Is Tourism Development a Sustainable Economic Growth Strategy in the Long Run? Evidence from GCC Countries

Abstract

The main objective of this study is to investigate the causal relationship between tourism development and economic growth in Gulf Cooperation Council (GCC) countries in a multivariate model, using panel data for the period 1995–2012. The study adopts a panel Granger causality analysis approach to assess the contribution of tourism to economic growth in GCC countries as a whole, and in each individual country. In the case of GCC countries as a whole, the results show a one-way Granger causality, from economic growth to tourism growth. Furthermore, Kuwait, Saudi Arabia, Qatar, and the United Arab Emirates follow the path of economy-driven tourism growth, as hypothesized. The reverse hypothesis (i.e., tourism-led growth hypothesis) holds true for Bahrain, while there is no causal relationship between tourism and economic growth in the case of Oman.

1. Introduction

The tourism industry has emerged as a key driver for a sustainable socio-economic development worldwide. The sector has become an increasingly important source of revenue and employment for many developing countries. According to the World Tourism Organization [1], due to the growing number of destinations opening up to and investing in tourism development, modern tourism has become a key driver for socio-economic progress. Tourism allows the creation of new jobs and enterprises, supporting infrastructure development, the protection of heritage and cultural values, and exporting earned revenues. The latest economic impact report provided by the World Travel and Tourism Council [2] indicated that the tourism industry generated US$ 7.5 trillion (9.8% of the world’s GDP) and about 277 million jobs (1 in 11 jobs) globally, in 2014. The report also predicts that the industry is expected to generate more than US$ 11.3 trillion (10.5% of the world’s GDP) and 357 million jobs for the world economy by 2025.

Given its increasing importance in the global economy, over the last few years, tourism has gained much attention among Gulf Cooperation Council (GCC) policymakers. GCC states, mostly rich oil countries, view the tourism industry not only as a source of revenue, but also, more importantly, as a way to diversify their economy and solve their unemployment problems [3].

The objective of this study is to investigate the relationship between tourism development and economic growth in GCC countries. It examines the causal relationship between tourism development and economic growth in GCC countries with a multivariate panel causality model, using panel data for the period 1995–2012. The remainder of the paper is organized as follows: Section 2 briefly presents the background of tourism development in GCC countries, Section 3 provides a brief literature review, Section 4 describes the estimation methodology and data sources, Section 5 presents the empirical results and Section 6 provides a summary, a brief conclusion, and some policy recommendations.

2. Tourism Development in GCC Countries: An Overview

Most GCC countries (Bahrain, Kuwait, Qatar, Oman, Saudi Arabia, and the United Arab Emirates) can be classified as newcomers to the global tourism agenda, especially in terms of attracting modern tourism, such as leisure, meetings, incentives, conventions and exhibitions, sport and adventure tourist activities. Over the last decade, massive projects in tourism infrastructure development have been undertaken in many GCC countries. All GCC countries have begun to establish national tourism development strategies and have tremendously expanded tourism through the adoption of positioning (re-positioning) strategies for tourist destinations and niche tourism initiatives [3,4]. Sport, leisure, event, hospitality infrastructures, and large-scale urban development have become a priority in such large-scale tourism initiatives. Ferrari Land, Formula One, Atlantis Dubai, Sharjah Biennale, the Asian Games, and the forthcoming Football World Cup are examples of these GCC tourism mega-projects. Today, more than 12 sites in GCC countries are listed on the UNESCO list of world cultural heritage [5]. Among them, the oasis of Al Ain in the Abu Dhabi emirate, Historic Jeddah (the Gate to Makkah) and At-Turaif District (ad-Dir’iyah) in Saudi Arabia, Qal’at Al-Bahrain, Bahla Fort, and Land of Frankincense in Oman, and Al Zubarah Archaeological site in Qatar.

In recent years, GCC countries have collectively played host to millions of tourists, with increasing number of visitors in most destinations. The number of international tourists increased from about 8.6 million in 1995 to almost 32 million in 2012 [6]. According to the economic impact reports on several GCC states provided by the World Travel and Tourism Council [2], travel and tourism directly supported 1,092,500 jobs in 2014 and around 1,597,000 jobs in 2015. However, the total contribution of travel and tourism to overall employment in GCC countries, including jobs indirectly supported by the industry, was about 1,957,500 jobs in 2014 and about 2,864,000 jobs in 2015. Indeed, tourism development provides great opportunities for GCC countries to diversify away from single resource economies. It also offers ways to help solve their socio-economic problems, such as unemployment [3].

The most recent tourism policy initiative in GCC countries was the development of intra-GCC tourism action plans [7]. The initiative was launched during the GCC tourism officials meeting in Doha on May 13, 2015. The focus of the meeting was on exploring mechanisms to enhance intra-GCC tourism, while building on the success of the Cruise Arabia Alliance, a regional collaboration that strengthens the Gulf’s position as a cruise destination. The meeting also encouraged GCC officials to accelerate a joint GCC action to ensure that all member states benefit from mega events taking place in the region over the next decade, such as the Dubai Expo 2020 and the FIFA World Cup 2022.

3. Literature Review

Many governments around the world have engaged in tourism development to promote long-run sustainable economic growth and development. In the literature, long-term influence of tourism on economic growth is known as the tourism-led growth hypothesis. Advocates of this hypothesis argue that the economic growth of a country not only is a function of labor, capital, export, and other factors in that country, but it can also be affected by the number of tourists entering that country. Tourism can act as a growth engine; it contributes to GDP growth, job creation, and foreign exchange generation. Researchers [8,9,10,11,12,13,14,15] suggested many reasons for tourism to be a positive long-run factor in the economic growth of a country. These reasons are as follows:

• Tourism produces foreign exchange earnings that are essential to import both consumer goods and capital and intermediate goods;
• Tourism facilitates the use of resources that are in line with the factor endowment of a country;
• Tourism creates job opportunities in tourist destinations;
• Tourism promotes improvements in a country’s infrastructures, benefitting not only tourists, but also residents;
• Tourism is considered a conduit for transferring new technological and managerial skills into an economy; and
• Tourism has the potential to create positive linkages with other sectors of the economy (e.g., agriculture, manufacturing, and other service industries).

On the other hand, economic growth also has a positive impact on tourism development. Economic growth can favor the tourism industry through the development of facilities and infrastructures, such as transportation development, information and communication technology development, electronic money, hotels, restaurants, and other entertainment services and facilities. The relationship between tourism development and economic growth has been a popular topic in the recent tourism-led growth literature. However, despite using research methodologies based on time series and/or panel data analytical techniques, empirical studies have shown mixed or even inconsistent results in terms of supporting the tourism-led economic growth hypothesis. Examples of studies that analyze the relationship between tourism and economic growth employing the Granger causality test with time series data analysis include Balaguer and Cantavella-Jorda [9], Durbarry [10], Gunduz and Hatemi-J [11], Brida, Carrera, and Risso [12], Chen and Chiou-Wei [13], Belloumi [16], Akinboade and Braimoh [17], Tang and Abosedra [18], and Tang and Tan [19]. These studies provide evidence in support of the tourism-led growth hypothesis. Furthermore, studies such as Dritsakis [14], Kim, Chen, and Jang [20], and Lee and Chien [21] support the validity of the feedback hypothesis, while Oh [22], Ozturk and Acaravci [23], Tang and Jang [24], and Payne and Mervar [25] present evidence of no relationship between the two variables. However, other empirical studies investigate the relationship between tourism development and economic growth using cross-section or panel data. Among these studies, Leitão [26], Tavares and Leitão [27], Leitão and Shahbaz [28], Sequeira and Campos [29], Lee and Chang [30], Po and Huang [31], Proenca and Soukiazis [15], Falk [32], Holzner [33], Ekanayake and Long [34], Lee and Brahmasrene [35], Aslan [36], Chou [37], Surugiu, Leitão and Surugiu [38], and Tugcu [39] obtain mixed results for the relationship between tourism and economic growth. Moreover, the results are sensitive to the specific group of countries under examination. For example, Balaguer and Cantavella [9] examined the tourism led-growth hypothesis for the first time. They used the Johansen-Juseliusco integration analysis in order to investigate the relationship between the tourism industry and long-run economic development in Spain, between 1975 and 1997. They reviewed the tourism led-growth hypothesis and concluded that during the study period, economic growth in Spain was strongly influenced by income from tourism. They also advocated that growth in the tourism industry had a positive long-run impact on economic development through a multiplier effect.

Eugenio-Martin, Morales and Scarpa [40] investigated the relationship between tourism and economic growth in 21 Latin American countries using GMM dynamic panel data analysis, over the period 1985–1998. Their results show that tourism has a positive and significant impact on the economic growth of Latin American countries. They also applied the Generalized Least Squares panel data model to explain foreign tourist arrivals, confirming the positive relationship between tourism and economic growth. The study suggested that developing infrastructure, education, and safety levels would increase tourism in Latin American countries, in general, and low income countries, in particular.

Lee and Chang [30] apply the heterogeneous panel co-integration technique to investigate the causal relationship between tourism and economic growth for two samples: OECD and non-OECD countries. They found evidence of panel co-integration between tourism development and GDP in both samples. Tourism development is shown to have a larger impact on GDP in non-OECD countries. In the long run, there is evidence of a unidirectional causal relationship from tourism development to economic growth in the case of OECD countries, and a bidirectional relationship in the sample of non-OECD countries.

Brida, Carrera and Risso [12] used the Johansen-Juselius econometrics method, the Granger causality test, and the impulse response analysis of shocks to investigate the long-run effect of tourism on economic growth in Mexico. The Johansen-Juselius test showed the presence of a co-integrated relationship among real GDP, tourism spending, and real exchange rate. The Granger causality test showed that causality goes unidirectional from tourism spending and RER to real GDP. They also found that a shock on tourism spending has a short-run, as well as a long-run positive effect on economic growth.

Chang, Khamkaev and McAleer [41] use a larger panel dataset that includes 159 countries, over the period 1989–2008. The panel threshold model shows a positive relationship between economic growth and tourism. However, the instrumental variable estimation indicates that tourism has a higher impact on economic growth in countries with a lower level of trade openness and investment.

Considering that GCC countries are new global destinations, where tourism-oriented development projects are on the rise, investigating if the tourism development strategy is a relevant growth factor for GCC economies would be of particular interest. To the best of the author’s knowledge, the causal relationship between tourism and economic growth in GCC states has never been studied in a panel context, without exception.

4. Data, Methodology and Results

4.1. Data

This empirical study uses annual data on six GCC countries (i.e., Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates) over the period 1995–2012. The variables used in this study are real international tourism revenue (receipts) and real gross domestic product. International tourism revenue is expressed in billions of U.S. dollars. Real gross domestic product is measured in constant 2005 U.S. dollars. The source of the data is the World Economic Outlook Database [6]. Real international tourism revenue (LTR) is used to measure tourism development and real gross domestic product (LGDP) is used as an economic growth indicator. Both variables are expressed in natural logarithms. Summary statistics are provided in

Table 1. Summary statistics of real GDP (constant 2005 U.S. dollars-unit: billion).

Country	Mean	Max.	Min.	Std. Dev.	Skew.	Kurt.	J–B
Bahrain	15.20	30.40	5.850	8.710	0.46	1.70	1.90
Kuwait	76.50	183.00	25.90	51.70	0.70	2.14	2.01
Oman	33.70	78.10	13.80	21.00	0.86	2.36	2.52
Qatar	58.50	192.00	8.140	59.30	1.07	2.87	3.44
Saudi Arabia	321.00	711.00	142.00	186.00	0.83	2.42	2.31
United Arab Emirates	179.00	384.00	65.700	105.00	0.59	1.94	1.90

Note: The sample period is from 1995–2012.

and

Table 2. Summary statistics of real international tourism revenue (unit: billion).

Country	Mean	Max.	Min.	Std. Dev.	Skew.	Kurt.	J–B
Bahrain	1.30	2.16	0.58	0.56	−0.02	1.42	1.88
Kuwait	0.46	0.78	0.29	0.14	0.68	2.44	1.64
Oman	0.72	1.78	0.19	0.48	0.86	2.76	2.25
Qatar	1.45	7.22	0.12	2.13	1.58	4.14	8.45
Saudi Arabia	5.13	9.32	3.42	2.04	0.66	2.00	2.08
United Arab Emirates	3.62	8.58	0.63	3.21	0.56	1.58	2.47

Note: The sample period is from 1995–2012.

.

The real GDP summary statistics in Table 1 indicate that Saudi Arabia and Bahrain had the highest and lowest average GDP: $321.0 and $15.2 billion dollars, respectively. The Jarque–Bera (J–B) normality test results indicate that the distribution of the real gross domestic products for these countries is approximately normal. Table 2 summarizes the descriptive statistics for international tourism revenue. Saudi Arabia and Kuwait received the highest and lowest mean tourism revenue, respectively. The Jarque–Bera (J–B) normality test results indicate that real international tourism revenues in these countries approximate a normal distribution.

4.2. Unit Root

One important step to investigate the causality between variables is to determine whether the series are integrated. To test for integration, the panel unit root tests developed by Levin, Lin, and Chu [42] and Im, Pesaran, and Shin [43] were used in this study. Both Levin, Lin, and Chu (LLC) and Im, Pesaran, and Shin (IPS) unit root tests consider the following panel ADF (Augmented Dickey–Fuller) specification:

$$\Delta y_{i,t}= {\rho }_i{y}_{i,t-1}+ {\displaystyle \sum }_{j=1}^{{\rho }_i}{\delta }_{i,j}\Delta y_{i,t-j}+ {\epsilon }_{i,t}$$

(1)

According to the LLC [42], the persistence parameters $({\rho }_i)$ are assumed to be identical across sections (i.e., ${\rho }_i=\rho$ for all $i$), whereas the lag order $({\rho }_i)$ may freely vary. The LLC tests the null hypothesis $(H_0: {\rho }_i=0$ for all $i$) against the alternative hypothesis $(H_1: {\rho }_i<0$ for all $i)$. Rejection of the null hypothesis indicates a possible panel integration procedure. The IPS [43] test differs from the LLC [42] test by assuming $({\rho }_i)$ to be heterogeneous across sections. The IPS tests the null hypothesis $(H_0: {\rho }_i=0$ for all $i$) against the alternative hypothesis $(H_1: {\rho }_i<0$ for all $i)$. Rejection of the null hypothesis suggests that the series are integrated. The LLC and IPS tests were performed on the data; both in levels and first differences, and the results are reported in

Table 3. Panel unit root test results.

Dependent		H0: All panels contain unit root
		LLC		IPS
	Variables	Constant	Constant + Trend	Constant	Constant + Trend
GCC Countries	GDP	0.606 (0.727)	−2.037 (0.021)	3.175 (0.999)	−4.326 (0.332)
	TR	0.021 (0.508)	−1.089 (0.138)	3.026 (0.998)	−0.050 (0.480)
	ΔGDP	−4.152 (0.000)	−3.520 (0.000)	−2.856 (0.002)	−1.193 (0.116)
	ΔTR	−5.730 (0.000)	−4.579 (0.000)	−5.599 (0.000)	−4.069 (0.000)

Note: Δ is the first-difference operator and p-values are in parentheses. The LCC test was performed using a Newey–West bandwidth selection with Bartlett Kernel. The Schwarz Bayesian Criterion was used to determine the optimal lag length.

. It is evident that not all variables are stationary in level, but they become stationary in their first-difference.

4.3. Cross-Sectional Dependence

Another crucial step in panel causality analysis is to take into account possible cross-section dependences across countries. It is well documented that a high degree of globalization, international trade, and financial integration make a country more sensitive to external economic shocks. Thus, ignoring cross-section dependency leads to substantial bias and size distortions [44,45]. To test for cross-sectional dependency, the Pesaran’s CD test statistic was used in this study [44]. The procedure for computing the CD test statistic is as follows:

$$CD= \sqrt{\frac{2T}{N(N-1)} \left({\displaystyle \sum }_{i=1}^{N-1} {\displaystyle \sum }_{j=i+1}^N\widehat{P}ij\right)}$$

(2)

where ($T$) is the time interval, ($N$) is the number of cross-section units, and ($\widehat{P}ij$) is the pair-wise correlation between cross-sections. According to the test results given in

Table 4. Cross-section dependency test results.

Test	Statistic
CDBP	248.19 (0.000)
CDLM	41.479 (0.000)
CDLMadj	41.302 (0.000)
CD	15.746 (0.000)

Note: p-values in parentheses.

, the null hypothesis is rejected for GCC countries at the 1% level of significance, meaning that a shock in one of these countries is transmitted to other GCC countries.

4.4. Panel Causality Analysis

The existence of both cross-section dependency and heterogeneity across the six GCC countries under analysis requires a panel causality method to account for these dynamics. The panel Granger causality test developed by Dumitrescu and Hurlin [46] was employed in this study. Dumitrescu and Hurlin [46] proposed a panel causality test based on the individual Wald statistic of Granger non-causality, averaged across the cross-section units. The testing procedure considers both the heterogeneity of causal relationships and the heterogeneity in the regression model used for testing the Granger causality. The liner panel regression model is as follows:

$$y_{i,t}= {\alpha }_i+ {\displaystyle \sum }_{j=1}^J{\delta }_i^j{y}_{i,t-j}+{\displaystyle \sum }_{j=1}^J{\beta }_i^j{x}_{i,t-j}+{\epsilon }_{i,t}$$

(3)

where ($y$) is real GDP, ($x$) refers to international tourism spending, and ($J$) is the lag length. Dumitrescu and Hurlin [46] state that “a homogenous specification of the relation between the variables $(x)$ and $(y)$ does not allow to interpret causality relations if any individual from the sample has an economic behavior different from that of the others.” Thus, they propose an average Wald statistic that tests the null hypothesis of no causal relationship for any of the cross-section units$(H_0: {\beta }_i=0, (i=1,\dots ,N))$, against the alternative hypothesis that a causal relationship exists for at least one subgroup of the panel $(H_1: {\beta }_i=0, (i=1,\mathrm{..}{N}_1)$; $({\beta }_{i }\ne 0, \left(i=N_1+1,N_1+2,\dots ,N\right))$. Rejection of the null hypothesis with $(N_1=0)$ indicates that $(x)$ Granger causes $(y)$ for all (i), whereas rejection of the null hypothesis with $(N_1>0)$ provides evidence that both the regression model and the causal relations vary from one individual country or sample to another. Under these circumstances, the average of the individual Wald statistics generated by Dumitrescu and Hurlin [46] assumes the following form:

$$W_{N,T}^{Hnc}=\frac{1}{N}{\displaystyle \sum }_{i=1}^N{W}_{i,T}$$

(4)

where $(W_{i,T})$ is the individual Wald statistic for the ὶ-th cross-section unit. The results from the panel Granger causality analysis are reported in

Table 5. Panel causality test results.

Cross-Section Units	Null Hypotheses
	H0: TR Does not Cause Growth		H0: Growth Does not Cause TR
	Wald	Decision	Wald	Decision
GCC Countries	3.410	Accept	9.625*	Reject
Bahrain	15.483*	Reject	4.884***	Reject
Kuwait	1.747	Accept	9.832*	Reject
Oman	2.769	Accept	2.498	Accept
Qatar	2.894	Accept	11.861*	Reject
Saudi Arabia	1.567	Accept	11.628*	Reject
United Arab Emirates	0.655	Accept	8.525**	Reject

Note: *, **, and *** indicate rejection of the null hypothesis at the 1, 5, and 10 percent levels of significance, respectively.

. In the case of GCC countries as a whole, the results show a one-way Granger causality running from economic growth to tourism development, statistically significant at the 1% level. For country specific hypotheses, the results show that Kuwait, Saudi Arabia, Qatar, and the United Arab Emirates follow the path of economic-growth-driven tourism; this hypothesis is significant at the 1% and 5% significance level, respectively. The feedback hypothesis is supported for Bahrain, implying that tourism receipts positively affect economic growth (the result is significant at the 1% level), and economic growth stimulates tourism growth (significant at the 10% level). Clearly, tourism growth is of great importance to the Bahraini economy. The empirical results show that there is no causal relationship between tourism revenue and economic growth in the case of Oman, in line with the results of previous studies.

With the exception of Oman, Table 5 exhibits some mixed results about the country specific hypotheses. These results are either theoretically or empirically consistent with the existing literature. As Chen and Chiou-Wei [13] state, the causal relationship between tourism and economic growth is not stable, and there might not be a unique relationship type among GCC countries. In this sense, Eugenio-Martin, Morales, and Scarpa [40], Chang, Khamkaev, and McAleer [41], Leitão [26], Tavares and Leitão [27], Chou [37], and Tugcu [39] in their empirical investigation argued that, depending on different dynamics, such as economic and political structures, trade openness, initial tourism conditions, investment rate, sectorial linkages, and historical background of the countries, this relationship may vary. Therefore, different studies could show different results in these countries in the context of tourism-led growth, even for the same sample.

The empirical results indicate that, for four GCC countries (Kuwait, Saudi Arabia, Qatar, and the United Arab Emirates), the economy-driven tourism growth hypothesis holds true. These results also support the tourism-led growth hypothesis in the case of the Bahraini economy, while no relation between these variables in found for Oman. There are several possible explanations for such mixed results for GCC countries. One of the explanations for the direction of causality, from economic growth to tourism development, in most GCC countries may lie in the history of their tourism industry structure. It is worth noticing that nations (such as GCC oil-rich countries) with no long history of tourism development initiatives rely on their endowments to maintain tourist inflows, and that economic growth is the only way to invest in tourism and sustain tourist inflows in such countries [26,38,39]. In absence of economic growth, the tourism industry shrinks. Another explanation may lie in the economic structure and the contribution of the tourism industry to the GDP of these countries. If this industry contributes only a small portion of the national GDP, their relationship is more likely to support an economy-driven tourism growth [14,19,25]. The direct contribution of the tourism industry to the GDP of these countries is very small, ranging from 1.5% for Kuwait to about 4.1% for UAE in 2015 [2]. Thus, GCC policymakers need to focus more on economy-oriented policies to promote tourism development as a potential source of economic growth. A third explanation may rely on the link between the tourism sector and the rest of the economy. Inter-sectorial linkages between tourism and the rest of the economy in GCC countries are either weak, or not strong enough. As Tugcu [39] points out, if linkages between the tourism industry and the rest of the economy are strong enough, economic growth may become an instrument to improve the efficiency of the tourism sector.

With respect to the Bahraini economy, one possible explanation for the impact of tourism on economic growth may lie in the tourism attraction policy adopted by Bahraini policymakers. Since the 1980s, Bahrain has become the regional hub for international banking and financial services and international leisure facilities [41,47]. It is well documented that countries with an established tourism attraction are likely to benefit from tourist inflows in the long run, which, in turn, fosters economic growth [25,38,39]. In the case of Oman, tourism development and economic growth do not seem to be sensitive to each other. One can attribute this to a relatively small contribution of tourism revenues to the overall economic output. Thus, tourism development has little or no impact on economic growth in this country. Therefore, the Omani policymakers need to develop travel and tourism policies that are not dependent on economic activities.

5. Conclusions

The main objective of this study is to investigate the causal relationship between tourism development and economic growth in Gulf Cooperation Council (GCC) countries in a multivariate model, using panel data for the period 1995–2012. The study adopts a panel Granger causality analysis approach to assess the contribution of tourism to economic growth in GCC countries as a whole, and in each individual country. In the case of GCC countries as a whole, the results show a one-way Granger causality, from economic growth to tourism growth. Furthermore, Kuwait, Saudi Arabia, Qatar, and the United Arab Emirates follow the path of economy-driven tourism growth, as hypothesized. The reverse hypothesis (i.e., tourism-led growth hypothesis) holds true for Bahrain, while there seem to be no causal relationship between tourism and economic growth in the case of Oman.

Given the fact that GCC countries have similar climate and natural resources, attracting international tourists may require some degree of specialization to offer unique services and activities. GCC policymakers intended to establish an intra-GCC tourism-oriented approach with the goal of enhancing GCC countries as a tourist destination. For the intra-GCC tourism policy to succeed, policymakers should formulate their tourism policy based on country-specific comparative advantages, rather than competitive advantages. In terms of recommendations for their economic policy, we argue that GCC countries should continue to promote the tourism sector and, at the same time, strengthen the link of the tourism industry with the rest of the GCC economy. To this end, a new GCC tourism framework should be designed, based on the notion of sustainable tourism in the region, and it should be managed following the three principles of sustainability. The economic, social, and cultural benefits of sustainable tourism need to be the focus of the sustainable development of the sector. Worldwide, the path towards sustainability in the tourism and hospitality industry has been a long process and remains an unfinished business, while in GCC countries this path has just begun. Therefore, GCC governments should play a major role in boosting and maintaining the path of sustainable tourism and hospitality industry within their regional destinations.