Essays on tourism and its determinants

Essays on Tourism and its Determinants Thaana Ghalia Department of Economics and Finance College of Business, Arts and Social Sciences Brunel University London, UK A thesis submitted for the degree of Doctor of Philosophy February 2016 Abstract This thesis is based on four essays dealing with tourism development and its determinants. Chapter Two explores the different definitions of ‘tourism’ and ‘tourist’, as well as the factors that influence tourism arrivals. We discuss traditional and more recent theories that underlie the study of the tourism industry. The third chapter examines the effect of tourism upon economic growth, investigating the effects of tourism specialization within tourism-exporting countries and nontourism-exporting countries annually over the period 1995–2007, applying paneldata methods in cross-sectional growth regressions. This study finds that tourism does not affect economic growth in either underdeveloped or developed countries. Moreover, tourism might cause Dutch Disease in tourism-exporting countries owing to their over-reliance on the exporting of non-traded goods. Chapter Four seeks to identify how institutional quality and aspects of infrastructure (internet access measured by size of country or per 100 people) influence tourist arrivals in a whole sample of 131 countries and in sub-samples comprising developed and developing countries (as defined by IMF criteria) using static and dynamic panel data. The findings indicate that internet access enhances the tourism industry, and most interestingly, that good governance is one of the most influential factors for improving and developing tourism. Chapter Five diagnoses the determinants of tourism flows using panel-data sets including 134 originating countries and 31 destination countries (selected depending on data availability) focusing on ICRG data for the period 2005–2009. The methodology makes use of basic and augmented gravity equations, together with the Hausman-Taylor and Poisson estimation techniques, whilst comparing the performance of the three gravity-equation methods. The results suggest that lower levels of political risk contribute to an increase in tourism flows. Furthermore, common language (positively), common currency (negatively) and political factors ii (particularly institutional quality) are the most prominent determinants in promoting (or deterring) tourism. Chapter Six gives concluding remarks. Acknowledgements I am grateful to many individuals for their care and support given during my the tumultuous task of completing doctoral studies. First and foremost, I would like to express my profound gratitude to Dr Jan Fidrmuc, my supervisor, for his enthusiastic encouragement, insightful advice, and invaluable suggestions. This work would not have been possible without him and his tremendous willingness to indulge thought provoking discussions. He has indeed been an inspiration and a driving force during this time. I am also thankful to Prof João Santos Silva, University of Surrey, and Dr. Maria Santana Gallego, University of Balearic Islands, for their kind help and constructive comments that helped me to conduct the fifth essay in my thesis. My deepest appreciation goes to the World Tourism Organization (WTO) for providing the tourists data for the study of gravity model in this thesis. The Brunel University Student Centre has also played a tremendous role during these past few years. As an overseas student at Brunel, I am deeply indebted to the centre's staff espacilly Mr Jose Sanchez for all the help and care they have shown me and many other students. I also would like to thank all the members of staff and my peers, Mohammad Tajik, Saeideh Aliakbari,Mohamad Helmi, Nahla Samargand,Francis Atsu in the Department of Economics and Finance at Brunel for their support and help during this journey. My grateful appreciation extends to Damascus University for the financial support they provided me during my doctoral studies. It is without a doubt that my family has always been my inspiration and my rock when I mostly needed support. My heartfelt gratitude goes to the most precious people in my life, my parents for their eternal love and moral support whenever I need it during my doctoral studies. The spirit of my brother Mohamed, who iii sacrificed his life defending our country, has always shadowed and looked over me as I completed my journey. To him I say, thank you, I love you, and I hope you are proud of me today. When times were tough, and the PhD cold, my sisters, brothers, sisters and brothers in law were my destination for a quick call at the crazy hours of the night, or to talk to as I had a meal. Thank you for always showing me love, support and encouragement. I also want to thank my dear friends Maher Alliwa, Kinana Jammoul, Osama Alraheb, Rami Younes ,Hayan Omran and Naima AlNaami for their help and support all the time. Last but not least, I would like to express my deepest gratitude to my relative Hind Ali who was there for me all the time. Indeed, no words can express my gratefulness and gratitude to her. Thank you all from the bottom of my heart Thaana Ghalia iv Dedicated to My country v Declaration I grant powers of discretion to the Librarian of Brunel University to allow this thesis to be copied in whole or in part without the necessity of contacting me for permission. This permission covers only single copies made for study purposes subject to the normal conditions of acknowledgment. vi Supporting Publications Journals 1. T. Ghalia and J. Fidrmuc, "The Curse of Tourism?," Journal of Hospitality & Tourism Research, December 2015. DOI: 10.1177/1096348015619414 2. M. Tajik, S. Aliakbari, T. Ghalia and S. Kaffash, "House prices and credit risk: Evidence from the United States," Economic Modelling, vol. 51, pp. 123-135, December 2015, DOI: 10.1016/j.econmod.2015.08.006 Conference Papers I have presented material from chapter 3 called “curse of tourism” at ESDS International Annual Conference,28 November, 2011 Also, I have presented the same paper above At 5th international conference on Advances in Tourism Economics, 6 ‐ 7 June 2013, Coimbra, Portugal. vii Contents A THESIS SUBMITTED FOR THE DEGREE OF ................................................................ I DOCTOR OF PHILOSOPHY .................................................................................................. I FEBRUARY 2016 ................................................................................................................... I ABSTRACT ........................................................................................................................................................ II DEDICATED TO MY COUNTRY ....................................................................................................................... V SUPPORTING PUBLICATIONS ...................................................................................................................... VII JOURNALS....................................................................................................................................................... VII CONFERENCE PAPERS.................................................................................................................................. VII CONTENTS...................................................................................................................... VIII LIST OF FIGURES .............................................................................................................XI LIST OF ACRONYMS ....................................................................................................XIV 1 INTRODUCTION ......................................................................................................... 15 2 TOURISM: CORE CONCEPTS ................................................................................. 22 2.1 INTRODUCTION.............................................................................................................................. 22 2.2 BASIC TOURISM CONCEPTS ......................................................................................................... 24 2.3 MARKET FOR TOURISM PRODUCTS ........................................................................................... 27 2.3.1 The Demand for Tourism Products ................................................................................. 27 2.3.2 Tourism Supply ........................................................................................................................ 28 2.4 FACTORS AFFECTING TOURISM .................................................................................................. 28 3 2.5 BENEFITS AND COSTS OF TOURISM ........................................................................................... 32 2.6 TOURISM AND ECONOMIC GROWTH .......................................................................................... 36 THE CURSE OF TOURISM ....................................................................................... 41 3.1 INTRODUCTION.............................................................................................................................. 41 3.2 LITERATURE REVIEW ................................................................................................................... 42 3.2.1 Tourism and Economic Growth ........................................................................................ 42 3.2.2 Dutch Disease........................................................................................................................... 43 3.2.3 Dutch Disease and Tourism ............................................................................................... 45 3.3 THE CURSE OF TOURISM? ........................................................................................................... 48 3.3.1 Data and Variables ................................................................................................................ 48 3.3.2 An Empirical Model of Economic Growth with Tourism ....................................... 52 3.3.3 Empirical Results and Discussion.................................................................................... 55 3.4 CONCLUDING REMARKS ............................................................................................................... 67 viii 4 TOURISM AND ITS DETERMINATES ................................................................... 68 4.1 INTRODUCTION.............................................................................................................................. 68 4.2 LITERATURE REVIEW ................................................................................................................... 70 4.3 METHODOLOGY ............................................................................................................................. 77 4.3.1 Data Set ....................................................................................................................................... 77 4.3.2 Model specification and econometric method............................................................. 84 4.4 EMPIRICAL ANALYSIS AND RESULTS ......................................................................................... 90 4.5 5 CONCLUSION ................................................................................................................................ 101 “INTERNATIONAL TOURISM AND INSTITUTIONAL QUALITY: EVIDENCE FROM GRAVITY MODEL” ..................................................................... 103 5.1 INTRODUCTION............................................................................................................................ 103 5.2 LITERATURE REVIEW ................................................................................................................. 106 5.3 DATA ............................................................................................................................................. 111 5.4 METHODOLOGY ........................................................................................................................... 123 5.4.1 Tradional gravity model : ............................................................................................... 123 5.4.2 Hausman Taylor model: ................................................................................................... 125 5.4.3 Poisson model ....................................................................................................................... 126 5.5 EMPIRICAL RESULTS .................................................................................................................. 128 5.5.1 Gravity variables as determinates of tourism flows .............................................. 128 5.5.2 Results from the OLS estimator...................................................................................... 128 5.5.3 Estimation results of the gravity equation origin and destination effects using OLS regression .......................................................................................................................................... 131 5.5.4 Estimation results of the gravity equation with country-pair effects.............. 133 5.5.5 Results from the Hausman-Taylor Model .................................................................. 144 5.5.6 Results of count Model (Poisson Model).................................................................... 147 5.6 CONCLUSION ................................................................................................................................ 149 6 CONCLUDING REMARKS ...................................................................................... 152 7 APPENDIX A .............................................................................................................. 157 8 7.1 HAUSMAN TEST ........................................................................................................................... 157 7.2 HAUSMAN TEST WITH TOAP....................................................................................................... 157 7.3 HAUSMAN TEST IN DEVELOPED COUNTRIES ............................................................................ 157 7.4 HAUSMAN TEST IN DEVELOPED COUNTRIES ............................................................................ 158 7.5 HAUSMAN TEST FOR LARGE POPULATION-SIZE COUNTRIES ................................................... 158 7.6 HAUSMAN TEST FOR SMALL POPULATION-SIZE COUNTRIES................................................... 158 APPENDIX B .............................................................................................................. 159 ix 8.1 9 CONTIGUITY CONTROLLED FOR YEAR AND COUNTRY FIXED EFFECTS ................................... 159 APPENDIX C.............................................................................................................. 161 9.1 161 9.2 162 9.3 164 10 APPENDIX D ............................................................................................................. 165 10.1 SOME.................................................................................................................................................. 165 10.2 SOME STATISTICS OF VARIABLES BY COUNTRY CODE. .............................. 173 REFERENCES ................................................................................................................... 183 x List of Figures FIGURE 2.1 TOURISM EMPLOYMENT ................................................................................. 34 FIGURE 2.2 ECONOMIC GROWTH AND INTERNATIONAL TOURIST ARRIVALS 1975–200537 FIGURE 4.1 INTERNATIONAL TOURIST ARRIVALS BY REGION, JAN 1995–MAR 2010 (%) 77 FIGURE 4.2 EIGENVALUES ................................................................................................ 80 FIGURE 5.1 COMPONENT LOADING FACTORS ................................................................. 115 xi LIST OF TABLES FIGURE 2.1 TOURISM EMPLOYMENT ................................................................................................................... 34 FIGURE 2.2 ECONOMIC GROWTH AND INTERNATIONAL TOURIST ARRIVALS 1975–2005 ................ 37 TABLE 3.1 VARIABLES USED IN THE PRESENT STUDY ....................................................................................... 49 TABLE 3.2 DESCRIPTIVE STATISTICS ...................................................................................................................... 50 TABLE ‎3.3 CROSS-CORRELATION BETWEEN VARIABLES, 1995-2007 ....................................................................... 50 TABLE 3.5 HAUSMAN TEST WITH LIFE EXPECTANCY ........................................................................................ 57 TABLE 3.6 HAUSMAN TEST WITH SECONDARY SCHOOL................................................................................... 57 TABLE 3.7 EFFECT OF TOURISM ON GROWTH: FIXED-EFFECTS MODEL AFTER REMOVING TOURISM FROM TRADE .............................................................................................................................................................. 59 TABLE 3.8 GDP TABLE 3.9 EFFECT OF TOURISM ON GROWTH: SPLIT SAMPLES DEPENDING ON TOURISM SHARE IN 60 FIXED-EFFECTS ESTIMATOR IN SUB-SAMPLES OF DEVELOPING AND DEVELOPED COUNTRIES ................................................................................................................................................................. 61 TABLE 3.10 FIXED EFFECTS IN WHOLE SAMPLE WITH TOURISM-TRADE INTERACTION (TTRADEP).. 62 TABLE 3.11 EFFECT OF TOURISM ON GROWTH DEPENDING ON SHARE OF TOURISM IN EXPORTS ..... 65 FIGURE 4.1 INTERNATIONAL TOURIST ARRIVALS BY REGION, JAN 1995–MAR 2010 (%) ................. 77 TABLE 4.1 EXPLANATORY VARIABLES .................................................................................................................. 79 TABLE 4.2 COMPONENT EIGENVALUES ................................................................................................................. 80 FIGURE 4.2 EIGENVALUES ...................................................................................................................................... 80 TABLE 4.3 DESCRIPTIVE STATISTICS OF VARIABLES ........................................................................................ 82 TABLE 4.4 DESCRIPTIVE STATISTICS OF GOVERNANCE INDICATORS ........................................................... 82 TABLE 4.5 CROSS-CORRELATION BETWEEN VARIABLES, 1995-2007 .......................................................... 83 TABLE 4.6 ESTIMATION RESULTS WITH STATIC AND DYNAMIC PANEL, IN WHOLE SAMPLE .................. 91 TABLE 4.7 ESTIMATION RESULTS OF STATIC AND DYNAMIC PANEL DATA WITH NET/SIZE IN THE WHOLE SAMPLE......................................................................................................................................................... 93 TABLE 4.8 ESTIMATION WITH NET/POP IN WHOLE SAMPLE ............................................................................. 94 TABLE 4.9 STATIC AND DYNAMIC ESTIMATION RESULTS WITH NET/POP IN WHOLE SAMPLE WITH DEPENDENT VARIABLE ARRIVALS/POP .............................................................................................................. 95 TABLE 4.10 FIXED-EFFECTS REGRESSION: DEVELOPED AND DEVELOPING COUNTRIES ....................... 96 TABLE 4.11 FIXED-EFFECTS REGRESSION: COUNTRIES WITH LARGE AND SMALL POPULATIONS ..... 98 TABLE 4.12 FIXED EFFECTS OF INDIVIDUAL GOVERNANCE INDICATORS ................................................. 99 TABLE 5.1 FIGURE 5.1 PRINCIPAL COMPONENTS (EIGENVECTORS) ...................................................................................114 COMPONENT LOADING FACTORS ................................................................................................115 TABLE 5.2 SCORING COEFFICIENTS ......................................................................................................................116 TABLE 5.3 SUMMARY OF VARIABLES USED IN THE MODEL ...........................................................................117 TABLE 5.4 DESCRIPTIVE STATISTICS OF POLITICAL RISKS .............................................................................119 xii TABLE 5.5 DESCRIPTIVE STATISTICS OF POLITICAL RISKS OF DESTINATION ............................................119 TABLE 5.6 DESCRIPTIVE STATISTICS OF POLITICAL RISKS OF ORIGIN ........................................................120 TABLE 5.7 DESCRIPTIVE STATISTICS OF OTHER EXPLANATORY VARIABLES ...........................................120 TABLE 5.10 BASIC AND AUGMENTED GRAVITY MODELS ...........................................................................129 TABLE 5.11 ESTIMATION RESULTS OF THE GRAVITY EQUATION ORIGIN AND DESTINATIONS EFFECTS 131 TABLE 5.112 ESTIMATION RESULTS OF THE GRAVITY EQUATION WITH COUNTRY-PAIR EFFECTS....135 TABLE 5.13 AUGMENTED GRAVITY MODEL WITH ICRG VARIABLES ....................................................138 TABLE 5.14 EXTENDED AUGMENTED GRAVITY MODEL WITH ICRG VARIABLES..............................140 TABLE 5.14 EXTENDED AUGMENTED GRAVITY MODEL WITH ICRG VARIABLES—CONTINUED ..142 TABLE 5.15 HAUSMAN-TAYLOR MODEL WITH ANALYSIS OF THREE PRINCIPAL COMPONENTS ......145 TABLE 5.16 RESULTS OF COUNT MODEL (POISSON MODEL) ......................................................................148 xiii List of Acronyms CEPII Centre d'Études Prospectives et d'Informations Internationales, Paris EGARCH exponential generalized autoregressive conditional heteroskedastic EGARCH-M EGARCH-in-mean FDI Foreign direct investment GDP Gross Domestic Product GMM Generalized method of moments H-T Hausman-Taylor (estimator) HTM Hausman-Taylor Model ICRG International Country Risk Guide/s ILO International Labour Organization LDC Less-Developed Country OECD Organization of Economic Co-operation and Development OLS Ordinary Least Square/s PCA Principal Component Analysis PPML Poisson Pseudo-Maximum Likelihood (estimator) PRIO International Peace Research Institute, Oslo SARS Severe Acute Respiratory Syndrome TOAR Tourist Arrivals UCDP Uppsala Conflict Data Project UNDESA United Nations Department of Economic and Social Affairs UNWTO United Nations World Tourism Organization USAID United States Agency for International Development VECM Vector Error Correction Model WESP World Economic Situation and Prospects WTTC World Travel and Tourism Council xiv 2. Tourism: Core Concepts 1 Introduction At any moment in time, there are many thousands of tourists beginning or ending their journeys to and from various venues around this planet. Furthermore, numerous meetings, exhibitions and conventions are in progress or being planned, whilst countless numbers of people are making their travel and holiday plans. In other words, travel and tourism is major global industry—big business that will continue to grow (Goeldner and Brent Ritchie, 2012). The World Travel and Tourism Council (WTTC) reported that the shares of world GDP and employment contributed by tourism and travel were 2.8 and 3.3 percent respectively during 2011– 2012 (WTTC, 2012). Moreover, the average global annual intake of international tourist arrivals grew at the rate of 4.6 percent compared to GDP growth rate of 3.5 percent between 1975 and 2000 (UNWTO, 2012). Tourism is defined by the UN World Tourism Organization (UNWTO) as the circulation of people who travel to or stay in places outside their home country (their normal place of residence and/or work) for short periods, usually no longer than a year, and for different purposes such as leisure, business, or any purposes other than formal employment (UNDESA, 2010 pp. 9–10). The destinations receiving tourist visits and activities can benefit in many ways, although distortions in the local and/or national economy can arise too, especially when they result from unplanned, uncontrolled or un-moderated dependence on the income and opportunities provided by tourism (Pleumarom, 1994). Tourism has the potential to generate increases in sales, outputs, labour earnings and employment in the host country, state or region (Garín-Muñoz and Montero-Martín, 2007; Ardahaey, 2011). The opportunities offered by the prospects of developing tourism in a particular locality are very attractive to business-minded people of all socio-economic backgrounds and conditions. As a result, tourism can give a valuable boost to the local economy of a destination, tourism can also become a foreign-exchange earner on a national scale, and thus an important source of exports especially for small and developing countries (Holloway et al., 2009). 15 2. Tourism: Core Concepts Not only does tourism increase external income and even foreign-exchange income, but also various studies have identified that it can also rapidly generate employment opportunities (Mathieson and Wall, 1982; Figini and Vici, 2009; Zortuk, 2009; Polat et al., 2010; Vellas, 2011). It is not surprising, then, that many governments at local and national level aim to achieve development in the tourism sector because of the various benefits that tourism is perceived to offer. Nor is it surprising that such bodies tend to regard economic benefits as the most important measure of tourism, as these can help achieve a positive balance of payments and stimulate the sectors dependent on tourism, thus tending to benefit the local area and the wider country as well (Ivanov and Webster, 2006; Polat et al., 2010). Many researchers have studied the importance and economic effects of tourism and business travel using a variety of approaches, such as Fletcher (1989), Archer (1995), Archer and Fletcher (1996), Dwyer et al. (2000a), Kweka et al. (2001, 2003), Sahli and Nowak (2007), Blake et al. (2008). Zortuk (2009) found a direct relationship between Gross Domestic Product (GDP) and tourist arrivals. One of the main drivers behind the growth in international tourism is the desire to experience the culture of tourist destinations. The increasing significance of cultural interest in generating tourism has been identified as an important ingredient in regional competitiveness for a considerable period of time (OECD, 2008). The OECD (2008) study reported a 45 percent increase in cultural travel from 1995 to 2007, with cultural trips comprising 40 percent of overall international tourism in 2007. Cultural organizations established by some developed countries, such as the Alliance Française (France, established 1883), the British Council (UK, 1934), the Goethe Institut (Germany, 1951), the Instituto Cervantes (Spain, 1991) and the Confucius Institute (CI) established by the Office of the Chinese Language Council International in 2004, have been playing a significant role in promoting the culture and language of their respective countries. The tourism industry, as one of the fastest growing industries in the world, is one of the main drivers in the world economy owing to its role in creating employment opportunities, generating income and export revenues. 16 2. Tourism: Core Concepts In 2009, according to WDI (2010) and WTTC (2011), tourism accounted for 3.2 percent of global domestic income, 5.5 percent of total exports, 2.8 percent of global employment, and 24 percent of service-industry exports. Further evidence reflecting the substantial growth in tourism over the recent years can be seen in the average annual increase of 6.6 percent in tourism receipts between 1950 and 2009 (UNWTO, 2010). Moreover, the average annual increase of 6 percent in the number of tourist arrivals between 1950 and 2009 (UNWTO, 2010) is expected to further increase at a rate well in excess of that until 2020 (WTTC, 2011). However, the non-economic benefits—such as social, environmental and other benefits — might not be so well identified (Pizam, 1978). It is true that cultural exchange between the host population and tourists can often generate social benefits (Armenski et al., 2011). In addition, tourism is often considered to be a “clean” industry for the environment, although many debates surround this issue (Jenner and Smith, 1992; Croall, 1995; Kreag, 2011; Bastola 2012). However, there may often be adverse effects if the tourist trade is not well-managed (Hjalager, 1996; Howie, 2003; Fennell and Ebert, 2004). In spite of the often-mentioned benefits, tourism can also exert negative effects such as causing deterioration of the environment through the physical impact of tourist visits and the over-exploitation of natural resources (Kuss et al., 1990; Cater and Goodall, 1992; Holzner, 2005; Capó et al., 2007; Holzner, 2010). Tourism can cause unwanted lifestyle changes that might have negative impacts on the traditions and customs of the host community (Pizam, 1978; Nowak et al., 2004; Cooper et al., 2008). Furthermore, many studies have investigated whether tourism causes the disruptive economic unbalancing phenomenon known as the ‘Dutch Disease’ (e.g. Corden, 1984; Chao et al., 2006; Capó et al., 2007; Nowak and Sahli, 2007; Mieiro et al., 2012). Moreover, previous studies such as that by Eilat and Einav (2004) suggest that there are many internal and external factors that might have an impact on tourism demand. These factors can include for example ethnic tensions, issues surrounding currency exchange rates, and internal or external conflicts. Over the last two decades the tourism literature has developed massively in response to the rapid growth in tourism flows worldwide. Research into various 17 2. Tourism: Core Concepts aspects of tourism and tourism demand has assumed a new significance. Recent literature mainly focuses on studying the factors affecting demand for tourism in particular countries and forecasting tourism demand according to stronger theoretical background while employing statistical approaches. However, despite the extensive research being conducted in these fields, there are still several fundamental questions that we attempt to address in this thesis, such as:  What impact does tourism specialization have on economic growth via trade?  Do host-country features such as the communications infrastructure have any effect on the performance of the tourism sector?  Do governance indicators affect tourism flows in the same way for countries at different levels of economic development, or do their effects depend on their population size?  What role do destination-country institutions play in determining cross-border tourism flows? Accordingly, this thesis aims to re-examine the different aspects of the relationship between tourism and economic growth in order to answer some of the above-mentioned questions that have not been sufficiently addressed in tourism studies, as well as to explore different aspects of tourism determinants using advanced econometric techniques. This thesis is based on four essays in the field of tourism determinants. Specifically, in Chapter 2 we discuss the basic concepts of tourism, while in Chapter 3, we investigate the linkage between economic growth and tourism of 32 “tourism-dependent” countries within a sample of 131 countries annually during the period 1995–2007, applying panel-data methods in crosssectional growth regressions for the countries. We then examine whether tourism specialization is a good option for underdeveloped countries whose GDP per capita is less than the average. After that, we seek to ascertain whether Dutch Disease exists in countries whose exports are dominated by tourism. The empirical results show that tourism specialization has no significant effects on economic growth and is related negatively to growth in the broad sample and in two smaller samples 18 2. Tourism: Core Concepts (tourism-dependent and non-tourism-dependent countries). Also, tourism specialization does not affect economic growth in two other samples: underdeveloped and developed countries. Finally, we show that tourism specialization might cause Dutch Disease in tourism-exporting countries owing to over-dependence on the exporting of non-traded goods. The empirical results suggest that tourism does not always lead to economic growth and it might even be considered detrimental. Chapter 4 revisits how governance and infrastructure quality affect tourism flows. A thorough review of previous studies reveals that no-one has yet dealt extensively with the issue of whether host-country features (such as the communications infrastructure) exert any effects on the performance of the tourism sector. Furthermore, the previous studies that are available have focused only on specific countries or groups of countries, such as those in southern Africa. In this chapter, we examine communications infrastructure (of internet and telephones relative to size and population) in terms of panel data and with division into subsamples. This chapter seeks to identify the most important determinants that have an impact on tourism (in terms of arrivals) in a whole sample of 131 countries and in sub-samples that comprise developed and developing countries. Sub-samples are also taken on the basis of the median population size of sample countries, and on the basis of World Bank indicators according to the 2012 classification. The determinants employed comprise economic, demographic, technological and political factors. In addition, we use governance indicators as a proxy for institutional quality applying principal component analysis (PCA). The six indicators of good governance comprise accountability of power, political stability, rule of law, regulatory quality, corruption, and government effectiveness, in respect of their effects upon tourism. For this analysis, our sample covers 131 countries over the period 1995 to 2007. Furthermore, we applied static and dynamic panel-data methodology to our analysis. The results of this chapter indicate that the governance of the destination is shown to be an important factor influencing the process of destination choice, in the case of both developed and less-developed countries, but 19 2. Tourism: Core Concepts some interesting differences arise between them with regard to the impact of conflict. In particular, in developing countries violent events have a more profound effect on tourism arrivals than is the case for developed countries. Murdoch and Sandler (2002) found that violent conflict is observed to be more of a detriment to economic growth in developing countries in the short-term, and that its negative impact via tourism can harm the economy as a whole. In addition, in the present study the level of technology available in or for the tourist destination is found to be the main universal factor explaining comparative advantage within the tourism market. Chapter 5 contributes to the literature on the tourism gravity model by using the ICRG data-set on institutional quality and political stability and examining the linkage between tourist flows and institutions in a global framework. This study applies the tourism gravity model while concentrating on the institutional enviorment which raised by the ICRG data set as they affect different countries. The gravity model (first posited by Tinbergen, 1962) is useful in making it possible to investigate trade flows between two countries by examining the distance between them and other factors that influence those flows. Many studies in tourism literature based on the gravity-model approach employ cross-sectional data, which is often the most appropriate form. However, the shortcoming of this approach lies in the possibility of its producing biased estimations owing to heterogeneity in the data drawn from different countries. To overcome this issue, panel data can be used instead of cross-sectional data (see for example Mátyás, 1997; Egger, 2000, 2002). In order to employ the classical panel-data estimation methods, the model is first transformed log-linearly and then the multiplicative gravity equation is estimated. This approach applies when using either cross-sectional or panel-data estimation methods. When the latter are used, they are naturally controlled for dataheterogeneity among countries. Either the fixed-effects or random-effects estimation methods are applied when panel data are being analysed. On the other hand, when using cross-sectional methods, the traditional ordinary least squares (OLS) technique is usually applied. According to Santos Silva and Tenreyro (2006), the estimation results based on the logarithmic-transformed model could be misleading in the 20 2. Tourism: Core Concepts presence of heteroskedasticity. They showed that in the presence of heteroskedasticity the assumptions are in general violated. This conclusion stems from the phenomenon highlighted by Jensen's inequality (Jensen, 1906) that states that the expected value of a logarithm of a random variable does not equal the logarithm of the expected value. The logarithmic transformation of the model is also beset by difficulties originating from the need to deal with zero-trade flows. In order to solve this problem, the gravity model should be estimated directly from the multiplicative form using the Poisson pseudo-maximum-likelihood estimation technique (Santos Silva and Tenreyro, 2006). This solution was first applied to cross-sectional data and later on to panel data. Westerlund and Wilhelmsson (2009) showed that, even when panel data are used, the presence of heteroskedasticity renders the traditional estimation biased and inconsistent when applying either of these two different approaches to gravity equation estimation on simulated and real data. In this chapter, owing to the need for a greater spread of data to yield relevant results, we applied our analysis on panel-data sets comprising 134 origin-countries and 31 destination-countries, which again were selected depending on relevant data availability. We estimated the gravity equation by using (a) traditional, (b) Hausman-Taylor, and (c) Poisson estimation techniques. We evaluated the performance of these three methods with respect to the theory of the gravity equation. The negative consequences of higher political risk for the tourism industry are highly important. To the best of our knowledge there exist no studies that investigate the effect of ICRG data using different methodology or gravity specifications. Thus this present study has undertaken to examine the various diseases of political instability (such as acts of terrorism, conflict, other forms of violence et cetera) that have negative effects on tourism. Finally, Chapter 6 summarizes the main findings of this thesis, and presents the major conclusions from the present research. It also offers some recommendations and suggestions regarding policy implications, and identifies and discusses the main limitations of this thesis. It ends with suggestions for future research that are beyond the scope of this thesis. 21 2. Tourism: Core Concepts 2 Tourism: Core Concepts 2.1 Introduction Many authors have provided various definitions the term ‘tourism’ according to their point of view. So, the UNTWO (2010) has defined tourism as activities involving travelling and staying by individuals in locations situated away from their habitual environment for relatively short periods (usually less than one year). In addition, the reason for such activities should be leisure, business or other purposes excluding being employed by a resident person or body in the country or location visited. The elements of time duration and purpose of visit are derived from the seminal work done by Hunziker and Krapf (1942) which will be discussed later. Defining the term ‘tourism’ has led also to variance in the meanings accepted regarding the supply and demand of tourists, the economic status of the origin/destination country, and to movement of people between countries. The varying definitions that have been put forward need to be reconciled to develop definitions that have more universal application. The task in reconciling these variances appears to be less straightforward, owing to the need for strong collaboration between the sectors managing hospitality, transport, retailing, and attractions in order to fully understand the workings of and the returns from the tourism sector (Buhalis and Cooper, 1998; Buhalis, 2000; Brent Ritchie and Crouch, 2003; Smith, 2007). Discussing the situation in Great Britain, Heeley (1980) defined tourism following two approaches. The first one reflected the essential nature of tourism, which can be related to four components: catering, transport, attractions and accommodation. The second approach addressed all the relationships involving the visitors who stay in the destination and do not exercise a major or permanent remunerated activity. This follows on from the definition of tourism made by Hunziker and Krapf (1942) that described it as the totality of different phenomena generated by visitors and the relationships existing between them and the host population, although their definition explicitly stipulated that the visit did not lead to permanent residence. Their definition also excluded any earning activity, and was 22 2. Tourism: Core Concepts one of the first to be generally accepted. Their influence on modern tourism studies continues to a certain extent because their ideas regarding relationships and economic status were incorporated in somewhat modified form into later ideas on tourism. However their technical definitions were weak because they had ideas concerning the social nature of tourism that depended heavily on the modes of visitation, visitor accommodation and the resultant relationships that were current at that time but have become outmoded (Shaw and Williams 2004). Burkart and Medlik (1974) recommended that it would be advisable to differentiate between the conceptual and technical or practical aspects of tourism definitions. They were eager to develop theoretical definitions of tourism to encompass its various characteristics. Their work was taken up in 1979 by the British Tourism Society, which stated ‘tourism is deemed to include any activity concerned with the temporary short-term movements of people to destinations outside the places where they normally live and work ,and their activity during the stay at these destinations’ (quoted by Vanhove, 2005, page 2). From the foregoing, we can see the significance attributed to activities involved in visiting and perhaps staying at the destination. This is very much a demand-side model. The International Association of Scientific Experts in Tourism (AIEST) at its conference in Cardiff in 1981 declared tourism to be ‘the entirety of interrelations and phenomena which result from people travelling to and stopping at places which are neither their main continuous domiciles nor place of work either for leisure or in the context of business activities or study’ (quoted by Vanhove, 2005, page 2). This definition includes both the spatial and dynamic aspects of tourism, but again is very much weighted to the demand side of the totality. However, to differing degrees tourism also involves the interaction between tourists and the local population. The definition should ideally make mention of the various interactions and outcomes arising from the relationships between the tourists, tourism suppliers, the government and the local people, thus introducing a supplyside aspect to the model. The International Government Conference on Travel and Tourism Statistics held by the World Tourism Organization (WTO) in Ottawa, Canada in 1991, stated ‘tourism comprises the activities of persons travelling to and staying in places outside their usual environment for not more than one consecutive 23 2. Tourism: Core Concepts year for leisure, business or other purposes’ (Holloway et al., 2009, page 6). Like the previous definitions, this is very much concentrates on the demand side of tourism, concentrating the main points on definitions of tourism, travellers, and tourists. The expression ‘usual environment’ is intended to ‘exclude trips within the person’s community of residence, as well as other usual trips, frequent and regular, between house and place of work’ (Page and Connell, 2006, page 12). Regarding the distinction between a person’s ‘usual environment’ and a tourism destination, many researchers have been concerned to include some element regarding the distances by travelled tourists away from their homes. For instance, in defining domestic tourism for statistical purposes in Australia, Stanford and McCann (1979) proposed that a tourist should had to have travelled at least 40 kilometres from the usual place of residence, and this definition is still used for determining ‘overnight stays’ by the Australian Bureau of Statistics (2003). In addition, governments and academics have variously defined the term ‘tourism’ according to the themes prevalent in such domains as geography, economics, sociology and cultural anthropology. Geographers are interested in the locational aspects of tourism and how it changes the natural and built environment. Economists are concerned with the contribution of tourism to the economy and they focus on demand/supply, foreign exchange and other financial aspects. Sociologists and cultural anthropologists investigate the travel and consumer behaviour of people as individuals and groups, located within the milieu in which they travel and stay. The habits and traditions of hosts as well as guests are taken into account (Theobald, 2005). Hence, tourism embraces a composite of activities, services, accommodation facilities, industries, transportation, travel experience and other hospitality services that involve both the consumer (the tourist) and the supplier (Williams, 1998; Smith, 2007). 2.2 Basic Tourism Concepts Tourism and tourists can be divided into various categories, and the anthropologist Valene Smith went so far as to identify five different types of tourism and seven different forms of tourist (Smith, 1989, pages 1–20). Taking a measured 24 2. Tourism: Core Concepts approach, we may discern the following types of tourism. Domestic tourism occurs when residents of a country visit destinations within their own country (subject to strictures about minimum distance travelled, et cetera). Another category is inbound tourism where visits to a country are made by non-residents. The combination of domestic tourism and inbound tourism may be called internal tourism although some authorities use this term to describe domestic tourism. Furthermore, outbound tourism consists of residents of a country visiting destination in other countries, which may be described as national tourism, although this would more strictly be applied to visits to a single political entity, whilst visits by the same person or group to more than one political entity might be called regional tourism. Finally, international tourism embraces the combination of inbound tourism and outbound tourism. An analysis of tourism must also include an analysis of the tourist. So, in attempting to define the tourist, one must look at the tourist activity itself. According to Burkhart and Medlik (1974), the tourist activity consists of two elements: the dynamic element and the static element .This means that the tourism activity involves tourists staying at a considerable distance from their original place of residence for at least one night, in addition to the time taken for the journey or trip (Smith, 2004; Williams et al., 2004). The discretionary options of using time and monetary resources are other factors which can distinguish the tourist from the day tripper. This can be seen clearly in the case of holiday tourists and can be applicable also to certain instances of business travel. For example, conferences outside the workplace are normal dayto-day activities for the employees who are participants, while the particular circumstances might also mark them out to be tourists (Leiper, 1979; Smith, 2007). The consumption of economic resources is also another factor that can characterize the tourist, as defined by the nature and measure of the expenditure behaviour of the person in question. One sort of business travel that cannot be considered as tourism, however, involves seasonal workers and commercial travellers who are engaged in performing their routine jobs; they do not exercise discretionary powers in the same way as tourists, so they are not tourists. Tourists do not normally travel for the sole purpose of remuneration and this differentiates them from travelling workers 25 2. Tourism: Core Concepts (Leiper, 1979; Pearce, 1993). Furthermore, geographical elements can define the tourist through the specific flow-patterns (Leiper, 1979; Smith, 2007), as elaborated below. (A) Tourist-generating regions are the ‘permanent residential bases of tourists, the places where tourists begin and end and in particular those features of the region which incidentally cause or stimulate the temporary outflow’ (Leiper, 1979, p. 396). This definition embraces the geographical and behavioural factors that drive the tourist-generating regions. So the generating regions also form part of the travel and tourism market industry, as the relevant business help generate demand for touristic travel to the destinations. The most important marketing activities for the tourism industry are the promotional aspects: advertising, retailing and wholesaling .Determining these functions can assess why a particular tourist region might experience a tourist exodus, in addition to the economic and social conditions in the region. (B) Tourist-destination regions are ‘locations which attract the tourist to stay temporarily and in particular those features which inherently contribute to that attraction’ (Leiper, 1979, p. 397). These attractions can be determined by the tourist in terms of several qualitative characteristics which he or she hopes to experience at the destination. The majority of tourist studies have assessed the tourist destinations in terms of location, accommodation, services, establishments, facilities and entertainment—in other words, where the most important aspects of the experience occur (Williams et al., 2004). (C) Transit routes region can be defined as ‘paths linking the touristgenerating region along with tourist travel’ (Leiper, 1979, p. 397). Transit routes are very important because their characteristics can affect the quality of access to a particular tourist destination. 26 2. Tourism: Core Concepts 2.3 Market for Tourism Products A market, in all forms, can be defined as a place where the buyers and sellers come into contact with one another (Diaz Ruiz, 2012). Thus, the market of tourism is the place where tourism demand and tourists meet tourism supply, together with the persons, firms and institutions that work in the domain of tourism services. 2.3.1 The Demand for Tourism Products ‘Tourism demand’, as the term is normally used, refers to a range of tourism products—goods and services—that the consumer is well-disposed towards and able to purchase during a specified time within the set of given conditions (Song et al., 2009, page 2). Tourism demand can be studied through various approaches. The economic approach examines tourism demand as the relationship between demand and price, or other factors. Meanwhile, the geographer takes into consideration the environment effect of the demand for tourism. The psychological approach studies many influences, not only on those who actually participate in tourism but also those who wish to (Cooper et al., 2008). Buhalis (2000) suggests that tourism demand can be conceptualized as three basic types which form the total of tourism demand. First, there is effective demand, represented by the actual number of tourists who complete their trips. The second type, suppressed demand, consists of people who do not travel for some particular reason, either because of personal circumstances or owing to external conditions that make travel impossible. The third type, called latent demand, refers to the potentiality of a location or some particular feature to generate demand. The demand for tourism differs from one place to another and from one specified period to another; these differences may be quantitative or qualitative because tourism demand is affected by a large group of economic or other factors. The most important factors are prices, incomes, price of other goods, fashion and tastes, advertising, leisure time and population. 27 2. Tourism: Core Concepts 2.3.2 Tourism Supply The definition of tourism supply faces a major problem owing to the wide variation in the spectrum of tourism businesses and organizations that are involved, from those that are wholly dedicated to servicing tourists to those that also serve local residents and other markets (Cooper et al., 2008, p.13). However, RossellóNadal et al. (2007) viewed the tourism supply as a set of tourism products and services for tourists to use and consume at certain destinations. This definition thus refers to services that have for the most part been planned privately to meet tourism demand (accommodation, shopping, sports facilities, et cetera). On the other hand, Gunn and Var (2002) introduced the idea that the supply of tourism consists of all planned programs and land uses provided for receiving tourists, and that these programs are controlled by the policies and practices of all three sectors (private enterprise, non-profit-making organizations, and governments). The quantity and quality of tourism supply differs from one country to another as a result of a group of factors such as technical improvements, prices, prices of other goods provided, taxes and subsidies, and other factors including wars, industrial relations and the weather. 2.4 Factors affecting Tourism In recent years the tourism sector has faced a number of problems and challenges generated by a range of factors, including economic, political, demographic, and technical factors, as well as threats and crises. The economic factors include the exchange rates, income levels, competition and efficiency of the national economy (Prideaux, 2005). Several studies have investigated the impact of exchange-rate movements on tourism services. For instance, the spending of overseas tourists declines in real terms in the UK when the UK pound is strong. A bivariate analysis shows significant effects reflected in the relationships between a country’s exchange rate and the expenditure-levels of overseas tourists in that country (Tse, 2001). This is seen in how the Asian financial crisis led to a growth of approximately 19.6 percent in Australian tourist flows to Indonesia during the 1997–1998 collapse in value of the Indonesian rupiah. Conversely, Indonesian tourist arrivals to Australia decreased by about 20 percent in the same period (Prideaux, 2005). Patsouratis et al. (2005) showed how exchange- 28 2. Tourism: Core Concepts rate fluctuations figure as an influential determinant of international tourism flows. Another important economic factor is the efficiency of the host economy—including the cost of public services and facilities, such as domestic transport costs, communication and cost of financial services. As the efficiency of the national economy improves, the demand of outbound tourism may increase (Prideaux, 2005). Political conditions, war, terrorism and political instability exert considerable influence on the decision-making processes of tourists. For example, Africa and Pakistan may have big game and majestic animals to hunt there, but the lack of personal safety acts as a serious deterrent to people unwilling to take such risks. The fear of terrorist activities likewise discourage people from making touristic visits to the areas affected. For example, after the September 11 attacks in the USA, the volume of cancellations made by private individuals and corporate bodies resulted in a loss of 2 billion dollars to the USA economy within the first month following the attacks (Goodrich 2002). Furthermore, no country’s tourist industry is immune to the effects of economic and financial crises elsewhere in the world. Papatheodorou et al. (2010) showed how the financial crises occurring in the summer of 2007 sent shock-waves that had grave consequences for national economies around the globe. The advanced economies showed a 7.5 percent decline in real tourism GDP during the last quarter of 2008 (IMF 2009a). The WTTC reported a drop in the growth rate of the travel and tourism industries to 1.0 percent in 2008 as a proportion of GDP. Various authors studied the possible effects that financial and economic crisis could exert on tourism. They showed that people who sustained income-loss during crisis conditions tended to finance their travelling plans from savings. If the economic downturn became lengthy, people would reduce their holiday expenditure (by taking shorter stays, visiting destinations closer to home, and so forth). If financial hardship became worse, they would cancel their plans (Smeral, 2009). Song and Lin (2010) predicted that the crisis would have a negative impact on both inbound and outbound tourism in Asia. In addition, Brent Ritchie et al. (2010) investigated the impact of economic crisis 2008–2009 on tourism in North America—Canada, the USA and Mexico. They found that tourism had been affected by the economic crisis because in Canada 29 2. Tourism: Core Concepts the decrease in disposable income had influenced the future travel plans of Canadians, whilst regarding the United States, the events of 9/11 and subsequent politics had exerted an effect more serious than the later economic crisis. In the case of Mexico, natural disasters, such as swine flu, had exerted a greater effect on Mexico’s tourism industry than the later economic crash. Similarly, Song and Lin (2010) show that the economic and financial crisis was bound to have a particular negative effect on both inbound and outbound tourism in Asia. In the meantime, many destinations outside Asia might attract more Chinese tourists and thus recover their tourism industry, if the Chinese economy were to remain strong during the next few years. The tourism industry has begun to exploit technology. This will change not only the type and scope of the services offered, but also the sort and extent of work within the industry (Buhalis, 1998; Werthner and Klein, 1999; Pease and Rowe, 2005). Technology as a factor facilitates the speed and efficiency with which the tourism industry operates. Information technology in the tourism sector can reduce considerably the costs of information handling, increasing the speed of information processing, whilst customers are enabled to interact more effectively within the whole process. It also affords flexibility in product-adaptation and greater reliability in the transferring of information (Hudson, 2008, pages 8–11). Also, advances in technology have huge effects on the operation of business tourism, including presentation technology in conferences. According to the UK Conference Market Survey (2002) 86 percent of organizers used PC–based facilities for presentations at conferences. Furthermore, the quality of services and facilities is an important factor in the conduct of business tourism, such as using valuable time in holding meetings, training staff, as well as having within a destination good quality of transport systems, accommodation and restaurants. Additionally, demographic change is one of the most influential drivers of developing trends in consumer behaviour in most European countries (Lohmann, 2004). Two important demographic trends are coming to prominence. The first in many countries is the rapid increase in the old-age population sector, owing to the rise in life expectancy particularly in the developed countries (e.g. Katz and Marshall, 2003; Bloom et al. 2011). Tomljenovic and Faulkner (2000), for example, 30 2. Tourism: Core Concepts found that older people are more favourably inclined toward tourism development than younger ones. Secondly, a declining number of children as result of falling fertility (particularly in industrial countries) in conjunction with the breakdown of traditional family structures. It is important to take account of this trend in forecasting a declining number of children as result of falling fertility particularly in industrial countries combined with the dissolution of traditional family patterns (Behnam, 1990). Also, in most cases people maintain the holidaying patterns acquired up to the mid-point of their lives. Therefore, their travel behaviours do not change simply when they enter their sixties or retire (Lohmann, 2004). This fact allows for predictions of much future tourist behaviour. The occurrence of diseases also affects tourist flows, such as Severe Acute Respiratory Syndrome (SARS) which appeared in some countries and hit tourism business in 2003 (Wilder-Smith, 2006). Pine and McKercher (2004) found that when SARS appeared on 12 February 2002 in Guangdong Province in China, it had a negative effect on tourism industry across the world, with business travel being especially depressed owing to the postponement of and/or decline in capital investment. The WTTC (2003) reported the enormous impact of SARS on Hong Kong and Singapore in reducing the tourism contribution to GDP by 41 percent in Hong Kong and 43 percent in Singapore. Min (2005) examined the impact of SARS on tourism in Taiwan and found that arrivals had been severely reduced by the SARS outbreak. Natural disasters such as the December 2004 tsunami also deterred tourists from travelling to the affected countries (EIU, 2005). The spread of human fatalities as a result of the H5N1 avian flu virus was also a great deterrent to tourists (McAleer et al., 2008). These problems adversely impacted on the image and the reputation of the affected destinations and caused many tourists to cancel their travel plans and remain in their own countries. Others problems negatively impacting on tourism are pollution and the rapid industrialization taking place in many cities and rural areas. Moscardo (1999) found that there was a relationship between environmental disasters and income hotel income along the Queensland Coast of Australia. Furthermore, the 1997 haze-related air pollution caused economic losses of about US$256 million to the tourism industry in Indonesia, Malaysia and Singapore (Anaman and Looi, 2000). Wang (2009) examined the impact of adverse 31 2. Tourism: Core Concepts events such as SARS and Asian financial crises events on the number of inbound tourism arrivals and found that any impact on safety—whether it be domestic or international—depresses tourism demand. Climate and weather are important factors in planning in the tourism industry. De Freitas (2003) found that effective information regarding climate conditions facilitates effective management and business planning. Gómez Martín (2005), Kozak et al. (2008), and Becken (2010) showed that climate plays an important role in motivating tourists to travel. Scott et al. (2008) added that the presence of a better climate in a person’s home region is related to a higher probability of domestic travel, whereas poor weather conditions increase the likelihood of a person’s undertaking international travel. 2.5 Benefits and Costs of Tourism Tourism has many benefits and it has a great impact on most countries. The impacts are economic, social, cultural and environmental, and their influence on tourism destinations might be positive and/or negative (Mathieson and Wall, 1982). Consequently, it is highly important for the tourism industry and destination residents to cooperate to plan for manageable growth and sustainable development (Buhalis and Cooper, 1998; Andereck and Vogt, 2000; Harrill, 2004; Dredge, 2010). Planning can help create a business sector with minimal costs to make tourism a blessing rather than a curse (Marzuki and Hay, 2013; Stylidis et al., 2014). According to UNWTO/ILO (2013), the economic benefits of tourism are derived through receipts from expenditure by visitors on accommodation, catering, and all the other services and goods generally required; these reached an estimated US$ 1159 billion (euro 873 billion) in 2012. Various researchers have studied the relationship between tourism and the economy (e.g. Fletcher, 1989, 1994; Zhou et al., 1997; Blake et al., 2001; Dwyer et al., 2003, 2004; Narayan, 2004). In his study of tourism in Spain, Blake (2000) found that an increase in tourism expenditure owing to an increase in tourist demand leads to an adjustment through a real exchange rate appreciation. Using a similar methodological approach, Narayan (2004) examined the economic impact of tourism on the economy of Fiji and concluded that tourism development leads to exchange- 32 2. Tourism: Core Concepts rate appreciation and to an increase in both domestic prices and wage rates. Tourism also helps some countries to withstand economic turbulence. For example, tourism in Cuba is a blessing because tourism has helped the Cuban economy to survive two disasters: the collapse of the USSR and the tightening of the US economic embargo on the island (Wilkinson 2008, page 981). In a different way, Cortés-Jiménez et al. (2009) consider both exports and tourism as potential factors for economic growth; they use inbound tourism as nontraded exports. Meanwhile, trade of goods was an engine of economic growth in two developed countries: Italy and Spain. The authors confirmed the long-run hypotheses of exports-led growth and tourism-led growth for both countries. Thus, tradable exports and inbound tourism can be considered as important channels for inducing economic growth. Employment opportunities are some of the most important benefits of tourism, since tourism is a large industry and can provide many jobs. This range of jobs suits many categories of people, including young people, as there are part-time and full-time jobs in tourism. Őnder and Durgun (2008) found that tourism had a positive effect on employment in Turkey, and that there is a mutual relationship between the two variables in the long term. Skene (1993a, 1993b) investigated the impact of tourism on employment in Australia. The studies also found that an increase in exports driven by tourism could offset an increase in imports, thus contributing to a balance of trade surplus. Therefore, tourism might be able to help the economy of the tourism destination to be less reliant on other sources such as agriculture and mining. This can be a benefit, and a significant one for regional rural communities. Figure 2.1 below shows that from 2000 to 2003 there was a decrease in employment and it reached its lowest point in 2003. Various changes of circumstances in subsequent years gave a stimulus to tourism and thus to employment, reaching a peak in 2007. Figure 2.1 shows the percentage of employment in tourism as part of economy-wide spending. 33 2. Tourism: Core Concepts Figure 2.1 Tourism employment Source: World Tourism Organization/International Monetary Fund Various authors found that, on the negative side, tourism also gives rise to economic costs, derived from fluctuations in tourist demand and supply (Ball, 1988; Song, 2010; Ardahaey, 2011; Marcussen, 2011). Furthermore, tourist activity might give rise to inflation when the buying capacity of the visitors is greater than that of the locals. The effects can be seen in rises in the prices of land, catering and services (Butler, 1996; Wagner, 1997; Lindberg et al., 2001; Archer et al., 2005). Moreover, tourism can reverse and cancel economic benefits through creating high dependency on external capital and distortions in the local economy through the centralization of economic activity in a single sector (Frechtling, 1994). Many studies have tended to focus on tourism as a source of wealth without taking into account the possibility that the tourism industry might also become a curse owing to its over-utilization of local and natural resources. For example, Nowak et al. (2004) investigated the impact of a tourism boom on structural adjustment, commodities, factor prices and residents’ welfare. They found that a tourist boom may cause the immiserization of residents if the beneficial impact which is caused by an increase in relative prices of non-traded commodities outweighs the negative effect which happens as a result of a loss of efficiency that occurs when returns increase to scale in the production and sale of manufactured goods. Capó et al. (2007) studied tourism as ‘Dutch Disease’. Their findings indicated the need to find a new export using natural resources to overcome the excessive 34 2. Tourism: Core Concepts dependence on tourism as an earner of external currency. They found that the tourist inflow boom of the 1960s induced a significant increase in wealth in Spain. But focusing on tourism and non-traded goods caused a lack of attention to industry and agriculture at the same time. Similarly, Chao et al. (2006) discussed the existence of Dutch Disease through a demand shock from a tourism boom using a dynamic framework, examining the impacts of tourism on accumulations of capital and welfare in an open dynamic economy. They showed that tourism can act to reduce local resident welfare as a result of the existence of externality which worsens the impact of industrialization. In addition, especially in a small island economy, the boom of inbound tourism may cause a loss of welfare when tourism activities and products use the coastal land areas intensively (Nowak and Sahli, 2007). Regarding social and cultural impacts on tourism, the interaction between tourists and the host community can likewise be positive or negative for the host community (Mathieson & Wall, 1982). Cooper et al. (2008) showed that tourism can improve the quality of life in a tourist destination, by increasing economic activity and offering a range of facilities initially aimed at visitors but that might also be used by locals. Tourism might also cause beneficial change in the traditions and customs of the host community by fomenting cultural exchange (Besculides et al., 2002; Carter and Beeton, 2004). Tourism might also help to preserve the cultural identity of the host location, by creating increased demand for the exhibition and exercise of local culture, which might otherwise have fallen into obscurity (Throsby, 1994; Quinn, 2009). Thus it can be argued that tourism can foment the conservation of cultural values and practices which might have been lost, if the locations had not been attractive to visitors (Richards, 1996). Conversely, the effects of tourism might act to suppress and destroy local tradition and culture through a disparate degree of economic power and prestige enjoyed by the tourists (Robinson, 1999; Throsby, 2001). In other cases, cultural preservation by commodification has cost the communities their authentic traditional customs, folklore, crafts, festivals—all of which have been grossly modified for consumption by visitors (Shepherd, 2002; Carter and Beeton, 2004; McLeod, 2006). The social difference between local population and tourists is another negative impact (Robinson, 1999). Where the gradient of difference is so steep as to put local 35 2. Tourism: Core Concepts residents at a gross disadvantage, then they become little more than servants for the tourists, thus creating a certain resentment among the local populations against the visitors (King et al., 1993). Tourism can thus establish a new form of colonialism based on local dependence on the income the foreign tourists bring (Teo and Leong, 2006; Wearing and Wearing, 2006). Such tourism-dependency can foster excessive drinking, alcoholism, gambling, crime, and drug-taking among the locals; it can cause these and other unwanted lifestyle changes that will lead to negative changes in traditions and customs (Cooper et al., 2008). Tourism also can cause cultural degeneration of the destination (Pizam, 1978). In this situation the local people might allow tourists to trespass upon or violate cultural practices or norms that have been current and cherished in the local community (Pandey et al., 1995). Otherwise they might try to adapt themselves to the customs and cultures of the visitors, and in that process they may possibly go so far as destroying the elements that underpinned the original attractiveness of the location for the tourist (Cohen, 1987). Other negative aspects of tourism include the impact exerted on the environment through pollution (airborne, water-borne, solid), degradation of the natural and open landscapes, and destruction of flora and fauna (Jenner and Smith, 1992; Croall, 1995). The invisible costs of tourism on the environment mount up, including landscapes that have been used to build hotels and airports, whilst pollution of waterways and the sea seriously undermines the welfare and stock levels of fish (Cater and Goodall, 1992; Wilkinson, 1992). Whilst Wilson (1997) addressed the problems facing Goa owing to rapid development of the tourism industry, Sawkar et al. (1998) showed that tourism has nevertheless delivered many benefits to Goa, as also to the Maldives, through the allocation of funds to protect parks and to support resource management research. A properly managed tourism can act to enhance a country’s appearance as well to preserve the environment. 2.6 Tourism and Economic Growth The large number of empirical studies relating to this topic can be divided into two main categories: the first one examines the relationship between tourism and economic growth using time-series techniques such as causality and co-integration in 36 2. Tourism: Core Concepts each country individually. The second one applies panel-data methods, considering many countries together. Using a panel-threshold model and measuring economic output by GDP, Chang et al. (2009:4; 2010) found that for the period 1975–2000 tourism growth ran at an average annual rate of 4.6 percent (UNWTO, 2008). During this period, the growth in tourism volume positively surpassed growth in economic output, although it fluctuated in line with GDP growth. When GDP growth was greater than 4 percent, tourism volume growth would be much higher than that, whilst in years when GDP growth was less than 2 percent, tourism growth was much depressed. They illustrated their point by plotting a graph showing the relationship between international tourism arrivals and GDP over the previous period (see Figure 2.2). Figure 2.2 Economic Growth and International Tourist Arrivals 1975–2005 Source: World Tourism Organization/International Monetary Fund In his study, Zortuk (2009) investigated the rapidly-developing tourism sector and the contribution that it was making to post-1980s economic growth. To this end he used the Granger Causality Test based on a Vector Error Correction Model (VECM) in order to examine the relationships between variables, the growth of GDP, tourist arrivals, and exchange rates in Turkey. The main results of his study showed a long-run relationship between Gross Domestic Product (GDP) and Tourist Arrivals (TOAR), as well as a unidirectional positive causal relationship. Using similar methodology, Balaguer and Cantavella-Jordá (2002), Dritsakis (2004), and 37 2. Tourism: Core Concepts Sanchez Carrera et al. (2008) analysed the effect of tourism on economic growth in Spain, Greece and Mexico respectively; they concluded that there is a positive relationship between tourism and economic growth. Furthermore, Barquet et al. (2009) also used causality-relationship techniques to study the link between economic growth and tourism expansion in Trentino-Alto Adige in Italy. In their study, GDP served to measure economic growth. They considered the relative prices between Trentino-Alto Adige and Germany as proxy variables for external competitiveness between 1988 and 2006. They concluded that international tourism expenditure positively impacted on the Trentino-Alto Adige economy; the relative prices produced a positive but slight effect. Furthermore, causality testing shows that the relative prices between Trentino-Alto Adige and Germany were weakly exogenous. Chen and Chiou-Wei (2009) applied an EGARCH-M model with uncertainty factors, examining the relationship between tourism and economic growth in two Asian countries: Taiwan and South Korea. The findings confirmed the hypothesis of tourism-led economic growth in Taiwan. For South Korea, they found a mutual causal relationship between the two variables under study. In contrast, Oh (2005) tested the causal relationship between economic growth and economic expansion in Korea. There were two main results: first, there is no relationship between the two variables of interest according to the co-integration test; second, the Granger Causality test implies a unidirectional relationship of economic-driven tourism growth. Using similar methods but with a different proxy (four industries related to tourism—airlines, casinos, hotels, and restaurants) Tang and Jang (2009) examined the relationship between tourism and GDP in the USA; the results showed that there was no co-integration between economic growth and the tourism industry. Moreover, the Granger causality test exhibited a unidirectional causality from GDP to the aforementioned four industries, which may represent a small portion of these industries to the whole economy. In addition, the causality tests displayed a temporal causal hierarchy; this temporal hierarchy might be used as a tool for the public and private sectors since it offers a guide for organizing industries according to their importance for the whole set of tourism and economic outputs. Furthermore, the airline and hotel industries seem to provide essential performances that might help to 38 2. Tourism: Core Concepts establish effective plans for using resources in these two industries rather than dividing resources equally among all four industries. With a different methodology, Blake et al. (2001) used the input-output approach to examine whether tourism is a key sector for the US economy. Kweka et al. (2001, 2003) did likewise for the Tanzanian economy. These studies found that tourism affects the economy significantly. Moreover, they identified tourism as a potential sector for driving an increase in economic growth. Eugenio-Martin et al. (2004) examined the relationship between tourism and economic growth in Latin American countries for the period 1985 to 1998 with an analysis based on a panel-data approach. The authors showed that an increase in the number of tourists per capita of local population has a positive effect on the economic growth of those countries having low and medium levels of income per capita, but not in the case of rich countries. This finding suggests that the increase in the number of tourist arrivals in a country offers an opportunity for economic growth for those countries that are still less-developed, but not for those countries that have become developed. Using similar methods, Sequeira and Nunes (2008) showed that tourism is a positive determinant factor of economic growth both in the total sample of countries and in poor countries of the sample. In addition, Figini and Vici (2009) made a cross-sectional analysis to show the relationship between tourism specialization and economic growth. They used data for more than 150 countries during the period 1980–2005. In contrast to the findings of Sequeira and Nunes (2008), they concluded that there were no grounds to suggest that tourism-based countries would generally have a higher growth-rate than nontourism-based countries. Arezki et al. (2009) also used a large cross-section of countries with instrumental variables techniques, covering the period between 1980 and 2002, to examine whether tourism specialization was a viable option for development. They defined a tool for tourism depending on the UNESCO World Heritage List. The result showed a positive relationship between the size of tourism specialization and economic growth. They supported this result with a great array of robustness checks. Furthermore, Po and Huang (2008) applied cross-sectional data for 88 countries over the period 1995–2005 to examine the non-linear relationship between tourism growth and economic growth. They recognized the indicator of 39 2. Tourism: Core Concepts tourism specialization (defined as receipts from international tourism as a percentage of GDP) as a threshold variable. The findings of the non-linearity tests shows that the countries fall into three different groups. The threshold regression results indicated a significant positive relationship between tourism development and economic growth when tourism specialization is less than 4.0488 percent (Group 1 = 57 countries) or over 4.7337 percent (Group 3 = 23 countries). However, if the threshold variable lies between these two values (Group 2 = 8 countries), there is no evidence for this significant relationship, owing to the low ratios of value added to the GDP by the service sector in Group 2 countries. Brau et al. (2003) investigated whether tourism specialization is a good option for many less-developed countries and regions. They made a comparison of growth performance for 14 tourism countries within a sample of 143 countries covering the period 1980–1995. The standard OLS cross-country growth regressions were included in their study. They found that the tourism countries showed significantly more rapid rates of growth in contrast to all the other countries in sample (OECD, Oil, LDC, and Small). On the other hand, Schubert et al. (2009) studied the impacts of international tourism demand on the economic growth of small tourism-dependent economies. They considered a large population of temporal optimizing agents as components of the dynamic model, and incorporated an AK technology (endogenous growth model) to present tourism production. The result of this model showed that the growth of tourism demand causes an increase in economic growth and trade. Fayissa et al. (2008) explored the potential contribution of tourism to economic growth and development in Africa within a neoclassical framework, using panel data of 42 African countries over the period 1995–2004. They concluded that the receipts from the tourism sector significantly affected current levels of GDP. In addition, these receipts impact on the economic growth of sub-Saharan countries in the form of investment in physical and human capital. Consequently, the African economies could increase their short-run economic growth by strengthening strategies in their tourism industries. 40 3. The Curse of Tourisms 3 The Curse of Tourism 3.1 Introduction Tourism is an important sector of most countries’ economies, and its significance will continue to increase (Goeldner and Brent Ritchie, 2012). Tourism bestows a number of social and economic benefits on the countries involved. Besides being a source of economic revenue, the process of cultural exchange between the host population and tourist visitors is often cited as a potential source of social benefits (Armenski et al., 2011). In addition, tourism can be considered as a ‘clean’ industry as regards the environment, although many serious debates surround this issue (Cater and Goodall, 1992; Jenner and Smith, 1992; Croall, 1995; Hjalager, 1996; Kreag 2011; Bastola 2012). The economic effects are amongst the most tangible outcomes of tourism, since the economic activity generated by tourism not only boosts the economy through an increase in foreign-exchange income but also helps to generate employment opportunities (Greffe, 1994; Briedenhann and Wickens, 2004; Ashley et al., 2007; Zortuk 2009; Polat et al. 2010) and stimulate the level of economic activity in the country (Ivanov & Webster 2006; Lee and Chang, 2008). According to the WTTC, the world tourism industry accounted for 10 per cent of the world’s GDP in 2004 (WTTC 2013). In spite of the aforementioned benefits of tourism, there is a possibility that tourism can also exert negative effects such as causing deterioration of the environment through the physical impact of tourist visits and leading to overexploitation of natural resources (Capó et al., 2007). Moreover, tourism can cause unwanted lifestyle changes that might have negative impacts on the traditions and customs of the host community (McLeod, 2006; Cooper et al., 2008). Furthermore, recent studies have been investigating whether tourism causes the disruptive economic unbalancing phenomenon known as Dutch Disease, which is discussed at length in Section 2 below. Capó et al. (2007) found that there is evidence of Dutch Disease in two tourism-oriented island areas of Spain, namely the Balearics and the 41 3. The Curse of Tourisms Canary Islands. Their findings indicate that the economic growth of these regions might indeed be compromised by their high dependence on tourism. The rest of this chapter is organized as follows. The next section will present a review of the literature on the relationship between economic growth and tourism; it also discusses Dutch Disease and the potential existence of Dutch-Disease effects in tourism-dependent economies. In Section 3 we describe the data, variables and methodology employed in this chapter. The empirical findings will be presented in Section 4, followed by concluding remarks in Section 5. 3.2 Literature Review 3.2.1 Tourism and Economic Growth Many studies have investigated the relationship between tourism and economic growth in the recipient countries. A considerable number of studies that have examined the relationship by concentrating on a single recipient country have reported findings that indicate positive effects. For instance, Dritsakis (2004) has found long-term positive effects exerted by tourism on economic growth in Greece. Similarly, Balaguer & Cantavella-Jordá (2002) found support for their hypothesis of positive effects for Spain’s economy. Studies on Turkey by Tosun (1999) and Guduz & Hatemi (2005) have also found empirical support for the tourism-led growth hypothesis. Other studies showing similar findings include that of Durbarry (2004) for Mauritius, Kim et al. (2006) for Taiwan, Mishra et al. (2011) for India, and Kadir & Karim (2012) for Malaysia. Moreover, Brau et al. (2003) discussed whether specializing in the tourism industry is a good option for less-developed countries and regions. They documented how tourism-specializing countries displayed significantly faster growth than any of the other sub-groups of countries within their sample (OECD, Oil, LDC, and Small Countries). In other words, the performance of tourism-specializing countries is positive, and is not apparently significantly based on the traditional variables of economic growth as put forth in the Mankiw-RomerWeil model (Mankiw et al., 1992). Tourism specialization appears to be an independent determinant. 42 3. The Curse of Tourisms To the contrary, however, Oh (2005) for (South) Korea, Payne and Mervar (2010) for Croatia, and Lee (2012) for Singapore, found no discernible link between tourism development and long-term economic growth. Figini and Vici (2009) conducted a cross-sectional analysis to investigate the relationship between tourism specialization and economic growth, and they concluded that tourism-based countries did not grow at a higher rate than non-tourism-based countries. In a panel analysis of African countries for the period 1995 to 2004, Fayissa et al. (2008) showed a positive relationship, with tourism receipts making a significant contribution to both GDP levels and general economic growth in sub-Saharan countries. A similar result was found by Eugenio-Martin et al. (2004) for a panel of Latin American countries from 1985 to 1998. Tourism is often viewed as an important engine of economic growth and development, especially for lessdeveloped countries (Brida and Risso 2009; Tang and Tan 2013), helping to increase the economic welfare of local populations. The discrepancies in these various findings might be explained by the choices made by the authors. Some studies suggest that certain variables are important regressors for explaining cross-country growth patterns and are more relevant than others (Sala-i-Martin 1997; Fernández et al., 2001). Similarly, the samples selected by Brau et al. (2003) might not have been wholly representative. If such be the case, then the use of ordinary least squares coefficients (OLS) adopted by Brau et al. (2003) in their analysis might have been particularly susceptible to bias (Ray and Rivera-Batiz 2002). 3.2.2 Dutch Disease The term ‘Dutch Disease’ was first introduced by The Economist (1977) describing the way in which the manufacturing sector in the Netherlands had gone into decline after the discovery of a large field of natural gas in 1959. Exploitation and exports of natural resources (in this case, gas) led to a considerable appreciation in the value of the Netherlands guilder, and this in turn made that country’s manufactured and value-added exports less competitive internationally. An increase in the revenues from natural resources pushes the value of a nation’s currency higher relative to that of other countries. Dutch Disease is thus defined as the negative 43 3. The Curse of Tourisms impact on an economy of foreign currency inflows, which leads to currency appreciation and to higher inflows of relatively cheap imports. In the longer term deindustrialization sets in owing to the difficulties encountered in selling the country’s exports. Corden and Neary (1982) developed the core model of Dutch Disease. The model comprised one non-traded goods sector (services, etc.) and two traded goods sectors, one of which is booming whilst the other is lagging. The booming sector usually arises from the exploitation of some natural mineral resource, typically petroleum and/or gas, although sufficient deposits of copper, gold and other precious metals can have the same effect. The lagging sector is almost always the manufacturing sector. Those industries that do not have any part in resource exploitation activities become uncompetitive and begin to atrophy. The condition is exacerbated by competition from similar industries operating in locations where labour costs are cheaper. Within the depressed sector of the country’s economy, job losses and wage stagnation constitute the push factors that help to drive the internal migration of labour to the more active or booming sector, which also exerts the pull factors of job opportunities and (potentially) higher wages. Furthermore, the additional income provided by the resource boom generates an increase in spending in the economy and leads to further labour-force losses from the manufacturing sector to the non-tradable goods (i.e. the service) sector. When income from naturalresource exploitation begins to decline, a country can find itself burdened with a very uncompetitive and unproductive manufacturing sector that is not able to generate much-needed export revenue. The term ‘Dutch Disease’ soon became applied to cases of varying degrees of similarity; models were devised that drew inspiration from the original, especially regarding situations involving financial dependence on aid or other income not arising from trade or activity across the broader range of a country’s economic sectors (e.g. Bruno and Sachs, 1982; Corden, 1984; Bandara, 1995; Rudd, 1996; Adenauer and Vagassky, 1998; Brahmbhatt et al., 2010; Fielding, 2010; IMF, 2011; Rajan and Subramanian, 2011). 44 3. The Curse of Tourisms 3.2.3 Dutch Disease and Tourism The literature on the links between tourism and Dutch Disease argues that tourism can be compared to an export boom following the discovery of natural resources. In relation to tourism, this phenomenon is sometimes also called the ‘Beach Disease’ (Holzner 2010). Capó et al. (2007) investigated whether tourism causes Dutch Disease in two different regions of Spain—the Balearics and the Canary Islands—both noted for the extremely high and long-standing incidence of tourism. They found that the tourist inflow boom of the 1960s induced a significant increase in wealth in Spain generally, whilst the increased focus on tourism and non-traded goods detracted from necessary attention to industry and agriculture at the local level in these two regions. Their study found that, whilst this change in production did lead to an increase in incomes, there is evidence that these two regions might not be able to maintain economic growth rates for much longer. The reduction in natural resources such as beaches or natural areas is not the sole driver of growth-decay. Rather, it is the heavy focus on the tourism sector that has led to the neglect of other sectors that might provide economic activity and employment during a recession in the tourism industry. The decline of the traditional sectors (manufacturing and agriculture) has deprived these tourism-dependent regions of much-needed economic diversity. The failure to introduce economic diversification into these regions could lead to their becoming mono-industrial areas whose populations might find it extremely difficult to gain competence in activities unconnected with tourism. The neglect of economic diversification, on-going education and training, combined with a lack of technological innovation at the local level are not only symptoms but also drivers of Dutch Disease for these regions. Using a theoretical model, Chao et al. (2006) discussed the existence of Dutch Disease through a demand shock from a tourism boom using a dynamic framework, examining the impacts of tourism on capital accumulation, sectoral output and resident welfare in an open dynamic economy. The authors realized that the expansion of tourism causes an increase in revenue and improvement in trade as a result of price rises in non-traded commodities. Nevertheless, the rise in the price of goods transfers the exploitation of resources from the manufacturing sector to 45 3. The Curse of Tourisms other sectors in the economy. Meanwhile, the demand for domestic capital declines, creating pressure on the manufacturing sector, which causes de-industrialization and leads to Dutch Disease. Thus, this model indicates that demand-induced Dutch Disease is likely to lead to a decline of the capital stock which in turn may cause a loss in resident welfare in the long-run, as a result of the existence of externality that impedes diversification in other economic sectors. Also using a theoretical model, Nowak et al. (2004) investigated the impact of a tourism boom on structural adjustment, commodities, factor prices and welfare. Their analysis used a hybrid of the specific-factors Ricardo-Viner-Jones model (Jones, 1971) and the factor-endowment Heckscher-Ohlin model (Ethier, 1972; Jones, 1987) under the assumption of full employment. In this open economy, the terms of trade were given exogenously. Three sectors represented the economy in the model: a non-traded goods sector, an agricultural sector producing an exportable good, and a manufacturing sector producing an importable good. They found that a tourist boom may cause the immiserization of residents: that is, that they may be rendered poorer than before the tourism boom. Tourist consumption consists largely of non-traded goods and services. When a tourism boom occurs, there is first an immediate, local and favourable effect owing to increases in the relative price of such non-traded goods. However, in the longer term a negative effect is encountered owing to efficiency loss that occurs in the presence of increasing returns to scale in manufacturing. Whenever this negative effect outweighs the initially positive effect, immiserization is the result. In a different way, Nowak and Sahli (2007) examined the relation between Dutch Disease and coastal tourism in a small island economy, applying the general equilibrium model. They found that the boom of inbound tourism may cause a loss of welfare when tourism activities and products make intensive use of the coastal land. Holzner (2005) examined whether Dutch Disease has an impact on the tourism sector in more than 100 countries. The results indicated a negative correlation between both real exchange rate variability/distortion and economic growth. In any case, the relationship between tourism and real exchange rate distortion is negative. One explanation given is that countries drawing high incomes from tourism tend to be more outward oriented. Tourism might generate high levels 46 3. The Curse of Tourisms of final-goods imports, such as those to which tourists are accustomed in their countries of origin and for which they create a demand in the tourism host country. This effect would strengthen import lobbies and the advocates of trade liberalization. In a similar later study, Holzner (2010) examined the impact of Dutch Disease on tourism-dependent countries. From the analysis of data over the period 1970 to 2007 covering 134 countries, the results showed that, when controlling for initial output level, physical capital and human capital, the countries with higher shares of tourism income in GDP enjoyed faster growth than other countries. His findings indicated that tourism-dependent countries do not face real-exchange-rate distortion and deindustrialization but higher-than-average economic growth rates. Investment in physical capital, such as transport infrastructure, is complementary to investment in tourism—higher economic growth, higher levels of investment and secondary school enrolment are associated with countries deriving high income from tourism. Furthermore, tourism-dependent countries are accompanied by low real exchange rate levels. The study employed cross-country and panel data analyses using the share of travel services exports in GDP as a proxy for tourism capital. Taking an extreme case of a tourism-dependent economy, Mieiro et al. (2012) investigated the presence of Dutch Disease in Macau owing to gaming tourism. Since the 19th century, gaming tourism has played an important role in Macau’s economy, but the 2002 liberalization of gaming provided the catalyst for the current gaming tourism boom in that country. Within a framework that takes this into account, impacts of selected tourism growth-indicators were tested econometrically to reveal the presence of Dutch Disease in Macau. Although the classic structural imbalances underlying Dutch Disease have been identified, Dutch Disease would only represent a serious economic problem to Macau if the territory were to lose its privileged gaming position. The authors thus propose ways of taking preventative measures to remedy such a future scenario by applying revenues from gaming tourism to build up sustainable development in educational and health investment. 47 3. The Curse of Tourisms 3.3 The Curse of Tourism? 3.3.1 Data and Variables In the present study, we include 133 countries depending on the availability of data. The analysis covers the period from 1995 to 2007. All variables and data were obtained from WDI online (World Bank 2013). We follow Figini and Vici (2009) who considered a broad sample and two smaller sub-samples on the basis of tourism specialization defined as the share of international tourism receipts in the GDP. Accordingly, we apply the term “tourism countries” to those countries in which the revenue from tourism as a share of GDP is greater than the average (5.72 per cent) over the period 1995–2007, while the term “non-tourism countries” applies to those countries with the share of tourism revenue in GDP smaller than this. To analyse the effect of tourism on economic growth, we have studied certain variables (listed in Table 3.1) that are commonly accepted in the economic growth literature as being robust determinants of growth. The dependent variable is the growth of GDP per capita at constant prices, denoted as ‘growth’. Tourism receipts as a share of GDP are calculated by using international tourism receipts as a percentage of exports multiplied by the ratio of exports of goods and services to GDP. The general government final consumption expenditure (GCE) is calculated as a percentage of GDP. Education, denoted as ‘school’, measures the share of population in secondary education. We use this variable as a proxy for investment in human capital. Gross fixed capital formation as a share of GDP (here denoted as I) measures investment in physical capital. The variable ‘trade’, defined as exportsplus-imports as a share of GDP, is used as a proxy for the openness of the economy (Sequeira and Nunes 2008). 48 3. The Curse of Tourisms Table 3.1 Variables used in the present study Denotation Definition growth growth of GDP per capita at constant prices GCE general government final consumption expenditure is proxed to estimate the effect of government consumption on growth gross fixed capital formation as percentage of GDP is used to measure I the physical capital investment, le life expectancy at birth (total years) is used as a proxy of health POP annual population growth rate school school = percentage of relevant-age population enrolled in secondary school is used as a proxy for human capital tourism receipts as a percentage of GDP is calculated TRP using the international recipts of tourism exports and the ratio of exports of goods and services to GDP exports and imports of goods and services as a share of GDP is trade index to measure the impact of openness of the economy on growth performance, ttradep interaction variable (tourism as GDP share) × (trade) In addition we have divided countries into two groups—developed and developing—on the basis of the UNDESA WESP classification (UNDESA, 2014). This distinction has been made to find whether tourism specialization helps underdeveloped countries to grow or not. Moreover, we have further divided countries into two groups on the basis of international receipts of exports. The two groups are: tourism-exporting countries whose receipts from tourism as a percentage of exports are greater than 8.90 per cent (the median share across all countries), and nontourism-exporting countries for which this figure is less than 8.90 per cent (in addition, we also use the average share of tourism in exports, 14.14 per cent, as an alternative threshold). Finally, we created an interaction variable (denoted ttradep), obtained by multiplying the tourism share in GDP with trade in order to investigate 49 3. The Curse of Tourisms the possible presence of Dutch Disease in tourism countries and non-tourism countries. Table 3.2 displays the relevant descriptive statistics. Table 3.2 Descriptive Statistics Variable Obs Mean Std. Dev. Min Max GCE 1695 15.68838 5.778925 3.364233 39.19374 I 1670 21.78991 6.848701 3.480034 64.14175 POP 1724 1.336841 1.219802 –3.93064 10.04283 trade 1680 86.16208 49.27788 14.77247 456.6461 TRP 1647 5.726809 8.121327 0.018056 66.11868 growth 1592 2.900116 4.036521 –29.6301 33.03049 le 1688 67.82511 9.830611 31.23919 85.16341 school 1191 74.75916 31.5836 5.177891 161.6618 Table ‎3.3 Cross-correlation between variables, 1995-2007 growth trade POP I GCE TRP school le growth 1 trade 0.1076* 1 POP -.2523* I 0.2258* 0.2814* -.1403* GCE -.0823* 0.1632* -.2038* 0.0701* 1 TRP 0.0478 0.3834* -0.04 0.2960* 0.1679* 1 school 0.1193* 0.1790* -.6184* 0.1292* 0.3898* 0.0586 1 le 0.0285 -0.0166 0.3308* 1 -.1426* 0.0431 1 -.1456* 1 50 -0.0414 -0.1147 3. The Curse of Tourisms Table 3.3 reports the cross-correlation matrix of variables used in this study. For the consistency of the correlation matrix with the regression analysis. The correlation matrix shows that highest correlation within varaibles is investment then trade. However, coefficients between of the rest variables and economic growth are rather different in terms of magnitude and significance level. Despite the large variation among the correlations reported in table 3.3, all coefficients are low and thereforemulticollinearity is not an issue here. In addition, we provide the descriptive statistics in more details to show that standard deviation between and within (see appendix 9.1). However, we can notice from the table that covariates have relatively reasonable standard deviations, indicating that growth characteristics vary among over time and sample. Initially we applied fixed effect . However, fixed effect estimate is biased for data for which within-cluster variation is minimal or for variables that change slowly over time (Reyna,2007, p10). Therefore, in our analysis, we applied both random and fixed effects to check robustness of results. Afterwards, we applied hasuman statistic to test which model fit our data best. The hasuman test does not provide enough evidence to reject null hypothesis that fixed effect model fits the data better than random effect. However, we are aware of the fact that there is less variation in some variables such as life expectancy within countries or across time variables. Therefore, we follow the theory to use these variables as index to measure human development. and then , we used school secondary variable for robustness check. We did not find a big difference in all estimations except for estimation in exporting and non-exporting countries., when we include the life expectancy tourism specialization affect economic growth positively. In order to show the variability we calculated the standard devotion and coefficients of variations by country code. (see 10.1 appendix ) 51 3. The Curse of Tourisms 3.3.2 An Empirical Model of Economic Growth with Tourism The standard Solow model of growth assumes output to be the product of labor and capital, Y=Kα(AL)1-α, where 0 < α < 1, K stands for the stock of physical capital, L represents labor and A is a catch-all parameter reflecting technological progress, quality of institutions and any other factors that increase output for given stocks of labor and capital. Mankiw, Romer and Weil (1992) use this basic formulation of the Solow model to derive a growth regression that can be estimated: 𝑌 𝑙𝑛 = 𝑎 + 𝑔𝑡 + 𝐿 𝛼 1−𝛼 ln(𝑠) + 𝛼 1−𝛼 ln(𝛿 + 𝑛 + 𝑔) + 𝜀 Eq. (3.1) where s is the savings rate, n is the rate of population growth,  is the depreciation rate, g is the rate of technological progress, and  is the error term;  and g are not observed but their sum is proxied as 0.05. This growth regression can be further augmented to add additional factors of production: Mankiw et al. (1992) add human capital, and Li, Liu and Rebelo (1998) include also foreign direct investment. Many other conditioning variables have been proposed in the literature. The initial output per capita helps account for the fact that countries that are relatively poor tend to grow faster: it is easier to catch up than to lead. Government consumption can be included to account for the distortionary effects of taxation and the dead-weight loss of government spending (see Barro, 1991, and others). Openness to trade has been shown to make countries more productive, holding other determinants of growth constant (Sachs and Warner, 1995). 1 Given their nature, as factors of growth augmenting the productivity of labor and capital, most of these variables can be seen as falling within the term A in the above production function. In our analysis, we build on this literature and include three basic factors of production, physical and human capital and labor; two productivity-augmenting parameters, government consumption and openness to trade, and our variable of interest, the share of tourism revenue in output. Therefore, we estimate the following baseline regression: 𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 = β0 + β1TRPit + β2Schoolit+ β3tradeit + β4Iit + β5GCEit + β6POPit + β7leit + ui + εit Eq. (3.2) 1 For a broad overview of these attempts, see Levine and Renelt (1992), and Sala-i-Martin (1997), and the subsequent replications of their assessments. 52 3. The Curse of Tourisms growth = the growth of GDP per capita at constant prices, TRP = tourist receipts as a percentage of GDP, school = percentage of relevant-age population enrolled in secondary school, trade = total of exports and imports of goods and services as a share of GDP, I = gross fixed capital formation as percentage of GDP, GCE = general government final consumption expenditure, POP = annual population growth rate, le = life expectancy at birth (total years), ui = country-specific fixed effects, it = error term. However, we have applied panel data in our investigation to estimate the parameters corresponding to variables of interest from the data under consideration. Thus, the usefulness of panel data models gives an estimation of large sample properties and increases the degrees of freedom. Furthermore, the panel data allows the reduction of endogeneity as result of consideration of specific country effects, omitted variables, reverse casualty and measurement error Campos and Sequeira (2005). Following Fayissa, Nsiah and Tadasse (2007) we consider the following panel data model with N cross-sectional units and T time periods: 𝑌𝑖𝑡 = 𝑋𝑖𝑡 𝐵𝑘 + 𝑍𝑖𝑡 𝛿 + 𝑢𝑖 + 𝜀𝑖𝑡 Eq.(3.3) i = 1, 2, ...,N; t = 1, 2, ..., T Where Yit is the dependent value measuring the growth of GDP per capita in country i at year t in our study. While X it is a vector of observable regressors of the explanatory variables (Gross fixed capital formation, Growth of Population, General 53 3. The Curse of Tourisms government consumption, Trade, Education and Tourism receipts) for country i = 1, 2…, m and at time t= 1, 2, …,T.  is a vector of unknown coefficients on x, Z i is the vector of variables which do not depend on time and are different just over individual countries,  is the vector of coefficients on z, u i is the individual-level effect.  it is the disturbance factor. Under assumption the ui aren’t correlated with  it whatever the u i correlated or uncorrelated with the regressors in X it and Z it , Baum (2006). Therefore, the random effects models are shown when u i are uncorrelated with regressors. However, the fixed effects models are known if the u i are correlated with regressors. If we assume each cross sectional unit has its own intercept over time and the time specific effects are not present, the one-way effect model is termed and the last model is often called the Least Squares Dummy Variables Model, Fayissa, Nsiah and Tadasse (2007). However, the LSDV model is charged with problems since this model infers an infinite number of parameters. To understand the FE model well, it can be removed the panel data averages. In other way removes additive effects within group transformation (LSDV) from each side of (3.3).then the Equation (3.3) becomes as follows: yit  y i  ( xit  x i )  ( zi  zi )  (ui  u) i   it   i Eq.(3.4) Where yi  (1 / T )t 1 yit xi  (1 / T )t 1 xit T  i  (1 / T )t 1  it T , T While ui and zi . are panel data averages. Then we get the Equation (3.5):    yit  ( xit )    it 54 and 3. The Curse of Tourisms Eq.(3.5) The last equation presents the consistent estimator  by the OLS on within –  transformed data. Then we can call this estimator  FE . Moreover, the estimated coefficients of the FE models cannot be prejudiced because the fixed effects model controls all time-constant difference between individuals. On the other side, the FE models cannot be used to examine time- invariant causes of dependent variables(Reyna, 2007). On the other hand, the random effects model specifies when the individual effects are assumed to be random and uncorrelated with the independent variables and overall disturbance term. yit  xit   zi   (ui   it ) Eq.(3.6) Where (ui   it ) is a compound error term and u i are the individual effects. The RE models can include time invariant variables, and this is consider from the advantages of the RE model. But disadvantage of random effects is that we need to define the individual chrematistics which may or may not affect the predictor variables. The problem with this is that some variables may not be available, therefore leading to omitted variable bias in the model. ( Reyna, p26. 2009). To determine the validity of the model, we use the Hausman Specification Test which shows whether a random-effects or fixed-effects model is to be preferred. In other words, this test examines whether the ui effects are correlated with the regressors, since the null hypothesis is that they are not. The Hausman Test supports the fixed-effects estimates, as will be seen in the discussion of the empirical results. 3.3.3 Empirical Results and Discussion We first apply the fixed-effects model and the random-effects model to the broad sample. The results are reported in Table 3.3 below. Table 3.4 Fixed and random effects with two different measures of human capital 55 3. The Curse of Tourisms (fixed effects with (random effects (fixed effects (random effects school) with school) with le) with le) VARIABLES growth growth growth growth GCE –0.231*** –0.193*** –0.191*** –0.133*** (0.0585) (0.0337) (0.0458) (0.0269) 0.129*** 0.139*** 0.132*** 0.127*** (0.0290) (0.0233) (0.0242) (0.0198) –1.004*** –0.946*** –0.489*** –0.691*** (0.236) (0.171) (0.174) (0.120) –0.0735 –0.0265 0.0213 0.00362 (0.0761) (0.0257) (0.0633) (0.0221) 0.0654*** 0.0144*** 0.0552*** 0.0112*** (0.0104) (0.00455) (0.00845) (0.00341) 0.0717*** 0.00799 (0.0171) (0.00744) 0.224*** 0.0146 (0.0631) (0.0176) I POP TRP trade school le –5.507*** 2.600*** –16.37*** 1.155 (1.792) (0.907) (4.202) (1.388) Observations 1,018 1,018 1,455 1,455 R-squared 0.140 Constant of 131 Number 0.104 131 132 132 countrycode Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 The results in Table 3.3 illustrate that both the fixed-effects and random-effects models suggest tourism specialization has no significant effect on economic growth. This result differs from the findings of Sequeira and Nunes (2008) and Arezki et al. (2009), which showed a positive impact of tourism on economic growth. The other explanatory variables (GCE, POP, I and trade) in both models have highly-significant effects on economic growth. Government consumption and population growth seem to be negatively related to economic growth, whilst investment and trade are positively 56 3. The Curse of Tourisms related to economic growth: these findings are in line with the economic growth literature. We used the Hausman Specification Test to check between fixed-effects and random-effects models, as shown in Table 3.4. The Hausman Test rejects the null hypothesis in favour of the fixed-effects (FE) models at (p<0.05). Thus, the countrylevel individual effects are not correlated with the regressors. We therefore adopt the FE models for the next analysis, in which we estimate fixed effects between economic growth and other explanatory variables. Table 3.5 | Hausman Test with life expectancy Coefficients (b) fixed GCE –0.3227108 I .1323041 POP –1.1859507 trade .0552824 TRP le (b–B) (B) random sqrt(diag(V_b–V_B)) Difference –0.05688 0.036514 0.004321 0.013598 0.202132 0.125887 .0112166 0.044066 0.007705 .0213419 .0035112 0.017831 0.059048 .224322 .015698 0.208624 0.059438 .1279829 chi2(6) = (b–B)'[(V_b–V_B)^(–1)](b–B) = 81.51 Table 3.6 Prob>chi2 = 0.0000 Hausman Test with secondary school Coefficients (b) (B) (b–B) sqrt(diag(V_b–V_B)) fixed random Difference S.E. GCE –0.23084 –0.19333 –0.03751 0.047879 I 0.128933 0.139141 –0.01021 0.017354 POP –1.00429 –0.94621 –0.05809 0.162481 trade 0.065401 0.014389 0.051012 0.009334 TRP –0.0735 –0.02653 –0.04697 0.071678 secschool 0.071651 0.007993 0.063658 0.015347 chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) 62.53 Prob>chi2 = 57 0.0000 3. The Curse of Tourisms Tourism is a part of a country’s exports and our regressions include trade already. Therefore, we do a re-estimation of the regressions after removing tourism from trade: specifically, we subtract tourism as share of GDP from trade as share of GDP. The results are similar to those obtained previously. The effect of tourism remains insignificant. On the basis of this finding, we might conclude that tourism does not enhance economic growth. The full regression results are shown in Table 3.6 below. 58 3. The Curse of Tourisms Table 3.7 Effect of tourism on growth: Fixed-effects model after removing tourism from trade (3) (4) VARIABLES growth growth GCE –0.231** –0.191** (0.0940) (0.0783) 0.129** 0.132*** (0.0581) (0.0380) –1.004*** –0.489 (0.374) (0.332) –0.00810 0.0765 (0.113) (0.0946) I POP TRP school 0.0717*** (0.0244) 0.224** le (0.102) 0.0654*** 0.0552*** (0.0144) (0.0121) –5.507** –16.37** (2.781) (6.470) Observations 1,018 1,455 R-squared 0.140 0.104 Number of countrycode 131 132 Trade (net of tourism) Constant Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Next, we apply the fixed-effects estimator to the tourism-countries sample (tourism specialization >5.72). The findings are given in Table 3.7 below. We conclude from the results given in Table 3.7 that tourism again has no significant effect on economic growth even among countries that rely heavily on tourism. Tourism appears not to be an important factor for enhancing economic growth in this group. This contrasts with previous studies such as Chang et al. (2010). 59 3. The Curse of Tourisms Table 3.8 Effect of tourism on growth: Split samples depending on tourism share in GDP (T-countries) (Non-T c’s) (T-countries) (Non-T c’s) VARIABLES growth growth growth growth GCE –0.507*** –0.217** –0.268*** –0.244*** (0.124) (0.106) (0.0955) (0.0855) 0.180*** 0.0961 0.115*** 0.141** (0.0498) (0.0825) (0.0378) (0.0569) –1.462*** –0.592 –1.042*** –0.397 (0.462) (0.369) (0.286) (0.314) –0.0757 –0.247 –0.0222 –0.199 (0.125) (0.272) (0.0902) (0.244) –0.000152 0.112*** 0.0309 0.0780*** (0.0191) (0.0212) (0.0203) (0.0184) 0.0888*** 0.0501* (0.0309) (0.0257) –0.103 0.234* (0.160) (0.122) I POP TRP trade school le 3.539 –6.526** 10.01 –17.27** (4.136) (3.202) (11.30) (7.206) Observations 253 765 354 1,101 R-squared 0.212 0.150 0.105 0.113 Number of country code 43 107 47 111 Constant Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 [T = tourism; c’s = countries] Trade also has a non-significant relationship with economic growth in countries dependent on tourism. In addition, government consumption and population growth are highly significant and again affect growth negatively. These results might lead to the conclusion that tourism is not a good option for these countries and is not a factor fostering economic growth. Secondly, we look at the non-tourism-dependent countries (tourism specialization < 5.72) which form 75 per cent of the whole sample. The results with different proxies of human capital in 60 3. The Curse of Tourisms Table 3.7 indicate that again tourism appears statistically insignificant in enhancing economic growth. Consequently, we found that tourism is not associated with higher growth rates in countries that specialize in tourism. This result supports the findings by Sequeira and Campos (2005) and Figini and Vici (2009). Table 3.8 shows a comparison for developed and developing countries. We find that tourism does not affect growth in either group of countries. Table 3.9 Fixed-effects estimator in sub-samples of developing and developed countries (Developed c’s (Developing (Developed (Developing growth) c’s growth) c’s growth 3) c’s growth) –0.446*** –0.227** –0.250* –0.190** (0.157) (0.0976) (0.137) (0.0825) 0.101 0.119* 0.0971 0.132*** (0.105) (0.0635) (0.0818) (0.0425) –0.963 –0.965** –0.907* –0.328 (0.945) (0.404) (0.453) (0.375) –0.364 –0.0836 –0.0448 –0.00228 (0.334) (0.122) (0.109) (0.107) 0.0548*** 0.0624*** 0.0560*** 0.0614*** (0.0119) (0.0202) (0.0110) (0.0193) –0.00873 0.119*** (0.00951) (0.0311) –0.323*** 0.304*** (0.0790) (0.108) VARIABLES GCE I POP TRP trade school le 6.641** –7.239** 24.45*** –21.73*** (2.946) (2.838) (6.045) (6.718) Observations 247 771 332 1,123 R-squared 0.271 0.144 0.265 0.105 Number of countrycode 28 103 29 103 Constant Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 61 3. The Curse of Tourisms Finally, we turn to examine the existence of Dutch Disease. We introduce the term ttradep, which is an interaction term combining trade as a share of GDP multiplied by the tourism specialization coefficient. The results obtained after augmenting the regression with this new interaction term are given in Table 3.9 below. We notice that both tourism and trade now both have positive and significant impact on economic growth. Their interaction (ttradep) is significant and negatively related to growth. Hence, while tourism and trade each have a positive effect, the countries that rely heavily on both tend to experience lower growth. Table 3.10 Fixed effects in whole sample with tourism-trade interaction (ttradep) (1) (2) VARIABLES growth growth GCE –0.251*** –0.221*** (0.0942) (0.0768) 0.129** 0.126*** (0.0594) (0.0374) –1.045*** –0.508 (0.374) (0.323) 0.224 0.303** (0.158) (0.134) 0.0833*** 0.0716*** (0.0161) (0.0142) I POP TRP trade 0.0631*** school (0.0232) ttradep –0.201*** –0.209*** (0.0584) (0.0595) 0.213** le (0.103) –6.425** –16.70** (2.801) (6.571) Observations 1,018 1,455 R-squared 0.148 0.112 Constant 62 3. The Curse of Tourisms 131 Number of countrycode 132 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 This result thus suggests that Dutch Disease might appear in the broad sample in those countries where there is a focus on tourism essentially as a main factor of economic growth. Surprisingly, tourism specilaztion affects economic growth positively and significantly. This might mean that if the country has a good life expectancy , tourism specialization affects economic growth more than good education. In Table 3.11 below, we therefore estimate again separate regressions for the tourism-dominated and non-tourism-dominated countries. Tourism is the main component (if not the only one) of exports of the nontraded sector. Therefore, we first examine if Dutch Disease exists in tourismdominated countries. The estimated coefficient of tourism is positive but insignificant (although it is close to being significant at the 10 per cent level). The effect of trade on growth is again positive and significant. The interaction term (ttradep) has again a negative effect on economic growth which is negative when we control for human capital using schooling. This means that tourism may be causing Dutch Disease in the countries focusing on tourism. This result might be due to the dependence of these countries on the exports of non-traded services: tourism. The foreign-currency receipts for these services would tend to cause the real exchange rate to appreciate, thus making the traded goods produced in the agriculture and manufacturing sectors less competitive in international markets. This suggests that the effect of tourism is negative in economies that are highly dependent on both exports and tourism. In Table 3.10 below, we give in Columns 1 and 3 the results from again estimating separate regressions for the tourism-dominated and non-tourismdominated countries. The estimated coefficient of tourism is positive but insignificant (although it is close to being significant at the 10 per cent level). The effect of trade on growth is again positive and significant. The interaction term (ttradep) has a significant effect on economic growth and is again negative. This means that tourism may be causing Dutch Disease in the countries focusing on tourism. This result might be due to the dependence of these countries on the export 63 3. The Curse of Tourisms of non-traded services, i.e. tourism. The foreign-currency receipts for these services would tend to cause the real exchange rate to appreciate, thus making the traded goods produced in the agriculture and manufacturing sectors less competitive in international markets. This suggests that the effect of tourism is negative in economies that are highly dependent on exports in non-tourism-dominated countries; we find that tourism has an insignificant impact. Therefore, there is no relationship between tourism and economic growth. Moreover, the coefficient ttradep is insignificant. So, we can say there is no evidence of Dutch Disease in these countries. In addition, after clearing tourism GDP from trade in both sub-samples, the findings were as follows: firstly, tourism does have an effect on economic growth in tourism-exporting countries but does not have an effect in non-tourism-dominated countries. Secondly, the interaction term ttradep still has a significant effect on economic growth in tourism-dominated countries. This means tourism causes Dutch Disease in these countries but there is no evidence for this in non-tourism-dominated countries, as evidenced by the 64 results given in Table 3.11 3. The Curse of Tourism Table 3.3 VARIABLES Effect of tourism on growth depending on share of tourism in exports (tourism (non-tourism (tourism (non-tourism (tourism (non-tourism (tourism (non-tourism exporting exporting exporting exporting exporting exporting exporting exporting countries) countries) countries) countries) countries) countries) countries) countries) Threshold: 8.9% (median tourism/exports) GCE I POP TRP trade ttradep school Threshold: 14.14% (average tourism/exports) growth growth growth growth growth growth growth growth –0.379** –0.204 –0.166 –0.277*** –0.469*** –0.236** –0.160 –0.279*** (0.152) (0.126) (0.131) (0.102) (0.114) (0.107) (0.165) (0.0866) 0.195*** 0.0771 0.134*** 0.143* 0.199*** 0.0776 0.137*** 0.122** (0.0438) (0.114) (0.0322) (0.0722) (0.0487) (0.0843) (0.0322) (0.0542) –1.145*** –1.015 –0.411 –0.866 –1.080*** –0.976 –0.466 –0.639 (0.392) (1.091) (0.387) (0.587) (0.391) (0.857) (0.429) (0.442) 0.204 –0.164 0.335** –0.628 0.0936 0.378 0.302** 0.0260 (0.174) (0.592) (0.132) (0.679) (0.190) (0.409) (0.141) (0.394) 0.0621* 0.0970*** 0.0780*** 0.0678** 0.0229 0.0954*** 0.0686** 0.0723*** (0.0338) (0.0328) (0.0229) (0.0268) (0.0333) (0.0238) (0.0328) (0.0220) –0.193** –0.225 –0.227*** –0.00910 –0.108 –0.292 –0.203*** –0.123 (0.0876) (0.305) (0.0526) (0.279) (0.0773) (0.228) (0.0613) (0.210) 0.0353 0.0626* 0.0526 0.0576** (0.0268) (0.0364) (0.0321) (0.0285) 65 3. The Curse of Tourism le 0.0827 0.281** 0.0239 0.248** (0.120) (0.130) (0.163) (0.125) –1.663 –6.651 –9.734 –18.64** 1.985 –6.530* –5.299 –17.34** (3.881) (5.111) (8.622) (7.641) (4.018) (3.822) (12.30) (7.276) Observations 509 509 731 724 333 685 485 970 R-squared 0.174 0.136 0.101 0.135 0.174 0.142 0.072 0.126 Number of countrycode 83 78 86 86 57 99 60 103 Constant Robust standard errors in parentheses *** p<0.01, ** 66 p<0.05, * p<0.1 3. The Curse of Tourism 3.4 Concluding Remarks In this study, we investigated the relationship between tourism and economic growth using annual data for 131 countries covering the period 1995 to 2007, by means of panel-data techniques. The fixed-effects model results suggest that tourism specialization has no significant effects on economic growth. The same results were obtained when we split the sample into underdeveloped and developed countries, as also into tourism-dependent and non-tourism-dependent countries. After adding an interaction term combining tourism and trade, we find that Dutch Disease might be an issue in countries that have both high exposure to trade in general and to tourism in particular. We find the same pattern in the sub-sample of countries with above-average reliance on tourism but not in the sub-sample of countries that do not have more than an average degree of reliance on tourism. These findings might be due to the relative dependence of these countries on the exports of the non-traded sector (tourism) which, in the case of countries overly-dependent on tourism receipts, may contribute to the real appreciation of the exchange rate that thus undermines the competitiveness of the traded sector (typically agriculture and manufacturing). Therefore, those countries whose trade relies heavily on tourism might experience Dutch Disease. In other words, excessive dependence on tourism might not enhance economic growth. In addition, if the tourism-exporting countries were to continue to rely on tourism as the main export resource, this could well cause the decline of the traded sector in favour of the non-traded sector. 67 4. Tourism and its Determinates 4 Tourism and its Determinates 4.1 Introduction Tourism has become a crucial factor driving economic growth for a number of countries. Worldwide, tourism accounts for 6 percent of world exports and 30 percent of service exports (UNWTO, 2015a). Consequently it is important to understand the relevant determinants of tourism, both in general and as they are applicable to any particular country or distinct area that attracts, or aims to attract, tourists. Most authors focus primarily on demand factors such as the level of income, relative prices, and exchange rates—either of the host-country currency or of the currency normally used or preferred by tourists. Other factors can also play significant roles in attracting or repelling tourists; internal tensions (ethnic, economic or of another nature) as well as external conflicts, often have a negative impact on tourist arrivals (see Crouch, 1994 a,b; Eilat and Einav, 2004; Garín-Muñoz, 2009). However, only a few authors have so far investigated the effects that local governance exerts upon tourism. The way in which local authorities deal with residents and visitors, the levels of efficiency that are perceived in the services that they deliver, the range of services that local authorities provide, their response to constant or incidental needs affecting infrastructure and other aspects of a locality, the perceived attitudes of local officials, the presence or absence of corruption, the levels of openness and accountability, and many other circumstances—all these impinge on what is known as the institutional quality of governance that exists in any particular tourism destination. Institutional quality and governance have significant impacts on tourism—both in attracting and retaining tourists. Therefore, it is important to study the interaction between governance and tourism, since different (and sometimes conflicting) groups seek to secure their favoured policy decisions in any particular locality (Dredge and Jenkins 2007), and the consequential effects of these can have significant repercussions upon the attractiveness of a locality for tourists. 68 4. Tourism and its Determinates The United States Agency for International Development (USAID, 2002), has defined good governance as a “complex system of interaction among structures, traditions, functions, and processes characterized by values of accountability, transparency, and participation. Such effective governance usually entails a need for appropriate institutions, decision-making rules and established practices” (Fayissa and Nsiah, 2010 p.2). Effective governance is a key prerequisite for making tourism sustainable and for laying the economic, socio-cultural and environmental foundations of sustainable development (Mowforth and Munt 2009). Whilst governance may in practice be effective, the perceived values as evidenced by the policies and actions of local authorities can also have an enhancing or deleterious effect. The Internet has become a major platform for consumer use in comparative tourism decision-making (Alrashid, 2012). The variety of information which it provides has enhanced management and e-commerce operations in the tourism industry. This has been achieved through facilitating promotional advertising to consumers, offering a variety of tourism products and services, and presenting enhanced value to both providers and consumers irrespective of their cultural orientations, nationality, or geographical location (Alrashid, 2012). Estimates made from the available statistics indicate that a number in excess of 75 million travellers world-wide are successfully engaging the internet in the process of planning their tourism activities (Hvidt, 2011). In the search for information regarding tourism over the internet, the most frequently researched categories include details regarding planned destinations such as climate, security, travel and accommodation, and core tourist attraction features (Hvidt, 2013). Many of the previous studies have focused only on specific countries or groups of countries, such as those in Africa (Naudé, 2005), or Asia-Pacific (Enright and Newton, 2005). This chapter seeks to identify those determinants that have the greatest impact on tourism (in terms of tourist arrivals) in a sample of 131 countries and in sub-samples that include developed and developing countries (with this categorization determined according to the IMF classification). Sub-samples are also formed based on the population, as classified by the World Bank (WB) in 2012. The 69 4. Tourism and its Determinates determinants employed in the current research comprise economic, demographic, technological and political factors. In addition, we use governance indicators as a proxy for institutional quality (Rios-Morales et al., 2011), applying principal component analysis (PCA). The six indicators of governance comprise accountability of power, political stability, the rule of law, regulatory quality, corruption, and government effectiveness. Our sample covers a total of 131 countries over the period 1995 to 2007. We have chosen these countries on the basis of the data that are available, and we have applied static and dynamic panel-data methodology in our analysis. This chapter aims to make a contribution to the study of determinants of tourism by focusing on the quality of institutions and communications infrastructure. To the best of our knowledge, no-one has yet dealt with the issue as to whether hostcountry communications infrastructure and quality of institutions have any effect on the performance of the tourism sector. This chapter is organized as follows. Section 2 reviews the literature relating to previous studies concerning the determinants of tourist arrivals. Section 3 presents the data and variables used in the analysis. Section 4 describes the model specifications and the econometric methodology, whilst the conclusions are presented in Section 5. 4.2 Literature Review In this section, we discuss the previous literature dealing with the main economic and non-economic determinants of tourist arrivals that can affect tourist flows and ‘destination image’. This latter term refers to those attributes that make a specific location appealing as a potential destination to travellers. Potential visitors can be made aware of these through official publications in the public media, or by private circulation of information—and especially by means of ‘word of mouth’, with the social media becoming an increasingly important forum for this (Litvin et al., 2008). Whilst the circumstances of tourism destinations vary from place to place, the fact that the information given is accurate is more likely to exert a favourable impression on visitors (Batinić, 2013). The flexibility and ease with which web- 70 4. Tourism and its Determinates pages can be corrected and/or updated makes the internet potentially the best source of information for prospective tourists. The internet thus has an important influence on the tourism industry through the facilities it provides for marketing, information, online booking, thus significantly impacting on the competition occurring among tourist destinations (Luo et al., 2004; Buhalis et al., 2011). Furthermore, in an exploration of the historic relationships between online interactions and performance, it was found that the European destinations offering online services have shown stronger performance in terms of arrivals and tourism revenue that has increased in line with the numbers of prospective tourists using those services (Tourism Economics, 2013). Indeed, the communications infrastructure is becoming daily more important in the promotion of tourism destinations. The internet and social media are exerting an ever-increasing influence upon the choices and decisions of prospective tourists by making information easily available to them. The success of a particular tourism destination increasingly depends on how well it is marketed through the electronic media (Buhalis, 1998; Buhalis and Law, 2008; Romanazzi et al., 2011). Indeed, where a tourism location has perhaps acquired a poor reputation, if an attractive and user-friendly online portal is made available it has the potential to help revive the fortunes of the location (Romanazzi et al., 2011). Certainly, the internet and its associated social media are gaining in importance in the world of marketing as consumers take to internet forums to air their views (Bickart and Schindler, 2001; Hennig-Thurau et al., 2004). So whilst increasing numbers of prospective tourists use the internet for their research and planning purposes (Cai et al., 2004; ParraLópez et al., 2011; Fotis et al., 2012), many tourists are also posting feedback regarding their experiences (Gretzel and Yoo, 2008; Litvin et al., 2008; Xiang and Gretzel, 2010). Consumer feedback is therefore also growing in importance, especially as a considerable proportion of it tends to be negative (Shea et al., 2004; Sen and Lerman, 2007). Regarding tourism, the main positive image attributes include pleasantness of climate, inexpensiveness of goods and services, safety issues, and similarity (or otherwise) of local lifestyle of the place to be visited (Gearing et al., 1974; Ritchie 71 4. Tourism and its Determinates and Zins, 1978; Schmidt, 1979). Gearing et al. (1974) proposed destination-image measures consisting of eight factors: 1. accessibility, 2. attitude towards tourists, 3. infrastructure, 4. price levels, 5. shopping and commercial facilities, 6. natural beauty, 7. climate, and 8. cultural and social characteristics. Ritchie and Zins (1978) identified four ‘features’ of the cultural image of a destination: (a) aspects of daily life, (b) remnants of the past, (c) quality-of-life conditions, and (d) compatible work habits of the local population. In their study of convention tourism, Var and Quayson (1985) investigated the effect of host image on tourist arrivals and found two crucial factors: firstly, accessibility, or how close a convention venue is to the home base of a delegate; and, secondly, the attractiveness of the convention location. Some authors have investigated the role and significance of the local transportation system in helping to improve destination image. The transportation system has been defined as the interaction between transport modes and all the means that support tourist movements entering into and departing from destinations, and moving around within the destinations (Prideaux, 2000). Studies by Khadaroo and Seetanah (2007, 2008) have indicated that the condition of the transport capital stock of a destination contributes directly (either positively or negatively) to its attractiveness, and the importance of transportation facilities subsists in the contribution that they make in adding value to the services offered to tourists and the experiences that tourists receive. Getz (1993) applied the framework of destination image on bringing tourism to the old ‘downtown’ business districts of Niagara Falls in an area stretching across the border from Canada to the USA at one of the oldest and most frequented locations for border-crossing and tourism between the two countries. He found that to be attractive as a tourism business district, a place should have three aspects: core attraction, central business zones, and supporting services. Kim (1993) derived six features in terms of selection criteria for tourists: (a) cultural attractiveness; (b) clean climate; (c) quality of accommodation and relaxation programmes; (d) familyoriented amenities and safety; (e) accessibility and overall country reputation; and (f) entertainment and recreational opportunities. In the same fashion, Chen and Hsu (2000) found that travel costs, quality of restaurants, local lifestyle, no language 72 4. Tourism and its Determinates restrictions, and availability of interesting places affect the choice of Korean tourists. Russo and van der Borg (2002) found that more attention should be paid to transportation facilities, access to information and quality of local human capital in order to enhance location attractiveness in the four European cities that they examined. Most studies have focused on factors that—either separately or jointly— determine tourist arrivals. For instance, Naudé and Saayman (2005) analysed how sociological and economic indicators, together with openness and governance indicators, affect tourist flows to Africa. They found that the most important determinants of travel to Africa included political stability, tourism infrastructure, marketing and information, and the level of development in the destination. Dhariwal (2005) attempted to analyse certain determinants of international tourist arrivals in India using annual data from 1966 to 2000. The results indicated that socio-political factors such as communism, terrorism and Indo-Pakistan tensions, seriously threaten the Indian tourism industry. In addition, Cho (2010) studied the impact of non-economic factors on tourism demand in four different continents. He identified that people from different areas have different preferences when selecting their destination. For example, Europeans and Asians prefer to visit a destination for its cultural heritage, whilst Americans like to visit places where there are numerous social events available. Görmüş and Göçer (2010) attempted to investigate the socio-economic determinants of international tourism demand in Turkey. They concluded that distance between the sending countries and Turkey negatively affects tourism demand. Meanwhile, the real income, relative prices, real exchange and trade value between Turkey and the sending countries play positive roles. Similarly, Ibrahim (2011) examined the main determinants of tourist flows to Egypt and showed that tourism in Egypt is very sensitive to price. He also showed that the real exchange rate and trade also have a significant impact, being related positively with tourist flows to Egypt. Zhang et al. (2009) developed the travel demand model for Thailand by performing a multiple-regression analysis. They showed that the factors that best 73 4. Tourism and its Determinates explained and had the most effect on tourist flows to Thailand are the exchange rate, promotional budget, Asian financial crisis and SARS. Furthermore, Proença and Soukiazis (2005) used a combination of timeseries and cross-sectional data to estimate the demand function of tourism in Portugal, considering Spain, Germany, France and the UK as the basic tourists to Portugal. Their analysis showed that per-capita income is the most important determinant of tourism demand while accommodation capacity is a very important factor for tourism supply. Examining the importance of the tourism industry in Croatia, Škuflić and Štoković (2011) sought to assess the determinants of tourism demand by using the GLS regression method. Their study yielded the following results: (1) income is positively related to the demand for tourism; (2) an increase in the prices for accommodation tebds to decrease the demand for tourism products. Some authors have studied the main determinants of tourist arrivals separately. For example, Keum (2010) and Fry et al. (2010) studied the effect of openness and economic factors on tourist arrivals. They found that there is a positive relationship between tourist flow and trade. Khan et al. (2005) and Khan (2006) showed theoretically that tourism might encourage international trade through tourists’ purchase of food, souvenirs, transportation et cetera in a foreign country. Thus, tourism has the potential to encourage trade. Travel might also lead to increased international trade through business visitors starting up new ventures or government agents negotiating trade agreements (Khan, 2006). The converse also applies, international trade could encourage tourism; when trade exists between two countries, there is likely to be an increase in business travel between those countries (Khan, 2006). Some authors have sought to explain the relationship between tourism and trade empirically. Thus, Shan and Wilson (2001) found that there is a two-way causality that operates between international travel and trade in China. Similarly, Santana-Gallego et al. (2007) found a long-term relationship between tourist flows and trade when applying causality techniques between trade and tourism for the OECD countries and the UK. In another study, on the relationship between international trade and tourism in small island regions, SantanaGallego et al. (2011) found that their results suggest a long-term bidirectional 74 4. Tourism and its Determinates relationship between tourism and trade, while the short-run link lies mainly in the trade generated by tourist arrivals. Al-Qudair (2004) focused on investigating the relationship between tourist arrivals in Islamic countries and different measures of trade—namely imports, exports and total trade. A long-term relationship was found between the number of tourists and trade in the cases of Benin, Egypt, Jordan, Syria and Tunisia, while the causality test indicated a unidirectional relationship between tourist flows and imports in the cases of Egypt, Syria and Malaysia. In terms of major determinants affecting the number of tourists, the most common variables are income and price (Lim, 1997a, 1997b). In this regard, tourism is considered to be a normal good, so that, when people’s incomes increase, they are better able and more likely to travel abroad. Moreover, there is evidence that international travellers are sensitive to price (Crouch, 1992). When Edwards (1995) studied the cost competitiveness of selected countries in the Asia-Pacific region he found that an increase in relative cost can be shown to result in a fall in market share in travel from every originating country. A decrease in relative cost is linked to an increase in market share (Dwyer et al., 2000a, 2000b, 2000c). Regarding the economic environment, Han et al. (2006) found that price competitiveness is a very important factor that influences the decision-making of American tourists. With regard to social factors, Gearing et al. (1974) and Schmidt (1979) considered social determinants to be important factors in destination image. Phakdisoth and Kim (2007) and Vietze (2008) explored whether good governance has a positive effect on tourism receipts per capita. Moreover, Eilat and Einav (2004) considered whether the political risk associated with a destination plays an important role in destination choice, for both developed and less-developed countries. They found that political risk is very important for tourism for both high- and low-GNP destinations. Daryaei et al. (2012) further explored the impact on the level of tourism exerted by good governance together with GDP growth, technology growth, the inflation rate as an indicator of economic infrastructure, and the improvement of education. Good governance includes accountability of power, political stability, the rule of law, regulatory quality, levels of corruption, and government effectiveness. 75 4. Tourism and its Determinates The results indicate that in both groups of countries improvements in the governance indicators were accompanied by positive effects on tourism. Various studies have examined the impact of political events on tourists’ destination choice (Hall and O’Sullivan, 1996; Sönmez, 1998; Seddighi et al., 2001; Neumayer, 2004; Fielding and Shortland, 2010) since, as suggested by many policy makers, both safety and stability play an important role in attracting tourists. It is expected that an increase in political violence leads to a decrease in tourist arrivals, if not immediately then at least in the long run, even though certain localities appear to attract tourists because of conflict (Timothy et al., 2004). Locations affected by high levels of political violence tend to have only a few unique characteristics, and these locations can be easily replaced by peaceful holiday destinations that have similar characteristics. The results confirm the belief that political violence affects tourist movements to affected countries and they also suggest that policy makers should be concerned about the negative effects of political violence not only within their own country but also within the wider region in which their country is situated (Ioannides and Apostolopoulos, 1999; Hitchcock and Putra, 2005; Issa and Altinay, 2006). Additionally, a report by UNWTO (2013) shows how the 2008–2009 global economy crisis influenced the tourism sector. The crisis peaked in 2009, with a 12 percent decrease in international tourist arrivals. During 2009, international tourist arrivals decreased by 4 percent at global level, coupled with a 6 percent decline in tourism receipts. In the last quarter of 2009, international tourist arrivals recovered and their growth rate turned positive, as shown in Figure 4.1. 76 4. Tourism and its Determinates 12000000 10000000 8000000 Africa Europe 6000000 Asia America 4000000 Oceanian 2000000 0 1994 Figure 4.1 1996 1998 2000 2002 2004 2006 2008 2010 2012 International tourist arrivals by region, Jan 1995–Mar 2010 (%) Source: Author’s calculations 4.3 Methodology 4.3.1 Data Set The data used in this chapter were obtained from two main sources: the World Bank’s World Development Indicators database (WDI, 2012), and the Central Intelligence Agency (CIA, 2012). The data cover the period between 1995 and 2007 for 131 countries. An attempt has been made to update the tourism arrivals data, but for most of the countries only the 2010 data are available. The dependent variable is the annual tourist arrivals. 77 4. Tourism and its Determinates The explanatory variables are:  The growth of GDP per capita: we use the growth of GDP per capita in constant 2000 US dollars in the host country as our measure of economic growth, (see Ivanov and Webster, 2006). We expect a well-functioning economy to attract more tourists.  Trade: used as proxy of openness and to check whether tourist arrivals are related to the economic interactions between the destination and its partners (see Song et al., 2003; Ibrahim, 2011)  Net/pop: the number of Internet users in a country for per 100 persons is used as a proxy to capture the effects of communication infrastructure on tourist flows  Net/size: the number of Internet users in a country divided by the country’s area is used as a proxy to capture the effects of communication infrastructure on tourist flows.  Health: this variable is the percentage of health expenditure in GDP, and is used as a proxy for the health quality and public sanitary conditions (see Su and Lin, 2014).  PPP: relative price variable which is normally used in demand models of tourism, for its likely impact on tourist’s decision to travel or not. 78 4. Tourism and its Determinates Table 4.1 Explanatory Variables Variable Explanatory notes Growth Growth of GDP per capita constant 2000 USD (WDI, 2013) Trade The ratio of the sum of imports and exports to GDP (WDI, 2013) Internet users /100 The number of internet connections per 100 person (WDI, 2012) persons Internet users/area The number of internet connections divided by the area of country square (WDI, 2012) and size from CIA (2012) Heath Expenditure Total health expenditure is the sum of public and private health expenditure to GDP (WDI, 2012) PPP The relative price competitiveness of the destination measured by the ratio of GDP in PPP to GDP by market exchange rate in the destination countries. (Zhang and Jensen, 2007; WDI, 2012) To find out how the quality of governance and institutions can impacts tourists’ arrivals, we use the World Bank’s World Governance Indicators (2007)— control of corruption (COC), voice and accountability (voice), rule of law (LAW), effectiveness of governance (EOG), and political stability (PS) (Vietze, 2009). The explanations of the governance indicators are given in Appendix 4.1 at the end of this chapter, where each indicator of governance is given in units of standard normal disturbance. These range from approximately –2.5 to 2.5. A higher value corresponds to better governance (Kaufmann et al., 2002; Rios-Morales et al., 2011). Whilst using these indicators we found a strong significant correlation amongst them which can cause multi-colinearity problems. Therefore, we used Principal Component Analysis in order to counteract the strong correlation between these measures and to allow us to derive one or more summary measures (“principal components”) from a set of indicators as explained in Table 4.2 below. We can see from table 4.2 that Component (1) explains 78.32 percent of the total variance. In addition, the Eigenvalue of Component 1 is higher than 1. Thus, if we wish to opt for a lower-dimensioned solution, we should keep Component 1 and we would then retain 78.32 percent of the variance in the original variables. 79 4. Tourism and its Determinates Table 4.2 Component Eigenvalues Component Eigenvalue Difference Proportion Cumulative Comp1 4.69898 4.22242 0.7832 0.7832 Comp2 0.476568 0.081584 0.0794 0.8626 Comp3 0.394984 0.179716 0.0658 0.9284 Comp4 0.215268 0.075 0.0359 0.9643 Comp5 0.140268 0.066342 0.0234 0.9877 Comp6 0.073926 . 0.0123 1 Source: Author’s calculations applying PCA method The advice indicated by the scree plot (see Figure 4.2.3 below) would be also to pick Component #1 because the elbow in the curve occurs at Component #2. This would suggest that one component accounts for a disproportionately large amount of the combined variance. 0 1 2 3 4 5 Scree plot of eigenvalues after pca 1 2 3 4 5 6 Number Figure 4.2 Eigenvalues Accordingly, after applying the PCA statistical technique to the World Bank governance indicators, we chose the first component because it captures as much as possible of the original variance in these indicators. 80 4. Tourism and its Determinates there is a little variation in instaitiutaional quality and this could be because of the way that each of six aggregate WGI measures are constructed. WGI measures are made by averaging data from the underlying sources that correspond to the criteria of governance being considered. The six composite WGI measures are useful as a tool for broad cross-country comparisons and for evaluating broad trends over time. For a full description of the WGI methodology and interactive data access, to the aggregate and individual indicators, please visit www.govindicators.org (see appendix 9.2&10.3). Next, by using data from the Uppsala Conflict Data Project (Gleditsch et al., 2002), we examine whether armed violent conflict exerts an impact on tourism arrivals. The UCDP defines armed conflict as “a contested incompatibility that concerns government or territory or both, where the use of armed force between two parties results in at least 25 battle-related deaths. Of these two parties at least one is the government of a state” (Gleditsch et al., 2002:619). The intensity of the conflict variable was coded into three categories, based on the criteria given in the UCDP/PRIO Armed Conflict Dataset Codebook (Version 4–2013, page 9). 0 = no conflict 1 = minor: between 25 and 999 battle-related deaths in a given year 2 = war: at least 1,000 battle-related deaths in a given year. In addition, we divided our sample into two groups, and we classified the countries according to IMF criteria (we use the IMF classification to categorize countries as developed or developing, see Nielsen, 2011, Table 4) which are based on the levels of development as shown in Appendix 4.2. Moreover, we divided countries into samples according to the sample-median of population; Tables 4.3 and 4.4 present the descriptive statistics. 81 4. Tourism and its Determinates Table 4.3 Descriptive Statistics of Variables Variable Obs Mean Std. Dev. Min Max netsize 1699 82.32184 439.5999 0 7936.984 GDPPER 1760 8042.055 11187.79 108.9024 67138.52 TOA 1680 5004512 1.04E+07 11000 8.09E+07 POP 1767 4.25E+07 1.45E+08 61700 1.32E+09 netpop 1699 13.99372 20.16298 0 88.90034 ppp 1728 0.574099 0.278601 0.140434 1.860173 Trade 1733 87.27659 51.59567 14.77247 438.9016 conflict 1767 0.160159 0.443702 0 2 healthEx 1722 6.219257 2.188463 0.137624 16.1524 Note: Values for descriptive statistics are in levels Table 4.4 Descriptive Statistics of governance indicators Variable Obs Mean Std. Dev. Min Max ps 1196 –.0196303 .9445767 –3.05644 1.57687 voice 1192 .1106886 .9278396 –1.95119 1.82669 reg 1186 .1648797 .9166428 –2.52663 2.02558 low 1191 .0784545 .9687808 –2.31285 1.96404 coc 1171 .1393674 1.020891 –2.48921 2.46656 eog 1179 .1403643 .9723042 –2.39408 2.23691 Note: Values for descriptive statistics are in levels 82 4. Tourism and its Determinates The table 4.5 perform a pairwise correlation analysis in this study; the degree of correlation varies among variables. The highest relationship is between tourism arrivals and population . in addition, the positive and significant correlation between tourism arrivals and institutional quality Table 4.5 Cross-correlation between variables, 1995-2007 TOA POP ppp Trade conflict pca health netsize netpop TOA 1 POP 0.3004 1 ppp 0.3526 -0.0978 1 Trade -0.1339 -0.2039 0.1084 1 conflict -0.0219 0.1582 -0.2307 -0.2308 1 pca 0.3171 -0.1164 0.8092 0.2773 -0.3051 1 health 0.3584 -0.0999 0.541 -0.1138 -0.1784 0.4819 1 netsize 0.033 -0.0149 0.1207 0.3297 -0.063 0.1253 -0.0775 1 netpop 0.2768 -0.0765 0.7155 0.2456 -0.1893 0.7064 0.4648 0.206 1 GDPPER 0.3361 -0.0668 0.8601 0.216 -0.1687 0.7829 0.4691 0.2188 0.7376 83 4. Tourism and its Determinates 4.3.2 Model specification and econometric method The bulk of empirical research on tourist-arrivals modelling has focused on determinants of tourist-flows in separate countries or regions. In this study, the data from 131 countries are used jointly to find out which determinants have an effect upon tourist arrivals in a multi-country setting. In view of the challenges facing the tourism industry and the need to formulate policy advice for supporting the tourism sector, it seems more appropriate to identify the long-run determinants of tourist arrivals. We therefore use panel data approaches that give better estimates for longrun relationships (as explained by Kennedy, 2003, p 308). We have already discussed panel-data techniques in detail previously in this thesis. 4.3.2.1 Static Panel Data Pooled ordinary least square regression (OLS) is employed at first because it yields a better understanding of the preliminary sign of each determinant of tourism flows (Su and Lin, 2014). This model assumes the pooled residual to be the sum of country-specific unobserved variables and the error term to be normally distributed. However, by omitting the unobserved variables, which may be correlated with other explanatory variables, the pooled OLS estimation with heteroskedasticity will cause severe problems of bias and inconsistency. To solve this problem, a panel-data model with fixed effects or random effects can be used. The fixed-effects model assumes that each country has its own unobserved country-specific variables and estimates a separate constant term for each country, while the random-effects model assumes that unobserved country-specific variables follow a normal distribution, for which one overall constant term is estimated. We employed both models and used the Hausman test to determine which model performs better. The null hypothesis is that the random-effects model performs better than the fixed-effects model. The rejection of the null hypothesis means that the fixed-effects model is the one to be used. Accordingly, the estimated model of tourist arrivals takes the following equation 84 4. Tourism and its Determinates 𝑙𝑛 𝑇𝐴𝑖,𝑡 = β0 + β1 𝑙𝑛𝐻𝑒𝑎𝑙𝑡ℎ𝑖,𝑡 + β2 ln 𝑇𝑅𝐴𝐷𝐸𝑖,𝑡 + β3 ln 𝑁𝑋𝑖,𝑡 + β3 𝑐𝑜𝑛𝑓𝑙𝑖𝑐𝑡𝑖,𝑡 + β5 𝑔𝑟𝑜𝑤𝑡ℎ𝑔𝑑𝑝i,t + β6 𝑙𝑛 𝑃𝑝𝑝𝑖,𝑡 + β7 𝑙𝑛 𝑃𝐶𝐴𝑖,𝑡 + 𝛾𝑡 + 𝜇𝑖,𝑡 + 𝜀𝑖,𝑡 Eq. (4.1) Equation 4.1 shows that there is a relationship between the variables under study. However, we need to specify the functional form of the model practically, since there are several forms that can be used to determine the tourist flows. It is appropriate to mention here that we did not take the logarithm of the PCA indicator—which is a composite measure combining the World Bank indicators—as this PCA indicator can be a negative number. 4.3.2.2 Dynamic Panel Data In the previous section, we discussed the classical static panel-data techniques, namely OLS and fixed effects. These estimates are likely to be biased since the estimators ignore dynamic effects. The fixed-effect estimates might be affected by the biases caused by the explanatory variable X (endogeneity) and the correlation that might appear between the lagged dependent variable and the error term (𝜺𝒊𝒕 ). To deal with this issue it has been recommended that the generalized method of moments (GMM) estimation method should be used (Arellano and Bond, 1991; Arellano and Bover, 1995; Blundell and Bond, 1998/2000). We employ a dynamic panel model where the parameters are estimated using the generalized methods of moments (GMM) following (Greene, 2012). The GMM Model can be illustrated using the following equation. 𝑦𝜇 = 𝜕 𝑦𝑡−1 + 𝛽𝑥𝑡 + 𝜋𝑖 + 𝜀𝑖𝑡 Eq. (4.2) There are two forms of GMM. One is known as the Balestra–Nerlove (1966) estimator , where the instruments for the lagged dependent variable are the current and lagged values of the exogenous variables. The second form of GMM is known as the Arellano and Bond (1991) estimator, where all the estimates are taken by applying the dependent variable as instruments, lagged by two and three periods to make the finite-sample biases less when we use too many instruments. The two-step 85 4. Tourism and its Determinates GMM would only be used to assess the validity of the model assuming that the second-order serial correlation does not exist. Arellano and Bond (1991) proposed removing the individual effects through the first difference transformation as below: 𝑦𝑡 − 𝑦𝑡−1 = 𝜕 (𝑦𝑡−1 − 𝑦𝑡−2 ) + 𝛽 (𝑥𝑡 − 𝑥𝑡−1 ) + (𝜖𝑡 − 𝜀𝑡−1 ) Eq. (4.3) Even though this transformation removes the individual specific effects, the regression model is still biased for two reasons; the first reason is that the bias comes from the high correlation of ∆𝑦𝑖,𝑡−1 with ∆𝜀𝑖𝑡 , while the second reason is the possible existence of endogeneity in other explanatory variables. So by following (Greene 2012; Mohammad Tajik,et al 2015, let us consider ∆𝑣𝑖 as being a vector of errors for countries 𝑖 in the first-difference model: 𝜈𝑖3 − 𝜈𝑖2 ∆𝑦𝑖3 − 𝛼∆𝑦𝑖2 𝜈𝑖4 − 𝜈𝑖3 ∆𝑦 − 𝛼∆𝑦𝑖3 ] ] = [ 𝑖4 ∆𝑣𝑖 = [ ⋮ ⋮ 𝜈𝑖𝑇 − 𝜈𝑖𝑇 ∆𝑦𝑖𝑇 − 𝛼∆𝑦𝑖𝑇 Eq. (4.4) Then let us consider the 𝐴𝐼 as the matrix of instruments for variables 𝑖. 𝐴𝑖 = 𝑦𝑖1 0 𝑦𝑖1 [ 0 0 ⋯ 𝑦𝑖2 0 0 ⋱ 0 𝑦𝑖1 𝑦𝑖2 0 0 ⋯ Eq. (4.5) 𝑦𝑖𝑇−2 ] where the rows in Equation 4.5 are in correspondence with Equation 4.3. The next equation (4.6) presents the orthogonality restrictions which give an initial requirements in estimating GMM model. This instrument matrix corresponds to the following moment conditions 𝐸(𝐴′𝑖 ∆𝜈𝑖 ) = 0 Eq. (4.6) where Equation 4.6 can be split into the two following equations, 𝐸[𝑦𝑖,𝑡−𝑠 ∆𝜐𝑖 ] = 0, 𝐸[𝑥𝑖,𝑡−𝑠 ∆𝜐𝑖 ] = 0, 𝑡 = 3, … , 𝑇 𝑎𝑛𝑑 𝑠 >2; 𝑡 = 3, … , 𝑇 𝑎𝑛𝑑 𝑎𝑙𝑙 𝑠; 86 Eq. (4.6.1) Eq. (4.6.2) 4. Tourism and its Determinates Explanatory variables can easily be strictly exogenous or predetermined or endogenous (see Roodman, 2006). Therefore when the x is strictly exogenous then the instruments are 𝐴𝑖 = [ 𝑦𝑖1 , 𝑥𝑖1 , … . , 𝑥𝑖4 0 𝐴𝑖 = [ 𝑦𝑖1 , 𝑥𝑖1 , 𝑥𝑖2 0 0 ] 𝑦𝑖1 , 𝑦𝑖2 , 𝑥𝑖1 , … . , 𝑥𝑖4 Eq. (4.7) but in the case where x is predetermined 0 𝑦𝑖1 , 𝑦𝑖2 , 𝑥𝑖1, , 𝑥𝑖2 , 𝑥𝑖3 ] Eq. (4.8) and when x is endogenous 𝑍𝑖 = [ 𝑦𝑖1 , 𝑥𝑖1 0 0 ] 𝑦𝑖1 , 𝑦𝑖2 , 𝑥𝑖1 , 𝑥𝑖2 Eq. (4.9) Thus 𝑥𝑖𝑡 are endogenous, 𝐸(𝜈𝑖𝑡 |𝑥𝑖𝑡 ) ≠ 0 and 𝐸(𝜈𝑖,𝑡−1 |𝑥𝑖𝑡−1 ) ≠ 0 . So, 𝑥𝑖𝑠 (s=1,2,…,t-2) can be taken as a valid instrument, as 𝐸(𝜈𝑖𝑡 |𝑥𝑖,𝑡−2 ) = 0. In addition, when 𝑥𝑖𝑡 are predetermined or weakly exogenous, 𝑥𝑖𝑠 (s=1,2,…,t-1) can be applied as instrument, indicating that there is information from 𝜈𝑖,𝑡−1 to 𝑥𝑖𝑡 then 𝐸(𝜈𝑖,𝑡−1 |𝑥𝑖𝑡 ) ≠ 0 but 𝐸(𝜈𝑖,𝑡−1 |𝑥𝑖𝑡−1 ) = 0. Since it is difficult to find good instruments for variables, Arellano and Bond (1991) suggested the use of what is called an internal instrument, which is based on the lagged values of explanatory variables. Two assumptions are considered: the error term is not serially correlated and the explanatory variables are uncorrelated with future realization of the error term. The GMM estimator uses the moment conditions mentioned earlier to estimate the parameters in two steps with consistency and efficiency. The one-step estimator minimizes: 87 4. Tourism and its Determinates ` 𝑁 ` −1 𝐽𝑁 = (1/ 𝑁 ∑𝑁 𝑖=1 𝐴𝑖 ∆𝑉𝑖 )`𝑤𝑁 (1/ 𝑁 ∑𝑖=1 𝐴𝑖 ∆𝑉𝑖 ) Eq. (4.10) where 𝑊𝑁 is a weight matrix. The one-step GMM estimator uses the weight matrix, but the one-step estimator results are consistent and robust; standard errors and autocorrelation are easily derived. Arellano and Bover (1995), and Blundell and Bond (1998/2000) then introduced the two-step GMM estimator, where the error term is assumed to be independent and homoskedastic across countries and over time (first step). In the second step, the residuals obtained from the first step are used to construct good estimates of the variance-covariance matrix. This two-step GMM is called the difference GMM estimator. The two-step GMM gives more general conditions, for example heteroskedasticity. Therefore, the weight matrix is calculated as, 1 ′ ′ 𝑊𝑁 (𝛼̂1 ) = ∑𝑁 𝑖=1 𝑍𝑖 ∆𝜈̂ 𝑖 ∆𝜈̂ 𝑖 𝑍𝑖 Eq. (4.11) 𝑁 ∆𝜈̂ 𝑖 = ∆𝑦𝑖 − 𝛼̂1 ∆𝑦𝑖,−1 If the lagged dependent and explanatory variables are a random walk, their levels are considered weak instruments which might affect the asymptotic and smallsample performance of the difference estimator. The difference GMM estimator decreases the signal-to-noise ratio. Also, there is another assumption which needs to be accounted for, which is that there is no correlation between the differences of these variables and the country-specific effect. Furthermore, Blundell and Bond (1998/2000) found out that the firstdifferenced GMM estimators might perform poorly if instruments are weak. When instruments are weak they become less informative, and the first-differenced GMM estimators suffer from finite sample-size distortion problems. To find the solution to this problem Blundell and Bond (1998/2000) suggested a new framework known as the system GMM, to estimate dynamic panel-data models by adding moment conditions if the intimal conditions satisfy 𝐸(𝜂𝑖 Δ𝑦𝑖2 ) = 0 Eq. (4.12) 88 4. Tourism and its Determinates Therefore, the 𝑇 − 2 additional to the moment conditions for the model in the first difference is (𝑢𝑖𝑡 Δ𝑦𝑖𝑡−1 ) = 𝐸((𝜂𝑖 + 𝜈𝑖𝑡 )Δ𝑦𝑖𝑡−1 )= 𝐸((𝑦𝑖𝑡 − 𝛼𝑦𝑖𝑡−1 )Δ𝑦𝑖𝑡−1 ) = 0 Eq. (4.13) Let 𝑢𝑖𝑡 be the vector of errors for individual i in the first-differenced and levels equation: 𝜈𝑖3 − 𝜈𝑖2 Δ𝑦𝑖3 − Δ𝑦𝑖2 ⋮ ⋮ 𝜈 − 𝜈 Δ𝑦 − 𝛼Δ𝑦 𝑖𝑇 𝑖𝑇−1 𝑖𝑇 𝑖𝑇−1 𝑢𝑖+ = 𝜂 + 𝜈 = 𝑦𝑖3 − Δ𝑦𝑖2 𝑖 𝑖3 ⋮ ⋮ [ 𝜂𝑖 + 𝜈𝑖𝑇 ] [ 𝑦𝑖𝑇 − Δ𝑦𝑖𝑇−1 ] and the 𝑍𝑖𝑠 matrix of instruments will be 𝑦𝑖1 0 𝑍𝑖𝑠 = 0 0 [0 0 𝑦𝑖1 𝑦𝑖2 0 0 0 0 0 ⋱ 0 0 0 0 0 𝑦𝑖1 𝑦𝑖2 … 𝑦𝑖𝑇−2 0 0 ′ Eq. (4.14) 0 0 0 Δ𝑦𝑖2 0 𝐸(𝑍𝑖𝑠 𝑢𝑖+ ) = 0 0 0 0 0 0 0 0 0 ⋱ 0 Δ𝑦𝑖𝑇−1 ] Eq. (4.15) These are the System GMM estimator moment conditions, a total of moment conditions, used to estimate 𝛼 by (linear) GMM. Thus, the model of Blundell and Bond (1998/2000) is employed to obtain the dynamic panel model. Then the Windmeijer (2005) finite-sample correction is used for whole-sample correction to fix the standard errors of Blundell and Bond (1998/2000). Moreover, the consistency of GMM system depends on there being no second-orders serial correlation in the residuals (m2 statistics). Therefore, the dynamic panel model is 89 4. Tourism and its Determinates valid if the Estimator is consistent and the instruments are well defined. The presence of a good instrument variable leads to good GMM estimates. Two tests are suggested as well: the Arellano-Bond (1991) test of autocorrelation and the Sargan test of overidentifying restrictions (Sargan, 1958). GMM is good for large N and small T samples, so we applied the method to the whole sample with net/pop or net/size, since the potential of endogeneity in tourism phenomena cannot be captured by static panel models (Khadaroo and Seetanah, 2008) and there are persistence effects that have influence on tourists’ choices when they prepare for holidays. Khadaroo and Seetanah (2008) state that tourists will return to a particular destination if they previously enjoy their stay in that locality. However, the discussion given above has not received much, if any, attention by authors. Therefore, the dynamic framework will enrich the analysis and provide important aspects within the argument.Hence, the specific linear dynamic model hereby used for our estimation can be defined as, 𝑝 𝑞 𝑦𝑖𝑡 = 𝑎0 + ∑𝑘=1 𝑎0 𝑌𝑖𝑡−𝑘 + ∑𝑗=1 𝛽𝑋𝑖𝑡−𝑗 + 𝜂𝑖 + 𝜆𝑡 + 𝜐𝑖𝑡 Eq. (4.17) where i=1,…,n and t=1,…,T, and where 𝑦𝑖𝑡 is the total number of tourists, 𝑌𝑖𝑡−𝑘 represents the autoregressive structure to reflect habit/persistence in the tourist’s choice of destination, and 𝑋𝑖𝑡−𝑗 are the current and lagged values of the matrix of regressors that could be strictly exogenous, or predetermined, or endogenous. Additionally, to capture the effect of common disturbances, 𝜐𝑖𝑡 is the error term, 𝜂𝑖 represents individual effects, and 𝜆𝑡 represents time-specific effects. 4.4 Empirical Analysis and Results Applying the static and dynamic panel data, we investigated the effect of PCA and communication infrastructure (net/size and net/pop) on international tourist flows whilst controlling other possible explanatory variables. Table 4.6 below shows the results of the estimation. 90 4. Tourism and its Determinates Table 4.6 static and dynamic panel results , in whole sample VARIABLES (OlS) logTOA (FE) logTOA (RE) logTOA (gmm) system) logTOA L.logTOA dgdp pca lognetsize loghealth logPOP logTrade logppp conflict Constant Observations R–squared Number of countrycode 3.903 (5.682) 0.322*** (0.0239) 0.085*** (0.0123) 0.182** (0.0870) 0.726*** (0.0156) 0.852*** (0.0780) 0.636*** (0.116) –0.218*** (0.0786) –1.235** (0.568) 8.036*** (2.987) 0.157*** (0.0418) 0.0744*** (0.0232) 0.00470 (0.175) 0.962** (0.480) 0.604*** (0.134) 0.599*** (0.0916) –0.102** (0.0454) –3.638 (7.989) 7.042** (2.938) 0.205*** (0.0353) 0.0818*** (0.0148) 0.0250 (0.165) 0.654*** (0.0459) 0.658*** (0.125) 0.644*** (0.0866) –0.105** (0.0503) 1.061 (1.206) 0.743*** (0.0711) 10.77*** (2.724) 0.0702*** (0.0225) 0.0181** (0.00747) 0.0593 (0.0411) 0.185*** (0.0527) 0.224*** (0.0759) 0.226*** (0.0662) –0.0616 (0.0390) –0.288 (0.265) 1,045 0.739 1,045 0.551 1,045 1,034 129 129 129 129 Number of instruments 22 AR(1) test, p.value 0 AR(2) test, p.value 0.431 Hansen test 0.185 The dependent variable is tourism arrivals. All models were estimated using the dynamic twostep system GMM estimator proposed by Blundell and Bond (1998) with Windmeijer’s (2005) finite-sample correction. Robust standard errors in parentheses, significance: *** p<0.01, ** 91 4. Tourism and its Determinates p<0.05, * p<0.1. To determine the validity of the model, we use the Hausman Specification Test which shows whether a random-effects or fixed-effects model is to be preferred. In other words, this test examines whether the ui effects are correlated with the regressors, since the null hypothesis is that they are not. The Hausman Test supports the fixed-effects estimates, as shown in Appendix 4.3 at the end of this chapter. Moreover, the results from the Hansen test of over-identifying restrictions for the autocorrelation in the first-difference residuals show that that the instruments are valid for the regressions and that the null hypothesis of no serial correlation cannot be rejected. The results show a positive elasticity (10.77) of growth as expected in all estimations, indicating that the level of development of a country is the most significant of the variables having effect on tourism flows. For our main variables— institutional quality and communication infrastructure—the results of static and dynamic panel-data analysis show that both of these two variables have significant effect. However, the estimated fixed and GMM values of PCA elasticity, at 0.157 and 0.0702, are higher than health expenditure and internet users, which means that if the country exhibits good governance, it will attract more tourist arrivals. In summary, the empirical results strongly support the hypothesis that institutional quality and infrastructure play a large role in determining tourism flows. Table 4.7 below shows the estimations results of different types of panel data, taking into account internet availability as a ratio to the area of the country (internet connections/country size) in the whole sample. The estimations for Table 4.6 also make use of population-weighted tourist inflows, in contrast to the use of overall tourist arrivals that was employed for the estimations shown in Table 4.5. 92 4. Tourism and its Determinates Table 4.7 static and dynamic panel data with internet users per kilometre squared in the whole sample VARIABLES (OLS) logtoap (RE) logtoap (FE) logtoap (gmm logtoap 0.829*** (0.0588) 0.0124** (0.00594) 0.0427** (0.0171) –0.0490*** (0.0169) 0.146** (0.0620) 0.0576* (0.0321) 0.143** (0.0622) –0.0449 (0.0354) 10.62*** (2.663) –0.185 (0.208) L.logtoap lognetsize pca logPOP logTrade loghealth logppp conflict dgdp Constant Observations R–squared 0.0856*** (0.0123) 0.322*** (0.0239) –0.274*** (0.0156) 0.852*** (0.0780) 0.182** (0.0870) 0.636*** (0.116) –0.218*** (0.0786) 3.903 (5.682) –1.235** (0.568) 0.0818*** (0.0148) 0.205*** (0.0353) –0.346*** (0.0459) 0.658*** (0.125) 0.0250 (0.165) 0.644*** (0.0866) –0.105** (0.0503) 7.042** (2.938) 1.061 (1.206) 0.0744*** (0.0232) 0.157*** (0.0418) –0.0381 (0.480) 0.604*** (0.134) 0.00470 (0.175) 0.599*** (0.0916) –0.102** (0.0454) 8.036*** (2.987) –3.638 (7.989) 1,045 0.783 1,045 1,045 0.472 1,034 129 129 129 Number of countrycode Number of instruments 22 AR(1) test, p.value 0 AR(2) test, p.value 0.495 Hansen test 0.133 The dependent variable is tourism arrivals/pop. All models are estimated using the dynamic two-step system GMM estimator proposed by Blundell and Bond (1998) with Windmeijer’s (2005) finite sample correction. Robust standard errors in parentheses, significance: *** p<0.01, ** p<0.05, * p<0.1. 93 4. Tourism and its Determinates We used population-weighted tourism inflows as the dependent variable in the calculations and we did not find any difference from the results given in Appendix 4.4. Next, we replaced internet connections divided by sthe area of the country by the ratio of internet connections per 100 persons. This modification has little effect on the results, as seen in Tables 4.8 and 4.9 below. Table 4.8 Estimation with internet users per 100 people in whole sample VARIABLES (OLS) logTOA (FE) logTOA (RE) logTOA L.logTOA dgdp loghealth logPOP pca lognetpop logTrade logppp conflict Constant Observations R–squared (gmmsystem) logTOA 0.965 (5.679) 0.0645 (0.0801) 0.725*** (0.0151) 0.297*** (0.0240) 0.161*** (0.0158) 0.827*** (0.0677) 0.562*** (0.114) –0.209*** (0.0754) –1.076** (0.512) 8.036*** (2.987) 0.00470 (0.175) 1.036** (0.461) 0.157*** (0.0418) 0.0744*** (0.0232) 0.604*** (0.134) 0.599*** (0.0916) –0.102** (0.0454) –4.851 (7.667) 6.826** (2.917) 0.0193 (0.164) 0.659*** (0.0440) 0.210*** (0.0344) 0.0843*** (0.0153) 0.672*** (0.123) 0.653*** (0.0869) –0.106** (0.0511) 0.903 (1.167) 0.825*** (0.0670) 9.122*** (1.962) 0.0242 (0.0359) 0.124** (0.0500) 0.0462* (0.0238) 0.0260** (0.0113) 0.131** (0.0662) 0.110*** (0.0427) –0.0275 (0.0232) –0.0760 (0.238) 1,045 0.751 1,045 0.551 1,045 1,034 Number of countrycode 129 129 129 Number of instruments 32 AR(1) test, p.value 0 AR(2) test, p.value 0.495 94 4. Tourism and its Determinates Hansen test 0.1 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table 4.9 static and dynamic estimation results with net/pop in whole sample with dependent variable arrivals/pop VARIABLES (OLS) logtoap (RE) logtoap (FE) logtoap (GMM) logtoap L.logtoap lognetpop pca logPOP logTrade loghealth logppp conflict dgdp Constant Observations R-squared 0.161*** (0.0158) 0.297*** (0.0240) –0.275*** (0.0151) 0.827*** (0.0677) 0.0645 (0.0801) 0.562*** (0.114) –0.209*** (0.0754) 0.965 (5.679) –1.076** (0.512) 0.0744*** (0.0232) 0.157*** (0.0418) 0.0362 (0.461) 0.604*** (0.134) 0.00470 (0.175) 0.599*** (0.0916) –0.102** (0.0454) 8.036*** (2.987) –4.851 (7.667) 0.0843*** (0.0153) 0.210*** (0.0344) –0.341*** (0.0440) 0.672*** (0.123) 0.0193 (0.164) 0.653*** (0.0869) –0.106** (0.0511) 6.826** (2.917) 0.903 (1.167) 0.858*** (0.0645) 0.0214* (0.0113) 0.0346* (0.0205) –0.0422** (0.0197) 0.105* (0.0588) 0.0232 (0.0324) 0.0923* (0.0519) –0.0347 (0.0296) 8.777*** (2.304) –0.0381 (0.197) 1,045 0.793 1,045 0.472 1,045 1,034 Number of countrycode 129 129 129 Number of instruments 28 AR(1) test, p.value 0 AR(2) test, p.value 0.401 Hansen test 0.0988 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 95 4. Tourism and its Determinates Political violence has a negative impact on tourism flows. The results of the static panel estimations indicate that conflict causes damage to tourism by decreasing tourist arrivals. The effects of political violence could be different for developing countries and developed countries. To examine this question we split the whole sample into two groups according to the IMF classification. The results of the Hausman Test shown in Appendices 4.5 and 4.6 at the end of this chapter, and in Table 4.11below showed preferences for applying fixedeffects rather than random-effects regression, so we report only the former. Table 4.10 VARIABLES dgdp Loghealth logPOP logTrade pca lognetpop logppp conflict Fixed-effects regression: developed and developing countries (whole (developed sample) countries ) (developing) (developed) (developing) logTOA logTOA logTOA logTOA logTOA 8.036*** 6.270 7.520** 6.270 7.520** (2.987) (6.474) (3.201) (6.474) (3.201) 0.00470 –0.0908 0.0635 –0.0908 0.0635 (0.175) (0.210) (0.195) (0.210) (0.195) 1.036** –0.804 1.263** –0.889 1.192** (0.461) (0.581) (0.536) (0.590) (0.558) 0.604*** 0.838*** 0.579*** 0.838*** 0.579*** (0.134) (0.248) (0.148) (0.248) (0.148) 0.157*** 0.145*** 0.157*** 0.145*** 0.157*** (0.0418) (0.0510) (0.0516) (0.0510) (0.0516) 0.0744*** 0.0851*** 0.0706*** (0.0232) (0.0209) (0.0264) 0.599*** 0.545*** 0.632*** 0.545*** 0.632*** (0.0916) (0.106) (0.114) (0.106) (0.114) –0.102** –0.0586 –0.0945* –0.0586 –0.0945* (0.0454) (0.0793) (0.0486) (0.0793) (0.0486) 96 4. Tourism and its Determinates lognetsize 0.0851*** 0.0706*** (0.0209) (0.0264) –4.851 24.12** –8.628 25.48*** –7.469 (7.667) (8.997) (8.936) (9.173) (9.312) Observations 1,045 274 771 274 771 R–squared 0.551 0.611 0.555 0.611 0.555 Number of countrycode 129 32 97 32 97 Constant Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 We can see from Table 4.9 that our main variables governance and infrastructure (with internet per 100 people and internet users divided by pop-size) have a significant effect in both samples. The governance of a country is found to be important for the process of making destination choice, in respect of both developed and developing countries. From the foregoing tables we see that there is no appreciable difference, except that with regard to the conflict variable, we can see that military conflicts have a detrimental impact on tourism arrivals in developing countries whereas no significant effect is observed in developed countries. Regarding to results of estimation in subsamples, we applied fixed effects in favour of random effects according to husman test. Therefore, In Table 4.10 we split the sample according to the median value of population. We define small countries as those with population below a threshold of 6,530,755 during the period 1995 to 2007. Armstrong et al. (1998) adopted a threshold of 3 million inhabitants, while the Commonwealth Secretariat and World Bank (CS/WB, 2000) adopted one of 1.5 million. 97 4. Tourism and its Determinates Table 4.11 Fixed-effects regression: countries with large and small populations VARIABLES (1) LC logTOA (2) LC logTOA (3) SC logTOA (4) SC logTOA dgdp 11.36*** (4.047) 0.116 (0.144) 2.316** (0.894) 0.547*** (0.136) 0.0890* (0.0449) 0.0207 (0.0274) 0.446*** (0.121) –0.112** (0.0474) 11.36*** (4.047) 0.116 (0.144) 2.337*** (0.869) 0.547*** (0.136) 0.0890* (0.0449) 2.918 (4.150) –0.206 (0.292) 0.211 (0.481) 0.621*** (0.227) 0.176*** (0.0544) 0.130*** (0.0355) 0.675*** (0.149) –0.314*** (0.0830) 2.918 (4.150) –0.206 (0.292) 0.341 (0.455) 0.621*** (0.227) 0.176*** (0.0544) loghealth logPOP logTrade pca lognetsize logppp Constant –27.83* (15.51) 0.446*** (0.121) –0.112** (0.0474) 0.0207 (0.0274) –28.19* (15.06) Observations R–squared Number of countrycode 548 0.608 68 548 0.608 68 conflict lognetpop 8.043 (7.936) 0.675*** (0.149) –0.314*** (0.0830) 0.130*** (0.0355) 6.129 (7.534) 497 0.570 65 497 0.570 65 Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 LC = countries with large populations> median 6,530,755); SC = countries with small populations< median 6,530,755) 98 4. Tourism and its Determinates Table 4.12 Fixed effects of individual governance indicators VARIABLES (1) logTO (2) logTOA (3) logTOA (4) logTOA (5) logTOA (6) logTOA dgdp 8.944* (2.893) 0.880* (0.470) 0.621* (0.128) 0.617* (0.0908 0.0189 (0.163) 0.0740 (0.0217 0.129* (0.0395 – (0.0445 9.285*** (2.955) 0.819* (0.486) 0.625*** (0.131) 0.627*** (0.0928) 0.0299 (0.168) 0.0719*** (0.0219) 9.712*** (2.917) 0.891** (0.438) 0.614*** (0.130) 0.607*** (0.0904) 0.0559 (0.163) 0.0747*** (0.0211) 9.306*** (2.926) 0.805* (0.474) 0.627*** (0.132) 0.618*** (0.0932) 0.0164 (0.167) 0.0756*** (0.0220) 8.740*** (3.014) 0.786 (0.484) 0.616*** (0.137) 0.649*** (0.0926) 0.0357 (0.172) 0.0747*** (0.0235) 8.051*** (3.023) 0.948** (0.479) 0.611*** (0.134) 0.625*** (0.0906) 0.0110 (0.168) 0.0700*** (0.0227) –0.111*** (0.0416) 0.0867 (0.0853) –0.127*** (0.0416) –0.121*** (0.0409) –0.114*** (0.0406) –0.112*** (0.0414) logPOP logTrade logppp loghealth lognetsize ps conflict voice reg 0.178** (0.0896) law 0.112** (0.0498) coc 0.116** (0.0455) eog Constant –2.370 (7.774) –1.425 (8.067) –2.628 (7.201) –1.203 (7.878) –0.860 (8.070) 0.137** (0.0633) –3.408 (7.943) Observations R–squared NOCC 1,078 0.542 130 1,074 0.531 130 1,068 0.552 130 1,073 0.541 130 1,057 0.538 129 1,064 0.535 130 Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 NOCC = Number of countrycode 99 4. Tourism and its Determinates Examining the results displayed in Table 4.11 above, we find that the effect of institutional quality on tourist arrivals is positive in both groups of countries, with the size of the effect being twice as large (and more strongly significant) in small countries. This is similar to the effect of communication infrastructure, which appears insignificant in large countries but positive and significant in small countries. Appendices 4.7 and 4.8 at the end of this chapter show the results of the Hausman Test. Accordingly, we applied the fixed-effects estimation to countries with high population-levels and the random-effects estimation to countries having small population-levels. Finally, as a robustness check, we replaced the composite indicator of institutional quality (pca) with the individual governance indicators, to test how they influence tourist arrivals separately, the results being displayed in Table 4.11 above. As can be seen in Table 4.11 above, all the variables have the expected signs and most are statistically significant. All governance indicators variables are positive and significant except for voice. These results reveal that improved governance indicators generate positive effects on the impressions gained by tourists regarding the security and governance of a destination country. Thus, as far as institutional quality is concerned, the greater the number of countries oriented toward achieving and maintaining good institutional systems, the greater will be the fostering effect on global tourism performance. In other words, “good governance” is one of the most effective factors for improving and developing the global tourism sector. The establishing of good governance practices is well known to support governments to build a higher-visibility legal and institutional system that exhibits transparency in order to improve a country’s image and thus attract more people to visit the country. 100 4. Tourism and its Determinates 4.5 Conclusion This study has sought to analyse the impact of the governance and communication infrastructure on tourism flows. We estimated the impact of six governance indicators (citizen voices and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law or state of law, and combating corruption) on tourism arrivals, using static and dynamic panel-data techniques in a sample of 131 counties during the period 1995 to 2003. We used the dynamic Generalized Method of Moments (GMM) of Blundell and Blond (1998/2000) to conclude that the significance of lagged dependent variables sheds light on persistence of tourist flows over time. A country can receive large numbers of tourists year after year, even if these tourists are always different people. Our estimations clearly indicated that the effect of governance and internet connectivity is positive and highly significant for tourism flows. However, the question is obviously more complex. In addition, the positive relationships between communication infrastructure and tourist inflows in our estimated model appear to corroborate the idea that the increasing levels of networks (internet networks in the tourism industry) have generated beneficial effects for the industry. We defined small countries as countries with median population of less than a threshold of 6,530,755 during 1995 to 2007, in order to identify those determinants that are the most important in having an impact on tourism (in terms of arrivals) in sub-samples taken on the basis of the median population size of sample countries. We found that good governance and infrastructure are key determinants of tourism flows in small population-level countries. Accessibility of internet networks is probably a key factor, with higher levels of population generating higher demand and more competition to internet access, which might put higher-population countries at a disadvantage with regard to accessibility per capita of population. It is also interesting that institutional quality barely affects tourism flows in the largecountry sample. This might arise because governments in such countries are able to build up good institutional frameworks more effectively owing to a lower population growth-rate. 101 4. Tourism and its Determinates The sub-sample analysis made on the basis of the level of development current in a destination country highlights the fact that the governance of the host country is shown to be important for the process of destination choice, for both developed and developing countries. The positive relationship between the information technology variable and tourist flows shows that an increase in technological endowment tends to promote the growth of the tourism industry. However, some interesting differences arise between countries with regard to conflict. In developing countries particularly, violent events have a more profound effect on tourism arrivals than is the case for developed countries. Violent conflict is well known as acting to the detriment of economic growth in less-developed countries in the short-term at least (Murdoch and Sandler, 2002), and its negative impact on tourism can harm the economy as whole. An explanation for the fact that for developed countries the results tend to indicate negative and insignificant impacts of conflict can be found in relation to the very few military conflicts that the developed countries experience. Even when such events happen in developed countries, they are very often of a territorial nature and thus limited in spatial extent. Furthermore, toursist have tended to have greater confidence in the ability of developed countries to deal effectively with such problems. The general level of development which is used as a proxy for technology in the present study is found to be the main universal factor behind explaining comparative advantage within tourism. In other words, “good governance” is one of the effective factors leading to improving and increasing tourism flows. Our results show that the technology proxy is an essential and comprehensive element for explaining the comparative advantage in the tourism industry. In addition, good quality of institutions is another of the most important factors that enhance tourism arrivals. 102 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5 “International Tourism and Institutional Quality: Evidence from Gravity Model” 5.1 Introduction Various scholars (Uysal and Crompton, 1984; Dwyer and Kim, 2003; Eilat and Einav, 2004; Naudé & Saayman, 2005; Song and Li, 2008; Culiuc, 2014) have conducted a series of studies analysing the determinants of tourist flows. In these studies, various techniques have been applied, including time-series data and panel data, as well as the gravity model (Prideaux, 2005). The gravity model concept was initially put forward by Tinbergen (1962) to explore flows of trade. The model was further developed and applied by Pöyhönen (1963). In its simplest form, it explains the trade flow between two countries by relating it to the economic mass of the two countries (using GDP as an indicator) and the distance between them. While the model was initially introduced as an empirical application, Anderson (1979) subsequently put forward a theoretical framework that supports this model. Since tourism constitutes trade in services, authors began to use the gravity approach to analysing the movements of international travellers and tourists soon after the model first emerged (Heanue and Pyers, 1966; Pyers, 1966; Quandt and Baumol, 1966; Wilson, 1967; Quandt and Young, 1969; Gordon and Edwards, 1973; Malamud, 1973; Durden and Silberman, 1975; Kau and Sirmans, 1977; Kliman, 1981). Using approaches based on the gravity model, efforts have been made to explore and identify the determinants of tourism arrivals. As mentioned, GDP was the indicator originally employed in gravity models for measuring the economic mass of the areas, countries or regions under consideration. However, some authors, such as Taplin and Qiu (1997), have used population instead of GDP as the basic indicator of a country’s “mass”. A large volume of studies has been published investigating the most appropriate econometric specification models for tourism (Um and Crompton, 1990; Witt and Witt, 1995; Wong, 1997a, 1997b; Eilat and Einav, 2004; Wong et al., 2006; Song et al., 2009; Massidda and Etzo, 2012; Etzao et al., 2013). Although there had been a tendency to neglect the gravity model in the more recent literature, it is 103 5. International Tourism and Institutional Quality :Evidence from Gravity Model coming back into use for modelling tourism demand particularly in the circumstances where there is a need to include and evaluate the role of structural factors (Morley et al., 2014). The gravity model is often operated on the basis of cross-sectional data, although this approach limits the analysis to a single time-period. Other researchers have used panel data instead of cross-sectional data to pass over this limitation of the model (Song, 2008). The application of panel-data estimation can help control for heterogeneity amongst countries. This makes it possible to employ fixed-effects or random-effects estimation methods, and also to apply classical estimation methods such as the traditional OLS (Mátyás 1997, 1998). This chapter considers a panel-data set comprising 134 countries of origin and 31 destination countries (selected depending on data availability). We estimate the gravity equation using three techniques: OLS, Hausman-Taylor, and Poisson estimation techniques. We compare the performance of the three approaches in relation to the gravity-equation theory. This chapter seeks to address the following questions:  Do economic factors play a role in determining tourist flows?  Do geographic factors in the origin and destination have an effect on tourism arrivals?  Finally, the central question for this chapter is: How does institutional quality impact on tourism flows? In order to answer the foregoing questions, this study employs the Gravity Model. At the outset it must be noted that, although many studies have used gravity equations as an instrument for the empirical modelling of tourism demand, the theoretical background to gravity modelling is still deficient in some aspects. Tourism-flows are movements of humans and not of merchandise. Tourists act as consumers when travelling for recreational purposes. Consumer demand and tastes often change suddenly. Humans as consumers often show a degree of randomness in behaviour that they might not normally display in other situations, particularly those related to work or business. Studies such as those by Turner and Witt (2001) and 104 5. International Tourism and Institutional Quality :Evidence from Gravity Model Cohen et al. (2014) have described the difficulties imposed on the forecasting powers of models by unpredictable changes in consumer demand. Consequently, between tourism and trade there are bound to be considerable and highly noticeable differences that will be encountered in the mechanisms and patterns of international flows, simply because these flows involve two very disparate classes of items or entities. In addition, there is no theoretical justification for incorporating policy instruments such as tourist taxes or promotional expenditures within the tourism gravity equation. Consequently, drawing inferences about the effects of tourism policies carries no guarantee of validity for the outcomes of such calculations. The negative consequences that higher political risk poses for the tourism industry are highly important. To the best of our knowledge, there are no studies that investigate the effect of the quality of institutions on tourism flows. Accordingly this study has undertaken to examine the various effects of political instability (such as acts of terrorism, conflict, other forms of violence, and so forth) that exert negative effects on tourism. To this effect, we use data from the International Country Risk Guide (ICRG, 2012) to account for institutional quality and political risk in the countries of origin and destination alike, and to measure the effect of institutions on tourist flows. This chapter is organized as follows. Section 2 reviews the literature relating to the determinants of tourist arrivals in the Gravity Model in general, as well as regarding the importance of political factors in particular. Section 3 presents the data and variables. Section 4 describes the model specifications, the econometric methodology and the results, whilst the conclusions are presented in Section 5. 105 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5.2 Literature Review The gravity model for international trade is studied by many authors (Anderson and wincoop, 2003). These researchers developed the framework of gravity to measure the bilateral trade and introduced the theoretical background of gravity equation in trade studies. Although there is a significant number of theoretical studies that support gravity trade model, there is a lack of studies supporting gravity tourism model. In addition, tourism bilateral data is not available as trade data. Therefore, it’s no surprise that gravity model for tourism was neglected in the literature. Keum (2010) identifies that gravity equation is valid to state the tourism arrivals by explaining different patterns of international tourism. Consequently, in this study we tried to propose some theoretical background of tourism gravity supported by some empirical evidence with the data in hand because there is few authors have been interested mainly in applying gravity models to answer questions concerning politics, institutions and financial flow nexuses. Papaioannou (2009) reported that institutions exhibiting poor performance (for example, legal inefficiency) can act as a barrier to foreign bank capital flows. In addition, he suggested that the quality of institutions might be a key consideration in the process of bank lending rather than income or human capital. Likewise, Bénassy-Quéré et al. (2007) used the gravity model approach when they focused on the role that quality of institutions plays in the process of foreign direct investment (FDI) allocation in source and recipient countries. The results of their study suggested that higher levels of good institutional quality were correlated with increased levels of FDI in recipient countries. However, this correlation was not apparent in respect of source countries in general. On the other hand, the gravity-model literature has emphasized the importance of institutional quality and political risk on trade. For instance, trade is significantly influenced by democracy (Milner and Kubota, 2005). According to Yu (2010) democratization affects trade in multiple ways. In particular, they concluded that a highly democratic country is likely to be an optimal actor in international trade owing to the likelihood of its exports being of a higher quality, and also because of the higher level of trust that trading partners are likely to give to such countries. Overall, 106 5. International Tourism and Institutional Quality :Evidence from Gravity Model they found robust evidence that hoigher levels of democratization are significantly reflected in increased levels of trade. In their study exploring the question of whether good institutions foster trade, Duc et al. (2008) adopted a gravity model that incorporated a Poisson Pseudo-Maximum Likelihood (PPML) estimator. Their results suggested that trade between open and democratic countries will in general (but not necessarily) tend to increase. Moreover, Moser et al. (2006) highlighted political risk as a “robust determinant” that impacts negatively on the flow of exports and international trade, and should be incorporated in empirical models of trade. A recent study by Mehchy et al. (2013) using gravity analysis examined the determinants of Syrian exports between 1995 and 2010. Their estimation indicated the importance of market size (measured by GDP) and population in attracting Syrian exports, whilst destination distance and the decline in Syrian institutional quality exert negative effects on Syrian exports. They listed the cultural similarities and trade agreements with Arab countries, with Turkey and with Europe that have previously played an effective role in determining Syrian exports. In addition, they clarified that changes in the nominal effective exchange rate did not affect Syrian exports significantly during the period 1995–2010. For the main conclusion of their study, they highlighted the decline in Syrian institutional performance as posing a grave threat to the Syrian export business and the national economy. Many researchers have chosen to study tourism flows using the gravity-model approach. For example, Prideaux et al. (2003) explored the limitations of forecasting models in crisis situations. Prideaux (2005), combined a review of the existing literature with an analysis of tourist-flow data using gravity-model techniques to examine the structure of bilateral tourism and identified multiple categories of factors that may affect the overall size of tourism flows (see Table 1 in Prideaux, 2005). Archibald et al. (2008) employed a dynamic tourism gravity model to measure the competitiveness of Caribbean tourism markets. Khadaroo and Seetanah (2008) used a gravity model to investigate the role of transport infrastructure in attracting tourists. Keum (2010) explored the gravity equation to assess how well it can explain tourism flows, and he undertook a general survey and exposition of the patterns of international tourism flows. Zhang et al. (2015) used the gravity-model approach to investigate the impacts of cultural values on tourism. The empirical evidence gathered 107 5. International Tourism and Institutional Quality :Evidence from Gravity Model by these and other studies supports the basic validity and usefulness of the Gravity Model for describing the flow of tourism as well as trade. Studies such as these and the paper by Morley et al. (2014) demonstrate that, within certain defined limits, the applicability and robustness of the gravity-model approach are well established. Lavallée (2005) applied a gravity model to assess the impact of the quality of governance in developing countries among 21 OECD countries. Lavallée’s results show that if a developing country has good governance policy, this will help it to import goods from industrialized countries. Corruption has been defined that is “an act in which the power of public office is used for personal gain in a manner that contravenes the rules of the game” (Aidt, 2003:F632, citing Jain, 2001). It has been argued that corruption tends to adversely affect the health of an economy (Méon and Sekkat, 2005). According to Poprawe (2015) corruption has a negative effect on tourism. However, the effect of corruption on tourism may be twofold (Dutt and Traça, 2007). Evidence shows that corruption may facilitate business activity, thus increasing the speed or ‘velocity’ of money and hence the rapidity of transacting business. In this respect, corruption may sometimes have positive side-effects for tourists, who may make arrangements that might not have been forthcoming except through the payment of bribes or generous tips. However, where such payments become expected, non-payment can have the effect of causing problems for the unwitting touridst. In view of this, it is relevant to ask whether assessments of the quality of judicial and governance frameworks could be developed as indicators to be applied to the evaluation of the state of democracy and corruption in a locality. The further question arises as to whether such indicators are sufficiently robust with respect to newer developments. The impact of institutional behaviour and quality on trade certainly needs to be addressed (Dearden, 2000; Duc et al., 2008). Furthermore, whilst tourism represents a vital contribution to economic development in many developing countries (Sinclair, 1998), however the developing countries have tended to represent the main locations of violence, often owing to conflicts over natural resources—access, ownership, and/or exploitation (Le Billon, 2001; Gleditsch et al., 2002; Piazza, 2006). Some writers have sought to study the effect of violence on tourism, since tourists are sensitive to the negative images that 108 5. International Tourism and Institutional Quality :Evidence from Gravity Model might be projected regarding any particular tourist destination (Neumayer, 2004). In fact, events of violence often have an impact on tourism not only contemporaneously, i.e. with immediate effect, but also with lagged, delayed effects. For example, the analysis by Enders et al. (1992) of the impact of terrorism on tourism in Spain and other western countries suggested that three to nine months could often pass before tourist arrivals decreased drastically. Countries whose image has suffered owing to violence often attempt by aggressive advertising campaigns to represent themselves as destinations that are wholly safe and secure for tourists (Sönmez et al., 1999), although these attempts may not be as effective as desired (Ahmed and Abdul-Kadir, 2013). The negative consequences of violence for the tourism industry are grave and highly important. However, the study-response has not been commensurate, as can be seen from the relative paucity of studies dealing with the impact of political violence on tourism. Accordingly, in this present chapter we have undertaken to examine the various diseases of political instability (such as acts of terrorism, revolution, armed conflict, other forms of violence, as well as the violation of personal integrity and rights) that have negative effects on tourism. More recently, Holder (2012) ran a model investigating the Arab Spring process. He found that in different countries, most of the outcomes depended on two factors: (a) the wealth of the dictator and (b) the provenance of the regime (either from a minority group, or from the majority). In addition, Cothran and Cothran (1998) have argued that political stability is crucial for developing Mexican tourism, even though certain tourists were attracted to Chiapas State to see the effects of the Zapatista uprising (Duffy, 2002). Archibald et al. (2008) showed that the importance of political stability as an indicator varies with regard to different destinations, with political stability as a consideration being more correlated to international tourists who travel from America and Europe. Communications infrastructure and level of development are important factors for tourist destinations in all continents. Neumayer (2010) adopted a gravity approach to examine the influence of visa restrictions on tourists. His finding showed that such restrictions reduced the numbers of bilateral travellers by between 52 and 63 percent on average. The study by Lien et al. (2014) estimated the effects of Confucius Institutes on inbound travel to China, processing panel data in a gravity model and using the Poisson Pseudo-Maximum Likelihood (PPML) estimator. The authors 109 5. International Tourism and Institutional Quality :Evidence from Gravity Model established the usefulness of these institutes in boosting tourism inflows in general, and particularly the inflows of business and worker tourists. In two different studies Fourie and Santana-Gallego (2011, 2013) used standard gravity models to investigate tourism flows. In the 2011 study they explored the impact of mega-events (cultural, sporting) on normal tourist inflows into the host-country/region. In the 2013 study they compared the determinants that drive tourism arrivals inbound to Africa from outside and between African countries, using a standard gravity model of 175 origin/destination countries. In the latter study they found that the factors affecting African-inbound and African-internal tourism are quite similar to factors affecting global tourist flows, such as income, distance and land area. Gil-Pareja et al. (2007a, 2007b) reported that common language, as well as the presence of embassies and consulates, are important factors attracting tourist arrivals from the G7 countries. In his study of the role of visas in determining cross-border travel, Tekleselassie (2014) found that GDP, population size, contiguity, common language, and previous colonial relationship also have a significant positive impact on cross-border travel. In addition, he found that geographical covariates such as distance and destination area negatively correlate with cross-border travel. Karemera et al. (2000) used a gravity-model approach to demonstrate how the population of source-country and the income-level of recipient country are the main factors that determine migration to North America. The high-population countries of Asia and Latin America provided the great bulk of migrants, whilst domestic restrictions on political and civil freedoms in source countries restrict migration from these countries to North America. Using an augmented gravity model that incorporated several measures for terrorism and similar violence, Nitsch and Schumacher (2004) investigated their effects on international trade. They identified terrorism as exogenous in their study and found that terrorism reduces trade, and a double increase would depress international trade by 4 percent. Similarly, Fratianni and Kang (2006) pointed out that terrorism tends to reduce trade flows by increasing trading costs and causing borders to become more rigid, particularly for countries that share a common land border. 110 5. International Tourism and Institutional Quality :Evidence from Gravity Model Yap and Saha (2013) investigated the negative effects exerted on tourism by political instability, terrorism and corruption. Their analysis of panel data for 139 countries led them to find that political instability and terrorism both exert a negative effect on tourism arrivals, even within UNESCO heritage areas. However, they found that terrorism had less of an impact than political instability and corruption also has a negative effect. Thompson (2011) investigated how the effects of terrorism upon tourism differ in developed and in developing countries. His analysis indicated that the impact of terrorism on tourism in greater in developing countries than it is in more developed ones. He suggested that the difference might be explained by the cushioning effects of welfare resources and a greater diversity in the economy enjoyed by developed countries, which have greater resources to invest in the tourism market. Drakos and Kutan (2003) examined regional effects of terrorism on tourism in Greece, Turkey and Israel—countries having high tourism potential and trade. They analysed the various elements of the resultant effects of terrorist incidents, in order to identify similarities and dissimilarities in the impacts that terrorist actions exert upon tourism in the different countries. They documented the ways in which each country’s share of the tourism market fluctuated in response to terrorist incidents. In this way they were able to map out the ‘contagion effects’ and the trends in how the patterns of tourist arrivals might shift from one country to another. A clear insight into the mechanism that affects the tourism flows between two countries is most valuable for identifying inefficiencies and obstacles, the need for remedial actions, as well as potential development areas. Additionally it is useful to identify the elements causing unequal bilateral flows by investigating areas such as GDP level, size of population, and issues arising from destination competitiveness. More detailed research should be conducted by examining the suggested gravity framework for particular bilateral pairs in order to recognize the deficiencies and marketing potentials between such countries. Further analysis in this area could be performed for multinational tourism frameworks. 5.3 Data This study uses tourist arrivals data from the UNWTO (2015b) dataset as the dependent variable. Following UNDESA (2008), the UNWTO defines ‘tourist’ as an 111 5. International Tourism and Institutional Quality :Evidence from Gravity Model overnight visitor, whereas ‘visitor’ refers to a broader concept, which includes both tourists and same-day visitors (excursionists, e.g. cruise passengers). A detailed review of tourism statistical concepts can be found in UNDESA (2008). The UNWTO takes great care to reconcile difference in national data collection on tourism to publish an annual summary of all tourism flows amongst countries. A set of macroeconomics indicators is drawn from the World Development Indicators published by the World Bank (2014). The gravity variables are provided by CEPII, including bilateral distance, and dummies for common culture and common borders (CEPII, 2014). Guiso et al. (2009) have indicated that the fact of sharing the same legal origin or background might reduce informational costs. In addition, we also include institutional quality. For institutional data this study adopts the International Country Risk Guide’s (ICRG) country risk composite score. The ICRG is the only agency to provide detailed monthly data for 140 developed, emerging and frontier markets, since December 2003 (Hoti et al., 2005). The ICRG ratings contain 22 variables explaining three components of country risk—economic, financial and political—where 12 variables represent the political component, while 5 variables represent each of the economic and financial components of risk. The scores range from zero to 12, with higher scores representing lower risks (and thus more favourable institutional environment). Regarding the effect of institutional quality on tourism flows this study uses the following political-risk indexes (ICRG, 2014)2 (1) Government stability (GS) (7) Internal conflict (IC) (2) Military in politics (MP) (8) Ethnic tensions (ET) (3) Socioeconomic conditions (SC) (9) External conflict (EC) (4) Religion in politics (RP) (10) Democratic accountability (DA) (5) Investment profile (IP) (11) Corruption (CC) (6) Law & order (LO) (12) Bureaucracy quality (BQ). 2 ICRG Variables definitions were taken from the International Country Risk Guide (ICRG) and available at ttp://www.prsgroup.com/PDFS/icrgmethodology.pdf 112 5. International Tourism and Institutional Quality :Evidence from Gravity Model The socioeconomic conditions (SC) composite refers to socioeconomic pressures in society caused by unemployment, consumer confidence, and poverty. The maximum score for SC is 12. Investment profile (IP, having a maximum score of 12) composite assesses the risks of expropriation, profit repatriation, and payment delays. Corruption (CC) assesses corrupt practices within the political system that undermine the security of foreign investment. Corruption may potentially distort the economic and financial environment, as well as reducing government and business efficiency when associated with the assumption of power through patronage rather by reason of competence. Actual corruption may also take the form of demands for special payments and bribes connected with import and export licenses, exchange controls, tax assessments, police protection, or loans. The maximum CC score is 6. Democratic accountability (DA) is a measure of how responsive a government is to the opinions and desires of its population. The maximum score is 6. Bureaucracy quality (BQ, with a maximum score of 4) measures the resilience of a country’s administration system, in other words how far the system has the strength and expertise to maintain day-to-day administrative functions without immediate drastic changes in policy or interruptions in delivering government services when a change occurs in the political complexion or identity of the ruling power in the government. Law & order (LO) assesses the resilience and impartiality of the legal system, as well as the extent to which popular observance of the law is maintained. The maximum LO score is 6. Government stability (GS) measures the ability of a government to undertake its declared program and stay in office. Such ability is assessed through governmental unity, legislative strength and popular support. The degree of popularity of a government is indicated by the degree of the population’s approval of its programmes and policies. The maximum GS score is 12. The ethnic tensions (ET) composite measures the degree of tension associated with divisions related to race, nationality, or language. The maximum score for ET is 6. Internal conflict (IC) measures political violence and its impact on governance. High scores indicate that there is no armed or unruly civil opposition to the government, and also that the government does not indulge in arbitrary violence 113 5. International Tourism and Institutional Quality :Evidence from Gravity Model (directly or indirectly) against the population. The maximum score for IC is 12. External conflict (EC) measures the risk to the incumbent government of war, crossborder conflict, and foreign pressures. The maximum score for EC is 12. Religion in politics (RP) measures the domination of society and/or governance by a single religious group that seeks to replace civil law by religious law and to exclude other religions from the political and/or social process. The maximum score is 6. Militaryin-politics (MP) assesses the degree of involvement of the armed forces in politics. Such involvement may diminish democracy or cause a threat to an elected civilian government. The maximum MP score is 6. Thus, with higher scores always give better performance. Applying Principal Components Analysis (PCA), followed by a varimax rotation to summarize the indicators from the ICRG political-risk index, we then run the regressions using these newly-created variables to represent the institutional framework of a country. On standard eigenvalue-based criteria, whereby we have to choose eigenvalues greater than 1, we see from Table 0.1 that three components exceed a value of 1, between them explaining almost 71 percent of total variance. Table 5.1 lists the principal components. Table 5.1 Principal components (eigenvectors) Component Eigenvalue Difference Proportion Cumulative Comp1 5.71942 4.11665 0.4766 0.4766 Comp2 1.60277 .4642 0.1336 0.6102 Comp3 1.13857 .216551 0.0949 0.7051 Comp4 .922017 .324344 0.0768 0.7819 Comp5 .597673 .158717 0.0498 0.8317 Comp6 .438956 .0365617 0.0366 0.8683 Comp7 .402394 .0577924 0.0335 0.9018 Comp8 .344602 .0903113 0.0287 0.9305 Comp9 .254290 .0224054 0.0212 0.9517 Comp10 .231885 .0290877 0.0193 0.9710 Comp11 .202797 .0581637 0.0169 0.9879 114 5. International Tourism and Institutional Quality :Evidence from Gravity Model Comp12 .144634 0.0121 1.0000 Source: Author’s calculations applying PCA method The first component we called ‘institutional quality’, which was correlated with factor loadings associated with socio-economic conditions, bureaucracy quality (with factor-loading greater than 0.4), investment profile, corruption, law and order (greater than 0.3), and military-in-politics. The second component represents cultural conflict, as it is highly correlated with religious tensions, ethnic tensions, internal and external conflicts/tensions. The last component is associated with democratic accountability, with a negative value, and with government stability. Hence, we can say that the higher values indicate a greater degree of government stability, but a lower degree of democratic accountability. The relative distribution of the component loadings is shown in Figure 5.1. Component loadings .5 GS ET IC RT EC 0 LO SC IPMP CC BQ -.5 DA 0 Figure 5.1 .1 .2 Component 1 .3 .4 Component Loading factors The scoring coefficients for the components are given in Table 5.2 below, whilst a summary of the variables used in the gravity model in this chapter is given in Table 5.3. The descriptive statistics of the political risks are displayed in Table 5.4, while Tables 5.5 and 5.6 show the specific values for destination and origin countries. 115 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.7 gives the descriptive statistics for other explanatory variables used in the model. Table 5.2 Scoring Coefficients Variable Comp1 Comp2 Comp3 Unexplained GS 0.0952 –0.0018 0.7247 .214 SC 0.4272 –0.0440 0.1222 .1828 IP 0.3862 –0.0019 0.0025 .2705 IC 0.1749 0.4145 0.0784 .2554 EC 0.1311 0.2750 0.0348 .6413 CC 0.3986 –0.0906 –0.0331 .3126 MP 0.2932 0.2094 –0.1169 .2349 RT –0.0911 0.6458 –0.1486 .2948 LO 0.3907 –0.0445 0.1689 .3152 ET –0.0199 0.5196 0.1962 .4156 DA 0.1897 0.0439 –0.5721 .2092 BQ 0.4115 –0.0873 –0.1458 .1928 116 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.3 Summary of variables used in the model Variable LnTR Definition Log of tourist arrivals to destination-country from the origincountry. Gravity variables dgdpcapita Log of gross domestic product per capita of the destination-country. ogdpcapita Log of gross domestic product per capita of the origin-country. Dist Log of the distance between countries in the pair as a proxy of transport costs. Geographic variables contig Dummy variable: both countries in the pair share a common land border. Social variables comlang_off Dummy variable: both countries in the pair have the same language. dpop Population size for destination-country. opop Population size for origin-country. Comleg Dummy variables: both countries have common legal features. comco Economic variables comcur Common colonizer between origin source of the tourist and hostcountry. Dummy variables: both countries have common currency Political variables Pc1 The first component, called the institutional quality. Pc2 The second component, called conflict culture. Pc3 The third component, representing public accountability and government stability. 117 5. International Tourism and Institutional Quality :Evidence from Gravity Model To explain tourist flows, the gravity variables population and income are appropriate (Llorca-Vivero ,2008). In general, a destination’s income and population. Can be viewed as indicators of potential supply, and the origin’s income and population as indicators of potential demand (Linnemann 1966). With population density, (pop) it is possible, to measure to which the size of a country can affect the number of tourism arrivals. While per capita GDP (gdpj) it is possible to test the extent to which wealth can positively affect the amount of tourism generated by a particular region. The distance between origin and destination (disti,j) is one of the baseline gravity variables and is measured in kilometers. Tourism arrivals (tourism) are used to proxy international tourism demand. Common border (contig) as a proxy of travel cost. Colonial ties(comco) examine the importance of colonial ties for International tourism. Regrading to ICRG variables , The first component (pca1)‘institutional quality how better institutions motivate tourism arrivals according countries. The second component represents (pca2) cultural conflict to examine the effect of conflict on tourism flows. The last component (pca3)is associated with democratic accountability, with a negative value, and with government stability. 118 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.4 Descriptive statistics of political risks Variable Obs. Mean Std. Dev. Min. Max. GS 685 8.504234 1.545796 4 11.5 SC 685 5.809839 2.625996 0 11 IP 685 8.942642 2.435511 1 12 IC 685 9.426730 1.608478 2.92 12 EC 685 9.859445 1.392781 3.75 12 CC 685 2.563109 1.188127 0 6 RT 685 4.621737 1.268134 1 6 LO 685 3.785241 1.289555 0.5 6 ET 685 4.037299 1.213140 1 6 DA 685 4.153182 1.712098 0 6 BQ 685 2.177489 1.115947 0 4 MP 685 3.902526 1.718879 0 6 Source: Author’s calculations Table 5.5 Descriptive statistics of political risks of destination Variable N Mean Sd Min. Max. p50 DGS 9965 8.104803 1.476566 5.08 11 8.04 DSC 9965 6.100653 2.28398 2 10.5 6 DIP 9965 9.333053 2.09101 4 12 9.5 DIC 9965 9.395184 1.333064 6.38 11.5 9.67 DEC 9965 9.882117 1.527061 5.38 12 10.33 DCC 9965 2.76312 1.04646 1 5.04 2.5 DMP 9965 4.039559 1.597821 0.5 6 4.5 DRP 9965 4.559708 1.310382 1 6 5 DLO 9965 3.705247 1.154911 2 6 3.5 DET 9965 3.570439 1.114916 1.5 6 3.5 DDA 9965 4.628722 1.292521 1.88 6 5 DBQ 9965 2.555822 1.040503 1 4 3 Source: Author’s calculations 119 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.6 Descriptive statistics of political risks of origin Variable N Mean Sd Min. Max. p50 OGS 9965 8.403343 1.515168 4.04 11.5 8.38 OSC 9965 6.378509 2.588036 0 11 6.5 OIP 9965 9.386249 2.462021 1 12 9.58 OIC 9965 9.531137 1.59308 2.92 12 9.79 OEC 9965 9.948426 1.350924 3.75 12 10 OC 9965 2.8682 1.314398 0 6 2.5 OMP 9965 4.24556 1.670833 0 6 5 ORP 9965 4.673695 1.267304 1 6 5 OLO 9965 4.019797 1.327139 0.5 6 4 OET 9965 4.079863 1.175955 1 6 4 ODA 9965 4.485473 1.665108 0 6 5 OBQ 9965 2.49041 1.133055 0 4 2 Source: Author’s calculations Table 5.7 Descriptive statistics of other explanatory variables Variable Obs. Mean Std. Dev. Min. Max. tourism 8208 164054.7 1115032 0 2.00e+07 odgdpcapita 9858 657850.4 2356413 275.453 2.40e+07 dgdpcapita 9907 262396.1 1139750 275.453 9200000 opop 9965 6.26e+07 1.88e+08 296734 1.30e+09 dpop 9965 4.78e+07 8.03e+07 329088 3.10e+08 dist 9965 7270.287 4211.778 111.0933 19711.86 contig 9965 .0361264 .1866141 0 1 comleg 9965 .3406924 .4739659 0 1 comcur 9965 .0200702 .1402477 0 1 comlang_off 9965 .185148 .3884371 0 1 comco 9965 .1063723 .308329 0 1 Source: Author’s calculations 120 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.8 Cross-correlation between tourism arrivals and the components analysis , 2005-2009 tourism dpc1 dpc2 dpc3 opc1 opc2 opc3 tourism 1 dpc1 0.1051* 1 dpc2 0.0287* 0.3748* 1 dpc3 -.0438* opc1 0.1060* -.0329* -0.0046 opc2 0.0380* 0.0057 0.0279* -.0283* 0.5315* 1 opc3 -.0463* -.2047* 0.1571* 1 0.0424* -.0424* -.0363* 1 0.0207 -.1940* -0.0057 1 From table above we can notice that highest positive significant correlation is between tourism and first component for both destination and origin, which presents institutional quality index. While there is a negative correlation between tourism arrivals and the third component. Regarding the correlation between tourism and the rest of control variables, table 5.9 shows that various social, economic and demographical variables are correlated with tourism. The positive correlation of tourism and contiguity indicates that if two countries share same border there will be more bilateral tourism.. Moreover, the highest relationship is the negative correlation observed between distance and tourism as it is expected. In addition, we can notice that there is a variation in most of variables while there is a little Variation in ICRG variables and this could be because of the way that each of 12 variables of political risk measures are constructed. Whereas, The Political Risk Rating includes 12 weighted variables covering both political and social attributes. (See appendix 9.3 )For a full description of the ICRG methodology and interactive data access, to the aggregate and individual composites ( see “International Country Risk Guide Methodology”.) 121 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.9 Cross-correlations between tourism and the rest of control variables, 2005-2009 tourism odgdpcapita dgdpcapita contig comlang_off comcol conflict comleg comcur dist opop tourism 1 odgdpcapita -0.0241* 1 dgdpcapita -0.0103 0.0092 1 contig 0.4025* 0.0220* -0.0073 1 comlang_off 0.0391* -0.0807* -0.0731* 0.0877* 1 comcol -0.0125 -0.0546* -0.0446* 0.0466* 0.2671* 1 conflict -0.0273 0.1874* 0.1862* 0.2960* -0.094 . 1 comleg 0.0574* -0.0609* -0.0242* 0.1332* 0.4369* 0.3701* -0.0604 1 comcur 0.2897* -0.0385* -0.0296* 0.1449* -0.0037 -0.0378* 0.1894* 0.0708* 1 dist -0.1594* 0.0597* 0.0038 -0.2724* 0.0162 -0.0204* -0.1095 0.0047 -0.1895* 1 opop 0.0313* 0.0076 0.0355* -0.0196 0.0383* 0.0139 -0.3064* -0.0166 -0.0222* 0.1122* 1 dpop 0.0855* -0.0017 0.2737* 0.0299* -0.0032 -0.1608* -0.023 -0.0265* 0.0101 0.0808* 0.0059 122 dpop 1 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5.4 Methodology3 5.4.1 Tradional gravity model : The gravity model has been used with a great degree of success to explain a number of economic phenomena, including international trade, migration, commuting, FDI flows and tourism (Cheng and Wall, 2004, 2005). The basic gravity function is specified as follows: 𝑇𝑜𝑑𝑡 = 𝐾. 𝑀𝑜 . 𝑀𝑑 (Eq. 5.1) 𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸 𝑜𝑑 where 𝑀𝑜 . and 𝑀𝑑 are the mass (economic size) of the origin and of the destination respectively, and 𝐷𝐼𝑆𝑇𝐴𝑁𝐶𝐸𝒐𝒅 denotes the distance between the location of origin and the location of destination, K is the proportionality constant, related to the frequency of the event. For example, if the same system of spatial interactions is considered, the value of K will be higher if the interaction were considered for one year, in comparison to the value of K if the interaction were considered for one week. Other authors (for instance Linnemann, 1966) include population as an additional measure of country size. o is used to index countries of origin, d to index countries of destination and t to index time. The dataset includes 134 origin countries and 31 destination countries (these numbers are determined by data availability), and the period under study is the decade 2005–2009. This yields 1993 country-pairs and 9965 observations in total. After taking logs, the gravity model of tourism thus takes the following form (Culiuc, 2014 p. 10): 𝑙𝑛𝑇𝑜𝑑𝑡 = 𝐵1 𝑙𝑛𝑌𝑜𝑡 + 𝐵2 𝑙𝑛𝑌𝑑𝑡 + 𝐵3 𝑙𝑛𝐷𝑜𝑑 + 𝐵𝐴 𝑋𝑜𝑑𝑡 + 𝜔𝑜 + 𝜂𝑡 + 𝜀𝑜𝑑𝑡 qt=1…T (Eq. 5.2) where 𝑇𝑜𝑑𝑡 is a measure of the tourism flow from country of origin o to destination d in year t while 𝑌𝑜𝑡 and 𝑌𝑑𝑡 are the gross domestic products per capita (measured in constant US$) of the origin- and destination-country respectively, 𝐷𝑜𝑑 is the distance 3 The following section is based on Santos Silva and Tenreyro (2006), Serlenga and Shin (2007), and Culiuc (2014). 123 5. International Tourism and Institutional Quality :Evidence from Gravity Model between the two countries, 𝑋𝑜𝑑𝑡 is a 1 × 𝑘 vector of other variables proxying other factors; and 𝜂𝑡 is a set of T year dummies capturing common time effects. However, the specification in Equation 5.2 suffers from omitted-variables bias as mentioned by Anderson and van Wincoop (2003) because it captures only the characteristics of o and d, without taking into account the reasons (the ‘attractiveness’) motivating the flows that occur from o to d as compared to flows going from o to other destinations. As bilateral flows are based on multilateral parameters, one way of dealing with the problem of multilateral parameters is to introduce dummies for origin countries and for destination countries. The specification then becomes 𝑙𝑛𝑇 𝑜𝑑𝑡 = 𝐵1 𝑙𝑛𝑌𝑜𝑡 + 𝐵2 𝑙𝑛𝑌𝑑𝑡 + 𝐵3 𝑙𝑛𝐷𝑜𝑑 + 𝐵𝐴 𝑋𝑜𝑑𝑡 + 𝜔𝑜 + 𝛿𝑑 + 𝜂𝑡 + 𝜀𝑜𝑑𝑡 (Eq. 5.3) in which 𝜔𝑜 𝑎𝑛𝑑 𝛿𝑑 are origin and destination dummy variables. But since there are time-invariant country variables such as geographical ones (distance, surface-area of country, etcetera) in the gravity equation, we are not able to estimate the coefficients of the mentioned variables. This problem can be addressed by using a fixed-effects approach where the panel variable is the country-pair. We introduce country-pair dummies 𝜑𝑜𝑑 :Therefore, the regression will be as follows: 𝑙𝑛𝑇𝑜𝑑 = 𝛼 + 𝛽1 ln 𝑃𝑂𝑃𝑡𝑜𝑡 + 𝛽2 ln 𝑃𝑂𝑃𝑡𝑑𝑡 + 𝛽3 ln 𝐺𝐷𝑃𝑡𝑜𝑡 + 𝛽4 ln 𝐺𝐷𝑃𝑡𝑑𝑡 − 𝛽5 ln 𝐷𝑖𝑠𝑡𝑜𝑑 + 𝛽6 ln 𝑐𝑜𝑚𝑙𝑒𝑔𝑜𝑑 + 𝛽7 ln 𝑐𝑜𝑛𝑡𝑔𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑙𝑎𝑛𝑔𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑐𝑢𝑟𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑐𝑜𝑙𝑜𝑜𝑑 + 𝜀𝑜𝑑𝑡 + 𝜂𝑡 + 𝜑𝑜𝑑 (Eq. 5.4) We introduce different fixed effects, first with time dummies are added to the regression, to account for the changing nature of the relationship over time. Then we run the regression associated with time-invariant origin and destination fixed effects and for time-varying origin and destination fixed effects. Finally, we present a specification where pair effects are also added. 124 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5.4.2 Hausman Taylor model: In addition, as an alternative to the country-pairs fixed effects models, Egger (2002, 2005) and Culiuc (2014) suggested using the Hausman-Taylor (1981) model (HTM). Whilst the HTM is being increasingly applied to gravity models of trade in goods, to the best of our knowledge it is only rarely applied in tourism studies. Therefore, the Hausman-Taylor (1981) estimator allows estimating coefficients on time-invariant variables by imposing assumptions on the endogeneity/exogeneity of each variable. Hence, the HTM estimator has advantages over the fixed- and randomeffects models, since it depends on instrument variables used for between and within variation of the strictly exogenous variables (Egger, 2002, 2005). On the other hand, one of the disadvantages of the H-T estimator is to be found in the problem of how one defines the endogeneity and exogeneity of variables. In the literature, GDP per capita is highlighted as likely to be an endogenous variable. Therefore, we have made various alternative endogeneity assumptions in the regressions (discussed in greater detail below). According to H-T we can divide the explanatory variables into four 1 categories: time varying ( 𝑋𝑖𝑡 ) uncorrelated with individual effects 𝛼𝑡𝑖 and time varying (𝑋𝑖𝑡2 ) correlated with 𝛼𝑖 , time-invariant (𝑍𝑖1 ) uncorrelated with 𝛼𝑡𝑖 and timeinvariant (𝑍𝑖2 ) correlated with 𝛼𝑡𝑖 (Rault et al., 2007) as follows: 1 𝑌𝑖𝑡 = 𝛽0 + 𝛽1 𝑋𝑖𝑡 + 𝛽2 𝑋𝑖𝑡2 + 𝑍𝑖1 𝛾1 + 𝑍𝑖2 𝛾2 + 𝛼𝑖 + 𝜃𝑡 + 𝜂𝑖𝑡 (Eq. 5.5) 𝛽1 and 𝛽2 are the coefficients for time-varying variables, 𝛾1 and 𝛾2 are the vectors of coefficients for time-invariant ones; 𝜃𝑡 is the time-specific effect common to all units and is applied to correct the impact of all the individual invariant determinants. 𝛼𝑖 is the individual effects that account for the effects of all possible time- invariant factors. 𝜂𝑖𝑡 is a zero mean idiosyncratic random disturbance uncorrelated within cross- sectional units. 125 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5.4.3 Poisson model Westerlund and Wilhelmsson (2009) investigated the influence of applying gravity equation estimations on both simulated and real data. They found theoretically that, even when panel data are used, the presence of heteroskedasticity causes traditional estimations to become biased and inconsistent. Santos Silva and Tenreyro (2006) discussed how the logarithmic transformation of the model is also beset by difficulties in dealing with zero-trade flows. They suggested an alternative way for estimating log-linearized regressions that comes from direct estimation of the multiplicative form of the gravity equation, pointing out that this is the most natural procedure without the need of any further information on the pattern of heteroskedasticity. The advantages of this model are that it deals with the zero-trade flows problem, it provides unbiased estimates in the presence of heteroskedasticity, all observations are weighted equally, and the mean is always positive. The disadvantage is that it may present limited-dependent variable bias when a significant part of the observations are censored (Santos Silva and Tenreyro, 2006; An and Puttitanun, 2009; Liu, 2009; Shepherd and Wilson, 2009; Siliverstovs and Schumacher, 2009; Westerlund and Wilhelmsson, 2009). Martínez-Zarzoso (2011) offers a cautionary view that has developed from an original paper (Martínez-Zarzoso et al., 2007) that was highly critical of Santos Silva and Tenreyro (2006) and was in turn critiqued by them. The subsequent paper by Martínez-Zarzoso (2011) is much toned-down. The cumulative distribution function of the standard Poisson probability model is expressed by Prob(V=j)=𝐹𝑝 (𝑗) = 𝑒 (−𝜆)𝜆𝑗 /𝑗! (Eq. 5.6) with 𝜆 = 𝑒 𝛽0+𝛽1𝑋1+𝛽2𝑋2+𝛽3𝑋3+𝛽4𝑋4+𝛽5𝑋5+𝛽6𝑋6+𝛽7𝑋7+𝛽8𝑋8+𝛽9𝑋9+𝛽10𝑋10+𝛽11𝑋11+𝛽12𝑋12+𝜀 where j denotes the possible values for tourism numbers (j=1,2…), 𝐹𝑝 (. ) is the cumulative distribution function of the standard Poisson probability model, and 𝜆 is the non-negative Poisson parameter to be estimated (Greene, 2001; Bettin et al., 126 5. International Tourism and Institutional Quality :Evidence from Gravity Model 2012), and the non-negative dependent variable. The volume of tourism flows is a count variable rather than a continuous variable. Santos Silva and Tenreyro (2006) present the gravity equation in its exponential form: 𝑇𝑖𝑗 = exp(𝑥𝑖𝑗 𝛽) + 𝜀𝑖𝑗 (Eq. 5.7) where 𝑇𝑖𝑗 represents the bilateral trade between country i and country j, and 𝑥𝑖𝑗 is a vector of explanatory variables some of which may be linear, some logarithmic, and some dummy variables. Therefore, we can introduce the PPML estimator as defined by Santos Silva and Tenreyro (2006) and Tenreyro (2007): 𝛽 ~ =arg 𝑚𝑖𝑛𝑏 ∑𝑛𝑖,𝑗[𝑇𝑖𝑗 − 𝑒𝑥𝑝( 𝑥𝑖𝑗 𝑏]2 which is used to solve the following set of first-order conditions: ∑𝑛𝑖,𝑗[𝑇𝑖𝑗 − 𝑒𝑥𝑝( 𝑥𝑖𝑗 𝛽 ~ )] exp(𝑥𝑖 𝐵 ~ ) 𝑥𝑖 = 0 (Eq. 5.8) We adopt the Santos Silva and Tenreyro (2006) specification to apply this estimator on cross-sectional data. However, this estimator has also been implemented in panel data environments. For this application, from Equation 5.5 we can derive the expected value of the log-linearized equation, which would be: 𝐸[𝑇𝑅𝑜𝑑𝑡 |𝑧𝑜𝑑𝑡 ] = exp[𝛽0 + 𝛽1 ln(𝐺𝐷𝑃𝑜𝑡 ) + 𝛽2 ln(𝐺𝐷𝑃𝑑𝑡 ) + 𝛽3 ln 𝑃𝑂𝑃𝑡𝑜𝑡 + 𝛽4 ln 𝑃𝑂𝑃𝑡𝑑𝑡 − 𝛽5 ln 𝐷𝑖𝑠𝑡𝑜𝑑 + 𝛽6 ln 𝑐𝑜𝑚𝑙𝑒𝑔𝑜𝑑 + 𝛽7 ln 𝑐𝑜𝑛𝑡𝑔𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑙𝑎𝑛𝑔𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑐𝑢𝑟𝑜𝑑 + 𝛽8 ln 𝑐𝑜𝑚𝑐𝑜𝑙𝑜𝑜𝑑 + 𝛽9 𝑃𝐶𝐴 + 𝜃𝑜 𝛿𝑜 + 𝜃𝑑 𝛿𝑑 +𝜃𝑡 𝛿𝑡 ) ( Eq.5.9) 𝑧𝑜𝑑𝑡 = [ln(𝐺𝐷𝑃𝑜𝑡 ) , ln(𝐺𝐷𝑃𝑑𝑡 ), ln 𝑃𝑂𝑃𝑡𝑜𝑡 , ln 𝑃𝑂𝑃𝑡𝑑𝑡 , ln 𝐷𝑖𝑠𝑡𝑜𝑑 , ln 𝑐𝑜𝑚𝑙𝑒𝑔𝑜𝑑 , ln 𝑐𝑜𝑛𝑡𝑔𝑜𝑑 , ln 𝑐𝑜𝑚𝑙𝑎𝑛𝑔𝑜𝑑 , ln 𝑐𝑜𝑚𝑐𝑢𝑟𝑜𝑑 + ln 𝑐𝑜𝑚𝑐𝑜𝑙𝑜𝑜𝑑 , 𝑃𝐶𝐴 + 𝛿𝑜 + 𝛿𝑑 + 𝛿𝑡 ] with an associated error term 𝜀𝑜𝑑𝑡 =𝑇𝑅𝑜𝑑𝑡 − 𝐸[𝑇𝑅𝑜𝑑𝑡 |𝑧𝑜𝑑𝑡 ]. As explained earlier, 𝑇𝑅𝑜𝑑𝑡 represents the tourism arrivals from origin o to destination d for each year during 127 the period 2005 to 2009, where 5. International Tourism and Institutional Quality :Evidence from Gravity Model 𝛿𝑜 and 𝛿𝑑 are country specific fixed effects and 𝛿𝑡 is the year-specific fixed effect capturing the business cycle, while 𝜃𝑜 , 𝜃𝑑 and 𝜃𝑡 are vectors of the parameters with sets of fixed effects. We compare the results of log-linear regression, Hausman-Taylor and Poisson models, and focus on the gravity equation with an extended set of political-risk ICRG controls. 5.5 Empirical Results 5.5.1 Gravity variables as determinates of tourism flows Three models have been applied in this study. Firstly, the OLS estimator is applied to three approaches to the Gravity Model: (a) the basic model, with main variable of gravity model for origin and destination countries (in our study it is population, as we are discussing tourist flows), together with distance; (b) an extended one with economic, geographical, social indicators; and (c) an extended gravity model with political controls in the country-pair sample for 134 “origin” and 31 “destination-host” countries during the 2005–2009 period with adjusted standard errors for heteroskedasticity. We start by using the OLS estimator for tourism. 5.5.2 Results from the OLS estimator We began by comparing the basic and extended gravity models by adding some of the geographical, historical and linguistic dummy variables, such as common colony, distance, etcetera. Then an extended model was introduced with political-risk variables (three components) as given in Table 5.10. 128 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.10 Basic and augmented Gravity Models VARIABLES logdist logdpop logopop (Traditional (Extended (Extended gravity) gravity ) with political risk) logtourism logtourism logtourism –1.323*** –1.069*** –1.446*** (0.0362) (0.0359) (0.0282) 0.822*** 0.777*** 0.832*** (0.0129) (0.0137) (0.00984) 0.565*** 0.539*** 0.828*** (0.0157) (0.0150) (0.0120) 0.202*** –0.0239* (0.0174) (0.0120) 0.0634*** 0.00685** (0.0115) (0.00878) 1.278*** 1.601*** (0.159) (0.127) 0.959*** 0.528*** (0.0750) (0.0504) –0.644*** –0.00463 (0.0965) (0.0792) –0.0982 0.401*** (0.0685) (0.0450) 3.197*** 0.236** (0.149) (0.0992) logdgdpcapita logogdpcapita contig comlang_off comcol comleg comcur dpc1 0.456*** (0.0108) dpc2 0.216*** (0.0158) –0.188*** dpc3 (0.0223) opc1 0.579*** (0.0103) opc2 0.199*** 129 gravity 5. International Tourism and Institutional Quality :Evidence from Gravity Model (0.0168) –0.236*** opc3 (0.0160) –3.747*** –3.463*** –8.415*** (0.410) (0.470) (0.353) Observations 8,208 8,078 8,078 R–squared 0.409 0.470 0.748 Constant Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 No time or country fixed effects included It can be seen from the first column in Table 5.8 that the basic variables of the gravity equation have strong effects on tourism flows. From pooled estimations, we find that the populations of origin and destination countries are a key determinant of tourism inflow in all three models. In addition, we notice that distance plays a substantial role on tourism flows, with increases in distance reducing tourist flows. Next, we augment the basic gravity application by adding geographical variables such as contiguity, economic time-invariant variables such as common currency, and social variables such as common language (results given second column in Table 5.8), which revealed that GDP per capita for source and receipt countries is an important determinant in tourism demand. So, the demographic factors are considered as more important for tourism flows. Therefore, the coefficients of population indicate that larger countries receive and send more tourists. In addition, common border, common currency and language exert positive influences between the source-country and the host-country, while the colonial relationship is less important. However, regarding the question as to whether the political-risk factors are important and key in explaining tourism flows, we can see from the third column in Table 5.8 that institutional quality and conflict have an important influence on tourism flows. The higher levels of ICRG components indicate better quality of institutions and accordingly lower risk. Regarding the third component, higher values are associated with lower degrees of democratic accountability, and it indicates that this will exert a negative effect on inflows of tourists among the countries. 130 5. International Tourism and Institutional Quality :Evidence from Gravity Model 5.5.3 Estimation results of the gravity equation origin and destination effects using OLS regression Table 511 below presents the results after controlling for origin and destination fixed effects. The table shows the estimation results for the theoreticallybased augmented Gravity Model (the Anderson–van Wincoop 2003 model) which introduces time, origin and destination fixed effects. Table 5.11 Estimation results of the gravity equation origin and destinations effects (or/de VARIABLES logdist logdpop logopop logdgdpcapita logogdpcapita contig comlang_off comcol comcur fixed (de fixed (or fixed effects) effects) effects) logtourism logtourism logtourism –1.405*** –1.405*** –1.435*** (0.0277) (0.0303) (0.0253) –0.0795 –0.308 0.848*** (0.949) (1.078) (0.00810) 0.421 0.769*** 0.643 (0.535) (0.0117) (0.568) 0.311 0.310 0.0219** (0.291) (0.356) (0.0109) 0.351 –0.00414 0.602* (0.338) (0.00831) (0.366) 1.419*** 1.440*** 1.560*** (0.127) (0.127) (0.126) 0.835*** 0.650*** 0.713*** (0.0560) (0.0522) (0.0533) 0.173** –0.165** 0.366*** (0.0771) (0.0804) (0.0754) –0.189* –0.327*** 0.137 131 5. International Tourism and Institutional Quality :Evidence from Gravity Model (0.105) (0.115) (0.102) 0.237*** 0.331*** 0.282*** (0.0394) (0.0411) (0.0422) 0.118** 0.0953* 0.487*** (0.0852) (0.0973) (0.00955) 0.102* 0.128 0.177*** (0.106) (0.115) (0.0143) 0.00964 0.0422 –0.113*** (0.0414) (0.0479) (0.0199) 0.0200* 0.539*** 0.0425 (0.0947) (0.0100) (0.106) –0.0300 0.170*** –0.00482 (0.0912) (0.0162) (0.0985) 0.0248 –0.201*** 0.0343 (0.0456) (0.0153) (0.0492) 3.759 6.893 –13.49 (16.79) (14.85) (10.69) Time effects yes yes yes Destination effects yes yes no Origin effects yes no yes Observations 8,078 8,078 8,078 R-squared 0.850 0.792 0.816 comleg dpc1 dpc2 dpc3 opc1 opc2 opc3 Constant Robust standard errors in parentheses p<0.01***, ** p<0.05, * p<0.1 The results indicate that geographical distance between the origin and partner country has a negative impact on bilateral tourism flows. Thus, we control for timeinvariant characteristics by adding origin-country and destination-country fixed effects. Adding fixed effects reduces the significance of institutional variables; this is not surprising as institutions, although not time-invariant, tend to change little from year to year. The coefficient on the first principal component of destination or origin 132 5. International Tourism and Institutional Quality :Evidence from Gravity Model institutional quality is significant and positive. In addition, in the second column after controlling the destination effects, we can see that the conflict culture is significant at the 1 percent level, which shows that if both source and host of countries have less religious tensions and less conflict, then the tourism inflows will be boosted between those countries. It can also be seen that when a source-country has a good level of institutional quality, there is greater opportunity for its population to engage in travel internationally. Moreover, for the geographical variables, we see that contiguity (where the origin and destination have a common border) encourages tourism flows among countries; having the same language is also good for bilateral flows. We next take into account common legal origins. Geographic distance has often tended to deter tourism arrivals, especially where cheap travel is not widely available. It is a commonplace that trade partners having adjacent borders exchange much more trade with each other, and even exchange much more than trade with each other. A similar situation often arises with tourism. Shared official language and colonial ties have almost the same impact on tourism. However, the more surprising result is the negative sign of common currency that is associated with a decrease in tourism arrivals in the model. In contrast to earlier findings (Santana-Gallego et al., 2010), however, this result might be an outcome of the specification. For example, according to basic and extended gravity models before controlling the destination and origin, we obtained significant and positive results. In Table 5.11 it can be seen that the results differ in line with which specification we applied. In addition, the same can be seen when we controlled for year and country fixed effects (see Appendix 8.1 at the end). 5.5.4 Estimation results of the gravity equation with country-pair effects We next control for time and country-pair effects jointly. Taking the countrypair effects into consideration is important since some time-invariants such as distance, contiguity, etcetera, do not fully account for trust and social linkages (Papaioannou, 2009). We chose to run the overall index of ICRG variables, since we calculated the sum of the 12 indicators for origin and destination(PCO and PCD) as shown in Table 5.12. 133 5. International Tourism and Institutional Quality :Evidence from Gravity Model We conclude that this index shows that the extent to which countries have sound political institutions, strong courts, and orderly succession of power really does serve to prompt higher levels of tourism arrivals. However, relating to the ICRG control variable, high values of its correlation coefficients with other variables mean that the ICRG risk-components have been an important determinant of macroeconomic variables. In addition, we notice that the ICRG index is more important for destination countries rather than originating countries, as shown in Table 5.10 below. 134 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.8 Estimation results of the gravity equation with country-pair effects VARIABLES logdpop logopop logdgdpcapita logogdpcapita (1) (2) (3) logtourism logtourism logtourism 1.169*** 0.795*** 1.161*** (0.302) (0.308) (0.303) 0.399* 0.415** 0.414* (0.230) (0.232) (0.230) 0.402*** 0.339* 0.338** (0.132) (0.139) (0.137) 0.916*** 0.905*** 0.936*** (0.118) (0.121) (0.119) dpc1 0.151*** (0.0308) dpc2 0.0486* (0.0278) –0.0436*** dpc3 (0.0124) opc1 0.00806 (0.0337) –0.0556 opc2 (0.0342) –0.000469 Opc3 (0.0145) PCD 0.617*** (0.301) –0.369 PCO (0.358) Time effects Yes Yes Country-pair effects Yes Yes 135 Yes Yes 5. International Tourism and Institutional Quality :Evidence from Gravity Model –32.06*** –24.89*** –33.60*** (7.407) (7.644) (6.065) Observations 8,078 8,078 8,078 R-squared 0.988 0.988 0.988 Constant Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Moreover, we notice that GDP per capita for origin and destination can be a factor that impairs the tourist flows between source and host-country. Regarding ICRG control variables, the results shown in Column 2 indicate that the first component to represent institutional quality is more important for destinationcountries than for origin-countries. That means that the countries having high levels of law and order, less corruption, and with good investment profiles are much more likely to enjoy enhanced tourists flows rather than other countries having less encouraging indicators. Interestingly, those economic and demographical factors have significant impact on both origin- and destination-countries, whereas political-risk (ICRG) variables are more important for destination-countries rather than origincountries. For destination-countries, reforms solely aimed at improving tourism flows may not be very useful if the authorities and planners take no steps to address and resolve political-risk factors. The results displayed in Column 3 show that low levels of political risk tend to have a significant effect in increasing tourist inflows from origin to destination. The coefficient of overall political indicators for recipient countries shows a positive and significant effect influence at the 1 percent level. An increase of 5 percent in the overall constricted indicator causes an increase of 61.7 percent in the numbers of tourists travelling to a particular destination. The most surprising aspect of the overall political indicators for origin-countries is the presence of the negative sign and insignificant impact. The crucial need for a tourist destination to provide effective regulatory institutions is demonstrated. The results highlight the 136 5. International Tourism and Institutional Quality :Evidence from Gravity Model stark fact that the success of a tourism destination in attracting tourists will be in great part determined by the degree of its success in removing political risks and improving the quality of its governance, institutions and other relevant public bodies and services. We next ran the ICRG variables separately, with the results shown in Tables 5.13 and 5.14below. 137 5. International Tourism and Institutional Quality :Evidence from Gravity Model Table 5.9 Augmented Gravity Model with ICRG variables (1) (2) (3) (5) (6) logtourism logtourism logtourism logtourism logtourism – – – – – 1.058*** 1.320*** 1.326*** 1.199*** 1.173*** VARIABLES logdist (0.0358) (0.0305) (0.0319) (0.0332) (0.0361) 0.764*** 0.764*** 0.846*** 0.856*** 0.797*** (0.0148) (0.00984) (0.0110) (0.0121) (0.0142) 0.529*** 0.712*** 0.729*** 0.808*** 0.608*** (0.0151) (0.0125) (0.0123) (0.0149) (0.0153) – – – – – 0.197*** 0.116*** 0.158*** logdpop logopop logdgdpcapita (0.0172) logogdpcapita comlang_off contig comcol comcur comleg DGS 0.0577** * (0.0130) –0.0117 (0.0141) (0.0165) – 0.0544** 0.0703** * * (0.0114) (0.00920) (0.00987) (0.0106) (0.0113) 0.939*** 1.224*** 0.651*** 1.045*** 1.021*** (0.0750) (0.0533) (0.0597) (0.0642) (0.0731) 1.305*** 1.708*** 1.955*** 1.273*** 1.211*** (0.159) (0.138) (0.144) (0.132) (0.152) – – – 0.599*** 0.681*** 0.498*** –0.182** – 0.658*** (0.0972) (0.0823) (0.0888) (0.0841) (0.0941) 3.254*** 1.212*** 0.689*** 2.291*** 2.825*** (0.148) (0.109) (0.111) (0.134) (0.142) –0.0968 0.272*** 0.228*** 0.115* –0.103 (0.0678) (0.0487) (0.0524) (0.0588) (0.0667) 0.0384** 0.127*** (0.0166) DSC 0.185*** (0.0128) 0.0239** (0.0189) OGS 0.0685*** 0.367*** (0.00844) 138 5. International Tourism and Institutional Quality :Evidence from Gravity Model OSC 0.525*** (0.00794) DIP 0.467*** (0.0108) OIP 0.506*** (0.0106) DIC 0.477*** (0.0174) OIC 0.645*** (0.0152) DEC 0.120*** (0.0178) OEC 0.294*** (0.0227) Constant – – – – – 2.043*** 10.14*** 14.65*** 20.49*** 8.714*** (0.529) (0.376) (0.421) (0.542) (0.580) Observations 8,078 8,078 8,078 8,078 8,078 R–squared 0.477 0.698 0.663 0.589 0.490 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 139 5. Gravity Modelling: International Tourism and Country Risk Table 5.10 Extended Augmented Gravity Model with ICRG variables VARIABLES logdist logdpop logopop logdgdpcapita logogdpcapita contig comlang_off comcol comleg comcur (1) (2) (3) (4) (5) (6) (7) logtourism logtourism logtourism logtourism logtourism logtourism logtourism –1.343*** –1.383*** –1.180*** –1.155*** –1.096*** –1.097*** –1.365*** (0.0319) (0.0311) (0.0337) (0.0335) (0.0353) (0.0317) (0.0302) 0.760*** 0.914*** 0.826*** 0.755*** 0.786*** 0.706*** 0.768*** (0.0105) (0.0115) (0.0132) (0.0116) (0.0134) (0.0118) (0.00950) 0.702*** 0.733*** 0.686*** 0.672*** 0.620*** 0.606*** 0.669*** (0.0131) (0.0131) (0.0149) (0.0142) (0.0149) (0.0140) (0.0125) –0.144*** 0.0326*** –0.117*** –0.122*** –0.174*** 0.0208 –0.216*** (0.0143) (0.0126) (0.0184) (0.0149) (0.0171) (0.0141) (0.0139) –0.0169 0.0376*** 0.101*** 0.0213** 0.0552*** 0.0561*** –0.0366*** (0.0103) (0.00991) (0.0109) (0.0106) (0.0111) (0.0102) (0.00925) 1.483*** 1.120*** 0.923*** 1.675*** 1.318*** 1.555*** 1.771*** (0.151) (0.119) (0.149) (0.158) (0.149) (0.147) (0.145) 0.544*** 0.774*** 0.795*** 0.876*** 0.926*** 0.469*** 0.478*** (0.0627) (0.0570) (0.0696) (0.0661) (0.0706) (0.0664) (0.0570) –0.185** –0.266*** 0.0642 –0.545*** –0.362*** –0.215** –0.511*** (0.0916) (0.0801) (0.0962) (0.0897) (0.0932) (0.0981) (0.0880) 0.245*** 0.181*** –0.00276 0.219*** –0.0453 0.257*** 0.231*** (0.0563) (0.0527) (0.0665) (0.0597) (0.0670) (0.0565) (0.0518) 1.003*** 1.148*** 3.022*** 1.716*** 2.849*** 1.712*** 1.146*** 140 5. Gravity Modelling: International Tourism and Country Risk (0.132) DCC (0.116) (0.141) (0.123) 0.650*** (0.0187) OC 0.955*** (0.0160) DMP 0.589*** (0.0156) OMP 0.724*** (0.0136) 141 (0.143) (0.141) (0.0997) 5. Gravity Modelling: International Tourism and Country Risk Table 5.14 Extended Augmented Gravity Model with ICRG variables—continued VARIABLES DRP (1) (2) (3) (4) (5) (6) (7) logtourism logtourism logtourism logtourism logtourism logtourism logtourism 0.159*** (0.0199) ORP 0.622*** (0.0209) DLO 0.399*** (0.0189) OLO 0.843*** (0.0181) DET 0.313*** (0.0220) OET 0.384*** (0.0215) DDA 0.757*** (0.0195) ODA 0.553*** (0.0140) DBQ 0.719*** (0.0205) OBQ 1.285*** (0.0168) 142 5. Gravity Modelling: International Tourism and Country Risk –7.937*** –14.06*** –10.76*** –10.00*** –7.654*** –11.62*** –6.870*** (0.389) (0.420) (0.553) (0.436) (0.526) (0.405) (0.358) Observations 8,078 8,078 8,078 8,078 8,078 8,078 8,078 R-squared 0.642 0.655 0.527 0.599 0.501 0.621 0.696 Constant 143 5. International tourism and institutions quality: evidence from gravity model The implications of the results shown in Table 5.10 were clear. Success in self-marketing will go hand-in-hand with success in self-improvement, which in turn will be a powerful determinant of the success of its tourist trade. Thus, it is expected that countries enjoying good institutional quality (including the rule of law and the control of corruption) will see a corresponding increase in beneficial tourist flows. These findings were corroborated when the ICRG variables were run separately to see how each affects tourism flows, with the results shown in Tables 5.13 and 5.14 above. It is apparent from these tables that the role of institutions/political-risk factors in attracting cross-border tourists is considerable. Moreover, institutional quality in both origin-countries and destination-countries has a positive effect on tourism flows and the size of the effect is, in most cases, of similar magnitude, too. We can see that all the ICRG variables have significant positive effects for both of origins and destinations countries. 5.5.5 Results from the Hausman-Taylor Model We used the Hausman-Taylor model with alternative endogeneity assumptions (see the preceding discussions) and these were applied to evaluate the effect of time-invariant variables on tourism and to check their comparability with some of the findings of the previous literature on the determinants of tourism. Table 5.15 below shows the results obtained by estimating the Hausman-Taylor model with the following specifications. The GDP per capita was treated as endogenous in all four specifications. In the first regression, we used the three principal components for origin and destination. Then we ran the regression using each component individually for robustness checks. We see that the first principal component (institutional quality) for destinations is significant at the 1 percent level, which means that countries with higher institutional quality attract more tourists. Conflict in origin-countries (second principal component) in specification #1 showed a positive effect which is significant at the 5 percent level. This might be explained by the tendency for greater numbers of tourists to originate from countries that enjoy low levels of tension in religion and conflict. The conflict culture marker for destination is positive but not significant, perhaps owing to the main variation in the first principal component. Sharing a common border, common language and 144 5. International tourism and institutions quality: evidence from gravity model common currency tend to help to increase tourism inflows. In three specifications, tourism increases when any two countries have the same colonial background. Table 5.11 VARIABLES logdpop logopop logdgdpcapita logogdpcapita dpc1 dpc2 dpc3 opc1 opc2 opc3 logdist comcol comcur contig comleg Hausman-Taylor model with analysis of three principal components (1) (2) (3) (4) logtourism logtourism logtourism logtourism 1.216*** 1.231*** 1.242*** 1.243*** (0.0622) (0.0627) (0.0623) (0.0630) 1.056*** 1.076*** 1.066*** 1.085*** (0.0695) (0.0700) (0.0699) (0.0704) 0.327*** 0.336*** 0.410*** 0.418*** (0.0763) (0.0758) (0.0724) (0.0734) 1.096*** 1.148*** 1.048*** 1.083*** (0.0915) (0.0891) (0.0880) (0.0887) 0.116*** 0.0898*** (0.0310) (0.0264) 0.0160 0.00311* (0.0296) (0.0278) –0.0177 0.00458 (0.0133) (0.0112) –0.0357 –0.0497* (0.0301) (0.0256) 0.0633** 0.0712*** (0.0288) (0.0270) –0.00711 0.00569 (0.0141) (0.0126) 0.129 0.134 0.220 0.193 (0.209) (0.212) (0.209) (0.210) 0.776** 0.878** 0.909** 0.968*** (0.357) (0.364) (0.362) (0.365) 5.513*** 5.605*** 5.844*** 5.800*** (0.692) (0.704) (0.688) (0.694) 3.437*** 3.410*** 3.577*** 3.534*** (0.582) (0.597) (0.591) (0.597) 0.102 0.0988 –0.0181 0.00422 (0.225) (0.231) (0.227) (0.230) 145 5. International tourism and institutions quality: evidence from gravity model comlang_off 1.008*** 1.031*** 1.141*** 1.148*** (0.258) (0.265) (0.261) (0.263) –46.92*** –48.19*** –48.65*** –49.22*** (3.198) (3.168) (3.136) (3.180) Observations 8,078 8,078 8,078 8,078 Number of paired 1,973 1,973 1,973 1,973 Sargen test 0.19 0.13 0.09 0.08 Constant Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1 For our main interest, it is to be noted that the presence of institutional quality shows positive and significant effects that are greater for host-countries than they are for origin-countries. On the other hand, economic factors (income) figure as an important determinant in origin-countries more than for destination-countries. It can be seen from Table 5.15 above that the coefficient of GDP per capita for origincountries (at 1.096) is considerably higher than that (0.327) for GDP per capita of destination-countries. This would be expected, as individuals tend to travel more frequently as they become more affluent. Moreover, a shared language will increase tourism inflows between two countries. It is reasonable to expect a common language to have a positive impact on trade in services (perhaps even more so than for trade in goods). Communication is greatly facilitated by a common language (Walsh, 2006). In addition, common colonial background, common currency, and common border likewise promote higher levels of tourism. Clearly, sharing a common border or having common history should make the flow of information easier; the common border dummy is positive and significant in the regression and these results agree with the findings of other studies (e.g. Fidrmuc and Karaja, 2013). Most studies tend to investigate the determinants of tourism with respect to economic factors. While few researchers have focused solely on political and institutional reforms, we find that institutional quality is important, together with culture and conflict in determining tourism flows. Distance has no significant 146 5. International tourism and institutions quality: evidence from gravity model influence on tourism flows using the HTM. As for contiguity, this may indicate that physical distances have little or no relevance for the movement of tourists. The Hausman-Taylor model was checked to see whether the instruments were valid or not. The first two specifications, according to the Sargan (1958) test, are valid but the last two are valid at just the 5 percent level. This result might arise owing to the main variation of loading in the first principal component. 5.5.6 Results of count Model (Poisson Model) Certain writers, including Santos Silva and Tenreyro (2006), have suggested performing an estimation of the model in levels rather than in logs by applying the Poisson Estimator with clustered standard errors, as the interpretation of coefficients as elasticities in log-linearized OLS can be highly biased in the presence of heteroskedasticity. Accordingly, we applied the Poisson Pseudo-Maximum Likelihood estimator to correct for the presence of heteroskedasticity, and to investigate the relationships between the three principal components, indices of ICRG and tourism, as well as common language and colonial ties. To test the hypothesis, we implemented heteroskedasticity-robust standard errors that allow for clustering within country-pairs; this addresses issues of over-dispersion associated with Poisson distributions as well as that of serial correlation. Table 5.16 below shows that the PPML estimation results are similar to the pooled OLS results. GDP per person and population size continue to have significant positive impacts on tourism flows although the coefficients in each case become smaller. Similarly, higher results from the principal component analysis indicate that greater numbers of tourists are willing to travel abroad. Language and contiguity are important for tourism in the PPML estimation in all five regressions. In addition, the results show that better institutional quality and the lack of conflict both produce more significance in terms of tourism flows, whilst GDP per capita is a better determinant for origin rather than for destination. Moreover, when the source- and recipient-countries share the same currency, this often leads to the generation of greater tourism inflows; this result is in line with Santana et al. (2010). 147 5. International tourism and institutions quality: evidence from gravity model Table 5.12 VARIABLES logdist logdpop logopop logdgdpcapita logogdpcapita contig comlang_off comcol comleg comcur dpc1 dpc2 dpc3 opc1 Results of count model (Poisson model) (1) (2) (3) (4) (5) Tourism Tourism Tourism Tourism Tourism flows flows flows flows flows –1.122*** –0.855*** –0.810*** –1.123*** –1.124*** (0.117) (0.102) (0.108) (0.120) (0.116) 0.730*** 0.618*** 0.611*** 0.715*** 0.716*** (0.0338) (0.0324) (0.0329) (0.0296) (0.0326) 0.574*** 0.605*** 0.584*** 0.662*** 0.551*** (0.0265) (0.0397) (0.0387) (0.0318) (0.0297) –0.0396 0.00163 –0.118*** 0.120*** –0.0256 (0.0244) (0.0475) (0.0293) (0.0381) (0.0258) 0.0445*** 0.0247 0.0174 0.0663*** 0.0525*** (0.0150) (0.0201) (0.0202) (0.0155) (0.0156) 1.113*** 1.388*** 1.480*** 1.027*** 1.081*** (0.304) (0.257) (0.284) (0.299) (0.289) 0.491*** –0.413** –0.381** 0.379*** 0.442*** (0.0936) (0.168) (0.165) (0.106) (0.114) –0.0488 1.237*** 1.016*** 0.415*** 0.297 (0.155) (0.208) (0.185) (0.158) (0.193) –0.0352 0.0931 0.148 –0.180 –0.103 (0.134) (0.131) (0.130) (0.148) (0.142) 1.398*** 0.469** 0.253 1.410*** 1.396*** (0.142) (0.206) (0.200) (0.145) (0.162) 0.228*** 0.291*** (0.0349) (0.0293) 0.240*** 0.418*** (0.0613) (0.0530) –0.0385 –0.0261 (0.0807) (0.0734) 0.391*** 0.426*** 148 5. International tourism and institutions quality: evidence from gravity model (0.0346) opc2 opc3 (0.0331) 0.113*** 0.340*** (0.0378) (0.0347) –0.0409 –0.242*** (0.0534) (0.0521) –18.88*** –20.46*** –19.08*** –21.99*** –18.52*** (1.364) (1.405) (1.242) (1.505) (1.346) Observations 8,078 8,078 8,078 8,078 8,078 R-squared 0.465 0.552 0.527 0.514 0.450 Constant Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 In general, we can conclude that GDP per capita for origin, population size, common language and common currency are important factors in increasing tourism flows according to the Poisson Model. The overall response to the importance of ICRG variables in the determining of arrivals inflows is that the first dimension of ICRG (institutional quality) was very positive and encourages people to choose to travel to the destination, while for the results from the OLS the PPML estimations indicate that the ICRG variables are the main determinants of tourism flows. These findings lead to the conclusion that corruption not only affects growth and investment, as confirmed by many authors, but that it also has a detrimental effect on the tourism sector. Since tourism contributes a great proportion of the GDP in developing nations in particular, a policy implication is that reducing public-sector corruption will increase an economy’s wealth in more ways than one: by increasing growth, investment and GDP—as shown by Mauro (1995) and others—and by increasing income-streams from tourism. 5.6 Conclusion This study examined the impacts of twelve major ICRG variables on tourism flows using gravity models estimated with standard OLS, the Hausman-Taylor model and the PPML technique. To investigate the relationship between the ICRG 149 5. International tourism and institutions quality: evidence from gravity model variables and tourism, we incorporated three principal components in the analysis, i.e. institutional quality, conflict culture and government effectiveness, in addition to some basic and extended gravity variables such as GDP per capita, population, common language, etcetera. We applied the traditional estimation technique based on log-linearization of the model, as well as the Hausman-Taylor estimator and the Poisson Pseudo-Maximum Likelihood estimation technique in order to compare the results of these three approaches. The estimated coefficients in the OLS model differed from our expectations. For instance, the OLS and other traditional models in this study have yielded results indicating that the effect of the partner-country’s GDP for source and destination is not a crucial determinant for tourism flows when we control for destination and origin effects. This result is contrary to our expectations, since the previous studies have suggested that GDP is an important determinant for tourism flows between two countries. However, the basic gravity estimation methods have predicted a significant positive impact for GDP per capita. In addition, the presence of a common language is found to be one of the most important determinants of tourism between two countries. The results are also consistent with earlier research, which showed that distance does not well explain flows in service-trade. Regarding the Hausman-Taylor estimator, we find that institutional quality is important, together with culture and conflict, in determining tourism flows. Distance has no significant influence on tourism flows according to the HTM. As for contiguity, this may indicate that physical distances have little or no relevance for the movement of tourists. The Poisson Pseudo-Maximum Likelihood estimator is the only estimation method that has been performing correctly to the specifications and has provided good estimation results according to the Gravity Model theory. Moreover, in order to answer the central question regarding how institutional quality affects tourism flows, we added the three principal components as explained in Section 5.3. We can report that institutional quality in both origin-countries and destination-countries encourages greater tourist flows. For the second component, absence of conflict and tensions have a positive and significant sign, which means that low levels of risk of conflict do attract tourists to visit such destinations, or at 150 5. International tourism and institutions quality: evidence from gravity model least do not deter them from doing so. The last component gives an indication that higher levels of government stability encourage tourist inflows even when there is lower democratic accountability. Therefore, political-risk levels (especially the first component) play an effective role and boost cross-border tourism. In summary, the results of this study suggest that lower levels of political risk contribute to an increase in tourism flows. Common language, common currency and political risk (particularly institutional quality) are highly important determinants in promoting tourism. Since tourism contributes a great proportion to national GDP especially in developing countries, a policy implication is that reducing the political risk and increasing the institutional quality will lead to an increase in an economy’s wealth in different ways. As previously stated, one example of institutional quality enhancement is the reduction of corruption; if this is successful it will lead to an increase in growth, investment, and GDP, and by increasing revenue from tourism (Mauro 1995). It would be most useful for future research to focus on analysing the longer time spans, in conjunction with additional data on tourism becoming available. The combination of the two features might well help address the issues encountered when attempting to estimate dynamic/system GMM models; at the commencement of this work, the dataset available covered only the period 2005– 2009. Future research could also expand the analysis by splitting the samples according to ICRG—those relevant for tourism to advanced economies, and those relevant for tourism to developing countries—which could not be incorporated into the present study owing to data limitations especially for destination countries. 151 6. Concluding Remarks 6 Concluding Remarks This thesis contributes to the literature on the concepts of tourism, the relationship between tourism specialization and economic growth, as well as the exploration and identification of the determinants influencing tourist flows. The stimulus for conducting research in this area arose from the fact that there is little or no convergence within the earlier and recent literature with respect to the studies, investigations and findings either regarding the effect of tourism on economic growth or about the main determinants of tourism. Smith (2001) remarked on how individuals in the 1980s were wondering whether tourism could be seen as a blessing or as a curse, and she observed how such questions are now academic, given that tourism is the world’s largest industry, and given its global role in generating jobs and in providing customers to support those jobs. While these questions are easily caught up in the wider debate surrounding the effects of globalization, the more important issue now in a scientific academic approach should be how to examine the role that tourism plays in the economy, and how to identify the benefits and disadvantages (in financial and non-financial terms) that proceed from tourism in any particular locality. Some studies support the argument that tourism can make a positive impact on economic growth (e.g. Khan et al., 1990; Copeland, 1991; Balaguer and Cantavella-Jordá, 2002; Lim and Cooper, 2009). However, others have indicated that the hypothesis that tourism leads to economic development is to be rejected (e.g. Oh, 2005; Sequeira and Campos, 2005). Therefore, in the second essay we discussed the most important factors that influence travel and tourism; amongst others these factors include diseases, natural disasters, pollution, climate and weather, and advances in technology. After that, we discussed the benefits and costs of tourism, presenting what previous studies have had to say about this issue. In the third essay, we performed an analysis to examine the relationship between tourism and economic growth; it covered 131 countries over the period 1995 to 2007, including 32 countries highly dependent on tourism during that period. Moreover, we divided countries into two groups according to their international receipts as share of exports—tourism-exporting countries, which 152 6. Concluding Remarks have international tourist receipts in excess of a 14 percent threshold against exports, and non-tourism-exporting countries which have receipts against exports below this threshold. The analysis made use of an interaction term, which represents trade as a share of GDP multiplied by a tourism specialization index. The general conclusion based on the empirical investigations carried out in this thesis is that the fixed-effects models suggest that tourism specialization has no significant effect on economic growth. Similar results were obtained that indicate that tourism does not play a role in fostering economic growth when we split the sample into underdeveloped and developed countries. After factoring in the mutual interaction, we discovered that Dutch Disease might appear in the broad sample if there is a dependence on tourism essentially as a main factor of economic growth in a particular country. In addition, we found that tourism affects trade positively, causing Dutch Disease through a dependence on tourism, but only in tourism-exporting countries. The main policy recommendation for countries that rely heavily on exporting tourism is that they should pay greater attention to diversifying their economies into manufacturing, and to investing in the infrastructure system quite apart from those facilities related to tourism and tourist commodities. During this study it has been found that tourism might indeed become a curse rather than a blessing for countries that export tourism. Therefore, study of the tourism industry faces challenges and opportunities, and further research should play its part in examining more deeply the dynamics of this industry. Whilst the magnitude of tourism increases for the economies of countries throughout the world, so the need for further research increases to understand the channels through which tourism exerts its effects on economic growth, and on environmental and social conditions. The main contributions and conclusions of the final two essays of this thesis can be summarized as follows. Firstly, Chapter 4 contributes to the literature on determinants of tourism considering the impact of institutional quality and infrastructure on tourism flows in the context of a whole sample of 131 countries during the period between 1995 and 2007. In addition, we split the sample according to IMF classification (developed and developing countries) and according to population size (small and large countries). In doing so, the investigation in this 153 6. Concluding Remarks chapter showed that there is a positive and highly significant relationship between governance, internet availability and tourism flows. However, the question is obviously more complex. We also note that the governance variables used vary little from one year to another owing to the short duration of the series (8 years), which in turn is related to the current lack of long-series data on governance. In addition, the positive relationships between the information technology variable and tourism flows in our estimated model appear to corroborate the idea that the increasing levels of technology in the tourism industry have generated beneficial effects for the industry, other things being equal (e.g. absence of conflict). Also, promoting the technology industry enhances the tourism industry. For the conflict factor, it is clear from the results that it is bad news for a country. For developed and developing countries, the findings support the view that institutional quality seems to be more relevant for international tourists, whilst issues surrounding the communications infrastructure (measured by internet availability and usage) give rise to important considerations for the levels of tourist arrivals in both samples. Our analysis suggests that policy-makers in tourist destinations are rightly concerned about safety and stability. From an economic policy perspective, it is useful to further develop the infrastructure, communications system, and the quality of institutions with a view to fostering the growth of tourism flows, given that the impact of infrastructure and governance indicators on the latter is positive and significant. This suggests that the nexus linking tourism flows, internet usage and institutional quality may be a fundamentally key determinant in the whole sample embracing developed/developing and large/small countries. Hence, in future research it would be well to conduct in-depth investigations into the impact of communications (internet availability/usage) and governance indicators upon tourist flows in different samples across the various regions. This would help to ascertain the extent to which the results of this chapter can be generalized to other naturalresource-dependent countries. We hope make-determined efforts to deal with these questions in the future. Secondly, Chapter 5 examines the tourism relationship by providing a comparison of country-risk ratings (ICRG) and how they affect tourism flows. The 154 6. Concluding Remarks main contribution of this chapter rests in using a variety of gravity-equation approaches to the relationship between ICRG data and tourism. When principal components analysis (PCA) was applied to the 12 political-risk variables from the ICRG, the PCA identified three components, which together explain more than 72 percent of total variance. Our research generated findings from the panel-data sets for 134 originating countries and 31 destination countries (selected depending on data availability) by running estimates of the gravity equation using three approaches—the traditional, the Hausman-Taylor and the Poisson estimation techniques. The findings of this chapter confirm that there are significant relationships between the various factors, whichever estimation technique is used. The findings also confirm that the impacts are significant and direct. In summary, the results of this study suggest that lower levels of political risk contribute to an increase in tourism flows. Furthermore and in particular, common language and political-risk (particularly institutional quality) act as the most highly important determinants in promoting tourism. Moreover, the three models confirm that institutional quality in both origin-countries and destination-countries encourages greater tourist flows. For the second component, absence of conflict and tensions have a positive and significant sign, which indicates that low levels of risk of conflict do attract tourists to visit such destinations, or at least do not deter tourists from visiting them. The last component gives an indication that higher levels of government stability encourage tourist inflows even when there is lower democratic accountability. Therefore, political-risk levels (especially the first component) play an efficient role and boost cross-border tourism. Regarding the main variable of gravity (population) in our estimation, the three models confirmed that population has no appreciable effect on tourism flows between host and home countries, but GDP per capita variables are positively significant. Since tourism contributes a great proportion of the GDP of many nations— developing countries in particular—one policy implication is that reducing the political risk and increasing the institutional quality of tourist destinations will increase the wealth of an economy in different ways. For example, the enhancement of institutional quality, especially through a reduction in local corruption and non- 155 6. Concluding Remarks user-friendly practices, will increase growth, investment and GDP (as shown by Mauro, 1995, 2004) by increasing the revenue derived from tourism, both directly and indirectly (through increases in tax revenues and in other ways). Future research could focus on analysing the longer time spans, in line with additional data on tourism becoming available. This would be of particular help in address the issues encountered when attempting to estimate dynamic/system GMM models; at the commencement of this work, the dataset available covered only the period 2005–2009. Future research could also expand the analysis in splitting the samples according to ICRG into two groups—{relevant for tourism within advanced economies} and {relevant for tourism within developing countries}—which have not been incorporated in the current study owing to data limitations especially for destination countries. Although the author believes that this thesis covers quite a lot of background, nevertheless it also has several limitations. One of the main limitations of this research is the data on tourism arrivals. In fact, in order to obtain a complete picture of the extent of tourism data for any country, several factors must be taken into account. These factors should reflect as fully as possible the particular local conditions of the tourism sector, not simply the generic ones apparently common to all destinations from a superficial inspection. For example, the analysis was hindered by limitations in obtaining data that reflect all aspects of the degree of penetration by tourism into the national and local economies in the samples of countries used in the analysis. In addition, the data that UNTWO provide require considerable work and effort to adapt and organize into a dataset that can be used for any particular analytical model. Therefore, it was necessary for the analysis conducted in this thesis to narrow the selection of the tourism-arrival indicators to the most widely-used measures that have been considered in the previous literature based on data availability. It is hoped, however, in spite of these limitations, that the essays in this thesis will make a fairly significant contribution to the literature on tourism studies. 156 7 Appendix A 7.1 Hausman Test fixed random Difference. S.E. 8.035526 7.041849 .9936779 .0925537 logPOP 0.961852 0.653729 .3081228 .1948208 logTrade 0.604268 0.657879 –.0536112 .0213743 logppp 0.599053 0.643769 –.0447167 .0111787 loghealth 0.004701 0.025037 –.020336 .0198164 lognetsize 0.074362 0.08183 –.0074672 .0044818 pca 0.157203 0.205254 –.0480507 .0128287 conflict –0.102120 –0.10535 .0032307 .0044926 dgdp chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 28.14 Prob>chi2 = 0.0004 7.2 Hausman Test with toap logPOP logTrade logppp loghealth lognetsize pca conflict Fixed fixed –0.0381486 0.6042673 0.5990525 0.0047014 0.0743623 0.1572033 –0.1021225 Random random –0.3462713 0.6578785 0.6437692 0.0250373 0.0818295 0.205254 –0.1053532 Difference Difference. 0.308123 –0.05361 –0.04472 –0.02034 –0.00747 –0.04805 0.003231 S.E. S.E. 0.1948208 0.0213743 0.0111787 0.0198164 0.0044818 0.0128287 0.0044926 chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 28.14 Prob>chi2 = 0.0004 7.3 Hausman Test in developed countries fixed Difference. –4.696071 . 0.7340281 –1.622847 0.350102 0.838009 0.7346406 0.1033688 0.017901 logppp 0.54505 0.484699 0.060351 . loghealth –0.09084 –0.141155 0.0503178 0.027164 lognetsize 0.08506 0.0568542 0.0282057 0.00451 dgdp 6.270304 logPOP –0.88882 logTrade random 10.96637 157 S.E. pca 0.145009 0.1686633 0.0236542 0.013203 conflict –0.05861 –0.121639 0.0630278 . chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 31.84 Prob>chi2 = 0.0001 7.4 Hausman Test in developed countries fixed fixed 7.51979 1.1923 0.579118 0.63164 0.063521 0.070579 0.156811 –0.09451 dgdp logPOP logTrade logppp loghealth lognetsize pca conflict random random 6.415809 0.6398926 0.6754121 0.6812586 0.0507297 0.0850127 0.2161995 –0.097857 Difference. difference 1.103981 0.5524071 – – 0.0127909 – – 0.0033441 S.E. S.E. . 0.230129 0.022559 0.005982 0.020496 0.005323 0.011458 . chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B)= 54.45 Prob>chi2 = 0.0000 7.5 Hausman Test for large population-size countries fixed fixed 11.35628 2.316391 0.547027 0.445947 0.115791 0.020743 0.089032 –0.11192 dgdp logPOP logTrade logppp loghealth lognetsize pca conflict random random 10.58598 0.8133799 0.6702962 0.5616915 0.2624211 0.0539659 0.1488039 –0.101853 Difference. difference 0.7703047 1.503011 – – – – – – S.E. S.E. . 0.286399 . . 0.005889 0.005926 0.013395 . chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 49.61 Prob>chi2 = 0.0000 7.6 Hausman Test for small population-size countries fixed dgdp logPOP logTrade logppp loghealth lognetsize random fixed 2.91815 0.21131 0.62050 0.67531 – 0.12978 random 2.66922 0.588837 0.594541 0.658624 – 0.120552 158 Difference. differen 0.248938 – 0.025966 0.016695 0.005100 0.009227 S.E. S.E 0.690288 0.249956 0.043838 0.023623 0.039544 0.006391 pca conflict 0.17638 – 0.222250 – – 0.027739 0.02178 0.041929 chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 7.05 Prob>chi2 = 0.5314 8 Appendix B 8.1 Contiguity controlled for year and country fixed effects VARIABLES logdpop logopop logdgdpcapita logogdpcapita comcur (1) (2) (3) logtourism logtourism logtourism 1.169*** 0.795*** 1.104*** (0.302) (0.308) (0.305) 0.399* 0.415* 0.412* (0.230) (0.232) (0.230) 0.402*** 0.339** 0.305** (0.132) (0.139) (0.135) 0.916*** 0.905*** 0.932*** (0.118) (0.121) (0.119) -0.826 -3.089 -1.593 (1.874) (1.932) (1.925) dpc1 0.151*** (0.0308) dpc2 -0.0486* (0.0278) dpc3 -0.0436*** (0.0124) opc1 0.00806 (0.0337) opc2 -0.0556 (0.0342) -0.000469 opc3 (0.0145) pco -0.0250 (0.0285) 159 pcd 0.0796*** (0.0247) Constant -32.06*** -24.89*** -30.16*** (7.407) (7.644) (7.581) Year fixed effects Yes Yes Yes Country pair effects Yes Yes Yes Observations 8,078 8,078 8,078 R-squared 0.988 0.988 0.988 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 160 9 Appendix c 9.1 Descriptive statistics for first chapter within more details. Variable gce i pop trade ttradep school growth Mean Std. Dev. Min Max Observations 15.68838 5.778924 3.36423 39.1937 N= 1695 between 5.446456 4.877276 28.99367 n= 133 within 2.010961 6.431003 27.17821 T-bar = 12.7444 6.848701 3.48003 64.1418 N= 1670 between 5.679105 8.269874 45.55693 n= 133 within 3.962142 0.999731 45.41389 T = 12.5564 1.219802 -3.93064 10.0428 N= 1724 between 1.104265 -1.33331 4.103721 n= 133 within 0.529338 -4.68191 7.27592 T-bar = 12.9624 49.27787 14.7725 456.646 N= 1680 between 51.55889 21.93957 412.4021 n= 133 within 11.925 22.00484 178.6359 T-bar = 12.6316 12.36387 0.004462 116.4984 N= 1638 between 12.65276 0.038268 103.6793 n= 133 within 2.841736 -12.9131 60.5484 T-bar = 12.3158 31.5836 5.17789 161.662 N= 1191 between 31.7321 5.73342 152.4844 n= 132 within 6.226507 48.94466 106.739 T-bar = 9.02273 4.036521 -29.6301 33.0305 N= 1592 between 2.202261 -2.54978 12.57449 n= 133 within 3.386093 -24.1802 23.35613 T = 11.9699 overall overall overall overall overall overall overall 21.78991 1.336841 86.16207 6.388372 74.75916 2.900116 161 le overall 67.65472 9.988338 31.2392 85.1634 N= 1701 between 9.835651 39.19711 81.21799 n= 133 within 1.649838 51.41642 78.82512 T-bar = 12.7895 9.2 Descriptive statistics for second chapter in more details Variable countr~e GDPPER TOA netpop POP ppp TELPOP Trade conflict Mean Std. Dev. Min Max Observations 98.74194 58.16915 1 197 N= between 58.39268 1 197 n=133 within 0 98.74194 98.74194 T-bar = 12.9926 11187.79 108.9024 67138.52 N= between 11189.57 112.9858 56073.74 n=133 within 1382.881 -3271.82 19106.83 T-bar = 12.9412 1.04E+07 11000 8.09E+07 N= between 1.01E+07 67333.33 7.27E+07 n=133 within 1816440 -9273873 2.54E+07 T-bar = 12.3529 20.16298 0 88.90034 N= between 14.9551 0.101163 56.63166 n=133 within 13.52761 -34.5956 64.4611 T-bar = 12.4926 1.45E+08 61700 1.32E+09 N= between 1.45E+08 62648.6 1.27E+09 n=133 within 6874659 -6.35E+07 1.46E+08 T-bar = 12.9926 0.278601 0.140434 1.860173 N= between 0.263239 0.277358 1.322858 n=133 within 0.091801 0.245572 1.123063 T-bar = 12.8955 20.31546 0.076353 90.36677 N= between 20.05261 0.215061 85.83148 n=133 within 3.150024 6.159036 37.2302 T-bar = 12.8456 51.59567 14.77247 438.9016 N= between 49.75758 22.36189 372.9004 n=133 within 12.82125 19.81148 179.7239 T-bar = 12.7426 0.443702 0 2 N= between 0.389997 0 1.846154 n=133 within 0.213859 -1.53215 2.006312 T-bar = 12.9926 overall overall overall overall overall overall overall overall overall 8042.055 5004512 13.99372 4.25E+07 0.574099 22.65227 87.27659 0.160159 162 1767 1760 1680 1699 1767 1728 1747 1733 1767 ps voice reg low coc eog pca LF dgdp overall 0.944577 -3.05644 1.57687 N= between 0.905607 -2.41734 1.422584 n=133 within 0.269317 -1.11669 1.267123 T-bar = 8.92537 0.92784 -1.95119 1.82669 N= between 0.911113 -1.72669 1.602511 n=133 within 0.170564 -0.94727 0.763866 T-bar = 8.89552 0.916643 -2.52663 2.02558 N= between 0.88953 -1.85087 1.836446 n=133 within 0.22565 -0.98218 1.532048 T-bar = 8.85075 0.968781 -2.31285 1.96404 N= between 0.947071 -1.94313 1.886921 n=133 within 0.199982 -1.29753 1.457386 T-bar = 8.88806 1.020891 -2.48921 2.46656 N= between 0.989716 -1.65981 2.335351 n=133 within 0.227465 -1.38999 1.306384 T-bar = 8.73881 0.972304 -2.39408 2.23691 N= between 0.951077 -1.67847 2.085287 n=133 within 0.196675 -0.97442 0.953262 T-bar = 8.79851 2.170471 -4.56873 4.622228 N= between 2.127116 -3.69779 4.462629 n=133 within 0.33157 -1.09406 2.087353 T-bar = 8.63433 9.239378 41.92988 82.50707 N= between 9.119442 43.22899 81.21799 n=133 within 1.409327 63.65468 77.9887 T-bar = 12.7721 0.005849 -0.03013 0.104125 N= between 0.003329 -0.00557 0.018218 n=133 within 0.004813 -0.0265 0.089968 T-bar = 11.9412 overall overall overall overall overall overall overall overall -0.01963 0.110689 0.16488 0.078455 0.139367 0.140364 0.266477 69.31159 0.004061 163 1196 1192 1186 1191 1171 1179 1157 1737 1624 9.3 Descriptive statistics for third chapter with in more details Variable tourism dgdpca~a odgdpc~a opop dpop Mean Std. Dev. Min Max Observations 164054.7 1115032 1 2.00E+07 N= 8208 between 1041357 1 1.96E+07 n= 1993 within 81949.11 -2235945 2504055 T-bar = 4.11841 1139750 275.453 9200000 N= 9907 between 1134526 294.1074 8500000 n= 1993 within 69909.37 -437604 962396.1 T-bar = 4.9709 2356413 275.453 2.40E+07 N= 9858 between 2348667 294.1074 2.16E+07 n= 1974 within 179166.5 -1942150 3057850 T-bar = 4.99392 1.88E+08 296734 1.30E+09 N= 9965 between 1.88E+08 309599 1.30E+09 n= 1993 within 5321644 2601266 1.03E+08 T= 5 8.03E+07 329088 3.10E+08 N= 9965 between 8.03E+07 341918.2 3.02E+08 n= 1993 within 1521579 3.78E+07 5.78E+07 T= 5 overall overall overall overall overall 262396.1 657850.4 6.26E+07 4.78E+07 164 10 Appendix d 10.1 Some statistics of variables by Country code. ountrycode stats GCE I POP trade gdpcon~t TRP secsch~l 1 mean 9.88599 22.7786 -0.09817 59.6566 1264.99 7.78296 72.7962 sd 1.26262 3.98167 0.536994 11.9657 270.434 3.88875 2.55748 cv 0.127718 0.174799 -5.4699 0.200577 0.213783 0.499649 0.035132 mean 12.6238 18.3948 1.06507 31.5012 7747.48 1.37065 83.983 sd 0.922605 3.51362 0.113425 10.9375 761.333 0.343026 4.40949 cv 0.073085 0.191011 0.106496 0.34721 0.098268 0.250266 0.052505 mean 10.8352 22.2481 -0.52754 73.2669 796.447 3.09959 88.596 sd 0.670148 7.37231 0.710797 7.52507 311.741 1.29296 2.93119 cv 0.061849 0.331368 -1.34739 0.102708 0.391414 0.417139 0.033085 mean 18.0908 24.4407 1.25348 40.6037 21246 3.22647 152.484 sd 0.199233 1.60846 0.138053 2.1194 1907.32 0.162654 5.12701 cv 0.011013 0.065811 0.110136 0.052197 0.089773 0.050412 0.033623 mean 18.5813 21.5331 0.366289 90.2885 23947.2 6.00505 100.653 sd 0.730085 0.864243 0.217581 12.3713 1703.14 0.374825 2.23733 cv 0.039291 0.040136 0.594014 0.13702 0.071121 0.062418 0.022228 mean 12.6746 32.537 0.93715 90.6847 869.94 1.74582 83.2466 sd 1.78226 11.5764 0.141608 17.3781 449.582 1.16768 4.57197 cv 0.140617 0.355792 0.151105 0.191632 0.516797 0.668845 0.054921 mean 18.4528 17.2356 2.25499 153.205 12498.6 11.0904 98.7615 sd 2.13108 4.78896 0.262688 13.7417 1186.12 1.22793 2.41543 cv 0.115488 0.277854 0.116492 0.089695 0.094901 0.11072 0.024457 mean 4.94416 22.6146 1.75443 35.0555 351.132 0.1081 44.7753 sd 0.481467 1.78686 0.168844 5.54961 49.6904 0.02014 2.14587 cv 0.097381 0.079014 0.096239 0.158309 0.141515 0.186312 0.047925 mean 20.2445 25.294 -0.40538 125.158 1440.09 1.14744 91.1879 sd 0.920623 2.69298 0.08094 13.3945 422.958 0.672454 4.49461 cv 0.045475 0.106467 -0.19967 0.107021 0.293702 0.586047 0.04929 mean 22.0564 20.0564 0.378437 156.37 22510.8 2.51299 134.368 sd 0.606706 0.759457 0.180741 14.5266 1584.74 0.573275 22.2132 cv 0.027507 0.037866 0.477598 0.092899 0.070399 0.228125 0.165316 mean 14.2114 20.6015 2.80877 114.876 3336.34 15.7502 75.1567 sd 0.75691 3.22435 0.565711 9.26104 383.158 2.87613 6.89143 cv 0.053261 0.156511 0.201409 0.080618 0.114844 0.182609 0.091694 mean 11.7332 17.9554 3.2038 43.5995 335.778 2.1584 22.543 sd 1.71166 0.852452 0.161818 3.90303 16.3111 0.330868 5.69205 cv 0.145882 0.047476 0.050508 0.08952 0.048577 0.153294 0.252498 2 3 4 5 6 7 8 9 10 11 12 165 13 14 15 16 17 18 19 20 21 22 23 24 25 26 mean 14.863 15.8305 2.02425 54.5723 1031.35 2.284 81.4151 sd 1.09044 3.32062 0.141043 10.2123 60.6152 0.818605 5.01272 cv 0.073366 0.209762 0.069677 0.187134 0.058772 0.358408 0.06157 mean 23.2665 22.6018 1.59981 85.4758 3643.45 4.31193 73.2333 sd 3.4475 2.68556 0.491066 5.89365 641.681 1.08097 6.15508 cv 0.148174 0.118821 0.306953 0.068951 0.176119 0.250692 0.084048 mean 19.9998 16.6672 1.36213 22.4086 3792.09 0.319714 104.045 sd 0.550107 0.826529 0.16093 5.18617 206.798 0.146566 3.47084 cv 0.027506 0.04959 0.118146 0.231436 0.054534 0.45843 0.033359 mean 16.4358 18.4513 -0.7416 116.946 1751.35 8.95051 97.3208 sd 2.20543 5.48184 0.494869 18.9673 344.735 2.57175 8.12029 cv 0.134185 0.297098 -0.6673 0.162188 0.19684 0.287331 0.083438 mean 20.9787 8.41494 1.93177 34.5786 111.25 0.205077 11.9325 sd 4.06942 3.50785 0.960933 12.4919 4.23081 0.047011 3.16324 cv 0.193979 0.416859 0.497438 0.361261 0.03803 0.229238 0.265095 mean 4.87728 16.6283 2.04608 109.088 327.914 9.17512 24.8717 sd 0.811619 2.70717 0.396622 26.9177 86.8 4.62424 8.42407 cv 0.166408 0.162805 0.193845 0.246752 0.264704 0.503997 0.338701 mean 9.56548 16.487 2.3267 41.3839 641.251 1.14262 26.8398 sd 0.507573 2.25471 0.17108 2.01662 41.9264 0.255904 2.62456 cv 0.053063 0.136757 0.073529 0.04873 0.065382 0.223964 0.097786 mean 19.5219 19.6896 0.958714 76.4302 23384.5 1.62391 104.155 sd 0.716404 1.17574 0.086813 5.228 2105.92 0.171645 3.20157 cv 0.036697 0.059714 0.090551 0.068402 0.090056 0.105698 0.030739 mean 17.5993 27.6222 1.79453 80.7875 1229.32 10.8719 69.6603 sd 3.48878 9.5167 0.236688 10.948 179.918 3.53416 7.04181 cv 0.198233 0.344531 0.131894 0.135517 0.146355 0.325071 0.101088 mean 11.5088 22.3062 1.22008 65.072 5094.02 1.65348 84.6607 sd 0.866132 2.80796 0.224815 8.69547 540.559 0.187462 6.66376 cv 0.075258 0.125882 0.184262 0.133628 0.106116 0.113374 0.078711 mean 14.7525 36.8807 0.775817 51.4831 1103.18 1.35724 66.336 sd 0.82225 3.6595 0.204319 14.3452 359.699 0.103383 8.38324 cv 0.055736 0.099225 0.26336 0.278639 0.326058 0.076171 0.126375 mean 18.5591 19.2764 1.57754 37.1887 2526.08 1.33864 72.7314 sd 1.96455 3.6206 0.210905 1.4883 186.921 0.189591 7.25755 cv 0.105853 0.187826 0.133692 0.04002 0.073997 0.141629 0.099786 mean 15.7746 24.7514 2.04442 132.185 1087.89 0.61024 40.0555 sd 3.57755 3.52401 0.316487 7.96033 62.0473 0.159218 5.53228 cv 0.226791 0.142376 0.154805 0.060221 0.057034 0.26091 0.138115 mean 13.6316 18.9032 2.11345 93.1759 4143.55 7.85112 68.7122 166 27 28 29 30 31 32 33 34 35 36 37 38 39 sd 0.653964 1.2307 0.37859 8.30856 463.121 0.942478 12.3941 cv 0.047974 0.065105 0.179133 0.089171 0.111769 0.120044 0.180376 mean 21.2293 21.3318 -0.42835 86.7208 5170.95 13.8986 87.263 sd 2.48159 3.82321 1.67591 7.19178 824.468 3.69736 3.63924 cv 0.116894 0.179226 -3.9125 0.08293 0.159442 0.266025 0.041704 mean 25.7328 19.9213 0.369942 84.5193 29804 2.21742 124.064 sd 0.491372 1.18217 0.106208 10.6986 1937.51 0.259573 4.05199 cv 0.019095 0.059342 0.287095 0.126582 0.065008 0.117061 0.032661 mean 28.9937 15.2498 2.38563 92.9205 797.061 1.23814 18.1548 sd 2.33443 9.65758 0.65697 14.5685 42.9414 0.296445 4.49574 cv 0.080515 0.633292 0.275387 0.156784 0.053875 0.239427 0.247633 mean 20.473 27.2118 0.040248 113.842 3721.61 19.1302 102.156 sd 1.76799 2.86466 0.575371 8.61022 233.034 1.45744 7.42307 cv 0.086357 0.105273 14.2957 0.075633 0.062616 0.076185 0.072664 mean 6.87549 17.9864 1.63042 75.9572 2749.19 11.2459 68.063 sd 1.20223 2.12585 0.176687 6.3881 384.211 1.61391 6.2563 cv 0.174858 0.118192 0.108369 0.084101 0.139754 0.143511 0.091919 mean 11.4029 20.3264 1.35452 57.7099 1427.55 1.6521 59.8904 sd 0.883198 1.86446 0.265975 6.31914 128.305 0.467446 5.73427 cv 0.077454 0.091726 0.196361 0.109498 0.089878 0.282941 0.095746 mean 11.6855 18.5573 1.88461 48.8143 1430.57 5.37633 84.7587 sd 0.815922 1.62855 0.01838 9.74747 140.471 1.11969 4.26764 cv 0.069824 0.087758 0.009753 0.199685 0.098192 0.208263 0.05035 mean 9.53234 16.7271 0.586959 67.046 2251.46 3.43119 57.8212 sd 0.812294 1.38056 0.289837 6.27545 197.775 1.55173 6.98992 cv 0.085215 0.082534 0.493794 0.093599 0.087843 0.452241 0.120889 mean 12.2206 21.7323 2.71886 37.969 133.821 2.93678 19.0699 sd 3.0908 2.64515 0.167709 8.51609 17.8642 0.886365 6.65436 cv 0.252916 0.121715 0.061683 0.224291 0.133493 0.301816 0.348947 mean 16.3168 18.797 0.729622 124.774 2151.39 21.2392 82.3354 sd 0.702822 3.37588 0.189855 5.33157 126.074 2.12752 2.45109 cv 0.043074 0.179596 0.26021 0.04273 0.058601 0.10017 0.02977 mean 21.6368 18.5827 0.296115 73.1769 23706.6 1.69552 118.364 sd 0.86117 1.01764 0.069015 6.67724 2971.43 0.126147 6.84617 cv 0.039801 0.054763 0.233068 0.091248 0.125342 0.0744 0.05784 mean 23.4199 19.0948 0.521646 51.2317 22342.1 2.33183 110.985 sd 0.376108 1.16928 0.196779 3.7661 1478.71 0.290813 2.15859 cv 0.016059 0.061236 0.377226 0.073511 0.066185 0.124715 0.019449 mean 10.6156 25.6417 2.31197 95.0429 4274.62 1.64203 50.1403 sd 1.91144 4.44612 0.362241 4.49161 320.921 0.773328 4.54959 167 40 41 42 43 44 45 46 47 48 67 68 69 cv 0.180059 0.173394 0.156681 0.047259 0.075076 0.470958 0.090737 mean 14.5866 19.5067 3.35412 100.109 325.024 12.8489 39.5613 sd 2.218 2.43925 0.345708 12.1711 17.7645 2.50063 11.5552 cv 0.152058 0.125047 0.10307 0.121578 0.054656 0.194618 0.292082 mean 12.0624 23.5383 -1.17762 69.39 755.54 4.17602 82.0465 sd 4.02413 6.66125 0.281897 14.2607 226.393 0.868289 3.68947 cv 0.333609 0.282996 -0.23938 0.205515 0.299643 0.207923 0.044968 mean 19.0371 19.6702 0.070671 65.51 22923.4 1.24276 100.236 sd 0.492858 1.74087 0.139735 12.4823 1209.3 0.148047 1.91954 cv 0.025889 0.088503 1.97726 0.190541 0.052754 0.119128 0.01915 mean 11.5597 24.8279 2.38102 90.6482 265.314 4.69627 40.0177 sd 1.33857 4.68379 0.17624 15.5321 25.1988 2.54041 4.89468 cv 0.115797 0.18865 0.074019 0.171345 0.094977 0.540941 0.122313 mean 16.4049 21.7061 0.453162 53.4257 12176.5 5.9806 95.4422 sd 1.10756 2.97348 0.156027 5.77054 1725.29 1.22096 5.21116 cv 0.067514 0.136988 0.344307 0.108011 0.14169 0.204153 0.0546 mean 16.5625 41.2187 0.603475 114.42 3793.32 22.3237 102.769 sd 1.20113 9.08227 0.327346 10.3851 453.194 3.37985 5.14177 cv 0.072521 0.220343 0.542436 0.090763 0.119472 0.151402 0.050033 mean 7.50396 17.4482 2.39613 56.7059 1715.49 2.69878 42.5794 sd 1.83162 2.14764 0.088201 12.0552 81.2276 1.04068 9.57321 cv 0.244087 0.123087 0.03681 0.212591 0.04735 0.38561 0.224832 mean 12.0455 18.9705 2.33858 68.3145 156.098 1.06406 17.7064 sd 3.85443 6.02907 0.121343 14.7861 28.9987 0.155 0.00191 cv 0.31999 0.317813 0.051888 0.216441 0.185772 0.145669 0.000108 mean 6.83651 15.6695 2.19335 52.7211 374.796 0.151941 23.5614 sd 0.713675 3.48941 0.456946 10.3353 19.8675 0.093105 8.74067 cv 0.104392 0.222688 0.208332 0.196038 0.053009 0.61277 0.370974 sd 1.09483 5.30118 1.10031 11.8221 458.402 0.552503 1.97807 sd 1.08828 1.86831 0.089219 6.64499 15.956 1.21331 5.32471 sd 1.09831 3.40521 0.33441 7.61366 1860.36 0.440532 3.41399 sd 5.8598 2.49278 0.898631 4.26503 2143.24 0.384535 3.85514 mean 18.301 17.5456 1.0954 93.3013 288.01 2.74152 86.4125 sd 0.913292 4.12009 0.26796 17.4904 36.9912 2.50996 2.11363 mean 20.5765 24.1892 -0.86804 98.9177 3847.17 2.65564 93.9333 sd 2.38037 5.64778 0.417875 8.00926 1245.7 0.894807 3.96282 cv 0.115684 0.233483 -0.4814 0.080969 0.323798 0.336946 0.042188 mean 27.7608 44.5959 1.2981 148.968 432.492 2.91616 33.069 sd 3.37177 14.2584 0.510236 13.1696 34.1946 0.605669 2.75236 cv 0.121458 0.319725 0.393065 0.088405 0.079064 0.207694 0.083231 168 70 71 72 73 74 75 76 77 78 79 80 81 82 83 mean 20.9817 22.2408 -0.61799 110.016 3814.4 3.74905 97.8927 sd 1.75563 2.41387 0.14336 10.9967 1029.13 0.864201 5.2735 cv 0.083675 0.108533 -0.23198 0.099955 0.269801 0.230512 0.05387 mean 7.47654 17.6256 2.92197 60.138 246.208 3.5177 21.3093 sd 1.64868 5.38158 0.132182 12.2739 10.2263 1.28993 5.22538 cv 0.220514 0.305328 0.045237 0.204095 0.041535 0.366698 0.245216 mean 11.7079 27.5554 2.14873 202.902 4165.92 6.68725 66.4468 sd 1.04394 9.04444 0.328375 12.0127 435.528 1.41933 4.25214 cv 0.089165 0.328227 0.152823 0.059205 0.104545 0.212243 0.063993 mean 20.6158 32.2242 1.79219 163.864 2521.3 63.2084 61.4559 sd 2.9512 7.70808 0.285265 7.6443 577.648 2.26574 15.1143 cv 0.143152 0.239202 0.159171 0.04665 0.229107 0.035846 0.245937 mean 10.617 22.8277 2.85662 65.242 255.369 2.17923 22.9108 sd 2.30409 2.95384 0.161749 7.17996 27.6244 0.995877 6.78629 cv 0.21702 0.129397 0.056623 0.110051 0.108174 0.456985 0.296205 mean 20.0313 22.4216 0.692208 177.53 9577.82 19.7521 94.177 sd 0.840468 4.34752 0.151942 12.9385 692.744 3.8797 6.92894 cv 0.041958 0.193898 0.219503 0.07288 0.072328 0.196419 0.073574 mean 13.4758 24.0232 0.958475 124.122 3878.39 17.8565 78.4725 sd 0.636761 1.78346 0.18948 6.70169 518.04 1.98502 8.18433 cv 0.047252 0.074239 0.197689 0.053993 0.133571 0.111165 0.104295 mean 10.8041 19.7136 1.24522 58.6709 5779.85 1.69763 77.7351 sd 0.75357 1.46839 0.26185 3.69435 486.404 0.334556 9.31925 cv 0.069748 0.074486 0.210285 0.062967 0.084155 0.197073 0.119885 mean 17.7198 20.6231 -1.33331 129.577 384.143 4.13556 84.6822 sd 5.22484 5.5392 0.420015 10.192 82.3036 0.481505 3.40322 cv 0.294858 0.268593 -0.31502 0.078656 0.214252 0.116431 0.040188 mean 14.6489 29.4345 1.15946 123.12 502.202 6.27257 75.5226 sd 1.92596 4.58576 0.217571 15.1706 84.2384 3.56907 13.839 cv 0.131474 0.155796 0.187648 0.123218 0.167738 0.568996 0.183243 mean 18.0727 24.9537 1.31943 63.1439 1409.15 7.15813 44.1469 sd 0.739108 3.21457 0.211256 7.686 167.543 2.25355 6.65922 cv 0.040896 0.128821 0.160112 0.121722 0.118897 0.314824 0.150842 mean 9.13493 21.8378 2.51325 61.8345 259.689 1.80666 10.1813 sd 1.63103 4.46656 0.369135 16.6823 52.2234 0.346773 4.60386 cv 0.178549 0.204533 0.146875 0.26979 0.2011 0.191942 0.452189 mean 24.4536 20.6073 1.96663 95.2278 2209.03 7.786 58.7932 sd 4.88074 2.10445 0.627523 9.69037 243.197 1.3104 1.58761 cv 0.199592 0.102121 0.319086 0.10176 0.110093 0.168301 0.027003 mean 8.98632 20.4812 2.19295 51.5756 224.897 3.60374 40.4486 169 84 85 86 87 88 89 90 91 92 93 94 95 96 sd 0.360378 1.18336 0.243086 6.95932 14.5186 1.27386 3.85992 cv 0.040103 0.057778 0.110849 0.134934 0.064556 0.353482 0.095428 mean 23.4588 20.7225 0.48453 125.688 23748.2 2.87599 122.501 sd 1.17676 1.33381 0.204608 9.13737 1824.71 0.184161 6.34301 cv 0.050163 0.064365 0.422283 0.072699 0.076836 0.064034 0.051779 mean 17.7473 21.815 1.1946 60.3284 13608.3 4.33339 116.239 sd 0.434834 1.26694 0.459207 4.43335 1097.27 0.657277 3.44375 cv 0.024501 0.058076 0.384401 0.073487 0.080633 0.151677 0.029627 mean 11.1182 26.1844 1.58468 75.8354 772.178 2.62125 59.4081 sd 0.876866 3.35344 0.335536 12.652 68.4221 0.577661 7.57408 cv 0.078867 0.12807 0.211738 0.166835 0.088609 0.220376 0.127492 mean 12.827 12.2458 3.53686 41.3743 167.608 1.15323 8.14356 sd 1.15625 3.41949 0.090767 1.95122 4.10957 0.319332 1.79452 cv 0.090142 0.279237 0.025663 0.04716 0.024519 0.276903 0.220361 mean 20.8908 19.7823 0.63398 72.533 37721.7 1.50032 114.671 sd 1.09802 2.17273 0.1504 2.71022 2959.59 0.202748 2.51336 cv 0.05256 0.109832 0.237231 0.037365 0.078459 0.135137 0.021918 mean 22.3624 16.2363 1.963 92.0209 8594.6 2.10691 82.1651 sd 2.307 3.62371 0.333657 6.36487 716.93 0.50274 7.43324 cv 0.103164 0.223186 0.169972 0.069168 0.083416 0.238614 0.090467 mean 9.83309 16.6097 2.36976 33.7594 556.074 0.790468 28.9135 sd 1.68706 2.17423 0.11945 3.30849 45.9323 0.138026 2.37609 cv 0.171569 0.130901 0.050406 0.098002 0.082601 0.174613 0.082179 mean 13.332 19.3801 1.87916 149.61 4051.77 6.38422 68.973 sd 0.995936 3.49966 0.136885 21.2587 492.347 1.70742 1.64245 cv 0.074702 0.18058 0.072844 0.142094 0.121514 0.267443 0.023813 mean 11.1903 19.3675 2.04965 101.291 1391.21 1.37573 60.194 sd 0.845638 2.10517 0.162739 13.9686 67.7778 0.296669 9.02555 cv 0.075569 0.108696 0.079399 0.137906 0.048719 0.215645 0.149941 mean 10.0947 20.5246 1.48068 36.8532 2167.93 1.62741 88.3268 sd 0.470443 2.45015 0.212417 6.83543 220.591 0.272113 7.42862 cv 0.046603 0.119376 0.143458 0.185477 0.101752 0.167206 0.084104 mean 11.5546 18.6979 1.99584 99.5123 1011.35 2.79111 80.7318 sd 1.38359 3.51075 0.124638 9.43789 93.6014 0.710671 3.92538 cv 0.119743 0.187762 0.062449 0.094841 0.092551 0.25462 0.048622 mean 18.8711 20.5733 -0.0849 62.8481 4567.1 3.47219 100.572 sd 0.384404 2.43995 0.208536 13.161 743.705 1.20648 2.30499 cv 0.02037 0.118598 -2.45641 0.209409 0.16284 0.347469 0.022919 mean 19.4945 24.1237 0.455807 66.4693 10678.7 5.07291 103.412 sd 1.18513 2.09546 0.167643 3.17596 728.868 0.283149 5.28034 170 97 98 99 100 101 102 103 104 105 106 107 108 109 cv 0.060793 0.086863 0.367794 0.047781 0.068254 0.055816 0.051061 mean 9.90209 21.6317 -0.41153 70.3327 1944.79 1.23713 83.39 sd 2.96115 2.93407 0.463426 8.42532 324.203 0.435458 3.50994 cv 0.299043 0.135637 -1.1261 0.119792 0.166703 0.351992 0.042091 mean 17.5166 18.2762 -0.33038 57.2026 1989.01 1.46905 86.5506 sd 1.77692 1.85168 0.153328 6.54141 446.761 0.401505 3.63071 cv 0.101442 0.101317 -0.4641 0.114355 0.224615 0.27331 0.041949 mean 11.4375 17.7347 4.10372 33.8483 236.547 1.38521 12.8841 sd 1.23358 2.8284 3.46797 2.89629 27.1356 0.606779 2.84453 cv 0.107854 0.159484 0.845078 0.085567 0.114715 0.43804 0.220778 mean 12.4372 23.5934 2.65053 65.8224 481.167 3.58351 18.8719 sd 1.57244 3.80282 0.044116 3.77349 31.8392 0.227731 3.97935 cv 0.12643 0.161181 0.016644 0.057328 0.066171 0.06355 0.210862 mean 25.3076 28.6388 1.07678 169.992 7137.54 36.2164 109.095 sd 4.02767 9.64901 1.13492 34.6501 574.223 4.1758 4.70964 cv 0.159149 0.336921 1.05399 0.203833 0.080451 0.115301 0.04317 mean 13.078 9.49561 2.32539 50.2845 199.723 4.34734 28.8779 sd 2.57445 4.50561 1.5851 10.3177 36.6473 2.40149 3.89482 cv 0.196853 0.474494 0.68165 0.205187 0.183491 0.552404 0.134872 mean 10.4461 29.4811 2.26328 412.402 23291.7 4.95571 . sd 1.15734 6.58478 1.58487 35.3733 3248.27 0.45883 . cv 0.110792 0.223356 0.700254 0.085774 0.13946 0.092586 . mean 20.4751 28.0973 0.071618 143.417 4137.89 2.73213 90.6233 sd 1.71643 3.67059 0.127478 19.9674 747.704 0.425487 4.37969 cv 0.08383 0.130638 1.77996 0.139227 0.180697 0.155735 0.048329 mean 18.6437 24.7355 0.112197 113.024 10387.9 5.24498 99.6393 sd 0.46881 1.67614 0.199152 13.0702 1621.75 0.433244 5.95012 cv 0.025146 0.067762 1.77502 0.115641 0.156118 0.082602 0.059717 mean 27.7039 8.26987 2.63535 80.6257 1116.46 1.50622 27.2568 sd 6.67111 3.11071 0.163767 19.3411 180.457 0.728636 4.26091 cv 0.240801 0.37615 0.062143 0.239887 0.161633 0.483751 0.156324 mean 18.9144 16.4954 1.71595 54.009 3180.81 2.89666 91.0126 sd 0.535032 1.55998 0.623168 6.46298 248.698 0.648726 4.82587 cv 0.028287 0.094571 0.363161 0.119665 0.078187 0.223956 0.053024 mean 17.6296 25.8875 1.0221 54.9416 14397.5 4.98265 114.635 sd 0.448271 3.30751 0.630205 4.81186 1429.99 0.408463 3.59753 cv 0.025427 0.127765 0.61658 0.087581 0.099322 0.081977 0.031382 mean 11.7852 23.9995 0.852549 77.8503 887.491 2.74522 84.2522 sd 1.99036 2.40865 0.262666 5.0468 130.73 0.620706 4.25156 cv 0.168886 0.100363 0.308094 0.064827 0.147303 0.226104 0.050462 171 110 111 112 113 114 115 116 117 118 119 120 132 133 mean 19.7492 45.5569 1.41201 117.809 7576.18 23.0114 96.5328 sd 1.03158 4.83204 2.1669 7.68855 614.351 4.10418 7.16054 cv 0.052234 0.106066 1.53462 0.065263 0.08109 0.178354 0.074177 mean 18.5515 25.4247 1.255 123.746 4265.19 39.4625 79.6027 sd 2.1562 3.67638 0.360908 8.91759 263.984 3.72874 8.05667 cv 0.116228 0.144599 0.287575 0.072064 0.061893 0.094488 0.101211 mean 21.1091 31.7215 0.066907 116.352 3221.49 23.4204 72.0802 sd 3.67373 2.29052 0.100933 7.87706 553.873 1.66433 3.46599 cv 0.174036 0.072207 1.50857 0.0677 0.171931 0.071063 0.048085 mean 9.73696 13.4042 2.28357 31.4436 376.872 0.19501 29.6355 sd 4.24886 4.57374 0.227883 10.3078 62.7195 0.235875 4.72827 cv 0.436364 0.341216 0.099792 0.327819 0.166421 1.20955 0.159548 mean 28.9872 20.4681 1.31836 64.8432 2096.26 1.7951 75.0105 sd 5.12793 5.11675 0.132382 11.3685 211.524 0.718718 2.90148 cv 0.176903 0.249987 0.100414 0.175323 0.100906 0.400378 0.038681 mean 17.2788 17.0705 1.49005 167.959 1371.04 2.68639 45.9608 sd 2.14273 2.54628 0.694998 25.3354 112.309 1.01493 3.75606 cv 0.124009 0.149162 0.466425 0.150842 0.081915 0.377804 0.081723 mean 26.6952 16.8787 0.315302 83.2166 27869.9 2.20928 128.08 sd 0.475509 1.00279 0.228516 7.83289 2984.01 0.392215 22.6044 cv 0.017813 0.059412 0.724753 0.094127 0.107069 0.177531 0.176487 mean 11.5617 21.7642 0.588501 81.4056 34509.8 3.1629 93.5177 sd 0.318729 0.794705 0.189986 9.02344 1658.82 0.2049 1.74636 cv 0.027568 0.036514 0.32283 0.110845 0.048068 0.064782 0.018674 mean 12.3913 21.9464 2.804 69.2306 1213.11 6.02557 54.0596 sd 1.50212 2.58161 0.404312 6.77765 32.9705 1.09719 13.3768 cv 0.121224 0.117632 0.144191 0.0979 0.027178 0.182089 0.247445 mean 11.9666 17.0504 2.69407 46.7814 285.015 6.08643 5.73342 sd 3.04034 1.4991 0.149799 7.35199 33.403 1.26708 0.386603 cv 0.254069 0.087922 0.055603 0.157156 0.117197 0.208182 0.06743 mean 11.2417 28.028 0.902641 119.833 2143.65 7.38874 70.9478 sd 0.945592 6.90913 0.208829 23.1127 235.27 1.05146 9.12302 cv 0.084115 0.246508 0.231354 0.192875 0.109752 0.142306 0.128588 mean 14.6275 22.259 3.13216 82.318 520.501 0.897612 44.5039 sd 1.43512 3.60932 0.357464 11.2464 27.77 0.340939 2.27987 cv 0.098111 0.162151 0.114127 0.136622 0.053352 0.379829 0.051229 mean 13.4032 18.288 2.53685 70.4469 328.024 1.69756 26.8694 sd 3.3372 4.85459 0.259898 5.45893 21.6676 0.912007 7.03199 cv 0.248986 0.265453 0.102449 0.07749 0.066055 0.537245 0.26171 172 10.2 Some statistics of variables by Country code. countrycode stats GDPPER TOA POP Trade netsize pca health 1 mean 1257.71 493846 3.10E+06 59.7779 3.112304 -1.27024 5.56163 sd 263.853 298682 33603.3 12.2184 5.449923 0.47567 1.13396 cv 0.209788 0.604808 0.010803 0.204396 1.75109 -0.37447 0.20389 mean 1882.41 1.00E+06 3.10E+07 59.7993 0.343345 -2.2719 3.65913 sd 178.356 403089 1.80E+06 8.29713 0.454276 0.637466 0.353496 cv 0.094748 0.398007 0.057927 0.13875 1.323088 -0.28059 0.096607 mean 10972.4 233846 78328.8 133.846 36.88907 1.780878 4.76853 sd 1395.45 16051.2 5666.09 26.2696 44.6108 0.159924 0.22075 cv 0.127178 0.06864 0.072337 0.196267 1.209323 0.089801 0.046293 mean 7749.95 3.10E+06 3.70E+07 31.5057 1.326459 -0.45807 8.51484 sd 770.334 665802 1.50E+06 10.9431 1.219623 0.80002 0.471647 cv 0.099399 0.211438 0.039078 0.347338 0.919458 -1.74649 0.055391 mean 796.044 164000 3.10E+06 73.691 2.73111 -1.02364 5.47235 sd 311.285 162713 50516.4 7.45298 2.583861 0.330873 0.751657 cv 0.39104 0.992155 0.016309 0.101138 0.946085 -0.32323 0.137355 mean 21923.7 4.80E+06 1.90E+07 39.4587 1.015873 3.907328 8.0531 sd 2030.44 593921 946772 1.93975 0.681519 0.245808 0.476005 cv 0.092614 0.124333 0.048659 0.049159 0.67087 0.062909 0.059108 mean 23811.3 1.80E+07 8.10E+06 90.7033 34.57082 4.045174 10.0713 sd 1920.1 1.30E+06 124250 12.9917 22.59921 0.075942 0.285589 cv 0.080638 0.0724 0.015373 0.143233 0.653708 0.018774 0.028357 mean 869.937 740000 8.10E+06 90.6847 3.829124 -2.25242 5.79957 sd 449.582 138071 278813 17.3779 5.779533 0.135469 1.13023 cv 0.516798 0.186582 0.034352 0.19163 1.509362 -0.06014 0.194881 mean 13042.2 2.80E+06 666817 153.866 130.515 0.591999 4.22562 2 5 6 7 9 10 11 13 173 14 15 16 17 18 19 20 22 23 sd 1239.35 1.20E+06 101629 13.3527 124.3686 0.486209 0.40537 cv 0.095026 0.434577 0.152409 0.086781 0.952907 0.8213 0.095931 mean 382.485 205769 1.30E+08 35.0555 3.741216 -2.11252 3.10434 sd 54.7854 40462.2 8.80E+06 5.54961 6.169389 0.512719 0.290905 cv 0.143235 0.196639 0.066576 0.158309 1.649033 -0.2427 0.093709 mean 9100.12 515923 267883 114.586 155.9405 3.032564 6.59338 sd 499.717 42496 2655.87 7.55033 165.6894 0.162782 0.480069 cv 0.054913 0.082369 0.009914 0.065892 1.062516 0.053678 0.072811 mean 1440.17 129308 1.00E+07 125.159 2.501504 -2.51029 6.44224 sd 423.115 94380.8 162699 13.3948 3.532511 0.173319 0.266275 cv 0.293796 0.729893 0.016347 0.107022 1.412155 -0.06904 0.041333 mean 22549.3 6.40E+06 1.00E+07 144.395 114.3548 3.316202 8.73093 sd 1620.84 443983 155882 11.2168 87.23407 0.146342 0.97777 cv 0.07188 0.068884 0.015098 0.077682 0.762837 0.044129 0.111989 mean 3316.82 195846 260260 114.91 0.611515 0.142528 4.25023 sd 383.004 41323.6 30545.5 9.15891 0.494484 0.35584 0.374913 cv 0.115473 0.211 0.117365 0.079705 0.808622 2.496642 0.08821 mean 343.926 139077 6.80E+06 43.1867 0.467921 -0.57548 4.54505 sd 19.0152 41298.6 802853 4.64414 0.460952 0.275765 0.16488 cv 0.055289 0.296948 0.118196 0.107536 0.985105 -0.47919 0.036277 mean 56073.7 321538 62648.6 65.6738 532.7282 2.796605 . sd 7449.51 49410.9 816.865 2.79051 251.9071 0.126592 . cv 0.132852 0.15367 0.013039 0.04249 0.472862 0.045266 . mean 1026 398077 8.50E+06 54.951 0.250062 -1.03186 5.36249 sd 50.2774 96429.1 650492 10.8612 0.265296 0.67737 0.742233 cv 0.049003 0.242237 0.076758 0.197653 1.060922 -0.65645 0.138412 mean 1495.6 183091 3.60E+06 108.266 6.098415 -1.21504 8.50939 174 24 25 27 28 29 30 32 33 37 sd 435.243 56339.1 191618 10.4228 8.065843 0.280731 1.30344 cv 0.291015 0.307711 0.052766 0.09627 1.322613 -0.23105 0.153177 mean 3343.67 1.10E+06 1.80E+06 86.6267 0.078551 1.771666 5.41527 sd 559.911 415769 107987 6.59201 0.058298 0.155446 1.37132 cv 0.167454 0.376157 0.060955 0.076097 0.742168 0.08774 0.253233 mean 3791.38 4.30E+06 1.80E+08 22.3619 2.249375 -0.0253 7.31086 sd 213.946 1.10E+06 9.20E+06 5.15551 2.471454 0.216654 0.637848 cv 0.05643 0.260965 0.052175 0.230549 1.098729 -8.56316 0.087247 mean 1783.98 3.70E+06 8.00E+06 113.653 7.701015 0.403058 6.41632 sd 374.673 984517 273327 13.075 7.666138 0.488982 1.04477 cv 0.210021 0.268835 0.034064 0.115043 0.995471 1.21318 0.16283 mean 220.627 173615 1.30E+07 35.5758 0.136318 -0.87295 5.53168 sd 24.5342 59773.9 1.40E+06 3.65035 0.138692 0.152613 1.12216 cv 0.111202 0.344289 0.111176 0.102608 1.017414 -0.17482 0.202861 mean 112.986 67333.3 6.70E+06 34.5786 0.612361 -3.46233 7.37471 sd 4.16446 61485.2 544027 12.4919 0.761742 0.44919 2.23175 cv 0.036858 0.913146 0.081518 0.361261 1.243943 -0.12974 0.302621 mean 331.32 1.40E+06 1.30E+07 109.087 0.153243 -2.08315 6.70815 sd 91.9266 377184 805063 26.9167 0.13835 0.153225 1.21266 cv 0.277455 0.260847 0.064098 0.246744 0.902814 -0.07355 0.180773 mean 23385.5 1.90E+07 3.10E+07 75.7626 1.625719 3.994459 9.36351 sd 2105.67 1.00E+06 1.10E+06 5.54791 0.932257 0.10815 0.478814 cv 0.090042 0.054871 0.036785 0.073228 0.573443 0.027075 0.051136 mean 1257.41 124462 442965 93.2919 4.282392 0.953644 4.99396 sd 267.609 78132 29031.4 15.5096 3.24691 0.297853 0.204721 cv 0.212825 0.62776 0.065539 0.166248 0.7582 0.312331 0.040994 mean 5088.94 1.80E+06 1.60E+07 65.0904 3.677139 -0.27842 7.00143 175 38 39 43 45 47 48 49 52 53 sd 537.306 314656 717009 8.72498 3.050929 0.149982 0.49997 cv 0.105583 0.177171 0.046041 0.134044 0.829702 -0.5387 0.07141 mean 1110.46 3.40E+07 1.30E+09 50.3442 6.135901 -1.762 4.42092 sd 372.861 1.10E+07 3.60E+07 13.5054 6.885508 0.280512 0.397345 cv 0.335772 0.323313 0.028714 0.268261 1.122167 -0.1592 0.089878 mean 2661.38 857385 4.00E+07 35.7285 2.476009 -2.27071 6.88329 sd 176.22 450896 2.60E+06 1.42641 2.85076 0.249977 1.28206 cv 0.066214 0.525897 0.063428 0.039923 1.151353 -0.11009 0.186258 mean 4155.83 1.20E+06 4.00E+06 93.1591 10.27146 0.800534 7.34551 sd 467.516 386790 320854 8.28845 8.981618 0.483335 0.662861 cv 0.112496 0.319053 0.080536 0.088971 0.874425 0.603766 0.09024 mean 5183.78 5.80E+06 4.50E+06 86.3515 13.03346 2.241868 6.90387 sd 836.003 2.20E+06 73291 6.67613 12.01281 0.169873 0.523698 cv 0.161273 0.384479 0.016345 0.077313 0.92169 0.075773 0.075856 mean 13448.9 2.30E+06 961190 102.259 21.8774 2.511885 5.86735 sd 1127.03 219052 68335.8 4.31341 16.23783 0.198812 0.559325 cv 0.083801 0.09326 0.071095 0.042181 0.742219 0.079148 0.095328 mean 5905.55 9.10E+06 1.00E+07 125.989 27.58109 -3.69779 6.61476 sd 760.945 426385 45428.2 17.2163 24.28686 0.481748 0.263172 cv 0.128852 0.047018 0.004424 0.13665 0.880562 -0.13028 0.039786 mean 29759.9 4.40E+06 5.40E+06 84.2756 61.81909 4.462629 9.0458 sd 1889.2 2.90E+06 69792.2 10.6069 41.18372 0.154454 0.715487 cv 0.063481 0.660194 0.013037 0.12586 0.666198 0.034611 0.079096 mean 2799.86 2.90E+06 8.70E+06 75.7776 11.30878 -0.70775 5.70578 sd 399.464 735115 523508 6.62874 11.66992 0.196822 0.514846 cv 0.142673 0.252084 0.059991 0.087476 1.031934 -0.2781 0.090232 mean 1417.43 666231 1.30E+07 58.0563 1.584681 -1.72843 5.63349 176 54 55 58 59 60 61 62 66 67 sd 117.044 163791 803436 6.71866 1.662597 0.407584 1.49743 cv 0.082575 0.245847 0.063832 0.115727 1.049168 -0.23581 0.265809 mean 1483.46 5.60E+06 6.90E+07 48.8142 3.12817 -1.16406 5.16103 sd 150.431 2.40E+06 4.90E+06 9.74732 4.285751 0.299877 0.59661 cv 0.101405 0.431837 0.070238 0.199682 1.370051 -0.25761 0.115599 mean 2253.11 757615 5.90E+06 66.5743 6.725471 -0.58826 7.41433 sd 198.428 351284 112078 5.74906 6.01894 0.198714 0.658446 cv 0.088069 0.46367 0.018846 0.086356 0.894947 -0.3378 0.088807 mean 4781.71 1.30E+06 1.40E+06 152.642 10.5338 2.331383 5.43565 sd 1436.45 504191 29653.3 9.94394 7.550958 0.355753 0.598217 cv 0.300406 0.395539 0.021608 0.065146 0.716831 0.152593 0.110054 mean 134.161 170077 6.70E+07 37.9451 0.075178 -2.41469 4.03192 sd 18.6462 68010.4 6.70E+06 8.49555 0.097114 0.241806 0.720579 cv 0.138983 0.39988 0.099646 0.223891 1.291792 -0.10014 0.178718 mean 2139.69 415923 809714 122.523 1.887113 -0.30227 3.42895 sd 140.083 90460.4 17132.2 6.04786 1.847529 0.476261 0.274671 cv 0.065469 0.217493 0.021158 0.049361 0.979024 -1.57561 0.080104 mean 23709.8 2.90E+06 5.20E+06 73.8531 8.284219 3.947882 7.84249 sd 3048.07 343919 55098.9 6.79653 4.463428 0.221629 0.402769 cv 0.128557 0.118642 0.010612 0.092028 0.538787 0.056139 0.051357 mean 21634.8 7.30E+07 6.10E+07 51.674 27.71329 3.21376 10.5713 sd 1341.01 6.20E+06 1.50E+06 3.83068 23.04148 0.105253 0.420312 cv 0.061984 0.085465 0.024236 0.074132 0.831423 0.032751 0.03976 mean 790.556 421462 4.40E+06 69.3823 1.586449 -1.67267 7.62329 sd 215.569 289742 121104 14.249 1.911713 0.784086 1.09703 cv 0.272681 0.687469 0.027248 0.20537 1.205026 -0.46876 0.143905 mean 22859.7 1.90E+07 8.20E+07 65.6458 89.09857 3.78207 10.5056 177 69 70 72 76 77 78 79 80 81 sd 1252.81 3.00E+06 262750 12.8931 66.0352 0.133652 0.232455 cv 0.054804 0.162601 0.003195 0.196403 0.741148 0.035338 0.022127 mean 12022.8 1.30E+07 1.10E+07 55.1727 11.81948 1.924874 8.89066 sd 1660.05 2.30E+06 169961 6.14528 10.14478 0.201384 0.562616 cv 0.138076 0.175667 0.015535 0.111383 0.85831 0.104622 0.063282 mean 4800.59 121231 101806 96.6438 30.17459 1.198515 6.75395 sd 747.978 12404 937.261 17.1113 27.89061 0.181231 0.636779 cv 0.15581 0.102317 0.009206 0.177055 0.924308 0.151213 0.094282 mean 1714.22 936231 1.20E+07 56.7649 3.257609 -1.18762 5.55174 sd 81.5219 362495 1.10E+06 12.1122 3.32857 0.098576 1.36269 cv 0.047556 0.387185 0.093987 0.213375 1.021783 -0.083 0.245454 mean 406.49 153385 8.80E+06 48.299 8.963947 -3.31105 5.73126 sd 17.3136 72004.1 563536 9.41375 10.61051 0.478263 0.562294 cv 0.042593 0.469435 0.064246 0.194906 1.183687 -0.14444 0.09811 mean 1197.61 505154 6.40E+06 115.603 1.899313 -1.41613 6.94447 sd 100.078 188563 510847 16.6407 2.083528 0.125892 1.20465 cv 0.083565 0.373279 0.080378 0.143947 1.09699 -0.0889 0.173469 mean 26798.8 1.10E+07 6.70E+06 312.638 2255.386 3.0158 . sd 3460.53 3.60E+06 206591 58.3542 1469.72 0.651374 . cv 0.12913 0.314391 0.031061 0.186651 0.651649 0.215987 . mean 4809.94 1.00E+07 1.00E+07 129.883 20.83761 2.348663 7.60837 sd 732.19 1.60E+06 94342.6 21.5714 20.38102 0.177405 0.575976 cv 0.152224 0.155602 0.00926 0.166083 0.978088 0.075534 0.075703 mean 31617.5 723417 285101 74.4049 1.644613 4.236406 9.50851 sd 3899.31 181669 13613.2 4.04101 0.906663 0.497543 0.690739 cv 0.123328 0.251126 0.047749 0.054311 0.551293 0.117445 0.072644 mean 473.69 3.00E+06 1.10E+09 30.8067 4.712219 -0.41645 4.20106 178 82 83 85 86 87 88 89 90 91 sd 92.2845 940619 6.80E+07 8.46708 4.966889 0.116893 0.212381 cv 0.19482 0.314929 0.063468 0.274845 1.054045 -0.28069 0.050554 mean 845.641 4.90E+06 2.20E+08 62.3468 2.412919 -1.75256 2.34445 sd 78.0971 338088 1.10E+07 11.8359 2.4098 0.500428 0.402625 cv 0.092353 0.06836 0.0499 0.18984 0.998708 -0.28554 0.171736 mean 1676.47 1.40E+06 6.60E+07 44.1403 1.685623 -2.21282 4.89846 sd 232.112 638218 3.80E+06 10.2775 1.783988 0.391039 0.721545 cv 0.138453 0.447703 0.057473 0.232838 1.058355 -0.17672 0.1473 mean 25378 6.50E+06 3.90E+06 157.125 14.63343 3.749591 6.82082 sd 4768.45 997059 244110 13.8621 12.33749 0.150789 0.622571 cv 0.187897 0.152457 0.062426 0.088224 0.843103 0.040215 0.091275 mean 19202.3 1.80E+06 6.40E+06 73.5058 56.10559 1.400764 7.7487 sd 1112.69 493802 533893 8.88627 43.10313 0.369379 0.169513 cv 0.057946 0.274124 0.083491 0.120892 0.76825 0.263698 0.021876 mean 19177.1 3.80E+07 5.80E+07 50.0905 42.16419 1.907549 8.11517 sd 883.656 3.60E+06 906229 3.92461 29.82961 0.313175 0.572418 cv 0.046079 0.096213 0.015746 0.07835 0.707463 0.164177 0.070537 mean 3629.37 1.30E+06 2.60E+06 99.2994 16.46964 -0.04489 4.77418 sd 138.467 180831 60357.6 8.57974 16.62478 0.234708 0.561025 cv 0.038152 0.134609 0.02327 0.086403 1.00942 -5.22839 0.117512 mean 37356.3 5.30E+06 1.30E+08 22.8333 1292.169 2.732524 7.62525 sd 1599.08 1.50E+06 834212 4.96012 901.4214 0.282092 0.520797 cv 0.042806 0.279659 0.00657 0.217232 0.697603 0.103235 0.068299 mean 1900.43 2.10E+06 4.90E+06 122.89 3.894627 0.167764 9.05174 sd 226.688 825041 469984 13.8277 4.010527 0.201496 0.607533 cv 0.119282 0.392015 0.095354 0.112521 1.029759 1.201068 0.067118 mean 1497.25 2.80E+06 1.50E+07 87.4216 0.08195 -1.70095 4.08968 179 92 94 96 97 98 99 100 103 104 sd 458.402 812195 311589 11.8713 0.079427 0.233908 0.463019 cv 0.306162 0.293768 0.020542 0.135794 0.969208 -0.13752 0.113216 mean 418.64 1.10E+06 3.20E+07 57.687 1.302877 -1.85003 4.39138 sd 16.2752 287385 3.30E+06 6.71083 1.816844 0.201761 0.180415 cv 0.038876 0.272701 0.101079 0.116332 1.394486 -0.10906 0.041084 mean 11956.8 5.00E+06 4.70E+07 71.133 209.8312 1.518261 4.83171 sd 1860.36 940080 1.10E+06 7.61366 153.917 0.387072 0.844665 cv 0.15559 0.187244 0.023636 0.107034 0.733528 0.254944 0.174817 mean 288.331 311615 4.90E+06 93.4926 1.404593 . 5.91602 sd 35.8586 455811 226597 17.9861 1.367304 . 0.770174 cv 0.124366 1.46274 0.045938 0.19238 0.973452 . 0.130185 mean 349.033 357154 5.40E+06 72.9962 0.12417 -2.62225 4.5249 sd 61.0354 326564 363599 9.8232 0.140268 0.375608 0.9065 cv 0.17487 0.91435 0.067468 0.134571 1.12965 -0.14324 0.200336 mean 3853.68 858538 2.40E+06 99.0515 8.377557 1.430829 6.32133 sd 1238.33 390211 66678.3 8.09682 8.005681 0.482456 0.326608 cv 0.321336 0.454506 0.028194 0.081744 0.95561 0.337186 0.051668 mean 4855.97 829615 3.80E+06 59.9996 27.45424 -1.08842 9.4848 sd 265.611 272610 221151 7.87114 22.39534 0.537582 1.33083 cv 0.054698 0.328598 0.058003 0.131187 0.815733 -0.49391 0.140312 mean 388.844 191429 2.00E+06 161.171 0.813119 -0.51332 7.48511 sd 31.4475 93555.4 100040 16.7826 0.895516 0.114773 0.63912 cv 0.080874 0.488722 0.05076 0.104129 1.101335 -0.22359 0.085386 mean 3816.59 1.40E+06 3.50E+06 109.214 10.21759 1.621409 6.03884 sd 1040.16 437523 79855.6 11.0439 9.773873 0.379701 0.359874 cv 0.272535 0.314747 0.022865 0.101122 0.956573 0.234179 0.059593 mean 46086.5 844462 442230 259.042 65.22876 4.07567 6.93459 180 106 107 108 109 110 112 115 116 120 sd 6399.64 61640.9 22630.3 38.3083 51.26056 0.395144 1.09183 cv 0.138861 0.072994 0.051173 0.147884 0.785858 0.096952 0.157447 mean 1762.5 164615 2.00E+06 97.0749 9.930593 -0.89592 8.68494 sd 160.726 40504.2 27317.6 14.9141 10.67137 0.349976 0.801498 cv 0.091192 0.246054 0.01358 0.153635 1.074595 -0.39063 0.092286 mean 244.111 170077 1.60E+07 61.3255 0.094136 -0.48666 3.54342 sd 9.93462 92590 1.90E+06 13.6478 0.075935 0.215334 0.381 cv 0.040697 0.544401 0.119383 0.222546 0.806654 -0.44247 0.107523 mean 151.927 354308 1.20E+07 66.6972 0.371435 -0.92914 6.24549 sd 5.96994 175350 1.20E+06 7.98373 0.393446 0.337344 1.40889 cv 0.039295 0.494908 0.10456 0.119701 1.059261 -0.36307 0.225586 mean 4122.03 1.20E+07 2.40E+07 202.793 19.99569 1.031231 3.39892 sd 410.661 4.90E+06 2.10E+06 11.983 16.3902 0.176871 0.487254 cv 0.099626 0.418446 0.086932 0.05909 0.819687 0.171514 0.143355 mean 2485.97 470231 277176 146.995 45.47278 0.259378 6.22556 sd 587.068 113770 17722 21.3178 49.41902 0.554724 0.556015 cv 0.236152 0.241944 0.063938 0.145023 1.086782 2.138672 0.089312 mean 9675.84 1.20E+06 389595 170.657 271.1646 2.744178 7.3538 sd 692.543 52092 13877.8 10.543 219.1735 0.489488 1.2287 cv 0.071574 0.045059 0.035621 0.061779 0.808268 0.178373 0.167084 mean 3862.23 650462 1.20E+06 124.74 57.152 0.931795 3.99847 sd 482.068 132675 45859.4 7.44017 41.02165 0.150071 0.553904 cv 0.124816 0.203971 0.038339 0.059645 0.717764 0.161056 0.138529 mean 5644.19 2.00E+07 1.00E+08 58.5766 4.706617 -0.14004 5.3732 sd 439.105 1.00E+06 5.40E+06 3.7269 4.318682 0.368477 0.443452 cv 0.077798 0.051133 0.053773 0.063624 0.917577 -2.63124 0.08253 mean 521.462 211308 2.40E+06 110.823 0.011254 -0.48042 4.66378 181 121 122 123 125 126 127 129 130 131 sd 89.245 123440 101098 15.6524 0.012867 0.079599 1.0081 cv 0.171144 0.584171 0.041338 0.141237 1.14337 -0.16569 0.216154 mean 1706.85 350000 633274 101.941 14.33252 -0.50475 8.02197 sd 199.825 323768 6825.68 16.6628 2.326695 0.353157 0.624668 cv 0.117073 0.925051 0.010778 0.163455 0.162337 -0.69966 0.07787 mean 1377.18 4.50E+06 2.90E+07 63.135 4.000252 -0.38925 4.60157 sd 166.166 1.50E+06 1.30E+06 7.66403 5.592874 0.315944 0.651081 cv 0.120657 0.333266 0.045275 0.121391 1.398131 -0.81166 0.141491 mean 258.194 541143 1.90E+07 61.9214 0.09446 -0.8432 5.68628 sd 50.2104 148303 1.90E+06 17.0019 0.098907 0.249416 0.830053 cv 0.194468 0.274055 0.101585 0.274572 1.047076 -0.2958 0.145975 mean 2177.17 655231 1.90E+06 95.1923 0.053151 0.684486 6.63978 sd 213.385 170209 162231 10.416 0.047274 0.235367 0.409735 cv 0.09801 0.25977 0.084452 0.10942 0.889428 0.34386 0.061709 mean 224.538 403385 2.50E+07 51.7568 0.756364 -1.40923 5.65344 sd 13.584 68690.8 2.20E+06 7.01769 0.894769 0.502068 0.416283 cv 0.060498 0.170286 0.088818 0.13559 1.182987 -0.35627 0.073634 mean 23800.1 9.20E+06 1.60E+07 125.342 230.5424 . 8.91418 sd 1905.99 1.40E+06 330443 8.58663 138.6754 . 0.978033 cv 0.080083 0.151629 0.020673 0.068506 0.601518 . 0.109717 mean 13804.6 2.00E+06 3.90E+06 59.2755 6.741813 4.315951 7.80819 sd 1137.45 343621 179421 4.24196 3.627232 0.117914 0.549664 cv 0.082396 0.168731 0.045569 0.071563 0.53802 0.027321 0.070396 mean 780.435 512231 5.10E+06 76.1421 0.643 -1.18342 7.70208 sd 74.4397 165067 296369 13.2027 0.565109 0.187469 0.718832 cv 0.095382 0.322251 0.05779 0.173396 0.878863 -0.15841 0.09333 mean 395.694 874846 1.30E+08 76.4144 2.385228 -2.82433 5.63909 182 134 sd 41.4394 170986 1.20E+07 6.92937 3.830391 0.208232 1.01142 cv 0.104726 0.195447 0.094074 0.090682 1.605881 -0.07373 0.179359 mean 37676.7 3.30E+06 4.50E+06 72.2441 7.27702 4.170915 9.06104 sd 2891.24 508287 109568 2.51695 4.797278 0.099266 0.520509 References Adenauer I, Vagassky L (1998). ‘Aid and the real exchange rate: Dutch Disease effects in African countries,’ Intereconomics: Review of International Trade and Development 33(July/August): 177–185. Ahmed TSAA, Abdul-Kadir AR (2013). ‘Impact of information sources on the decision-making process of travel to the Egyptian tourist destination after January 25, 2011’. Tourism 61(4):395–423. Aidt TS (2003). ‘Economic Analysis of Corruption: A Survey’. The Economic Journal 113:F632–F652 [https://www.academia.edu/7868133/ECONOMIC_ANALYSIS_OF_CORRUPTION_A _SURVEY]. Al-Qudair KHA (2004). ‘The Causal Relationship between Tourism and International Trade in some Islamic Countries’. Economic Studies 5(19):1–20. [http://colleges.ksu.edu.sa/Papers/Papers/The%20relatioship%20between%20touris m%20and%20trade.pdf accessed 10 April 2014]. Alrashid SA (2012). ‘Internet Adoption in Gulf Cooperation Council’s Tourism Industry’. New Media and Mass Communication 3:36–42. 183 An G, Puttitanun T (2009). ‘Revisiting McCallum's Border Puzzle’. Economic Development Quarterly 23: 167–170 [http://edq.sagepub.com/content/early/2009/03/11/0891242408328604.full.pdf]. Anaman KA, Looi CN (2000). ‘Economic impact of haze-related air pollution on the tourism industry in Brunei Darussalam. Topics in Economic Analysis and Policy 30:133–144. Andereck KL, Vogt CA (2000). ‘The Relationship between Residents’ Attitudes toward Tourism and Tourism Development Options’. Journal of Travel Research 39(1)27–36. Anderson JE (1979). ‘A Theoretical Foundation for the Gravity Model’. American Economic Review 69(1):106–116. Anderson JE, Marcouiller D (2002). ‘Insecurity and the pattern of trade: An empirical investigation’. The Review of Economics and Statistics 84(2):342–352, MIT Press. Anderson JE, van Wincoop E (2003). ‘Gravity with gravitas: A solution to the border puzzle’. American Economic Review 93(1):170–192. Archer B (1995). ‘Importance of Tourism for the Economy of Bermuda’, Annals of Tourism Research 22:918–930. Archer B, Cooper C, Ruhanen L (2005). ‘The positive and negative impacts of tourism’, Chapter 5 pp. 79–102 in Theobald WF (ed.) Global Tourism {third edition]. Burlington, Mass.: Elsevier Butterworth-Heinemann. Archer BH, Fletcher J (1996). ‘The Economic Impact of Tourism in the Seychelles. Annals of Tourism Research 23(1):32–47. Archibald X, LaCorbinière J, Moore W (2008). ‘Analysis of tourism competitiveness in the Caribbean: A Gravity Model approach’. Paper presented at the 29th Annual Review Seminar of the Research Department of the Central Bank of Barbados, 28– 31 July 2008. Ardahaey FT (2011). ‘Economic Impacts of Tourism Industry’. International Journal of Business and Management 6(8):206–215 [doi:10.5539/ijbm.v6n8p206]. 184 Arellano M, Bond SR (1991). ‘Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations’. Review of Economic Studies 58(2):277–297. Arellano M, Bover O (1995). ‘Another look at the instrumental variable estimation of error-components models’. Journal of Econometrics 68(1):29–51. Arezki R, Cherif R, Piotrowski J (2009). ‘Tourism Specialization and Economic Development: Evidence from the UNESCO World Heritage List’. International Monetary Fund, IMF Institute and Fiscal Affairs Department: IMF Working Paper WP/09/176 [available at http://www.imf.org/external/pubs/ft/wp/2009/wp09176.pdf accessed 01 March 2014]. Armenski T, Dragičević V, Pejović L, Lukić T, Djurdjev B (2011). ‘Interaction between Tourists and Residents: Influence on Tourism Development’. Polish Sociological Review 173:107–118. Armstrong HW, de Kervenoael RJ, Li X, Read R (1998). ‘A Comparison of the Economic Performance of Different Micro-States and between Micro-States and Larger Countries’. World Development 26:639–656. Asadi R, Daryaei M (2012). ‘Recognition and Prioritization of Pull Factors of Azerbaijan as a Destination for Iranian Tourists’. Journal of American Science 8(8):189–194. Ashley C, Boyd C, Goodwin H (2000). ‘Pro-poor tourism: Putting poverty at the heart of the tourism agenda’ [available from Overseas Development Institute website http://www.odi.org.uk/resources/download/2096.pdf]. Ashley C, De Brine P, Lehr A, Wilde H (2007). ‘The Role of the Tourism Sector in Expanding Economic Opportunity’. Cambridge, Mass.: Harvard University— Kennedy School of Government, Corporate Social Responsibility Initiative Report No.23[www.hks.harvard.edu/mrcbg/CSRI/publications/report_23_EO%20Tourism%20Final.pdf]. Australian Bureau of Statistics (2003). ‘9502.0.55.001–Framework for Australian TourismStatistics’[http://www.abs.gov.au/ausstats/[email protected]/Latestproducts/C63787 B7F00A3EE6CA256DA900835CC3?opendocument]. 185 Balaguer J, Cantavella-Jordá M (2002). ‘Tourism as a long-run economic growth factor: The Spanish case’. Applied Economics 34(7):877–884. Balestra P, Nerlove M (1966). ‘Pooling Cross Section and Time Series Data in the Estimation of a Dynamic Model: The Demand for Natural Gas’. Econometrica 34:585–612. Ball RM (1988). ‘Seasonality: a problem for workers in the tourism labour market?’ Service Industries Journal 8(4):501–513. Baltagi BH, Bresson G, Pirotte A (2003). ‘Fixed effects, random effects or Hausman–Taylor? A pre-test estimator’. Economics Letters 79:361–369. Bandara JS (1995). ‘“Dutch” Disease in a developing country: the case of foreign capital inflows to Sri Lanka’. Seoul Journal of Economics 8 (Fall): 314–329. Barquet A, Brida JG, Risso WA (2009). ‘Causality between economic growth and tourism expansion: Empirical evidence from Trentino-Alto Adige’. TOURISMOS: An International Multidisciplinary Journal of Tourism 5(2):87–98. Bastola RK (2012). The Impact of the Development of the Tourism Industry on the Lifestyle of the Host Community: Case Study in Pietarsaari. Karleby, Finland: Centria University of Applied Sciences; Thesis, Degree Programme in Tourism, December 2012. Batinić I (2013). ‘The Role and Importance of the Internet in Contemporary Tourism in Travel Agencies’ Business’. International Journal of Cognitive Research in Science, Engineering and Education (IJCRSEE) 1(2) [www.ijcrsee.com/index.php/ijcrsee/article/view/63/180]. Baum, C. F. (2006). `An Introduction to Modern Econometrics Using Stata`. College Station, TX: Stata Press Becken S (2010). ‘The Importance of Climate and Weather for Tourism: A Literature Review’. New Zealand: Lincoln University—Land, Environment and People Series. Behnam D (1990). ‘An International Inquiry into the Future of the Family: A UNESCO Project’. International Social Science Journal 42(4) (126):547–552. 186 Bénassy-Quéré A, Coupet M, Mayer T (2007) ‘Institutional determinants of Foreign Direct Investment’. The World Economy 30(5):764–782. Besculides A, Lee M, McCormick P (2002). ‘Residents’ perceptions of the cultural benefits of tourism’. Annals of Tourism Research 29(2):303–319. Bettin G, Lucchetti R, Zazzaro A (2012). ‘Endogeneity and sample selection in a model for remittances’. Journal of Development Economics 99(2):370–384. Bickart B, Schindler R (2001). ‘Internet forums as influential sources of consumer information’. Journal of Interactive Marketing 15(3):31–40. Blake A (2000). ‘The Economic Effects of Tourism in Spain’. Tourism and Travel Research Institute Discussion Papers, No. 2000/2. Blake A, Arbache JS, Sinclair MT, Teles V (2008). ‘Tourism and Poverty Relief’. Annals of Tourism Research 35(1):107–126. Blake A, Durbarry R, Sinclair MT, Sugiyarto G (2001). ‘Modelling Tourism and Travel using Tourism Satellite Accounts and Tourism Policy and Forecasting Models’. Tourism and Travel Research Institute Discussion Papers, No. 2001/4. Bliss H, Russett B (1998). ‘Democratic trading partners: The liberal connection, 1962–1989’. Journal of Politics 60(4):1126–1147. Bloom DE, Canning D, Fink G (2011). ‘Implications of Population Aging for Economic Growth’. Cambridge, Mass.: National Bureau of Economic Research, Working Paper 16705. Blundell R, Bond S (1998). ‘GMM estimation with persistent panel data: an application to production functions’. Paper presented at the Eighth International Conference on Panel Data, Göteborg University, 11–12 June 1998. Later published as Blundell and Bond (2000). Blundell R, Bond S (2000). ‘GMM estimation with persistent panel data: an application to production functions’. Econometric Reviews 19(3):321–340. Boisso D, Ferrantino M (1997). ‘Economic distance, cultural distance, and openness in international trade: Empirical puzzles’. Journal of Economic Integration 12(4):456–484. 187 Brahmbhatt M, Canuto O, Vostroknutova E (2010). ‘Dealing with Dutch Disease’. World Bank, Poverty Reduction and Economic Management Network (PREM): Economic Premise June 2010, Number 16. Brau R, Lanza A, Pigliaru F (2003). ‘How fast are small tourism countries growing? The cross-country evidence’. Paper prepared for the Conference Tourism and Sustainable Economic Development, Chia, Sardinia 19–20 September 2003 [FEEM Working Paper no. 85; doi: 10.2139/ssrn.453340]. Brent Ritchie JR, Crouch GI (2003). The Competitive Destination: A Sustainable Tourism Perspective. Wallingford, Oxon/Cambridge, Massachusetts: CABI Publishing. Brent Ritchie JR, Molinar CMA, Frechtling DC (2010). ‘Impacts of the World Recession and Economic Crisis on Tourism: North America’. Journal of Travel Research 49(1):5–15. Brida JG, Risso WA (2009). ‘Tourism as a factor of long-run economic growth: An empirical analysis for Chile’. European Journal of Tourism Research 2(2):178–185. Briedenhann J, Wickens E (2004). ‘Tourism routes as a tool for the economic development of rural areas—vibrant hope or impossible dream?’ Tourism Management 25(1):71–79. Brown F, Hall D (2008). ‘Tourism and development in the Global South: The issues’. Third World Quarterly 29(5):839–849 [doi: 10.1080/01436590802105967]. Bruno M, Sachs J (1982). ‘Energy and Resource Allocation: A dynamic model of the “Dutch Disease”’ Review of Economic Studies 49:845–859. Buhalis D (1998). ‘Strategic use of information technologies in the tourism industry’. Tourism Management 19(5):409–421. Buhalis D (2000). ‘Marketing the competitive destination of the future’. Tourism Management 21(1):97–116. Buhalis D, Cooper C (1998). ‘Competition or co-operation: The needs of Small and Medium sized Tourism Enterprises at a destination level’ pp. 324–346 in Laws E, 188 Faulkner B, Moscardo G (eds.), Embracing and Managing Change in Tourism. London: Routledge. Buhalis D, Law R (2008). ‘Progress in information technology and tourism management: 20 years on and 10 years after the Internet—The state of eTourism research’. Tourism Management 29(4):609–623. Buhalis D, Leung D, Law R (2011). ‘eTourism: Critical Information and Communication Technologies for Tourism Destinations’, pp. 205–224 in Y. Wang and A Pizam (eds.) Destination Marketing and Management. Wallingford UK: CAB International. Burkart AJ, Medlik S (1974). Tourism. Past, Present and Future. London: Heinemann. Butler R (1996). ‘Impacts, Carrying Capacity, Control and Responsibility in Tourist Destinations’. Progress in Tourism and Hospitality Research 2(3):283–294. Cai LA, Feng R, Breiter D (2004). ‘Tourist purchase decision involvement and information preferences’. Journal of Vacation Marketing 10(2):138–148. Capó J, Font AR, Nadal JR (2007). ‘Dutch Disease in Tourism Economies: Evidence from the Balearics and the Canary Islands’. Journal of Sustainable Tourism 15(6):615–627. Carter R, Beeton R (2004). ‘A model of cultural change and tourism’. Asia Pacific Journal of Tourism Research 9(4):423–442. Caselli F, Esquivel G, Lefort F (1996). ‘Reopening the convergence debate: A new look at cross-country growth empirics’. Journal of Economic Growth 1(3):363–389. Cater E, Goodall B (1992). ‘Must tourism destroy its resource base?’ in Mannion A, Bowley S (eds.) Environmental Issues in the 1990s. Chichester: Wiley. CEPII (2014). ‘Databases’. CEPII, Paris, France [available at http://www.cepii.fr/cepii/en/bdd_modele/bdd.asp]. Chang C-L, Khamkaew T, McAleer M (2009). ‘A Panel Threshold Model of Tourism Specialization and Economic Development’. University of Tokyo—Centre 189 for International Research on the Japanese Economy (CIRJE), Faculty of Economics: CIRJE Discussion Papers, Discussion Paper CIRJE-F-685. Chang C-L, Khamkaew T, McAleer M (2010). ‘IV Estimation of a Panel Threshold Model of Tourism Specialization and Economic Development’. Kyoto University, Institute of Economic Research: KIER Working Papers #708. Chao C-C, Hazari BR, Laffargue J-P, Sgro PM, Yu ESH (2006). ‘Tourism, Dutch disease and welfare in an open dynamic economy’. Japanese Economic Review 57(4):501–515 [available at SSRN: http://ssrn.com/abstract=942104 or doi:10.1111/j.1468-5876.2006.00400.x.] Chen C-F, Chiou-Wei SZ (2009). ‘Tourism expansion, tourism uncertainty and economic growth: New evidence from Taiwan and Korea’. Tourism Management 30(6):812–818. Chen JS, Hsu CHC (2000). ‘Measurement of Korean Tourists' Perceived Images of Overseas Destinations’. Journal of Travel Research 38(4):411–416. Cheng I-H, Wall HJ (2004, 2005). ‘Controlling for heterogeneity in gravity models of trade and integration’. Working Paper 1999-010E [http://research.stlouisfed.org/wp/1999/1999-010.pdf], revised July 2004, published 2005 in Federal Reserve Bank of St. Louis Review 87(1):49–63. Cheptea A (2007). ‘Trade liberalization and institutional reforms’. Economics of at Transition15(2):211–255[available http://papers.ssrn.com/sol3/papers.cfm?abstract_id=976785]. Cho V (2010). ‘A study of the non-economic determinants in tourism demand’. International Journal of Tourism Research 12(4):307–320. Chok S, Macbeth J, Warren C (2007). ‘Tourism as a tool for poverty alleviation: A critical analysis of ‘pro-poor tourism’ and implications for sustainability’. Current Issues in Tourism 10(2–3):144–165 [doi: 10.2167/cit303]. Chul H, Muzaffer U, Weaver PA (1995). ‘Product bundles and market segments based on travel motivations: A canonical correlation approach’. International Journal of Hospitality Management 14(2):123–137. 190 CIA (2012). ‘The World Factbook’. Central Intelligence Agency [www.cia.gov/library/publications/the-world-factbook/geos/vm.html]. Cohen E (1979). ‘A Phenomenology of Tourist Experiences’. Sociology 13(2):179– 201. Cohen E (1987). ‘Alternative Tourism—A Critique’. Tourism Recreation Research 12(2):13–18 [doi: 10.1080/02508281.1987.11014508]. Cohen SA, Prayag G, Moital M (2014). ‘Consumer behaviour in tourism: Concepts, influences and opportunities’. Current Issues in Tourism 17(10):872–909. Cooper CH, Fletcher J, Fyall A, Gilbert D, Wanhill S (2008). Tourism: Principles and Practice [fourth edition]. Harlow: Pearson Education Limited. Copeland BR (1991). ‘Tourism, Welfare and De-Industrialization in a Small Open Economy’. Economica New Series 58(232):515–529. Corden WM (1984). ‘Booming Sector and Dutch Disease Economics: Survey and Consolidation’. Oxford Economic Papers 36:339–380 [http://www.jstor.org/stable/2662669 accessed 24 Feb 2014]. Corden WM, Neary JP (1982). ‘Booming Sector and De-industrialisation in a Small Open Economy’. The Economic Journal 92 (December):825–848 [doi: 10.2307/2232670]. Cortés-Jiménez I, Pulina M, Riera i Prunera C, Artis M (2009). ‘Tourism and Exports as a means of Growth’. University of Barcelona, Research Institute of Applied Economics: Working Paper 2009/10 [http://core.ac.uk/download/pdf/6225089.pdf]. Cothran DA, Cothran CC (1998). ‘Promise or political risk for Mexican tourism’. Annals of Tourism Research 25(2):477–497. Croall J (1995). Preserve or Destroy: Tourism and the Environment. London: Calouste Gulbenkian Foundation/CABI. Crouch GI (1992). ‘Effect of income and price on international tourism’. Annals of Tourism Research 19(4):643–664. 191 Crouch GI (1994a). ‘The Study of International Tourism Demand: A Survey of Practice’. Journal of Travel Research 32(4):41–55 [April 1994]. Crouch GI (1994b). ‘The Study of International Tourism Demand: A Review of Findings’. Journal of Travel Research 33(1):12–23 [July 1994]. CS/WB (2000). Small States: Meeting Challenges in the Global Economy, Report of the Commonwealth Secretariat/World Bank Joint Task Force on Small States, London: Commonwealth Secretariat; Washington D.C.: The World Bank. Culiuc A (2014). ‘Determinants of International Tourism’. International Monetary Fund Working Paper WP/14/82 [http://www.imf.org/external/pubs/ft/wp/2014/wp1482.pdf]. Daryaei A, Aliashrafipour M, Eisapour K, Afsharian M (2012). ‘The Effect of Good Governance on Tourism industry Development’. Advances in Environmental Biology 6(7):2046–2052. Dearden SJH (2000). ‘Corruption and Economic Development’. DSA European Development Policy Study Group Discussion Paper No. 18, October 2000 [available at www.devstud.org.uk/downloads/4bb9998f27b48_Dp18.doc]. De Freitas CR (2003). ‘Tourism climatology: Evaluating environmental information for decision making and business planning in the recreation and tourism sector’. International Journal of Biometeorology 48(1):45–54. De Groot HLF, Linders GJ, Rietveld P, Subramanian U (2004). ‘The institutional determinants of bilateral trade patterns’. Kyklos 57(1):103–123. De Sousa J, Disdier A-C (2006). ‘La qualité du cadre juridique constitue-t-elle une barrière au commerce? Application aux économies en transition’. Revue économique 57(1):135–151. Dhariwal R (2005). ‘Tourist arrivals in India: how important are domestic disorders?’ Tourism Economics 11(2):185–205. Diaz Ruiz CA (2012). ‘Theories of markets: Insights from marketing and the sociology of markets’. The Marketing Review 12(1):61–77. 192 Dieke P (2003). ‘Tourism in Africa’s Economic Development: Policy Implications’. Management Decision 41(3):287–295 [doi: 10.118/00251740310469468]. Dilanchiev A (2012). ‘Empirical analysis of Georgian trade pattern: Gravity Model’. Journal of Social Sciences 1(1):75–78. Drakos K, Kutan AM (2003). ‘Regional Effects of Terrorism on Tourism in Three Mediterranean Countries’. The Journal of Conflict Resolution 47(5):621–641. Dredge D, Jenkins J (2007). Tourism Planning and Policy. Brisbane: John Wiley and Sons. Dredge D (2010). ‘Place change and tourism development conflict: Evaluating public interest’. Tourism Management 31(1):104–112. Dritsakis N (2004). ‘Tourism as a long-run economic growth factor: An empirical investigation for Greece using causality analysis’. Tourism Economics 10(3):305– 316. Duc C, Lavallée E, Siroën J-M (2008). ‘The gravity of institutions’. Economie international 113:95–113 [available from http://www.cairn.info/zen.php?ID_ARTICLE=ECOI_113_0095]. Duffy R (2002). A Trip too Far: Ecotourism, Politics and Exploitation. London & New York: Earthscan. Durbarry R (2004). ‘Tourism and economic growth: the case of Mauritius’. Tourism Economics 10(4):389–401. Durbarry R, Sinclair MT (2003). ‘Market shares analysis: The case of French tourism demand’. Annals of Tourism Research 30(4):927–941. Durden G, Silberman J (1975). ‘The determinants of Florida tourist flows: A gravity model approach’. Review of Regional Studies 5(3):31–41. Dutt P, Traça D (2007). ‘Corruption and Bilateral Trade Flows: Extortion or Evasion?’ INSEAD No. 2007/37/EPS. Dwyer L, Forsyth P, Madden J, Spurr R (2000a). ‘Economic Impacts of Inbound Tourism under Different Assumptions regarding the Macroeconomy’. Current Issues in Tourism 3(4):325–363. 193 Dwyer L, Forsyth P, Rao P (2000b). ‘The price competitiveness of travel and tourism: A comparison of 19 destinations’. Tourism Management 21(1):9–22. Dwyer L, Forsyth P, Rao P (2000c). ‘Sectoral analysis of price competitiveness of tourism: An international comparison’. Tourism Analysis 5(1):1–12. Dwyer L, Forsyth P, Spurr R, Ho T (2003). ‘The Contribution of Tourism to a State Economy: a Multi-regional General Equilibrium Analysis’. Tourism Economics 9(4):431–448. Dwyer L, Forsyth P, Spurr R (2004). ‘Evaluating Tourism's Economic Effects: New and Old Approaches’. Tourism Management 25(3):307–317. Dwyer L, Kim C (2003). ‘Destination Competitiveness: Determinants and Indicators’, Current Issues in Tourism 6(5):369–414. Economist (1977). ‘The Dutch Disease’. The Economist, 26 November 1977, pp. 82– 83. Edwards A (1995). Asia–Pacific travel forecasts to 2005. Research Report, Economist Intelligence Unit, London. Egger P (2000). ‘A note on the proper econometric specification of the gravity equation’. Economics Letters 66: 25–31. Egger P (2002). ‘SUR Estimation of Error Components Models with AR(1) Disturbances and Unobserved Endogenous Effects’. Paper presented at the 10th International Conference on Panel Data, Berlin, July 5-6, 2002 B6-3, International Conferences on Panel Data [http://econpapers.repec.org/cpd/2002/39_Egger.pdf]. Egger P (2005). ‘Alternative Techniques for Estimation of Cross-Section Gravity Models’. Review of International Economics 13(5):881–891. Eichengreen B, Irwin DA (1998). ‘The Role of History in Bilateral Trade Flows’, pp.33–57 in Jeffrey A. Frankel (ed.) The Regionalization of the World Economy. Chicago: University of Chicago Press. Eilat Y, Einav L (2004). ‘Determinants of International Tourism: A ThreeDimensional Panel Data Analysis’. Applied Economics 36:1315–1327. 194 EIU (2005). ‘Special Report—Asia’s tsunami: The impact’. London: Economist Intelligence Unit [http://graphics.eiu.com/files/ad_pdfs/tsunami_special.pdf]. Enders W, Sandler T, Parise GF (1992). ‘An econometric analysis of the impact of terrorism on tourism’. Kyklos 45:531–554. Ethier W (1972). ‘Nontraded Goods and the Heckscher-Ohlin Model’. International Economic Review 13(1):132–147. Etzo I, Massidda C, Piras R (2013). ‘Migration and Inbound Tourism by Purpose of Visit: An Italian perspective’. Department of Economics and Business, University of Cagliari, Italy [http://www.webmeets.com/files/papers/IEA/2014/248/Migration%20&%20Inbound %20Tourism%20by%20purpose%20of%20visit_Etzo-Massidda-Piras_oct2013.pdf]. Eugenio-Martin JL, Morales NM, Scarpa R (2004). ‘Tourism and economic growth in Latin American countries: A panel data approach’. Fondazione Eni Enrico Mattei. Fayissa B, Nsiah C (2010). ‘The Impact of Governance on Economic Growth: Further Evidence for Africa’. Middle Tennessee State University Department of Economics & Finance, Working Paper Series, December 2010 [http://core.ac.uk/download/pdf/6328268.pdf]. Fayissa B, Nsiah C, Tadasse B (2008). ‘The impact of tourism on economic growth and development in Africa’. Tourism Economics 14(4):807–818. Fennell D, Ebert K (2004). ‘Tourism and the Precautionary Principle’. Journal of Sustainable Tourism 12(6):461–479. Fernández C, Ley E, Steel MFJ (2001). ‘Model uncertainty in cross-country growth regressions’. Journal of Applied Econometrics 16(5):563–576. Fidrmuc J, Karaja E (2013). ‘Uncertainty, informational spillovers and policy reform: A gravity model approach’. Brunel University, Economics & Finance Working Paper Series: Working Paper No. 12-03. Fielding D (2010). ‘Aid and Dutch Disease in the South Pacific and in other small island states’. Journal of Development Studies 46(5):918–940. 195 Fielding D, Shortland A (2010). ‘How do Tourists React to Political Violence? An Empirical Analysis of Tourism in Egypt’. DIW Berlin Discussion Paper #1022 [www.diw.de/documents/publikationen/73/diw_01.c.358263.de/dp1022.pdf]. Figini P, Vici L (2009). ‘Tourism and growth in a cross-section of countries’. Working Paper Series 01_09, The Rimini Centre for Economic Analysis, republished in Tourism Economics 16(4):789–805. Fletcher JE (1989). ‘Input-Output Analysis and Tourism Impact Analysis’. Annals of Tourism Research 16(3):514–529. Fletcher JE (1994). ‘Assessing the Economic Impacts of Travel and Tourism— Introduction to Travel Economic Impact Estimation’ Chapter 31, pp. 359–365 in Cooper CP (ed.) Progress in Tourism, Recreation and Hospitality Management. London and New York: Belhaven Press. Fotis J, Buhalis D, Rossides N (2012). ‘Social media use and impact during the holiday travel planning process’, pp. 13–24 in Fuchs M, Ricci F, Cantoni L (eds.) Information and Communication Technologies in Tourism 2012. Vienna: Springer Vienna. Fourie J, Santana-Gallego M (2011). ‘The impact of mega-events on tourist arrivals’. Tourism Management 32(6):1364–1370. Fourie J, Santana-Gallego M (2013). ‘The determinants of African tourism’. Development Southern Africa 30(3):347–366 [doi: 10.1080/0376835X.2013.817302]. Fratianni MU, Kang H (2006). ‘International Terrorism, International Trade, and Borders’. Research in Global Strategic Management February 2006 [doi: 10.1016/S1064-4857(06)12010-0]. Frechtling DC (1994). ‘Assessing the economic impacts of travel and tourism— Introduction to travel economic impact estimation’ in Brent Ritchie JR, Goeldner CR (eds.) Travel, Tourism and Hospitality Research: A Handbook for Managers and Researchers [second edition]. New York: John Wiley and Sons Inc. Fry D, Saayman A, Saayman M (2010). ‘The Relationship between Tourism and Trade in South Africa’. South African Journal of Economics 78(3):287–306. 196 Garín-Muñoz T (2006). ‘Inbound international tourism to the Canary Islands: A dynamic panel data model’. Tourism Management 27(2):281–291. Garín-Muñoz T, Montero-Martín LF (2007). ‘Tourism in the Balearic Islands: A dynamic model for international demand using panel data’. Tourism Management 28(5):1224–1235. Garín-Muñoz T (2009). ‘Tourism in Galicia: Foreign and domestic demand’. Tourism Economics 15(4):753–769. Gearing CE, Swart WW, Var T (1974). ‘Establishing a Measure of Touristic Attractiveness’. Journal of Travel Research 12(4):1–18. Getz D (1993). ‘Planning for tourism business districts’. Annals of Tourism Research 20(3):583–600. Gil-Pareja S, Llorca-Vivero R, Martínez-Serrano JA (2007a). ‘The effect of EMU on tourism’. Review of International Economics 15(2):302–312. Gil-Pareja S, Llorca-Viveroa R, Martínez-Serrano JA (2007b). ‘The Impact of Embassies and Consulates on Tourism’. Tourism Management 28(2):355–360. Gleditsch NP, Wallensteen P, Eriksson M, Sollenberg M, Strand H (2002). ‘Armed conflict 1946–2000: A new dataset’ Journal of Peace Research 39(5):615–637 [originally presented as a paper at the PRIO/Uppsala/World Bank conference Identifying Wars: Systematic Conflict Research and its Utility in Conflict Resolution and Prevention, Uppsala,08–09 June 2001]. Goeldner CR, Brent Ritchie JR (2012). Tourism: Principles, Practices, Philosophies [12th edition]. Hoboken, New Jersey: John Wiley and Sons, Inc. Goh C, Law R (2003). ‘Incorporating the rough sets theory into travel demand analysis’. Tourism Management 24(5):511–517. Gómez Martín, M (2005). ‘Weather, Climate and Tourism: A Geographical Perspective’. Annals of Tourism Research 32(3):571–591. Goodrich JN (2002). ‘September 11, 2001 attack on America: A record of the immediate impacts and reactions in the USA travel and tourism industry’. Tourism Management 23(6):573–580. 197 Gordon IR, Edwards SL (1973). ‘Holiday trip generation’. Journal of Transport Economics and Policy 7:153–168. Görmüş Ş, Göçer İ (2010). ‘The socio-economic determinant of tourism demand in Turkey: A panel data approach. International Research Journal of Finance and Economics 55:88–89. Greene W (2001). ‘Estimating Econometric Models with Fixed Effects’. Department of Economics, Stern School of Business, New York University [http://www.stern.nyu.edu/~wgreene/fixedeffects.doc]. Greene WH (2012). Econometric Analysis [seventh edition]. Harlow: Pearson Education Limited/Prentice Hall. Greffe X (2004). ‘Is rural tourism a lever for economic and social development?’ Journal of Sustainable Tourism 2:23–40. Gretzel U, Yoo KH (2008). ‘Use and Impact of Online Travel Reviews’, pp. 35–46 in Information and Communication Technologies in Tourism 2008 (Volume 2): Vienna: Springer Vienna. Guduz L, Hatemi A (2005). ‘Is the tourism-led growth hypothesis valid for Turkey?’ Applied Economics 12(8):499–504. Guiso L, Sapienza P, Zingales L (2009). ‘Cultural biases in economic exchange’. The Quarterly Journal of Economics 124(3):1095–1131. Gunn CA, Var, T (2002). Tourism Planning: Basic, Concepts, Cases [fourth edition]. New York: Routledge. Hall CM, O’Sullivan V (1996). ‘Tourism, Political Stability and Violence’, pp. 105– 121 in A. Pizam and Y. Mansfield (eds.) Tourism, Crime and International Security Issues. New York: Wiley. Han Z, Durbarry R, Sinclair MT (2006). ‘Modelling US tourism demand for European destinations’. Tourism Management 27(1):1–10. Hanafiah MHM, Harun MFM (2010). ‘Tourism demand in Malaysia: A crosssectional pool time-series analysis’. International Journal of Trade, Economics and Finance 1(2):200–203. 198 Harrill R (2004). ‘Residents’ attitudes toward tourism development: A literature review with implications for tourism planning’. Journal of Planning Literature 18(3):251–266. Harrison D (2008). ‘Pro-poor tourism: A critique’. Third World Quarterly 29(5):851–868 [doi: 10.1080/01436590802105983]. Hausman JA, Taylor WE (1981). ‘Panel data and unobservable individual effects’. Econometrica 49:1377–1398. Heanue KE, Pyers CE (1966). ‘A comparative evaluation of trip distribution procedures’. Highway Research Record 114:35. Heeley J (1980). ‘The definition of tourism in Great Britain: Does terminological confusion have to rule?’ The Tourist Review 35(2):11–14 [http://dx.doi.org/10.1108/eb057811]. Hennig-Thurau T, Gwinner K, Walsh G, Gremler D (2004). ‘Electronic word-ofmouth via consumer-opinion platforms: What motivates consumers to articulate themselves on the Internet?’ Journal of Interactive Marketing 18(1):38–52. Hitchcock M, Putra IND (2005). ‘The Bali bombings: Tourism crisis management and conflict avoidance’. Current Issues in Tourism 8:62–76. Hjalager A (1996). ‘Tourism and the environment: The innovation connection’. Journal of Sustainable Tourism 4(4):201–218. Holder R (2012). ‘The political economics of the Arab Spring’. Munich: CESifo, Working Paper 4023. Holloway JC, Humphreys C, Davidson R (2009). The Business of Tourism [eighth edition]. London: FT Prentice Hall/Pearson Education. Holzner M (2005). ‘Fear of Croatian Disease: Is there a danger of a Dutch Disease Effect with respect to a boom in the tourism sector in Croatia in the long run—“The Croatian Disease”?’ Doctoral thesis, WU Vienna University of Economics and Business. Holzner M (2010). ‘Tourism and Economic Development: the Beach Disease?’ The Vienna Institute for International Economic Studies (wiiw) Working Paper #66. 199 Hoti S, McAleer M, Shareef R (2005). ‘Modelling international tourism and country risk spillovers for Cyprus and Malta’. Edith Cowan University, Perth, Western Australia: School of Accounting, Finance and Economics and FIMARC Working Paper Series, Working Paper 0516 [available at http://core.kmi.open.ac.uk/download/pdf/6408171.pdf]. Howie F (2003). Managing the Tourist Destination. London: Continuum. Hunziker W, Krapf K (1942). ‘Grundriss der Allgemeinen Fremdenverkehrslehre’ [Outline of the general teaching of tourism]. Seminars für Fremdenverkehr und Verkehrspolitik an der Handels-Hochschule St. Gallen 1. Zurich: Polygraphischer Verlag AG. Hvidt M (2011). ‘Economic Reforms in the Arab Gulf Countries: Lip Service or Actual Implementation?’ pp. 39–54 in Legrenzi M, Momani B (eds.) Shifting GeoEconomic Power of the Gulf: Oil, Finance and Institutions. Farnham: Ashgate. Ibrahim MAMA (2011). ‘The Determinants of International Tourism Demand for Egypt: Panel Data Evidence’. European Journal of Economics, Finance and Administrative Sciences Issue 30:50–58. ICRG (2012). ‘The International Country Risk Guide’. Published by the Political Risk Services Group [available at https://www.prsgroup.com/about-us/our-twomethodologies/icrg]. ICRG (2014). ‘International Country Risk Guide Methodology’ [http://www.prsgroup.com/wp-content/uploads/2012/11/icrgmethodology.pdf] accessed 13 Dec 2014. IMF (2009) ‘Tourism Specialization and Economic Development: Evidence from the UNESCO World Heritage List’. IMF Working Paper WP/09/176 prepared by Arezki R, Cherif R, Piotrowski J [www.imf.org/external/pubs/ft/wp/2009/wp09176.pdf]. IMF (2011). ‘Oil-Price Boom and Real Exchange Rate Appreciation: Is There Dutch Disease in the CEMAC?’ International Monetary Fund Working Paper WP/11/268 (prepared by Juan P. [https://www.imf.org/external/pubs/ft/wp/2011/wp11268.pdf]. 200 Treviño) Ioannides D, Apostolopoulos Y (1999). ‘Political Instability, War, and Tourism in Cyprus: Effects, Management, and Prospects for Recovery’. Journal of Travel Research 38:51–56. Ismail K (2010). ‘The Structural Manifestation of the “Dutch Disease”: The Case of Oil Exporting Countries’. International Monetary Fund, Washington, DC: Working Paper 10/103. Issa I, Altinay L (2006). ‘Impacts of political instability on tourism planning and development: The case of Lebanon’. Tourism Economics 12(3):361–381. Ivanov S, Webster C (2006). ‘Measuring the Impact of Tourism on Economic Growth’, pp 21–30 in Proceedings of GEOTOUR 2006 Conference, Kosice, Slovakia 7–8 October 2006. Jain AK (2001). ‘Corruption: a review’. Journal of Economic Surveys 15(1):71–121. Jenner P, Smith C (1992). ‘The tourism industry and the environment’. London: Economic Intelligence Unit. Jensen JLWV (1906). ‘Sur les fonctions convexes et les inégalités entre les valeurs moynennes’. Acta Mathematica 30(1):175–193 [doi: 10.1007/BF02418571]. Jones R (1971). ‘A Three-factor Model in Theory, Trade and History’, pp. 3–21 in Bhagwati J, Jones RW, Miundell RA, Vanek J (eds.) Trade, Balance of Payments and Growth: Essays in Honor of C.P. Kindleberger. Amsterdam: North-Holland. Jones R (1987). ‘Heckscher-Ohlin Trade Theory’, pp. 620–627 in Eatwell J, Milgate M, Newman P (eds.) The New Palgrave Dictionary of Economics. New York: Stockton Press. Kadir N, Karim MZA (2012). ‘Tourism and economic growth in Malaysia: Evidence from tourist arrivals from ASEAN-5 countries’. Ekonomska Istrazivanja-Economic Research 25(4):1089–1100. Karemera D, Oguledo V, Davis B (2000). ‘A gravity model analysis of international migration to North America. Applied Economics 32:1745–1755. Katz S, Marshall B (2003). ‘New sex for old: lifestyle, consumerism, and the ethics of aging well’. Journal of Aging Studies 17:3–16. 201 Kau JB, Sirmans CF (1977). ‘The Functional Form of the Gravity Model: A new technique with empirical results’. University of Illinois at Urbana-Champaign: College of Commerce and Business Administration: Faculty Working Papers #359. 202 Kaufmann D, Kraay A, Zoido-Lobatón P (2002). ‘Governance Matters II: Updated Indicators for 2000-01’. The World Bank Development Research Group and the World Bank Institute, Governance, Regulation, and Finance Division. February 2002. Kennedy P (2003). A Guide to Econometrics [fifth edition]. Oxford: Blackwell. Keum K (2010). ‘Tourism flows and trade theory: A panel data analysis with the Gravity Model’. Published online 2008, and in The Annals of Regional Science 44(3):541–557, June 2010. Khadaroo J, Seetanah B (2007). ‘Transport infrastructure and tourism development’. Annals of Tourism Research 34(4):1021–1032. Khadaroo J, Seetanah B (2008). ‘The role of transport infrastructure in international tourism development: A gravity model approach’. Tourism Management 29(5):831– 840. Khan H (2006). ‘International trade and tourism: An overview’. The University of Nottingham: U21Global 5(7). [www.u21global.com/PartnerAdmin/ViewContent?module=NEWSLETTER&oid=1 57293 accessed 12 April 2014]. Khan H, Phang S-Y, Toh RS (1995). ‘Singapore’s Hospitality Industry: The Multiplier Effect’. Cornell Hotel & Restaurant Administration Quarterly 36(1):64– 69. Khan H, Seng CF, Cheong WK (1990). ‘Tourism multiplier effects on Singapore’. Annals of Tourism Research 17:408–418. Khan H, Toh RS, Chua L (2005). ‘Tourism and trade: Cointegration and Granger causality tests’. Journal of Travel Research 44(2):171–176. Kim HJ, Chen M-H, Jang SS (2006). ‘Tourism expansion and economic development: The case of Taiwan’. Tourism Management 27:925–933. Kim S, Yoon Y (2003). ‘The Hierarchical Effects of Affective and Cognitive Components on Tourism Destination Image’. Journal of Travel & Tourism Marketing 14(2):1–22. 203 Kim Y (1993). ‘The analysis of visitor's behaviour in Sobaeksan National Park’. Applied Ecosystem Studies in Korea 6(2):218–228. Kimura F, Lee H-H (2006). ‘The gravity equation in international trade in services’. Online paper published by the Kiel Institute [doi: 10.1007/s10290-006-0058-8]. King B, Pizam A, Milman A (1993). ‘Social impacts of tourism: Host perceptions’. Annals of Tourism Research 20(4):650–665 [doi: 10.1016/0160-7383(93)90089-L]. Kliman ML (1981). ‘A quantitative analysis of Canadian overseas tourism’. Transportation Research 15:487–497. Kosnan SSA, Ismail NW (2012). ‘Demand factors for international tourism in Malaysia: 1998–2009’. Prosiding Persidangan Kebangsaan Ekonomi Malaysia Ke, VII, pp.44–50 [available at www.ukm.my/fep/perkem/pdf/perkemVII/PKEM2012_1A5.pdf]. Kosnan SSA, Ismail NW, Kaliappan SR (2013). ‘Determinants of international tourism in Malaysia: Evidence from Gravity Model’. Jurnal Ekonomi Malaysia 47(1):131–138. Kozak N, Uysal M, Birkan I (2008). ‘An analysis of cities based on tourism supply and climatic conditions in Turkey’. Tourism Geographies 10(1):81–97. Kreag G (2011). ‘The Impacts of Tourism’. University of Minnesota: Publication T13, Minnesota Sea Grant.[www.seagrant.umn.edu/tourism/pdfs/ImpactsTourism.pdf, accessed 25 Feb 2014]. Kuss FR, Graefe AR, Vaske JJ (1990). Recreation Impacts and Carrying Capacity [Vols. I & II]. Washington DC: National Parks and Conservation Association. Kweka J, Morrissey O, Blake A (2001). ‘Is tourism a key sector in Tanzania: Input– output analysis of income, output, employment and tax revenue’. Christel DeHaan Tourism and Travel Research Institute, Discussion Papers 2001/1 [www.nottingham.ac.uk/ttri/pdf/2001_1.pdf]. Kweka J, Morrissey O, Blake A (2003). ‘The Economic Potential of Tourism in Tanzania’. Journal of International Development 15: 335–351. 204 Lavallée E (2005). ‘Governance, Corruption and Trade: A North-South Approach’. Paris Dauphine University: Economics Papers from University Paris Dauphine [persistent link: http://EconPapers.repec.org/RePEc:dau:papers:123456789/4090]. Lavallée E (2006). ‘Similarité institutionnelle, qualité des institutions et commerce international’. Économie internationale 108:27–58. Lavallée E, Razafindrakoto M, Roubaud F (2008). ‘Corruption and trust in political institutions in sub-Saharan Africa’. Paris: DIAL, Working Paper DT/2008-07 [available at www.en.dial.ird.fr/content/download/49582/380080/version/1/.../200807.pdf]. Laws E, Agrusa JF, Scott N, Richins H (2011). ‘Tourist Destination Governance: Practice, Theory and Issues’, Chapter One pp. 1–16 in E. Laws, H. Richins, J.F. Agrusa (eds.) Tourist Destination Governance: Practice, Theory and Issues. London: CABI Publishing. Leask A, Spiller JE (2002). ‘U.K. Conference Venues: Past, Present and Future’. Journal of Convention and Exhibition Management 4(1)29–54. New York/London: The Haworth Press. Le Billon P (2001). ‘The political economy of war: Natural resources and armed conflicts’. Political Geography 20(5):561–584. Lee C-C, Chang C-P (2008). ‘Tourism development and economic growth: a closer look at panels’. Tourism Management 29:180–192. Lee CG (2012). ‘Tourism, trade, and income: Evidence from Singapore’. Anatolia: An International Journal of Tourism and Hospitality Research 23(3):348–358. Lee C, Kwon K (1995). ‘Importance of secondary impact of foreign tourism receipts on the South Korean economy’. Journal of Travel research 34:50–54. Leiper N (1979). ‘The Framework of Tourism: Towards a Definition of Tourism, Tourist, and the Tourist Industry’. Annals of Tourism Research 6(4):390–407. Leitão NC (2010). ‘Does Trade Help to Explain Tourism Demand? The Case of Portugal’. Theoretical & Applied Economics 17(3):63–74. 205 Lew A (1987). ‘The English-Speaking Tourist and the Attractions of Singapore’. Singapore Journal of Tropical Geography 8:44–59. Levine, R., & Renelt, D. (1992). A sensitivity analysis of cross-country growth regressions. American Economics Review, 82, 942-963. Li G, Song H, Witt SF (2005). ‘Recent developments in econometric modelling and forecasting’. Journal of Travel Research 44:82–99. Lien D, Ghosh S, Yamarik S (2014). ‘Does the Confucius Institute impact international travel to China? A panel data analysis’. Applied Economics 46(17):1985–1995. Lim C (1997a). ‘Review of international tourism demand models’. Annals of Tourism Research 24:835–849. Lim C (1997b). ‘An econometric classification and review of international tourism demand models’. Tourism Economics 3(1):69–81. Lim C (1999). A meta-analysis review of international tourism demand’. Journal of Travel Research 37:273–284. Lim CC, Cooper C (2009). ‘Beyond Sustainability: Optimising Island Tourism Development’. International Journal of Tourism Research 11:89–103. Lindberg K, Andersson TD, Dellaert BGC (2001). ‘Tourism Development: Assessing Social Gains and Losses’. Annals of Tourism Research 28(4):1010–1030. Linnemann H (1966). An Econometric Study of International Trade Flows. Amsterdam: North-Holland Publishing Company. Litvin SW, Goldsmith RE, Pan B (2008). ‘Electronic word-of-mouth in hospitality and tourism management’. Tourism Management 29(3):458–468. Liu X (2009). ‘GATT/WTO promotes trade strongly: Sample selection and model specification’. Review of International Economics 17(3):428–446. Lo J TY-Y, Pan S (2014). ‘Confucius Institutes and China’s soft power: Practices and paradoxes’. Compare: A Journal of Comparative and International Education [doi: 10.1080/03057925.2014.916185]. 206 Lohmann M, Danielsson J (2004). ‘How to get the future of tourism out of today’s consumer surveys—Prospects for senior and kids travel in Germany’. Paper, Theme 2 – Consumer Surveys, 7th International Forum on Tourism Statistics, Stockholm, Sweden, 9-11 June 2004 [www.scb.se/grupp/produkter_tjanster/.../lohmanndanielsson_pap12.doc}. Luo M, Feng R, Cai LA (2004). ‘Information Search Behavior and Tourist Characteristics’. Journal of Travel & Tourism Marketing 17(2-3):15–25. Malamud B (1973). ‘Gravity model calibration of tourist travel to Las Vegas’. Journal of Leisure Research 5(4):23–33. Mankiw NG, Romer D, Weil DN (1992). ‘A Contribution to the Empirics of Economic Growth’. The Quarterly Journal of Economics 107(2): 407–437 [http://links.jstor.org/sici?sici=00335533%28199205%29107%3A2%3C407%3AACTTEO%3E2.0.CO%3B2-5]. Marcussen HC (2011). ‘Seasonality in tourism—Separating the natural and institutional causes’. Bornholm, Denmark: Centre for Regional and Tourism Research [available: http://www.crt.dk/media/Seasonality_in_tourism_natural_institutional_causes_Marc ussen_24.pdf]. Martínez-Zarzoso I, Nowak-Lehmann F, Vollmer S (2007). ‘The log of gravity revisited’. University of Göttingen: Centre for Globalization and Europeanization of the Economy (CeGE) Discussion Paper 64 [http://wwwuser.gwdg.de/~cege/Diskussionspapiere/64.pdf]. Martínez-Zarzoso I (2011). ‘The Log of Gravity Revisited’ [http://repositori.uji.es/xmlui/bitstream/handle/10234/39421/49226.pdf?sequence=1] . Marzuki A, Hay I (2013) ‘Towards a Public Participation Framework in Tourism Planning’. Tourism Planning & Development 10(4):494–512 [doi: 10.1080/21568316.2013.804432]. Massidda C, Etzo I (2012). ‘The determinants of Italian domestic tourism: A panel data analysis’. Tourism Management 33: 603–610. 207 Mathieson A, Wall G (1982). Tourism: Economic, physical and social impacts. Harlow, UK: Longman. Mátyás L (1997). ‘Proper econometric specification of the gravity model’. The World Economy 20(3):363–368. Mátyás L (1998). ‘The Gravity Model: Some Econometric Considerations’. The World Economy 21(3):397–401. Mauro P (1995). ‘Corruption and Growth’. The Quarterly Journal of Economics 110(3):681–712. Mauro P (2004). ‘The Persistence of Corruption and Slow Economic Growth’. International Monetary Fund: IMF Staff Papers, Vol. 51, no. 1. [http://www1.worldbank.org/publicsector/anticorrupt/feb06course/mauro.pdf]. McAleer M, Huang B-W, Kuo H-I, Chen C-C, Chang C-L (2008). ‘An Econometric Analysis of SARS and Avian Flu on International Tourist Arrivals to Asia’. Working Paper EI 2008-21 [repub.eur.nl/pub/.../SARS%20and%20Avian_flu_working%20paper_.pdf.] McCallum J (1995). ‘National borders matter: Canada–US regional trade patterns’. American Economic Review 85(3):615–623. Macleod D (2006). ‘Cultural commodification and tourism: A very special relationship. Tourism Culture & Communication 6(2): 71–84. Mehchy Z, Nasser R, Schiffbauer M (2013). ‘Trade determinants and potential of Syria: Using a gravity model, with an estimation of the Syrian Crisis’ impact on exports’. Giza, Egypt: Economic Research Forum, Working Paper 773 [http://www.erf.org.eg/CMS/uploads/pdf/773.pdf]. Méon P-G, Sekkat K (2005). ‘Does Corruption Grease or Sand the Wheels of Growth’. Public Choice 122(1/2):69–97. Mieiro S, Nogueira Ramos P, Alves J (2012). ‘Gaming Tourism Boom, Foreign Currency Inflows, and Dutch Disease Effects: an Empirical Model for Macau’. International Journal of Trade, Economics and Finance 3(6):421–427. Milner HV, Kubota K (2005). ‘Why the move to Free Trade? Democracy and trade 208 policy in the developing countries’. International Organization 59(1):107–143. Min JCH (2005). ‘The Effect of the SARS Illness on Tourism in Taiwan: An Empirical Study’. International Journal of Management 22(3):497–508. Mishra PK, Rout HB, Mohapatra SS (2011). ‘Causality between Tourism and Economic Growth: Empirical Evidence from India’. European Journal of Social Sciences 18(4):518–527. Mo C-M, Howard DR, Havitz ME (1993). ‘Testing an international tourist role typology’. Annals of Tourism Research 20(2):319–335. Morley C, Rosselló J, Santana-Gallego M (2014). ‘Gravity models for tourism demand: theory and use’. Annals of Tourism Research 48(1):1–10. Moscardo G (1999). ‘Supporting ecologically sustainable tourism on the Great Barrier Reef: The importance of visitor research’, pp. 236–253 in Molloy J, Davies J (eds.) Tourism and Hospitality: Delighting the senses. Canberra: Bureau of Tourism Research. Moser C, Nestmann T, Wedow M (2006). ‘Political Risk and Export Promotion: Evidence from Germany’. Frankfurt-am-Main: Deutsche Bundesbank, Discussion Paper Series 1: Economic Studies No 36/2006. Mowforth M, Munt I (2009). Tourism and Sustainability: Development, globalisation and new tourism in the Third World. London: Routledge. Murdoch JC, Sandler T (2002). ‘Economic Growth, Civil Wars, and Spatial Spillovers’. Journal of Conflict Resolution 46(1):91–110. Mushtaq A, Zaman K (2013). ‘Impact of macroeconomic factors on tourism receipts: evidence from SAARC region’. European Economic Letters 2(2):38–43. Narayan PK (2004). ‘Economic Impact of Tourism on Fiji’s Economy: Empirical Evidence from the Computable General Equilibrium Model’. Tourism Economics 10(4):419–433. Naudé WA, Saayman A (2005). ‘Determinants of tourist arrivals in Africa: a panel data regression analysis’. Tourism Economics 11(3):365–391. [First appeared as a 209 paper prepared for the International Conference, Centre for the Study of African Economies, St. Catherine’s College, University of Oxford, 21–22 March 2004]. Neumayer E (2002). ‘Do we trust the data?: On the validity and reliability of crossnational environmental surveys’. Social Science Quarterly 83(1):332–340. Neumayer E (2004). ‘The impact of political violence on tourism: Dynamic crossnational estimation’. Journal of Conflict Resolution 48(2):259–281. Neumayer E (2010). ‘Visa restrictions and bilateral travel’. The Professional Geographer 62(2):1–11. Nielsen L (2011). ‘Classifications of Countries based on their Level of Development: How it is done and how it could be done’. International Monetary Fund, IMF Working Paper WP/11/31 [https://www.imf.org/external/pubs/ft/wp/2011/wp1131.pdf]. Nitsch V, Schumacher D (2004). ‘Terrorism and international trade: an empirical investigation’. European Journal of Political Economy 20:423–433. Nowak J-J, Sahli M, Sgro P (2004). ‘Tourism, Trade and Domestic Welfare’. FEEM Working Paper 24, [available at SSRN: http://ssrn.com/abstract=504445]. Nowak J-J, Sahli M (2007). ‘Coastal tourism and “Dutch disease” in a small island economy’. Tourism Economics 13(1):49–65. OECD (2008). ‘The Impact of Culture on Tourism’. Paris: Organization for Economic Co-operation and Development. [doi: 10.1787/9789264040731-en]. Oh C-O (2005). ‘The contribution of tourism development to economic growth in the Korean economy’. Tourism Management 26(1):39–44. Olson O, Guthrie J, and Humphrey C (1998). ‘International experiences with ’new’ public financial management (NPFM) reforms: New world? small world? better world?’ pp. 17–48 in Olson O, Guthrie J, Humphrey C (eds.) Global Warning: International Financial Management Changes. Bergen, Norway: Cappelan Akademisk Forlag. Őnder K, Durgun A (2008). ‘The effects of the tourism sector on employment in Turkey: An econometric application’. [http://epoka.edu.al/new/icme/26.pdf]. 210 Page SJ, Connell J (2006). Tourism: A Modern Synthesis [second edition]. London: Thomson Learning. Pandey RW, Chettri P, Kunwar RR, Ghimire G (1995). ‘Case study on the effects of tourism on culture and the environment in Nepal’. Bangkok: UNESCO Principal Office Asia & Pacific: RACAP Series on Culture and Tourism in Asia, No. 4. Papaioannou E (2009). ‘What drives international financial flows? Politics, institutions and other determinants’. Journal of Development Economics 88(2):269– 281 (published from the CEPR Discussion Paper 7010, 2008). Papatheodorou A, Rosselló J, Xiao H (2010). ‘Global Economic Crisis and Tourism: Consequences and Perspectives’. Journal of Travel Research 49(1) 39–45. Parra-López E, Bulchand-Gidumal J, Gutiérrez-Taño D, Díaz-Armas R (2011). ‘Intentions to use social media in organizing and taking vacation trips’. Computers in Human Behavior 27(2):640–654. Patsouratis V, Frangouli Z, Anastasopoulos G (2005). ‘Competition in Tourism among the Mediterranean Countries’. Applied Economics 37(16):1865–1870. [doi: 10.1080/00036840500217226]. Payne JE, Mervar A (2010). ‘Research Note: The Tourism-Growth Nexus in Croatia’. Tourism Economics 16(4):1089–1094. Pearce PL (1993). ‘Fundamentals of tourist motivation’, pp. 113–134 in Pearce D, Butler W (eds.), Tourism and research: Critiques and challenges. London: Routledge. Pease W, Rowe M (2005). ‘An Overview of Information Technology in the Tourism Industry’. University of Queensland [https://eprints.usq.edu.au/245/1/Pease.pdf]. Pedregal P (1992). ‘Jensen's Inequality in the Calculus of Variations’. Carnegie Mellon University Research Report No. 92-NA-016, May 1992. Phakdisoth L, Kim D (2007). ‘The Determinants of Inbound Tourism in Laos’. ASEAN Economic Bulletin 24(2):225–237. Piazza JA (2006). ‘Rooted in poverty?: Terrorism, poor economic development, and social cleavages’. Terrorism and Political Violence 18(1):159–177. 211 Pine R, McKercher B (2004). ‘The impact of SARS on Hong Kong’s tourism industry’. International Journal of Contemporary Hospitality Management 16(2):139–143. Pizam A (1978). ‘Tourism’s Impacts: The Social Costs to the Destination Community as Perceived by its Residents’. Journal of Travel Research 16(4):8–12. Pleumarom A (1994). ‘The Political Economy of Tourism’. Ecologist 24(4):142– 148. Po W-C, Huang B-N (2008). ‘Tourism development and economic growth—A nonlinear approach’. Physica A: Statistical Mechanics and its Applications 387(22):5535–5542. Polat E, Turkan S, Gunay S (2010). ‘Relationship between Tourism and Trade in Turkey’. Paper presented to the International Conference on Applied Economics– ICOAE 2010, Athens, 26–28 August 2010 [http://kastoria.teikoz.gr/icoae/?p=389, accessed 25 February 2014]. Poprawe M (2015). ‘A Panel Data Analysis of the Effect of Corruption on Tourism’. Applied Economics 47(23):2399–2412. Pöyhönen P (1963). ‘A tentative model for the volume of trade between countries’. Weltwirtschaftliches Archiv 90(1):93–100. Prideaux B (2000). ‘The role of the transport system in destination development’. Tourism Management 21(1):53–63. Prideaux B (2005). ‘Factors affecting bilateral tourism flows’. Annals of Tourism Research 32:780–801. Prideaux B, Laws E, Faulkner B (2003). ‘Events in Indonesia: exploring the limits to formal tourism trends forecasting methods in complex crisis situations. Tourism Management 24:475–487. Proença SA, Soukiazis E (2005). ‘Demand for Tourism in Portugal: A Panel Data Approach’. University of Coimbra: Centro de Estudos da União Europeia (CEUNEUROP), Faculdade de Economia da Universidade de Coimbra. [Discussion 212 Paper (FEBRUARY) Nº 29 [accessed from www4.fe.uc.pt/ceue/working_papers/isaraelias29.pdf 12 April 2014]. Pyers CE (1966). ‘Evaluation of intervening opportunities trip distribution model’. Highway Research Record 114:76–77. Quandt RE, Baumol WJ (1966). ‘The demand for abstract transport models: Theory and measurement’. Journal of Regional Science 6:12–26. Quandt RE, Young KH (1969). ‘Cross-sectional travel demand models: Estimates and tests’. Journal of Regional Science 9(2):201–214. Quinn B (20009). ‘Festivals, events and tourism’. Dublin Institute of Technology: School of Hospitality Management and Tourism [http://arrow.dit.ie/cgi/viewcontent.cgi?article=1000&context=tfschhmtbook]. Rajan RG, Subramanian A. (2011). ‘Aid, Dutch Disease, and manufacturing growth’. Journal of Development Economics 94(1):106–118. Rault C, Sova R, Sova A (2007). ‘Modelling International Trade Flows between Eastern European Countries and OECD Countries’. IZA Discussion Paper No. 2851. Ray J, Rivera-Batiz FL (2002). ‘An Analysis of Sample Selection bias in crosscountry growth recessions’. Columbia University, Department of Economics, Discussion Paper Series; Discussion Paper #:0102-10. Richards G (1996). ‘The scope and significance of cultural tourism’, pp. 19–46 in Richards G (ed.), Cultural tourism in Europe. Oxford: CABI Publishing. Rios‐Morales R, Gamberger D, Jenkins I, Smuc T (2011). ‘Modelling investment in the tourism industry using the World Bank's good governance indicators’. Journal of Modelling in Management 6(3):279–296. Ritchie JRB, Zins M (1978). ‘An Empirical Evaluation of the Role of Culture and its Components as Determinants of the Attractiveness of a Tourism Region’. Annals of Tourism Research 5(2):252–267. Robinson M (1999). ‘Cultural conflicts in tourism: Inevitability and inequality’, pp. 1–32 in Robinson M, Boniface P (eds.), Tourism and cultural conflicts. Oxford: CABI Publishing. 213 Romanazzi S, Petruzzellis L, Iannuzzi E (2011). ‘“Click & experience. Just virtually there.” The Effect of a Destination Website on Tourist Choice: Evidence from Italy’. Journal of Hospitality Marketing & Management 20:791–813. Roodman D (2006). ‘How to do Xtabond2: An Introduction to Difference and System GMM’. Stata (December 2006). Centre for Global Development Working Paper No. 103. Available at SSRN http://ssrn.com/abstract=982943 or http://dx.doi.org/10.2139/ssrn.982943 Rosselló-Nadal J, Riera-Font A, Capó-Parrilla J (2007). ‘The Contributions of Economic Analyses to Tourism: A Survey’, pp. 79–108 in Tavidze A (ed.) Progress in Economics Research Volume 11. New York: Nova Science Publishers. Rudd D (1996). ‘An Empirical Analysis of Dutch Disease: Developing and Developed Countries’. Illinois Wesleyan University: Honors Projects. Paper 62 [http://digitalcommons.iwu.edu/econ_honproj/62]. Russo AP, van der Borg J (2002). ‘Planning considerations for cultural tourism: a case study of four European cities’. Tourism Management 23(6):631–637. Sahli M, Nowak J-J (2007). ‘Does Inbound Tourism Benefit Developing Countries? A Trade Theoretic Approach’. Journal of Travel Research 45:426–434. Sala-i-Martin X (1997). ‘I Just Ran Two Million Regressions’. American Economic Review 87(2):178–183. Sanchez Carrera EJ, Brida JG, Risso WA (2008). ‘Tourism's Impact on Long-Run Mexican Economic Growth’ (September 2007). Economics Bulletin 23(21):1– 8[available at SSRN: http://ssrn.com/abstract=1076225 or http://dx.doi.org/10.2139/ssrn.1076225]. Santana-Gallego M, Ledesma-Rodríguez FJ, Pérez-Rodríguez JV (2007). ‘On the Relationship between Trade and Tourism’. Proceedings of the First Conference of the International Association for Tourism Economics, 25-27 October 2007. Santana-Gallego M, Ledesma-Rodríguez FJ, Pérez-Rodruígez JV, Cortés I (2010). ‘Does a Common Currency Promote Countries’ Growth via Trade and Tourism?’ The World Economy 33(12):1811–1835. 214 Santana-Gallego M, Ledesma-Rodríguez FJ, Pérez-Rodríguez JV (2011). ‘Tourism and trade in small island regions: the case of the Canary Islands’. Tourism Economics 17(1):107–125. Santos Silva JMC, Tenreyro S (2006). ‘The Log of Gravity’. The Review of Economics and Statistics 88(4):641–658. Santos Silva JMC, Tenreyro S (2008). ‘Comments on “The log of gravity revisited” ’ [http://personal.lse.ac.uk/tenreyro/mznlv.pdf]. Sargan JD (1958). ‘The Estimation of Economic Relationships using Instrumental Variables’. Econometrica 26(3):393–415. Sawkar K, Noronha L, Mascarenhas A, Chauhan OS, Saeed S (1998). ‘Tourism and the Environment: Case Studies on Goa, India, and the Maldives’. World Bank [http://citeseerx.ist.psu.edu/viewdoc/versions;jsessionid=AE608CAE94B65DF6415 FD0BA7526DFFA?doi=10.1.1.201.1963]. Scheyvens R (2007). ‘Exploring the tourism-poverty nexus’. Current Issues in Tourism 10(2,3):231–254 [doi: 10.2167/cit318.0]. Schmidt CJ (1979). ‘The Guided Tour: Insulated Adventure’. Journal of Contemporary Ethnography 7:441–467. Schubert SF, Brida JG, Risso WA (2011). ‘The impacts of international tourism demand on economic growth of small economies dependent on tourism’. Tourism Management 32(2):377–385 [doi: 10.1016/j.tourman.2010.03.007]. Scott D, Jones B, Konopek J (2008). ‘Exploring the impact of climate-induced environmental changes on future visitation to Canada’s Rocky Mountain National Parks’. Tourism Review International 12:43–56. Seddighi HR, Nuttall MW, Theocharous AL (2001). ‘Does cultural background of tourists influence the destination choice? An empirical study with special reference to political instability’. Tourism Management 22(2):181–191. Sen S, Lerman D (2007). ‘Why are you telling me this? An examination into negative consumer reviews on the Web’. Journal of Interactive Marketing 21(4):76– 94. 215 Sequeira TN, Campos C (2005). ‘International Tourism and Economic Growth: A Panel Data Approach’, The Fondazione Eni Enrico Mattei Note di Lavoro Series Index: [www.feem.it/feem/pub/publication/Wpapers]. Sequeira TN, Nunes PM (2008). ‘Does tourism influence economic growth? A dynamic panel data approach’. Applied Economics 40(18):2431–2441 [doi: 10.1080/00036840600949520]. Serlenga L, Shin Y (2007). ‘Gravity Models of Intra-EU Trade: Application of the CCEP-HT Estimation in Heterogeneous Panels with Unobserved Common Timespecific Factors’. Journal of Applied Econometrics 22:361–381 (paper originally produced in mimeograph, University of Edinburgh, February 2004). Shakya M (2014). ‘Social Capital, Tourism and Socio-Economic Transformation of Rural Society: Evidence from Nepal’. Ruhr-Universität Bochum: IEE Working Papers Number 208. Shan J, Wilson K (2001). ‘Causality between trade and tourism: empirical evidence from China’. Applied Economics Letters 8(4):279–283. Shaw G, Williams AM (2004). Tourism and Tourism Spaces. London: SAGE Publications. Shea L, Enghagen L, Khullar A (2004). ‘Internet diffusion of an e-complaint: A content analysis of unsolicited responses’. Journal of Travel and Tourism Marketing 17:105–116. Sheldon PJ (1997). Tourism Information Technologies. Wallingford UK: CAB International. Sheldon PJ, Var T (1985). ‘Tourism forecasting: A review of empirical research’. Journal of Forecasting 4(2):183–195. Shepherd R (2002). ‘Commodification, culture and tourism’. Tourist Studies 2(2): 183–201. Shepherd B, Wilson JS (2009). ‘Trade facilitation in ASEAN member countries: Measuring progress and assessing priorities’. Journal of Asian Economics 20(4): 367–383. 216 Siliverstovs B, Schumacher D (2009). ‘Estimating gravity equations: to log or not to log?’ Empirical Economics 36(3):645–669. Silva JMC Santos, Tenreyro S (2006). ‘The Log of Gravity’. The Review of Economics and Statistics 88(4):641–658. Sinclair MT (1998). ‘Tourism and economic development: A survey’. Journal of Development Studies 34(5):1–51. Skene J (1993a). ‘The economic impact of the growth in visitor expenditure: A quantitative assessment’. Canberra: Bureau of Industry Economics Working Paper No 92. Skene J (1993b). ‘Some short run effects of an increase in visitor expenditure’. Canberra: Bureau of Industry Economics Working Paper No 94. Škuflić L, Štoković I (2011). ‘Demand Function for Croatian Tourist Product: A Panel Data Approach’. Modern Economy 2(1):49–53. Smeral E (2009). ‘The Impact of the Financial and Economic Crisis on European Tourism’. Journal of Travel Research 49(1) 39–45. Smith SLJ (2004). ‘The measurement of global tourism: Old debates, new consensus and continuing challenges’, in Lew A, Hall CM, Williams A (eds.) A Companion to Tourism. Oxford: Blackwell. Smith S (2007). ‘Duelling definitions: Challenges and implications of conflicting international concepts of tourism’, pp. 123–138 in Airey D, Tribe J (eds.) Developments in Tourism Research. Oxford: Elsevier. Smith VL (1989). Hosts and Guests: The Anthropology of Tourism [second edition]. Philadelphia: University of Pennsylvania Press. Smith VL (2001). ‘The Nature of Tourism’, Chapter 4 in VL Smith and MA Brent (eds.) Hosts and Guests Revisited: Tourism Issues of the 21st Century. Elmsford, New York: Cognizant Communication Corporation. Song H (2010). ‘Tourism demand modelling and forecasting: how should demand be measured?’ Tourism Economics 16(1):63–81. 217 Song H, Li G (2008). ‘Tourism demand modelling and forecasting: A Review of recent research’. Published online by the University of Surrey, UK [epubs.surrey.ac.uk/1085/1/fulltext.pdf]. Song H, Lin S (2010). ‘Impacts of the Financial Crisis on Tourism in Asia’. Journal of Travel Research 49(1): 16-30. Song H, Witt SF, Li G (2003). ‘Modelling and forecasting the demand for Thai tourism’. Tourism Economics 9(4):363–387. Song H, Witt SF, Li G (2009). The Advanced Econometrics of Tourism Demand. New York: Routledge. Sönmez SF (1998). ‘Tourism, terrorism and political instability’. Annals of Tourism Research 25(2):416–456. Sönmez SF, Apostolopoulos Y, Tarlow P (1999). ‘Tourism in crisis: Managing the effects of terrorism’. Journal of Travel Research 38(1):13–18. Stanford J, McCann B (1979). Tourism in the Australian Economy: Some preliminary estimates. Canberra: Bureau of Industry Economics, Working Paper No. 5. Stylidis D, Biran A, Sit J, Szivas EM (2014). ‘Residents' support for tourism development: The role of residents' place image and perceived tourism impacts’. Tourism Management 45:260–274 [doi: http://dx.doi.org/10.1016/j.tourman.2014.05.006]. Su Y-W, Lin H-L (2014). ‘Analysis of international tourist arrivals worldwide: The role of world heritage sites’. Tourism Management 40:46–58. Tajik, M., Aliakbari, S., Ghalia, T., & Kaffash, S. (2015). House prices and credit risk: Evidence from the United States. Economic Modelling, 51, 123-135. Tang C-HH, Jang SS (2009). ‘The tourism-economy causality in the United States: A sub-industry level examination. Tourism Management 30(4):553–558. 218 Tang CF, Tan EC (2013). ‘How stable is the tourism-led growth hypothesis in Malaysia? Evidence from disaggregated tourism markets’. Tourism Management 37:52–57. Taplin JHE, Qiu M (1997). ‘Car trip attraction and route choice in Australia’. Annals of Tourism Research 24(3):624–637. Tekleselassie TG (2014). ‘Institutional Determinants of Cross-Border Travel: The Role of Visas’. Presentation made on 05 December 2014 [www.sussex.ac.uk/webteam/gateway/file.php?name=tekleselassie.pdf&site=24]. Tenreyro S (2007). ‘On the trade impact of nominal exchange rate volatility’. Journal of Development Economics 82(2):485–508. Teo P, Leong S (2006). ‘A Postcolonial Analysis of Backpacking’. Annals of Tourism Research 33(1):109–131. The Economist—see under Economist. Theobald WF (2005). ‘The Meaning, Scope, and Measurement of Travel and Tourism’, Chapter 1, pages 5–24 in Theobald WF (ed.) Global Tourism. Burlington, MA: Elsevier Butterworth-Heinemann. Thompson A (2011). ‘Terrorism and tourism in developed versus developing countries’. Tourism Economics 17(3):693–700. Throsby D (1994). ‘The Production and Consumption of the Arts: A View of Cultural Economics’. Journal of Economic Literature, American Economic Association 32(1):1–29. Throsby D (2001). Economics and Culture [Chapter Three: ‘Cultural Capital and Sustainability’]. Cambridge: Cambridge University Press. Timothy DJ, Prideaux B, Seongseop-Kim S (2004). ‘Tourism at borders of conflict and (de)militarized zones’, pp. 83–94 in New Horizons in Tourism: Strange Experiences & Stranger Practices. Cambridge, MA: CABI. Tinbergen J (1962). Shaping the World Economy: Suggestions for an International Economic Policy. New York: The Twentieth Century Fund. 219 Tomljenovic R, Faulkner B (2000). ‘Tourism and older residents in a sunbelt resort’. Annals of Tourism Research 27(1):93–114. Tosun C (1999). ‘An analysis of contributions of international inbound tourism to the Turkish economy’. Tourism Economics 5(3):217–250. Tosun C, Timothy D (2001). ‘Shortcomings in planning approaches to tourism development in developing countries: The case of Turkey’. International Journal of Contemporary Hospitality Management 13(7):352–359. Tourism Economics (2013). ‘The Impact of Online Content on European Tourism’. Oxford: Tourism Economics. [http://sete.gr/_fileuploads/entries/Online%20library/GR/131204_The%20Impact%2 0of%20Online%20Content%20on%20European%20Tourism.pdf]. Torres-Reyna, O. (2007), “Panel data analysis: fixed and random effects using STATA”, Princeton University, available at: http://dss.princeton.edu/training (accessed 10 March 2012). Tse RYC (2001). ‘Estimating the impact of economic factors on tourism: Evidence from Hong Kong’. Tourism Economics 7:277–293. Turner L, Witt SF (2001). ‘Forecasting tourism using univariate and multivariate structural time series models’. Tourism Economics 7(2):135–148. UCDP/PRIO (2013). ‘UCDP/PRIO Armed Conflict Dataset Codebook Version 4– 2013’ [www.pcr.uu.se/digitalAssets/124/124920_1codebook_ucdp_prio-armed- conflict-dataset-v4_2013.pdf]. Um S, Crompton JL (1990). ‘Attitude Determinants in Tourism Destination Choice’. Annals of Tourism Research 17:432–448. UNDESA (2008). ‘International Recommendations for Tourism Statistics’. United Nations, Department of Economic & Social Affairs, Statistics Division: Studies in Methods, Series M, No. 83/Rev.1 [http://unstats.un.org/unsd/publication/SeriesM/seriesm_83rev1e.pdf]. UNDESA (2010). ‘International Recommendations for Tourism Statistics 2008’, Studies in Methods Series M No. 83/Rev.1; New York: United Nations Department 220 of Economic & Social Affairs. [http://unstats.un.org/unsd/publication/Seriesm/SeriesM_83rev1e.pdf#page=21]. UNDESA (2014). ‘WESP Country Classification’. [http://www.un.org/en/development/desa/policy/wesp/wesp_current/2014wesp_coun try_classification.pdf]. UNWTO (2008). ‘Tourism and the world economy’, in the Facts & Figures section [available online at http://www.unwto.org/index.php]. UNWTO (2010). ‘UNWTO Annual Report 2010: A Year of Recovery’ [http://dtxtq4w60xqpw.cloudfront.net/sites/all/files/pdf/final_annual_report_pdf_3.p df]. UNWTO (2012). ‘UNWTO Annual Report 2012’. [http://dtxtq4w60xqpw.cloudfront.net/sites/all/files/pdf/annual_report_2012.pdf]. UNWTO/ILO (2013). ‘Economic Crisis, International Tourism Decline and its Impact on the Poor’. Madrid UNWTO [www.ilo.org/wcmsp5/groups/public/--ed_dialogue/---sector/documents/publication/wcms_214576.pdf]. UNWTO (2015a). ‘UNWTO Tourism Highlights 2015 Edition’. [http://www.eunwto.org/doi/pdf/10.18111/9789284416899]. UNWTO (2015b). ‘Yearbook of Tourism Statistics, Data 2009–2013’, 2015 Edition [http://www.e-unwto.org/doi/book/10.18111/9789284416363]. USAID (2002). USAID Supports Good Governance’. [Available online at http://www.docstoc.com/docs/673298/USAIDSupports-Good-Governance]. Uysal M, Crompton JL (1984). ‘Determinants of demand for international tourist flows to Turkey’. Tourism Management 5(4):288–297. Vanhove N (2005). The Economics of Tourism Destinations. Oxford/Burlington, MA: Elsevier Butterworth-Heinemann. Var T, Quayson J (1985). ‘The multiplier impact of tourism in the Okanagan’. Annals of Tourism Research 12(4):497–514. 221 Velasquez MEB, Oh J (2013). ‘What determines international tourist arrivals to Peru? A gravity approach’. International Area Studies Review 16(4):357–369 [doi: 10.1177/2233865913505103] Vellas F (2011). ‘The Indirect Impact of Tourism: An economic analysis’. Paper submitted to the Third Meeting of T20 Tourism Ministers, Paris, France, 25 October 2011 [http://dtxtq4w60xqpw.cloudfront.net/sites/all/files/111020- rapport_vellas_en.pdf]. Vietze C (2008). ‘Cultural effects on inbound tourism into the USA: a gravity approach’. Jena Economic Research Papers 2008-037. Jena: Friedrich-SchillerUniversity Jena, Max-Planck-Institute of Economics. Vietze C (2009). ‘What's pushing international tourism expenditures?’ Jena Economic Research Papers 2009-014. Jena: Friedrich-Schiller-University Jena, MaxPlanck-Institute of Economics. Wagner JE (1997). ‘Estimating the Economic Impacts of Tourism’. Annals of Tourism Research 24(3):592–608. Walsh K (2006). ‘Trade in Services: Does Gravity Hold? A Gravity Model Approach to Estimating Barriers to Services Trade’. Dublin, Trinity College: Institute for International Integration Studies, Discussion Paper No. 183. Wang Y-S (2009). ‘The impact of crisis events and macroeconomic activity on Taiwan's international inbound tourism demand’. Tourism Management 30(1):75– 82. Warr B, Ayres R (2006). ‘Economic growth, technological progress and energy use in the US over the last century: Identifying common trends and structural change in macroeconomic time series’. Paper for INSEAD Session 49 in Helsinki 2006. WDI (2010) see World Bank (2010). WDI (2012) see World Bank (2012) WDI (2013) see World Bank (2013) 222 Wearing S, Wearing M (2006). ‘Rereading the Subjugating Tourist’ in Neoliberalism: Postcolonial Otherness and the Tourist Experience. Tourism Analysis 11(2):145–162. Werthner H, Klein S (1999). ‘ICT and the changing landscape of global tourism distribution’. Electronic Markets 9(4):256. Westerlund J, Wilhelmsson F (2009). ‘Estimating the Gravity Model without gravity using panel data’. Applied Economics 43:641–649. [available at http://folk.uio.no/rnymoen/Estimating%20the%20gravity%20model.pdf: The paper was originally presented at the 2006 spring meeting of the Midwest International Economics Group and at a seminar at Lund University]. Wilder-Smith A (2006). ‘The severe acute respiratory syndrome: Impact on travel and tourism’. Travel Medicine and Infectious Disease 4(2):53–60. Wilkinson PF (1992). ‘Tourism: Development imperative and environmental problems, pp. 22–32 in Carden F (ed.) Discussion Forum II on the Graduate Programme in Development Studies at the Bandung Institute of Technology (Research Series Paper 30). Waterloo: University Consortium on the Environment, University of Waterloo. Wilkinson W (2008). ‘Cuba's Tourism ‘Boom’: a curse or a blessing?’ Third World Quarterly 29(5):979–993 [doi: 10.1080/01436590802106189]. Williams AM, Hall CM, Lew A (2004). ‘Conclusions: Contemporary themes and challenges in tourism’, pp.611–618 in Lew A, Hall CM, Williams AM (eds.) Companion to Tourism. Oxford: Blackwell’s. Wilson A (1967). ‘A statistical theory of spatial distribution models’. Transportation Research 1(3):253–269. Wilson D (1997). ‘Paradoxes of Tourism in Goa’. Annals of Tourism Research 21(1):52–75. Windmeijer F (2005). ‘A finite sample correction for the variance of linear efficient two-step GMM estimators’. Journal of Econometrics 126(1):25–51. 223 Witt SF, Witt CA (1995). ‘Forecasting tourism demand: a review of empirical research’. International Journal of Forecasting 11:447–475. Wong KF (1997a). ‘An Investigation of the Time-Series Behaviour of International Tourist Arrivals’. Tourism Economics 3(2):185–199. Wong KF (1997b). ‘The Relevance of Business Cycles in Forecasting International Tourist Arrivals’. Tourism Management 18(8):581–586. Wong KF, Song H, Chon KS (2006). ‘Bayesian models for tourism demand forecasting’. Tourism Management 27(5):773–780. World Bank (2010). ‘World Development Indicators 2012’; Washington, DC: The World Bank [http://data.worldbank.org/sites/default/files/wdi-2010-final.pdf]. World Bank (2012). ‘World Development Indicators 2012’; Washington, DC: The World Bank [http://data.worldbank.org/sites/default/files/wdi-2012-ebook.pdf]. World Bank (2013). ‘World Development Indicators 2013’; Washington, DC: The World Bank [http://databank.worldbank.org/data/download/WDI-2013-ebook.pdf]. World Bank (2014). ‘World Development Indicators 2014’ [available at http://data.worldbank.org/data-catalog/world-development-indicators]. WTTC (2003). ‘Singapore: Special SARS Analysis—Impact on Travel and Tourism’. London: WTTC. WTTC (2011). ‘The Comparative Economic Impact of Travel and Tourism’ [http://www.wttc.org//media/files/reports/benchmark%20reports/the_comparative_economic_impact_of_tr avel__tourism.pdf]. WTTC (2012). ‘Travel and Tourism Economic Impact 2012’. World Travel and Tourism Council [available at www.wttc.org/site_media/uploads/downloads/world2012.pdf accessed 23 February 2015]. WTTC (2013). ‘Benchmarking Travel & Tourism—Global Summary’ (November 2013) World Travel and Tourism Council. 224 [http://wttc.org/site_media/uploads/downloads/2013_Global_summary_1.pdf accessed 25 February 2014]. Xiang Z, Gretzel U (2010). ‘Role of social media in travel information search’. Tourism Management 31(2):179–188. Xiong B, Chen S (2014). ‘Estimating Gravity Equation Models in the Presence of Sample Selection and Heteroskedasticity’. Applied Economics 46(24):2993–3003. Yakop M, van Bergeijk PAG (2011). ‘Economic diplomacy, trade and developing countries’. Cambridge Journal of Regions, Economy and Society 4(2):253–267. Yap G, Saha S (2013). ‘Do political instability, terrorism, and corruption have deterring effects on tourism development even in the presence of UNESCO heritage? A cross-country panel estimate’. Tourism Analysis 18(5):587–599. Yu M (2010). ‘Trade, democracy, and the gravity equation’. Journal of Development Economics 91:289–300. Zahidi S (2014). ‘The Human Capital Index’; Chapter 12, pp 92–96 in Education and Skills 2.0: New Targets and Innovative Approaches. Geneva: World Economic Forum [available at http://www3.weforum.org/docs/GAC/2014/WEF_GAC_EducationSkills_TargetsInn ovativeApproaches_Book_2014.pdf accessed 10 March 2014]. Zhang J, Jensen C (2007). ‘Comparative Advantage: Explaining Tourism Flows’. Annals of Tourism Research 34(1):223–243. Zhang Y, Qu H, Tavitiyaman P (2009). ‘The Determinants of the Travel Demand on International Tourist Arrivals to Thailand’. Asia Pacific Journal of Tourism Research 14(1):77–92. Zhang Y, Wu T, Li X (2015). ‘The Impacts of Cultural Values on Bilateral International Tourism Flows’. Tourism Travel and Research Association: Advancing Tourism Research Globally. Paper 15 [http://scholarworks.umass.edu/ttra/ttra2015/Academic_Papers_Oral/15]. Zhou D, Yanagida JF, Chakravorty U, Leung PS (1997). ‘Estimating Economic Impacts from Tourism’. Annals of Tourism Research 24(1):76–89. 225 Zortuk M (2009). ‘Economic Impact of Tourism on Turkey’s Economy: Evidence from Cointegration Tests’, International Research Journal of Finance and Economics 25:231–239. 226