Analyzing the influence of transportation and macroeconomic determinants on Chinese inbound tourism: a Markov switching model using Lasso estimation

Abstract

This empirical study examines the influence of transportation and Macroeconomic determinants on Chinese inbound tourism from 1995 through 2018 by using the Markov switching model based on Lasso estimation. Due to sustainable economic growth, the further implementation of the Reform and Opening policy (1978) and the substantial increase in people's personal income, China's tourism industry has experienced unprecedented development in recent years. The huge outbound market (outbound tourism of Chinese tourists) has attracted worldwide attention. The domestic market has expanded steadily. However, the development of the inbound market (foreigners travelling to China) is lower than the outbound market and the domestic market. There are many factors affecting Chinese inbound tourism. The development of inbound tourism may be affected by factors such as international exchange rates, national income, and consumption levels. At the same time, some unexpected or major events may also promote or hinder inbound tourism. This paper focuses on analyzing the impact of transportation and macro factors on Chinese inbound tourism. To sum up, the growth rate of China's inbound tourism in 2018 was 1.2 %, while China's outbound tourism was growing at 14.7%. Following this trend, the scissors gap between outbound and inbound tourists will continue to widen. From an economic point of view, the number of outbound tourists far exceeds the number of inbound tourists, which will lead to a continuous deficit in tourism trade and affect the balance of payments. Briefly, the purposes of this study are as follows: 1) to measure the cycle of China's inbound tourism and find the high-growth stage and low-growth stage of inbound tourism. 2) identify the transportation and macroeconomic factors affecting the top ten source countries for Chinese inbound tourism. We considered the top 10 source countries for China's inbound tourism demand as of 2018. As data on Myanmar tourists and Vietnamese tourists are not available, we collected South Korea, Japan, Russia, USA, Mongolia, Malaysia, Philippines, Singapore, India and Canada. We used RLO (China railway length in operation), HWL (China highway length), CPL (China coastal major port length of quay), RPL (China river major port length of quay), NIAR (China number of international air route), LIAR (China length of the international air route), APT (China number of airport), IAPT (China international air passenger traffic), RFAI (China fixed-asset investment on the railway), NLR (China navigable length of the river) and ADU (China aircraft daily utilization) to measure China's transportation development level. The macro-factor independent variables are GDPR (China's GDP per capita growth rate), UNPR (unemployment rate), INFR (inflation rate), and EXR (official exchange rate (USD / CNY). The main contribution of this paper is to estimate the Markov switching model based on Lasso estimation. On the other hand, most of the existing studies use the total number of arrivals or total inbound tourism revenue to measure the business cycle of the tourism industry of one country. This paper collects the top 10 countries' tourist arrivals to China and aims to find out the tourism business cycle of each source country. The empirical results show that the nonlinear model is more suitable for analyzing the tourism business cycle than the linear model. The estimation results of the two nonlinear models are roughly the same, but the MS-Lasso model is better than the MS-Ridge model in the goodness of model fitting. Feature selection screened out the main factors that affect China's inbound tourism. They are InHWL (China highway length), InIAPT (China international air passenger traffic), InNLR (China navigable length of the river), InADU (China aircraft daily utiliation), GDPR (China's GDP per capita growth rate) and UNPR (China's unemployment rate). For China's inbound tourism business cycle, the possibility of being in a high-growth stage is large, with long persistence and stability. The possibility of being in a low-growth stage is small, with short persistence and variability.