The Role of Tourism in Development: A Dilemma Between Economic Growth and Mangrove Forest Degradation

Development Studies of Interdisciplinary Faculty, Satya Wacana Christian University, Jl. Diponegoro No. 52-60, Salatiga, Indonesia 50711 Received , 2019/Accepted , 2019 October 22 November 29 Elia Radianto Titi S Prabawa Wilson M. A. Therik Gatot Sasongko , , , , usilowati uther Marthen L Ndoen * (A Case Study of Regencies/Cities in North Maluku Province) The Role of Tourism in Development: A Dilemma Between Economic Growth and Mangrove Forest Degradation * author, email: eliaradianto@yahoo.com Corresponding The establishment of the Morotai Island Regency as one of the 10 National Tourism Strategic Areas and Special Economic Zones is a central government strategy to accelerate the development of regencies/cities in North Maluku Province, which are still classified as underdeveloped regions in Eastern Indonesia. This study aims to analyze the influence of tourist arrival rate, price, human development index, and tourism promotion policy on economic growth by using regression of panel data of fixed-effect model (FEM) based on the feasible generalized least square (FGLS) VI method in eight regencies/cities in North Maluku Province during the period of 2012 2017. This finding shows that tourism development was closely related to economic growth, both in the short and long term. However, the development of tourism facilities and other supporting tourism facilities on the coast by coastal landfill had sacrificed the growth of coastal mangrove forests. Therefore, the local government is encouraged to issue regional regulations on the implementation of sustainable tourism businesses and educate the local community and visitors to play a role in protecting the environment of tourist destinations sustainable. –


Introduction
Apart from the tourism popularity as reported by the WTTC , in-depth studies are highly needed, specifically on the role of the tourism sector to improve the economic growth in a region or a country. In 2005 Oh carried out a research that applied the two-stage Enggle and Granger and bivariate ector utoregression (VAR) model and used quarterly data (during the first quarter of 1975 to the first quarter of 2001). As a result, Oh (2005) found that the tourist (2019) , v a However, state that there were probably two methodological weaknesses in the study, which caused the results of an analysis of economic growth led by tourism to be insignificant. First, the priority of the approach was not motivated by the economic theory but on the econometric approach. Second, researches generally only focused on individual countries so that tourism did not have marginal explanatory power. Therefore, by using the quantile regression model according to the least-squares method based on a sample of 109 countries during the period of 1995 2011, it could be proven that the hypothesis of economic growth led by tourism (number of tourist arrivals, gross capital formation, average school age as a proxy for source quality human resources, and percentage of expenditure for R & D) could be accepted according to theory. Similarly, by applying east quare ummy ariable (LSDV) method, Nunes and Sequira (2011) found that the Ng and Du (2011) l s d v arrival rate did not influence economic growth. Ohs' hypothesis was stating that the economic growth led by tourism does not support the Korea economics context due to a rapid economic expansion, which caused a tendency in the tourist arrival only in the short term.
Meanwhile, a previous research by Pin et al. (2016), which used a quantile regression model in 109 countries during 1995 2011, added that besides the number of tourist arrivals, the price level (foreign exchange rate) variable was also significant for the economic growth. In addition, the study concluded that the hypothesis on tourism-led economic growth was acceptable according to theory. Further, a study of Guerrero et al. (2015), which used a multiple regression analysis in Andalucía, Spain in the period of 2005 2012, found that the tourist arrivals, number of starred hotels, average tourism industry income, and Pound sterling exchange rate variables were significant and positive, except the hotel price index which was significant. However, the coefficient was negative, according to the theory. Dwyer et al. (2016) used a dynamic panel data model in four countries in the period of 1990 2008 and added that the price coefficient (that was shown through rice ompetitiveness --p c i p p p ndicator/PCI using urchasing ower arity/PPP) was negative but not significant. It implied that the intended outcome as a result of several things, including the possibility that tourists were very price sensitive as a result of a large number of destinations that tourists could visit and poor travel infrastructure availability.
In contrast, previous research by Anggraeni (2017), which used panel data of the Fixed Effect Model in eight ASEAN countries during the period of 2000 2012, found that the international tourist arrival variable significantly influenced the economic growth, which was following the hypothesis of economic growth led by tourism. Sadiku et al. (2017) took a sample of 6 Western Balkan countries during 1998 2014, and according to panel data regression, it was found that the tourist arrival rate, tourist reception, private investment, and government spending had a positive effect on real per capita income in the long run. A study by Zeren (2018), which used Dynamic Common Correlated Effects Mean Group Estimator in 81 provinces during the period of 2000 2015 in Turkey, found that tourist arrival and length of stay variables were positively related to the economic growth as the higher the tourist arrival rate and length of stay, the more the spending and the more positive the impact on economic growth.
---tourism arrival rate (tourist arrival rate/number of population) and income variable (share of tourist income on gross domestic product) positively influenced the economic growth, both for big or small countries (had less than 5 million population) and poor countries with below-average GDP per capita during the period of 19802002. -Similarly, Sadeghi et al. (2011 used a panel vector autoregressive (P-VAR) approach in a sample of 20 developing countries during the period of 1995 2008 and found that the tourist arrival variable was significant for the economic growth. It was concluded that the tourism-led growth hypothesis in developing countries could be confirmed, and it justified the need for government intervention in promoting tourist destinations to increase the -In contrast to previous researches, Indonesia is currently trying to increase its economic growth through the tourism sector as a driver of development to exploit the potential of tourism in regions through the establishment of 10 National Tourism Development Strategis Areas (k s p n awasan trategis ariwisata asional/KSPN Morotai Island Regency in North Maluku Province is considered as an under-developed regency in Eastern Indonesia but has a variety of potential natural, cultural, and world-class tourism excellence in the form of World War II Asia Pacific heritage. It has been famous as the "pearl on the lips of pacific" and attracts the Indonesian government to set it in the 10 KSPN. In fact, to accelerate the economic development of the surrounding area, the Central Government also set it as a Special Economic Zones ( KEK It is expected that it can accelerate the economic growth through the tourism sector and other associated sectors as well as provide added value in overcoming poverty and job creation, which in turn will increase the prosperity of the community in Morotai Island Regency and other regencies/cities in North Maluku Province. On the other hand, Butar-Butar et al. (2015), found that the reduced number and function of mangrove forests as a buffer zone for coastal zones in Indonesia was triggered by loosening of licenses issued by regency/city or provincial governments to the private sector for exploitation permits for coastal locations to become tourist destinations that were not environmentally friendly. More specifically, by using SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis, Fithor et al. (2018) found that efforts to bring mass tourism to the opening of Maron Beach in Semarang had caused negative impacts on the mangrove ecosystems in the form of (1) reducing the function of mangroves in protecting the coast from coastal abrasion, (2) reducing the ability of wind abrasion protection, (3) weakening environmental conditions, and (4) reducing fish production. Figure 1 shows that the existence of KSPN and KEK policies, through the construction of various tourism support facilities as well as tourism promotion funded by the central government through the calendar of events since 2012 2017, has been able to increase the number of tourist arrivals in almost all regencies/cities except West Halmahera Regency which experienced a decrease in tourist arrivals from 17.812 to 6.992 tourists (60.75% declining rate). Likewise, there has also been an increase in changes in the human development index, and price measured using urchasing ower arity p p p number of tourist arrivals to increase the economic growth. Correspondingly, used Granger Bashir and Suhel (2018), c m ausality odel and simultaneous equations in South Sumatera Province, Indonesia during the period of 2000 2015 and found that government spending policies to increase tourism promotion were significant to the economic growth and concluded that public policy initiatives for development tourist destinations would have a positive impact on the economic growth. Province. The existence of changes in positive growth in the four variables referred to is quite influential on the occurrence of changes in economic growth in each regency/city except East Halmahera Regency, which has a negative growth of 6.42%. While other regencies/cities experienced positive growth changes (13.27% for West Halmahera, 8.51% for Central Halmahera, 18.44% for Sula Islands, 14.78% for North Halmahera, 13.86% for Morotai Island, 23.90% for Ternate City, and 19.78% for Tidore City).
presents all the details. Figure 1 However, the policy of bringing mass tourism to pursue economic growth with a loose policy to carry out construction of large-scale tourism facilities (for example construction of recreational areas, culinary centers and hotels by making landfill on the coast of Morotai Island) in Morotai Island Regency as one of the KSPN and KEK greatly influenced the growth of mangrove forests. Meanwhile, the mangrove forests helped in maintaining marine and coastal ecosystems and could also be used as an attractive tourist destination. Therefore, further study would be highly needed.
According to the data and explanations mentioned earlier, this study would utilize the panel data model test (pooled data) during the period of 2012 2017 to estimate the influence of tourist arrival rate, spending, human development index, and tourism promotional policies to per capita ross egional omestic roduct in regencies/cities in North Maluku Province. However, it seemed that the efforts to pursue economic growth by bringing mass tourism through the construction of tourism facilities such as beach g r d p recreation parks, culinary centers, and hotels on the coast by coastal landfill had sacrificed the growth of mangrove forests on the coast. This could be seen in the example of the case in the Morotai Island Regency, which showed that in 2005, the area of mangrove forests was 15.618 a. However, in 2017, the area of mangrove forests had decreased to 1.833 Ha or had decreased by 88.26%. Based on the data referred to and in relation to the demands of development done by the government, it is important to understand the dilemma between the effort to teach coastal tourism development and construction of coastal tourism facilities on the coast of the forest mangroves, so it is interesting to do research. h

Research esign d
This study used an economic growth model developed by Solow (1956) and was based on several previous studies which examined the contribution of tourism to long-term economic growth (Nunes & Sequeira, ;Zhang, 2015;Pin et al., Erbas et al. 2017). Compared to previous empirical researches, this study examines the influence of four independent variables (tourist arrival rates, spending, human development index, and tourism promotional policies) on ross egional omestic roduct per 2011 2016; g r d p capita as the dependent variable. For this reason, an econometric analysis was conducted using panel data (pooled data) or a combination of time series and cross section data, which allows not only an increase in degrees of freedom and number of samples with a better estimation but also overcome the lack of data by considering specific country or regional effects, have large variability and reduce collinearity between explanatory variables, and produce  Greene (2012) and Ekananda (2016) added that the use of dynamic linear models using the VI instrument variable could be used to analyze the long-term influence of the variables used. Therefore, by considering the opinion referred to, to test the hypothesis of economic growth led by tourism, in the long run, the dynamic panel data analysis would be used as a reference.

M s ethod
However, Radianto (1995), Insukindro and Radianto (1995), Radianto (2001), Petit and Seetaram (2012), and Leitão (2015) stated that in relation to the use of the best model according to econometric criteria, the use of the model must be able to consider the current conditions depending on the conditions of the previous year and should be considered capable in covering a limited amount of data for a long-term analysis as expected by the theory. In other words, the model referred to should be able to calculate the extent of the current visit depending on the number of previous visits by considering the dissemination of information carried out by tourists in previous years.

Empirical variables and model used
This study used the ross egional omestic roduct (GRDP) per capita variable (as the dependent variable) and ourism articipation ndex (TPI) (as the independent variable, TPI is a ratio between the number of tourist arrivals and the number of population in the destination area). The spending was measured by using t p i Then, from the theoretical model of quation , it can be derived into the general form of economics as is expressed in Equation : Y economic growth TPI level of tourist arrivals P price of commodities consumed by tourists HDI uman evelopment ndex and DTPP ummy of ourism romotion olicy. Then, i = 1, 2, ... N; t = 1, 2, ..., T, or in other words, the i-index shows the dimensions of the crosssection, and the t-index represents the time which indicates the time-series dimensions. While β is a constant, β By observing the data and variables used in Table 1, a basic model would be derived to test the growth hypothesis led by tourism (Prasad et al., 2010;Ng & Du, 2011;Nunes & Sequeira, 2011;Pin et al., 2016;Sadiku et al., 2017;Bogdan et al., 2018;Zeren, 2018) to the quation E .
[1] p p p urchasing ower arity (PPP) as a proxy of the price (P). Others analyses used in this study are uman evelopment h d i h r q d t p p ndex (as a proxy for uman esources uality/HDI) and ummy ourism romotional olicy (DTPP). The DTPP is the governments' policy to promote tourism in the form of an annual tourism festival event carried out and/or participated by regencies/cities through government funds (State Budget and Regional Budget) (as a proxy of tourism promotional policy). While the data used in this study were time-series data from 2012 to 2017 with eight out of ten regencies/cities. The consideration for choosing the sample was because the two regencies (South Halmahera and Taliabu Island) were newly expanded regions. Therefore, the researchers were faced with data availability problems, and adjustments were made to the availability of data in the eight regencies/cities in question. Table 1   The next step before lowering the empirical model estimated. There were two-phase procedures for panel data testing, which aimed to choose the best panel data model in estimating the empirical model to be used. The intended procedures were Chow and Hausman Test which aimed to obtain the best model among the three panel data models (common effect model or pooled ordinary least squares PLS, random effect model REM, and fixed effect model FEM) (Sequeira & Nunes, 2011;Petit & Seetaram, 2012;Susanti, 2013;Susilawati et al., 2014;Ekananda, 2016;Yuliana & Sitorus, 2018;et al., 2019 Whereas the F table is obtained from:

Hausman est t
The second step used to determine which model was more suitable between the or the . This should be done for the selection of models with a constant and random effect using the Hausman test. In this test, the effect of the rror terms could always be random. However, in the dummy variable model, the nature of randomity was limited in the sample data used. By using the error components model, the obstacles were assumed to be random for the entire population, whereas the dummy variable model did not assume that. Therefore, the dummy variable model could be used more freely. However, if the randomity assumption occurred, then the assumption increased the efficiency of the estimation.
In , β , β β β 1 2, 3 4 , and were not zero and statistically significant, it indicated that TPI, P, HDI, and DTPP had a as explained below (Baltagi 2003(Baltagi 2005(Baltagi 2008Ekananda, 2016) ; ; as shown as Equation Furthermore, this equation substituted in the equation above will be: In other words, the conditions that must be met are as follows: Based on quation , a natural logarithmic transformation was performed for all variables except the tourism promotions' policy dummy variable as follows If a panel data model had been found with a constant and random effect throughout the observation period as the two tests mentioned referred to, then quation could be derived into a dynamic economic growth function that could be estimated as an empirical model according to the following model: Chi tat > Chi-table  was not zero and statistically significant, there was a long-term effect on Y in estimating the long-term effects of TPI, P, HDI, and DTPP variables on Y. Furthermore i = 1, 2, ... N; t = 1, 2, ..., T, or in other words, the i-index showed dimensions of the crosssection, and the t-index represented the time which indicated the time-series dimensions. While β was an intercept, the β β 5 t-1 0 1 , β , β β β 2 3, 4, 5 and were the measured parameters. Whereas µ was an error term, and the underlying assumption was that the independent variable was non-stochastic, and the error term followed the classical assumption .

Results and Discussion
Selecting the FEM or PLS odel by Chow est m t To determine the use of the FEM or PLS, which was most appropriate to use in estimating panel data, a Chow Test was carried out whose calculations can be seen in  Table 4.
For this reason, heteroscedasticity testing using the estimation method of eighted east quare (WLS) or often referred to as eneralized east quare (GLS). Table 5 w l s g l s shows that the value of um quare esidual in eighted tatistics (SSRWS) is smaller than the value of um quare esidual in s s r w s s s r u s nweighted tatistics (SSRUS) (0.006981 < 0.008983). Thus, it can be said that the FGLS VI model avoided heteroscedasticity.

A ong-erm nalysis by sing FEM with FGLS VI l t a u
To obtain the best and efficient model in accordance with the LSDV method, the FGLS method with instrument VI variables needs to be tested for heteroscedasticity and autocorrelation. According to Hsiao (2003), Baltagi (2005), Baltagi (2008), Greene (2012), and Hassler et al. (2012), heteroscedasticity and autocorrelation often appeared in timeseries data, as panel data sets had proximity to cross-location data on one side and also used time-series data. The existence of heteroscedasticity would result in a consistent estimate of the regression coefficient, but it had an inefficient estimate. Therefore, there would be an error in the estimation standard and would be biased.
Autocorrelation detection should be done to avoid the frequent variation of data in time series data on panel data so that it caused a consistent regression estimation, but the estimation coefficient was not efficient, and the error standard was biased (Baltagi 2005(Baltagi 2008Greene, 2012). However, the ; Table 3 Results of selecting common effect model or fem selection by how test C  In the long-term hypothesis, it can be seen in the adjustment variable or the economic growth lag variable Table 6 shows that the R value is 0.997359. It indicated that variations in changes in the economic growth could be explained by tourist arrival rate, spending, human development index, and tourism promotion policy variables as much as 99.74%, while the rest 0.26 % was influenced by other variables, which were not included in the model. 2 results show that the Durbin Watson (DW) test with n = 40 and 5 explanatory variables obtain DW of dl = 1. 048 and du = 1. 684 at . 05. This indicated that there were autocorrelations (DW value of 2. 370237 value of 4 -dl (4 -1. 048 = 2. 952) and 4 -du (4 -1. 684 = 2. 316) or 2. 279 > DW < 2. 622), Therefore, it can be said that there was no autocorrelation which showed that the estimation with FGLS VI had been met.  Table 6 presents the results of the analysis.
In the short term, according to Table 6, there were three variables with highly significant hypotheses of a 99% confidence level. They were hypotheses of TPI, HDI, and DTPP, while the price variable (P) was not significant. Thus, based on the FGLS VI method, the results reveal that the TPI, HDI, and DTPP variables had a positive effect on economic growth. In other words, if there were an increase in the TPI, HDI, and DTPP, it would stimulate the economic growth for Morotai Island Regency and other regencies/cities in North Maluku Province. Although the price (P) variable was not significant, the results were negative and following the theory. It indicated that an increase in prices would cause a decrease in the number of tourist arrivals, which in turn would reduce economic growth.
(Y ), which is significant at the 99% confidence level. Thus, it could be continued with the calculation of the long-term coefficients of each significant independent variable: TPI (0.0073), HDI (3.447), and DTPP (0.1142). In other words, if there were an increase in the tourist arrival rate, HDI, and t-1 Pin et al.
stated that tourism did not grow separately, but depended on efficient infrastructure in supporting the movement of goods and people, quality of human resources, level of technology, broader and more diverse development policy strategies, and a tendency to support tourist arrivals to increase the economic growth. (2016), Therefore, an in-depth and comprehensive study of the stipulation of Morotai Island Regency as KSPN and KEK by the Indonesian government was needed to make it an engine of economic growth for the surrounding regencies / cities in North Maluku Province, especially the one related to tourist arrival rate, prices consumed by tourists, quality of human resources, and tourism promotion policies variables, and the sustainability of the environmental ecosystem.  (2018), these findings imply that there was a relationship between the tourist arrival rate and economic growth, both in the short and long term. There was a significant influence on the tourist arrival rate variable supported by 143 marine tourism objects (53.76%) and 123 natural, cultural, and historical tourism objects (46.24%) of the total attractions of 266 in eight regencies/cities observed in North Maluku Province. Table 7 Ng and Du (2011) (2011), (2016), presents the number of tourism objects by type in regencies/cities in North Maluku Province.
The HDI variable, as a proxy for the quality of human resources, was significant for the economic growth during the observation period of 1995 to 2011. This is in line with the finding of Ng and Du (2011) that the average school age, as a proxy for human resource quality, was significant according to theory and added that it was necessary for the government to provide training to prepare the skilled and creative workforce in supporting the tourism sector. However, this study did not only use the average school-age as a proxy for the quality of human resources, but all components of HDI ( ife xpectancy ge/LEA, chool ength xpectancy/SLE, l e a s l e a l s e c verage ength of chool/ALS, and xpenditures per apita/EPC) to measure the quality of human resources. Table 8. shows that there is an increase in HDI of regencies/cities in North Maluku Province from 63.93 in 2012 to 67.20 in 2017 or an average increase of 1.00%.
Although the HDI growth of Morotai Island Regency, as a KSPN and KEK region, is above the average regencies/cities in North Maluku Province, its HDI position in 2017 is the lowest and in a 'moderate' HDI group (60 ≤ 60.71 < 70). It was a result of a low SLE, which reaches 12.17 years, ALS of 6.89 years, and EPC adjusted to IDR 6.17 million annually (Table 9). This implied that efforts were needed to improve the dimensions of education and decent living standards, which in turn would support the development of tourism and the growth of a sustainable economy.
Source: BPS of North Maluku Province, 2018.   The findings show that the tourism promotional policy applied in each regency/city by the government aimed to bring in as many tourists as possible to increase economic growth. This is in line with a statement by Bashir and Suhel The construction of a Rehabilitation Place and Culinary Center, as well as Hotel buildings in the Morotai Island Regency in the coastal area by making landfill have obtained permission from the local government. The permission was not the results of an Environmental Impact Assessment. Similarly, there are no regional regulations governing the "Mangrove Ecosystem Management," which can be used to determine the areas that can be used as coastal tourism businesses. Therefore, we cannot stop legally any development in mangrove forests (M., November 2019). Figure 2 shows that before the development on the coast of Morotai Island, most of its beaches are surrounded by mangrove forests. However, after the development, as explained earlier, it has caused a reduction in mangrove forests (Figure 3, Figure 4).
Unfortunately, on a large scale, the policy had not yet considered the environmental aspects which affected the marine and coastal ecosystems. For example, in Morotai Island Regency, it was very easy to get a tourism business permit because without going through an Environmental Impact Analysis. A key informed who worked at the Morotai Island tourism agency that: (2018), that public policy initiatives were needed for the development of tourist destinations to increase domestic and international tourist demand, and would have a positive impact on the economic growth in the future.
The above statement had empirical evidence which showed that there were constructions of parks, culinary centers, and hotel buildings in the coastal area of Daruba City by making landfill in the coast of the mangrove forest area, causing the mangrove forests to be damaged and eventually died.
There is a negative effect in the form of pollution from the building construction (Figure 3, Figure 4). It damaging some of the living mangrove plants next to the northern part of the buildings -which were usually used by fishermen for shelter     -and some were also dead. Figure 5 picturizes the trail of mangrove forests in the form of dried tree trunks that used to grow in the sea.
The loosen rules, as stated by the key informant and shown by the empirical evidence, showed that it profoundly affected the mangrove forests, and there was a decline in the mangrove forest area and their types. In a research by Ahmad (2014), the data used was from the Central Statistics Agency of Morotai Island since 2005-2017, and it was found that there had been degradation mangrove forest or a decline in the mangrove forest area of 88.26% or with an average decay per year of 21.18%, and also species of mangrove forests decreased by 26.92% or with an average annual decrease of 3.43. Table 11 explains the changes affected.
The above findings are similar to the findings of Butar-Butar et al. (2015) and Fithor et al. (2018) who found that efforts to bring mass tourism (mass tourism) with a loosen licensing policy to the private sector to exploit the coastal area into tourist destinations (which was not environmentally friendly) would atly impact the mangrove ecosystem negatively in the form of (1) reduce the function of mangroves in protecting the coast from coastal abrasion, (2) reduce the ability to protect wind abrasion, (3) weaken (2) n the long run, a legal organization was needed by involving the legislative role in formulating sustainable tourism development policies by issuing a i Regional Regulation on the Implementation of Tourism Businesses. environmental conditions, and (4) reduce fish production.
This implied that a policy was needed to minimize the negative effects on the coastal tourism destinations, especially in protecting marine and coastal ecosystems in the form of mangrove forests, seagrass beds and coral reefs. Similarly, by considering the fact that Morotai Island Regency had the potential of fisheries sector in the form of the potential of the Indonesian Archipelagic Sea Lanes III -a tuna migration pathway, it was the right time to take over the development of coastal tourism objects by considering: (1) n the short term, a study involving relevant technical agencies to issue a which regulated the operation of tourism businesses by considering aspects of environmental and ecosystem sustainability, was needed.
i Circular Letter ( t 3) he role of the government was needed to educate the local community and visitors to protect the environment so that the enchanting beach tourist destinations could have a dual effect in the form of potential attraction of the beauty of tourist destinations and could maintain the potential of  Figure 5 A photo of the remains of dried mangrove trees but can still be used by fishermen to tie their boats to the north side of the park, culinary center building and hotels on the coast of Daruba City, the capital of Morotai Island Regency , . On the other hand, although the results show that the coefficient of price variable was negative according to the theory, but it was not significant and it was similar to a finding by Dwyer et al. (2016). However, according to Guerrero et al. (2015), although the price variable was negative, but if there was a support for quality services by the private sector, and supported with the availability of qualified infrastructure, it would greatly support an increase in the number of tourists, which in turn would increase the economic growth.
Based on the results of hypotheses testing of the four independent variables, it shows that there are three significant independent variables ( ourism articipation ndex, uman t p i h d i evelopment ndex, and a dummy of tourist promotion policies) except the price variable (P). Similarly, the adjustment variable (Y ) was significant. It indicated that the FGLS VI method could be used to predict the effect of each independent variable on the economic growth in the long run. Therefore, it could be concluded that the hypothesis of economic growth led by tourism could apply to Morotai Island Regency as an area of KSPN and KEK, and also other regencies/cities in North Maluku Province. The policy should be reviewed and considered seriously by the government. For example, in the KSPN and KEK areas in the Morotai Island District, the loosen licensing policies had caused the extent of mangrove forests to decrease along the coast of Morotai Island by 88.26% and also species of mangrove forests decreased by 26.92%. This implied that a policy to minimize the negative effects on the coastal tourism destinations, especially in protecting marine and coastal ecosystems in the form of mangrove forests, seagrass beds, and coral reefs was needed in the form of: (1) in the short term, a study involving relevant technical agencies to issue a which regulated the operation of tourism businesses by considering aspects of environmental and ecosystem sustainability was needed, (2) in the long run, a legal organization was needed by involving the legislative role in formulating sustainable tourism development policies by issuing a t-1 C L ircular etter Regional Regulation on the Implementation of Tourism Businesses, and (3) the role of the government was needed to educate the local community and visitors to protect the environment, so that the enchanting beach tourist destinations could have a dual effect in the form of potential attraction of the beauty of tourist destinations and could maintain the potential of marine fisheries sustainable. There was also a need of synchronization of programs among regencies/cities in North Maluku Province with KSPN and KEK policies in Morotai Island Regency, so that the two intended government programs can be the engines of economic growth, both in the short and long term for regencies/cities in North Maluku Province and can be a source of foreign exchange earnings for Indonesia.