earthquake, the seasonal fluctuations expressed by the Gini coefficient seem to vary from year to year in spite of the general variation of the number of tourists. Several points were worth considering through the analysis by estimated entropy measure for the Japanese selected 10 prefectures in the sample period from the year 2008 to 2016. First, there is the one-time rapid increase in seasonality in tourism demand for major four prefectures in 2010 and rapid reversal in next year. In addition, the other two prefectures, have the same pattern in 2011 although the levels of fluctuations are relatively low. The examination by utilizing the estimated Theil measure shows that the highest concentration of seasonality is recorded for Tokyo and for the other prefectures located in the east and west areas in Japan just like the case of entropy measure. In addition, Miyagi and Fukushima follow the same pattern in next year. Okinawa prefecture have the highest seasonality in 2014 during the period under our study. The evolution patterns of seasonality for the prefectures described by the estimated relative Theil measure are almost the same as the ones measured by the two indicators described above. The analysis by the entropy decompositions at the intra- and the inter-monthly levels are also conducted. The estimated seasonality of Japanese selected 10 prefectures that could be explained by the intra-monthly part of entropy measure revealed that the seasonality for the half of the total number of prefectures reflects a relative increase in 2010, and the one for the other two prefectures follows the same pattern in next year. The graphs of the estimated inter-monthly part of entropy measure described the increase in tourism demand seasonality in 2010 like the other measures. Tokyo experienced the relatively rapid reduction in seasonality in next year as a reactionary fall. Okinawa had a certain degree of increase in seasonality again in 2013, and had a fluctuation of seasonality during the following periods. Finally, consideration of the tourism seasonality based on the relation between interest groups - origins and destinations - by applying the entropy theory are implemented. The estimated entropy measures for annual amount of arrivals from origin ( i ) to destination ( j ) for the selected major six courses are examined. In particular, the seasonality for the routes Hokkaido→Tokyo, Kyoto→Tokyo, and Okinawa→Tokyo decreased rapidly in 2009, and increased drastically in 2010 as a rebound. In short, the routes from the three major areas to Tokyo indicate temporal growth of seasonality. All the measures based on the entropy theory show a very similar pattern in that the tourism seasonality has the temporal rapid growth in 2010 and the reactionary fall in 2011. Probably, the financial and economic crisis of global economy occurred around 2008 and 2009 had a negative influence on Japanese tourism in 2010. By the effect of the crisis, the number of visits or visiting frequency of tourism in Japan decreased, and its downturn might generate the concentration of tourism in specific season or month as a result of selective behavior by tourists during the hard times.