Abstract
This study examines the effects of economic uncertainty on tourism demand in India for the period 2006:01 to 2018:04. This study has two major contributions. First, using the nonlinear autoregressive distributed lag technique, the asymmetric effects are tested and second, it provides evidence for both long- and short-run effects. The estimated results have confirmed that economic uncertainty has asymmetric effects on tourism demand. Specifically, it is shown that increasing uncertainty has more penetrating negative effects than the positive effects caused by decreasing uncertainty.
Introduction
A good number of studies have tested the effects of the political economy and economic policy on tourism (e.g. see Bianchi, 2018) and have found that the interlinkage is worth investigating. However, the measurement of political and economic uncertainty has always been a challenging task. Recently, Baker et al. (2016) have developed a system index of economic policy uncertainty that is called the economic policy uncertainty (EPU) index. Taking advantage of this development, we attempt to test the effect of economic policy (uncertainty) on the tourism demand for India. We select India for examination not randomly. According to the World Travel and Tourism Council, India’s travel and tourism were growing at 6.7% and accounting for 9.2% of the total economy in 2018. The country is the largest market in South Asia and has been the key driver for the growth of the sector in the region. In 2019, the industry is expected to grow by 8.5% to contribute Indian rupees 18.34 trillion (US$260 billion) to the economy and account for 9.3% of the gross domestic product (GDP). On the other side, it is also noteworthy India is a multiparty federal parliamentary democratic republic and economic policies are often changed when the government changes. In fact, the economic policy shift is a way of life in India.
Economic policy uncertainty can affect tourism in a variety of ways. First, economic uncertainty greatly influences international trade and foreign investment. While, the business travel demand, both inbound and outbound, is a function of international trade and foreign investment (e.g. see Endo, 2006 and Kulendran and Wilson, 2000). Thus, the level of economic uncertainty can affect tourism through international trade and foreign investment channels. Second, at the time of economic and financial crisis, both inbound and outbound tourism is negatively affected. Foreign tourists avoid going to countries that are suffering from the crisis. Song and Lin (2010) have shown that the financial and economic crisis has a negative effect on both inbound and outbound tourism in Asia. It is noteworthy that to some extent, the business cycle depends on economic policies. Third, the robustness of price level and the exchange rates are heavily dependent on economic policy certainty and both factors are key components of the tourism demand function (see Dogru et al., 2017). Finally, economic policy uncertainty on taxes, subsidy, interest rate, inflation, land, natural resource extraction, and employment often leads to political protests and violence in developing countries, which deters foreign tourists to visit the affected country (Neumayer, 2004).
We use nonlinear autoregressive distributed lag (NARDL) model of Shin et al. (2014) to ascertain the possible nonlinear link between economic and political uncertainty and tourism demand. NARDL is an extension of the ARDL of Pesaran et al. (2001) in the context of asymmetric relationships. The selection of the method is important as NARDL offers an error correction (EC) process that incorporates the asymmetries into the long-run cointegration. The method facilitates testing of the asymmetric responses of the tourism demand to negative and positive changes in political uncertainty. The positive and negative movements in the uncertainty may cause different effects, which we attempt to ascertain in this study. Specifically, by using this technique, we are interested to know whether the asymmetric changes, that is, positive and negative, in policy uncertainty have the same effect or a different impact on Indian tourism. Also, the technique is most appropriate for our analysis since it yields long-run as well as short-run asymmetries estimates.
Demir and Gozgor (2018) and Madanoglu and Ozdemir (2018) using the EPU index attempted to test the effect on tourism. However, these studies suffer on two accounts. First, they used annual data, while the EPU index is designed to capture the short-run uncertainty more appropriately. A long-run (annual) analysis using these data may not be quite suitable as the short-run variations are settled down in the long run. Second, the study has ignored the asymmetric effects on tourism, which is likely to be present. Two recent studies, Tsui et al. (2018) and Ersan et al. (2019) published in this journal have also examined the effects of economic uncertainty. Both studies have utilized monthly data and conducted panel analysis. However, the asymmetric effect of uncertainty is not tested in both studies. We take care of these issues in this article. Specifically, this study tests the asymmetric effects of economic policy uncertainty on Indian tourism. To the best of our knowledge, this study is one of the earliest attempts using the NARDL framework for uncovering a complex asymmetric association between EPU and tourism demand in long as well as short run.
The model and data
To test the effect of economic uncertainty on the tourism, following Martins et al. (2017), we specify the tourism demand function of India as follows:
We specify the following asymmetric long-run equation of the tourism demand model as follows:
where LTA is foreign tourist arrival (TA), LWGP is the world GDP to capture the income effect or business cycle effect, and LEPU is an economic uncertainty index. LREER is real effective exchange rate. All variables are logarithmic form. LTA and world gross domestic product (WGDP) are seasonally adjusted.
For the purpose of our empirical analysis, we use monthly data of the number of foreign TA in India from 2006:01 to 2018:04. To measure policy-related economic uncertainty in India, we utilize EPU index (see Baker et al. (2016) for a detailed discussion on the index). The EPU index on India is constructed the index considering the information from seven leading Indian newspapers. After the normalization of the information, an index on EPU is created. The literature shows that relative price and exchange rates are potential determinants of foreign tourism demand; however, they are obviously correlated that causes multicollinearity problem. To avoid such a problem, we estimate the effects of exchange rate and relative price by using real effective exchange rate (REER). This is the trade-weighted indices constructed using 36 countries’ information on currency exchange and prices against Indian Rupee. To capture the income effect on tourist demand, we use the index of WGDP 1 . TA data series comes from the Ministry of Tourism database, Government of India, while REER and WGDP are retrieved from the Handbook of Statistics on Indian Economy by Reserve Bank of India and International Financial Statistics data set by IMF, respectively.
Empirical results
The results of equation (2) are presented in Table 1. Panel A presents the results of NARDL equation in which LTA is dependent variable and all explanatory variables are at level. Panel B shows results of the long-run form in which dependent variable is the first difference of LTA and explanatory variables are included both at the level and at the first difference and panel C presents results of the short run (the error-correction model equation). The estimated statistics of F-bounds test shows that variables in consideration form a long-run relationship, aka, cointegrated. Panel A shows that the positive movement in EPU leads to a fall in tourism demand in t−2 period, while in the same period, the negative movement causes a positive effect. Importantly, the positive effect is dominating over negative effect suggesting that an increase in uncertainty hurts more than what a reduction in uncertainty pushes the demand. It is noteworthy that t−2 period results signs are reversed to that of t−1 indicating that economic uncertainty decreases the demand but with lags but not instantly. Panel B is also validating these results. Finally, panel C confirms that the results hold in the short run as well. It noteworthy that EC is negative and statistically significant suggesting that the system of cointegration is working well and any deviation in short run is quickly corrected, with a rate of 47% every month. Results of coefficients of REER and WGDP are found to be positive and statistically significant. Thus, suggesting that increasing income of world and real depreciation in the Indian currency have an increasing effect on India’s inbound tourism. This is very much on the theoretical lines.
Results of effects of tourism demand function: Application of NARDL.
Note: NARDL: nonlinear autoregressive distributed lag; REER: real effective exchange rate; EC: error correction; WGDP: world gross domestic product; EPU: economic policy uncertainty; TA: tourist arrival; ECM: error-correction model. Standard errors in parentheses. Method: NARDL (2, 1, 2, 0, 0); POS and NEG indicate positive and negative impact, respectively. Δ indicates first difference. Period of analysis is 2006:01 to 2018:04.
*p < 0.10; **p < 0.05.
ARDL type set up has several advantages; however, the models have a difficult lag structure, with contemporaneous values, lags, first differences, and lagged first differences of the independent variables entering in the model specification. This makes the interpretation of the estimated coefficients difficult in different time horizons. Very recently, Jordan and Philips (2017) have proposed dynamic simulations of the ARDL model, which are helpful in understanding the forecasting ability of the model in different time horizon, that is, long and short run. We utilize it and produce dynamic simulations of our model (equation (2)). The simulation result is shown in Figure 1, which suggests that our model has both short- and long-run forecasting ability as the forecasted demand is statistically significant across the time horizon. This, in turn, shows that the model works well in both time horizons.
Dynamic simulations of autoregressive distributed lag models.
Conclusion
This study has examined asymmetric effects of economic uncertainty on tourism demand in India. We have found that economic uncertainty has asymmetric effects on tourism demand. Importantly, it has been estimated that the negative effect of the increasing uncertainty has a more penetrating negative effect than the positive effect of the decreasing uncertainty.
The effect of economic policy uncertainty on investment, corporate profit, and economic growth is widely investigated and well established in the standard literature. This study is important in showing that the tourism industry, which is increasingly becoming vital for employment generation and foreign exchange earning in developing economies like India, is also greatly affected by economic policy uncertainty. Our results reveal that the uncertainty of the policy hurts foreign tourism in India. The implications of our results are that the policymakers should provide a credible policy plan and robust policy frameworks that can favorably influence tourism. When governments adjust economic policies to stabilize the business cycle or mop up higher revenue, they should consider the negative effects of the economic policy uncertainty caused by frequent changes of economic policies and maintain transparency and stability of the policies. Future research may focus on the asymmetric effects of the uncertainty on tourism for countries across the world.
Footnotes
Acknowledgements
The author thanks an anonymous referee for his/her useful comments and helpful suggestions on the previous version of this article. Any errors or omissions are solely of the author’s.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
