Regime-switching effect of tourism specialization on economic growth in Asia Pacific countries

Chiang, Geng-Nan; Sung, Wei-Ying; Lei, Wen-Guu Article Regime-switching effect of tourism specialization on economic growth in Asia Pacific countries Economies Provided in Cooperation with: MDPI – Multidisciplinary Digital Publishing Institute, Basel Suggested Citation: Chiang, Geng-Nan; Sung, Wei-Ying; Lei, Wen-Guu (2017) : Regime-switching effect of tourism specialization on economic growth in Asia Pacific countries, Economies, ISSN 2227-7099, MDPI, Basel, Vol. 5, Iss. 3, pp. 1-14, https://doi.org/10.3390/economies5030023 This Version is available at: https://hdl.handle.net/10419/197026 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ economies Article Regime-Switching Effect of Tourism Specialization on Economic Growth in Asia Pacific Countries Geng-Nan Chiang 1, Wei-Ying Sung 2,* and Wen-Guu Lei 3 1Department of Finance, Feng Chia University, Taichung City 407, Taiwan; [email protected] 2Institute of Public Affairs Management, National Sun Yat-sen University, Kaohsiung City 804, Taiwan 3Department of Tourism and Travel Management, Da-Yen University, Changhua County 515, Taiwan; [email protected] *Correspondence: [email protected]; Tel.: +886-912-153-997 Academic Editor: Juan Ignacio Pulido Fernández Received: 30 March 2017; Accepted: 20 June 2017; Published: 27 June 2017 Abstract: In the past 30 years, many studies have focused on exploring the relationship between tourism development and economic growth. However, there has been no consensus reached concerning of the relationship. This study will attempt to clarify the relationship between tourism development and economic growth. The purpose of this study is to analyze the relationship between tourism development and economic growth. This study applies the Panel Smooth Transition Regression Model (PSTR) proposed by Gonzalez et al. (2005) to investigate the regime-switching effect of tourism specialization on economic growth in Asia Pacific countries over the period 1996–2009. The results are as follows: (a) there were regime-switching effects of tourism specialization on economic growth; (b) the tourism specialization on economic growth has a better explanation for the effects of non-linear PSTR than linear PLS (Panel Least Squares); (c) in medium degree of tourism specialization countries (the value is between 0.0123~0.01663), tourism development has a significantly positive influence on economic growth, but consumption ability and investment ratios have a significantly negative influence on economic growth; (d) in low or high degree of tourism specialization countries (the value is below 0.0123 or above 0.01663), tourism development has a reduced influence on economic growth, and significantly positive influence on consumption ability and investment ratios. On the basis of these results, this study presents policy recommendations and areas for future research. Keywords: tourism specialization; real international tourism receipts growth (% of GDP); economic growth; regime-switching effect; Asia Pacific countries; Panel Smooth Transition Regression Model (PSTR) JEL Classification: R11 1. Introduction 1.1. Background and Motivation Tourism has been one of the key factors influencing economic growth in most countries ( Dwyer et al. 2004 ). According to surveys done by the World Tourism Organization, the number of global tourists reached 1.186 million and international tourism receipts amounted to 1260 billion US dollars in 2015 (World Tourism Organization 2016). Among all the regions in the world, the Asia Pacific countries have the fastest development of tourism. With the number of global tourists up to 279 million and international tourism receipts reaching 418 billion US dollars in 2015 Economies 2017,5, 23; doi:10.3390/economies5030023 www.mdpi.com/journal/economies Economies 2017,5, 23 2 of 14 ( World Tourism Organization 2016 ), Asia Pacific countries have become the key indicators for global tourism development and economic growth. Tourism development can facilitate the development of related industries and further boost the overall economic growth of a country (Lee and Chang 2008). Therefore, many countries take tourism development into account when making important policies on economic growth (Chou 2013; Chen and Chiou-Wei 2009 ;Oh 2005). In academia, more and more researchers have recently begun to examine the relationship between tourism development and the overall economic growth of a country (Chou 2013;Chen and Chiou-Wei 2009;De Vita and Kyaw 2016a;Dritsakis 2012;Lee and Chang 2008; Wang 2012a). Previous studies indicate that the tourism development of a country has a direct influence on its economic growth (Dritsakis 2004; Lee and Chang 2008 ;Oh 2005;Yen 2012). An investigation of the relationship between tourism development and economic growth found that general economic factors, such as price level and investment ability, have an indirect effect on the economic growth of a country (Po and Huang 2008). Regarding tourism development and economic growth, Lee and Chang (2008) suggest that the international tourism receipts of a country have a significant influence on its GDP growth. Chao et al. (2006) suggest that, while the promotion of tourism development might bring substantial international tourism receipts to a country, certain crowding-out effects may take place in its economy. Several studies show that the relationship between tourism development and economic growth have different effects on countries with a range of economic developments (Bilen et al. 2017; De Vita and Kyaw 2016a ;Eugenio-Martin et al. 2004;Holzner 2011;Lee and Chang 2008; Wang 2012b ), i.e., tourism development does not necessarily have a positive influence on the economic growth of a country; in fact, it may even have a negative impact, to certain degree. The situation may depend on the economic development and tourism specialization of a particular country (Chou 2013; De Vita and Kyaw 2016b;Sequeire and Campo 2005;Brau et al. 2007;Po and Huang 2008). Lee and Chien (2008) indicate that there has been no consensus regarding whether tourism development has a positive or negative effect on the economic growth of a country. Through a literature review, this study found that most studies have used a linear model to explore the relationship between tourism development and economic growth; few studies have adopted a non-linear model for their examination. However, the linear model fails to eliminate such problems as short-term economic fluctuations and structural changes, when used to investigate the relationships among variables ( Po and Huang 2008 ). The non-linear model should be applied not only to eliminate the problems, but also to better understand the effects between two variables. Therefore, this study used the Panel Smooth Transition Regression Model (PSTR) to investigate the regime-switching effect of tourism specialization on economic growth in Asia Pacific countries, where tourism has developed the fastest rates in recent years. In sum, through both a linear and a non linear model analysis, this study explored the relationship between tourism development and economic growth among different stages of tourism specialization. Moreover, this study compared the prediction effects of the traditional linear model and non-linear model to present policy recommendations and potential areas for future research. 1.2. Purpose Based on the aforementioned research background and motivations, the research purposes are as follows. 1. To understand the regime-switching effect of tourism specialization in Asia Pacific countries on their economic growth. 2. To compare the prediction effects of the linear model and the non-linear model in regard to the influence of the tourism development in Asia Pacific countries on their economic growth. Economies 2017,5, 23 3 of 14 2. Literature Review In the past 30 years, many studies have focused on exploring the relationship between tourism development and economic growth (Chou 2013) and various hypotheses have been proposed. Some have suggested that tourism development would influence economic growth (Wang 2012b), some have indicated that economic growth would affect tourism development (Narayan 2004), and still others have suggested that tourism development and economic growth affect each other (Bilen et al. 2017; Chen and Chiou-Wei 2009;Lee and Chang 2008). However, there is no real consensus regarding the relationship between tourism development and economic growth (Oh 2005). Reviewing the previous studies, the authors of this study found that the relationship between tourism development and economic growth can be divided into three categories, as follows. The first category of studies is devoted to the exploration of causality in the relationship between tourism development and economic growth of a single country. Balaguer and Cantavella-Jorda (2002) used Granger’s causality test to investigate the relationship between tourism development and economic growth in Spain, and discovered that tourism development has a positive influence on economic growth. Dritsakis (2004) explored the relationship between tourism development and economic growth in Greece and found that they have mutual influence, with the actual exchange rate and the international tourism receipts having the most significant effects on economic growth. Durbarry (2004) analyzed the influence of tourism development in Mauritius on the country’s economic growth, through the gross domestic fixed capital formation, trade exports and real international tourism receipts. The results showed that all three factors had a significantly positive influence on the country’s economic growth. Narayan (2004) analyzed the influence of tourism development in Fiji on its economic growth and found that tourism development leads to the appreciation of the local currency’s exchange rate and rising prices of goods. Kim et al. (2006) analyzed the relationship of tourism development and economic growth in Taiwan; the results showed that economic scale and trade openness affect economic growth. Most studies in the first category adopted Granger’s causality test to examine the relationship between tourism development and economic growth; however, this method may lead to biased estimates because of insufficient sample data, short-term economic fluctuations, and the inability to show features of different countries. To solve this problem, some researchers have started to apply panel data in examining the relationship between tourism development and economic growth among various countries, which forms the second category of studies. For example, Bilen et al. ( 2017 ) analyzed the relationship between tourism development and economic growth in Mediterranean countries. The results showed a bidirectional causality between tourism and economic growth. Eugenio-Martin et al. (2004) analyzed the relationship between tourism development and economic growth in Latin American countries and found that tourism development has a positive influence on economic growth in low and middle income countries, but no effect on economic growth in high income countries. However, when Lanza et al. (2003) analyzed the relationship of tourism development and economic growth in 13 OECD countries, they discovered that tourism development positively influenced economic growth in developed countries. Lee and Chang (2008) further compared the influence of tourism development on OECD countries and non-OECD countries. Their results showed that international tourism receipts have a greater influence on the GDP of non-OECD countries than that of OECD countries, and that the currency exchange rate significantly affects the economic growth in both OECD and non-OECD countries. Yen (2010) investigated the top nine most-visited countries and found that tourism development did not influence economic growth. Wang (2012b) used the threshold effect to examine the relationship between tourism development and economic growth in 10 countries (growth rate of international tourism receipts as threshold variable). The results showed that currency exchange rates have a positive influence on the economic growth in high-threshold countries, and that inflation suppresses economic growth in both high-threshold and low-threshold countries. Economies 2017,5, 23 4 of 14 From the second category of studies, it is clear that although researchers tried to solve the problem of insufficient sample data by adding panel data, short-term economic fluctuations and structural changes could not be eliminated (Po and Huang 2008). As a result, some researchers started trying the non-linear model to solve the aforementioned problem, which forms the third category of studies. For example, Po and Huang (2008) adopted the threshold vector autoregressive model in which tourism specialization was used as the threshold variable to analyze the relationship between tourism development and economic growth in 88 countries. The results suggested that those 88 countries could be divided into three regimes. In Regime 1 and Regime 3 (lower than the low threshold and higher than the high threshold), tourism development had a positive influence on economic growth. In Regime 2 (within the low threshold and high threshold), although tourism development did not have a significant influence, further analysis showed that tourism development still had a positive influence on economic growth. Chang et al. (2010) adopted the Panel Threshold Regression Model in which tourism specialization was used as the threshold variable to examine the relationship between tourism specialization and economic growth in 131 countries. The results showed that those 131 countries could be divided into three regimes. Among them, tourism development would have a significantly positive influence on economic growth in low-regime and middle-regime countries, while it would not significantly impact the economic growth in high-regime countries. Yen (2012) also adopted the Panel Threshold Regression Model to analyze the relationship between tourism development and economic growth in 84 countries. The results showed that those 84 countries could be grouped into two categories: high-threshold countries and low-threshold countries. Tourism development had a positive influence on economic growth in both high-threshold countries and low-threshold countries, while trade openness had a negative influence on their economic growth. De Vita and Kyaw (2016a) used the system generalized methods-of-moments (SYS-GMM) estimation methodology to investigate the tourism-growth relationship for a large panel of 129 countries. The results showed that they could be divided into three categories: low-income countries, middle-income countries and high-income countries. Among them, tourism development had a significantly positive influence on economic growth in low-income countries, middle-income countries and high-income countries. From the third category of studies, it can be concluded that the non-linear model can solve the problem of biased estimates by eliminating insufficient sample data, short-term economic fluctuations and structural changes. The authors of this paper reviewed the previous studies and found that few studies applied the non-linear model to explore the relationship between tourism development and economic growth. This may be one of reasons why the previous studies failed to clearly define the relationship between tourism development and economic growth. As the result, this study aims to adopt tourism specialization as the threshold variable, and use cross-sectional data and the Panel Smooth Transition Regression Model (PSTR) to examine the relationship between tourism development and economic growth. Moreover, this study will compare the differences of prediction effects of both the traditional linear model, and the non-linear model. In the past, there has been no consensus reached concerning the relationship between tourism development and economic growth, which may be related to the nonlinear relationship between them. Therefore, this study has adopted the non-liner model to examine the relationship between tourism development and economic growth. This study further clarifies the influence of different stages of tourism development and economic growth. On the basis of the results found in this study, the authors provide policy recommendations and areas for future research. 3. Methodology 3.1. Data Sources The research data in this study came from the World Development Indicators database (WDI) from the World Bank (2014 ); the International Financial Statistics (IFS) from the International Monetary Fund (2014) ; Economies 2017,5, 23 5 of 14 the Economic Data (ED) from the World Travel & Tourism Council (WTTC) and the United Nations (2014) . To maintain the consistency and completeness of the data in this study, the researcher eliminated countries and time periods with missing values, and selected 33 Asia Pacific countries over the period 1996–2009 as data for the research sample. 3.2. Definition and Measurement of Variables 3.2.1. Threshold Variables Tourism specialization is used as the threshold variable in this study; it is defined as the percentage of international tourism receipts in GDP. From the literature review, it was found that tourism specialization has often been used as the threshold variable. For example, Sequeire and Campo (2005) , Brau et al. (2007), Po and Huang (2008), Chang et al. (2010), Yen (2012) and Kung (2013) all used tourism specialization as the threshold variable in their studies, despite having different definitions. Sequeire and Campo (2005) used “the percentage of tourism receipts in GDP”, “the percentage of tourism receipts in the export of goods and labor” and “the percentage of global tourists in the total population” to measure the tourism specialization of a country. Chang et al. (2010) adopted “the percentage of real tourism GDP in real GDP” to measure the tourism specialization of a country. Brau et al. (2007) , Po and Huang (2008) and Yen (2012) defined tourism specialization as “the percentage of international tourism receipt in GDP”. Considering that most previous studies defined tourism specialization as the percentage of international tourism receipts in GDP, and such a definition is more suitable for this study’s examination of economic growth of countries in a specific region, this study adopted the definition used in Brau et al. (2007), Po and Huang (2008) and Yen (2012) . 3.2.2. Explanatory Variables This study used the growth rate of international tourism receipts (TRG), the percentage of the gross fixed capital formation in GDP (I) and inflation rate ( π ) as explanatory variables (see Table 1). It was found through the literature review that the aforementioned variables have often been used as explanatory variables of economic growth (Kung 2013;Yen 2012). For example, Yen (2012) adopted “the growth rate of international tourism receipts”, “the percentage of the gross fixed capital formation in GDP” and the “inflation rate” as explanatory variables to predict economic growth, when exploring the relationship between tourism development and economic growth of various countries. Table 1. Definitions of explanatory variables. Variables Definitions Remarks EG economic growth “The growth rate of the GDP per capita” from the World Development Indicators (WDI) of the World Bank was used to measure the economic growth of a country in this study. TRG growth rate of international tourism receipts “The real international tourism receipts” was used to measure the tourism development of a country in this study. πinflation rate “Inflation, GDP deflator” from the WDI of the World Bank was used to measure the price level of a country in this study. Ithe percentage of the gross fixed capital formation in GDP “The percentage of the gross fixed capital formation in GDP” from the WDI of the World Bank was used as the proxy variable of the real capital investment in this study. qdegree of tourism specialization “The percentage of international tourism receipts in GDP” was used to measure the tourism specialization of a country in this study. Data source: The World Development Indicators database (WDI) of the World Bank (2014). Economies 2017,5, 23 6 of 14 3.3. Model Estimation and Tests 3.3.1. Panel Data Least Squares Regression (PLS) The study applies the Panel Data Least Squares Regression (PLS) by Hsiao (1986) proposed to investigate the effect of tourism specialization on economic growth in a general linear model. In order to reduce the heterogeneity bias of a single sample, and control the autocorrelation between time serious and individual differences in panel data, the study used the fixed effect of PLS to analyze panel data as follows: yit =αi+ k ∑ k=1 βkXkit +εit (1) where irepresents different countries, i= 1 . . . N;tis the observation time state, t= 1 . . . T;kis the number of countries, k= 1 . . . k;yis the vector; B is the regression coefficients of explanatory variables; χ is the vector of explanatory variables; εit error item; αi : intercept, individual effect, not change with time, which had different effect in other units. 3.3.2. Panel Smooth Transition Regression (PSTR) Model This research intends to understand the regime-switching effect of tourism specialization on economic growth if had a smooth transition threshold effect. This study constructs the PSTR model by Gonzalez et al. (2005), and can be defined as: yit =µi+β0 0Xit +β0 1Xitg(qit;γ,c)+εit (2) where i= 1,.......,N,t= 1,........,T, and Nand Tstand for the cross-section and time dimensions of the panel, respectively. The dependent variable y it is a scalar; irepresents the fixed individual effect; X it is a k-dimensional vector of time-varying exogenous variables; µit is the residual term. The transition function g(q it ; γ ,c) is a continuous function of the observable variable q it . It is normalized to be bounded between 0 and 1; these extreme values are associated with regression coefficients β0 0 and β0 0 + β0 1 . The value of qit determines the value of g (qit;γ,c) and thus the effective regression coefficients β0 0+β0 1.g(qit;γ,c)for any individual iat time t. Following Granger and Teräsvirta (1999), Teräsvirta (1994), and Jansen and Teräsvirta (1996), we have formulated the transition function as follows: g(qit;γ,c)=(1+exp"−γ m ∏ j=1qit −cj#)−1 with γ>0and c1≤c2≤. . . ≤cm(3) where c (c 1 ,...,c m ) is an m-dimensional vector of location parameters and the slope parameter determines the smoothness of the transitions. In general, it is sufficient to consider m= 1 or m= 2, as these values allow for commonly encountered types of variations in the parameters. In the case of m= 1, the model specifies that the two extreme regimes are associated with low and high values of qit with a single monotonic transition of the coefficients from β0 0 to β0 0 + β0 1 as qit increases, such that the change is centered around c 1 . In the case of m= 2, the transition function has its minimum at (c 1 +c 2 )/2 and reaches the value 1 at both low and high values of qit . When this approaches infinity, the PSTR model reduces to a three-regime panel threshold regression (PTR) model with identical outer regimes and a different middle regime (Gonzalez et al. 2005). The multi-level PSTR model is a generalization of the PSTR model that allows for more than two different regimes; it can be formulated as: yit =µi+β0 0Xit + r ∑ j=1 β0 0Xitgjqj it;γj;cj+εit (4) Economies 2017,5, 23 7 of 14 where the transition functions gj(qit;γ,c) , j= 1, ..., γ depend on the slope parameters γj and on location parameters C j . If r= 1, qj it =qit , and γj→∞ for all j= 1,..., γ then the transition function becomes an indicator function, with I[A] = 1 when event A occurs, and I[A] = 0 otherwise; in such a case, the model in Equation (4) becomes a PTR model with r + 1 regimes. As a result, the multi-level PSTR model can be viewed as a generalization of the multiple regime panel threshold model (PTR) in Hansen (1999). 3.3.3. Building the Panel Smooth Transition Regression Model The PSTR model building procedure consists of specification, estimation and evaluation stages. Specification includes tests for homogeneity, and selection of the transition variable qit . If the tests fail to show homogeneity, then specification includes the determination of the appropriate form of the transition function; the form is dictated by the value of m in Equation (3). A nonlinear least square method is used for parameter estimation. At the evaluation stage the estimated model is subjected to misspecification tests to check whether it provides an adequate description of the data. The null hypotheses to be tested at this stage includes parameter constancy, absence of remaining heterogeneity and absence of autocorrelation in the errors. Finally, the number of regimes in the panel must be specified, which means that a value must be assigned to r in Equation (4). 4. Data Analysis 4.1. Basic Descriptive Statistical Analysis of the Variables It is clear from Table 2, that the average EG of the 33 Asia Pacific countries is 3.87%, the average TGR is 18.74%, the average π is 9.6%, the average I is 23.53% and the average q is 3.68%. Except for the obviously higher number of the average TGR, the results in this study are similar to those of Yen (2012)’s analysis of the relationship between tourism development and economic growth of all countries in the world. It is suggested that such a difference may arise from the different sampling sizes. Yen (2012) adopted 84 countries in the world as the research sample data while this study only targeted Asia Pacific countries. Since the Asia Pacific countries have had such rapid tourism development over recent years, the international tourism receipts in this region are higher than those in most places of the world (Europe was ranked first, Asia Pacific was ranked second) (World Tourism Organization 2016). Table 2. Basic descriptive statistical analysis. Variables Mean Std. Dev Variance Max Min Kurtosis Skewness EG 3.8697 4.6711 1.2071 33.0305 −14.3851 5.5259 0.3277 TRG 18.7434 92.5802 4.9394 1820 −50.8772 309.55 16.293 π9.6028 15.3534 1.5988 137.9649 −21.4438 22.328 3.8647 I 23.5348 11.4895 0.4882 63.0487 −89.8562 30.949 −3.3547 q 3.6767 4.67 1.2702 35.167 0.0879 11.498 2.9909 4.2. Correlation Analysis of the Variables From Table 3, it is obvious that the correlation coefficient of all explanatory variables falls between − 0.2800 and 0.1172; the correlation coefficient of EG and TRG is 0.0449; the correlation coefficient of EG and π is − 0.0136; the correlation coefficient of EG and I is 0.1172. The results indicate that there is a significantly positive correlation between EG and I (r= 0.1172, p< 0.05) at the 10% confidence level, while there is no significant correlation between EG and other explanatory variables. It can be concluded that despite the low correlation between economic growth and the explanatory variables in the study, the correlation is similar to that of previous studies (Yen 2012;Kung 2013). Economies 2017,5, 23 8 of 14 Table 3. Variables correlation analysis. EG TRG ΠI EG 1 TRG 0.0449 1 Π−0.0136 −0.0150 1 I 0.1172 * −0.0187 −0.2800 ** 1 Note: *, ** and *** denote significance at the 10%, 5% and 1% level, respectively. 4.3. Analysis of Panel Unit Root Test To enhance the accuracy of the results, data used in an econometric model should be confirmed to be stationary before the model is established and estimated through a time series ( Nelson and Plosser 1982 ). The structure of the sample data used in this study is balanced panel data. Therefore, a panel unit root test of the research data should be done before data analysis. This study adopted the most used Fisher-type augmented Dickey–Fuller test (ADF) to test the stationary state of the variables in this study (Maddala and Wu 1999). Moreover, because the ADF unit root test does not take into consideration the autocorrelation and ARCH/GARCH of residuals, this study also adopted the PP–Fisher test, proposed by Phillips and Perron (1988), to enhance the test results. From Table 4, the results from the three tests all reject the null hypothesis of the unit root, i.e., the sample data in this study are stationary. Thus, an analysis of the linear regression model and the panel smooth transition regression model could be conducted. Table 4. Panel unit root test. Augmented Dickey–Fuller (ADF)–Fisher Test (p-Value) PP–Fisher Test (p-Value) EG 135.6643 *** 120.6493 *** (0.0000) (0.0000) TRG 136.2843 *** 221.1713 *** (0.0000) (0.0000) π115.9583 *** 185.2743 *** (0.0000) (0.0000) I92.18593 *** 75.18653 ** (0.0005) (0.00194) Note: *, ** and *** denote significance at the 10%, 5% and 1% level, respectively. 4.4. Analysis of the Ordinary Least-Squares (OLS) Regression for Panel Data From the test results in Table 5, it is clear that only the influence of the gross fixed capital formation in GDP ( β = 0.0505, p< 0.05) on economic growth reached a significant level. TGR, the most important representative of tourism development, did not achieve a significant influence on economic growth ( β = 0.0024, p> 0.05), thereby showing that tourism development does not necessarily have a direct impact on economic growth. This result differs from the conclusion in most previous studies that tourism development has a direct influence on economic growth (Balaguer and Cantavella-Jorda 2002; Dritsakis 2004;Kung 2013;Lee and Chang 2008;Oh 2005), but is similar to Yen (2010)’s conclusion that tourism development does not have a significant influence on economic growth. This result also presents the reason why the previous studies failed to clearly define the relationship between tourism development and economic growth. It suggests that the linear model may not fully explain the relationship between tourism development and economic growth. It may take a non-linear model to further understand the relationship between tourism development and economic growth.