MEASURING TOURISM PENETRATION

IN SMALL ISLANDS*















Jerome L. McElroy

Professor of Economics

Dept. of Business and Economics

Saint Mary's College

Notre Dame, IN 46556

jmcelroy@saintmarys.edu



and



Klaus de Albuquerque

Professor of Sociology

Department of Sociology

College of Charleston

Charleston, SC 29424

albuquerque@cofc.edu



















*An earlier version of this article was presented to the 23rd Annual Caribbean Studies Association Meeting, St. John's, Antigua, (May 26-30, 1998).







Introduction

Two postwar forces have transformed the world of small islands, those less than one million in population and 10,000 square kilometers in area. They are the global spread of international tourism and the restructuring of insular economies. Since 1950 tourism has grown above five percent per year into the largest global industry representing roughly ten percent of world GDP and employment and seven percent of capital spending (Vellas and Becherel, 1995). This same period has witnessed the reorientation of small island economies away from traditional export staples like sugar and copra toward mass tourism development, related construction and financial services.

Nowhere has this transformation proceeded further than in Mediterranean and Caribbean destinations, the so-called "Pleasure Periphery" of Europe and North America respectively (Turner and Ash 1976). Recent research, however, has emphasized the socio-environmental damage of over-rapid mass tourism development particularly in the smaller more accessible and popular resort islands (Briguglio and others 1996a). In the Mediterranean, such negative impacts have included deforestation, beach alteration, near-shore pollution and reef destruction (Priestley and others 1996, Williams and Shaw 1991), and paralyzing summer crowding and other socio-cultural intrusions that threaten insular lifestyles and identity (Lanfant and others 1995). In the Caribbean, the scarring of mountains with condominium developments and road networks has caused widespread erosion and wildlife extinction (McElroy and others, 1990). The concentration of large infrastructure and resort complexes along delicate coastlines has destroyed mangroves and beaches and caused lagoon pollution from sand mining, dredging, and sewage dumping (Wilkinson 1989). Even in popular Pacific destinations, delicate mangroves have been harvested for construction material and reefs severely damaged by visitor trampling and collecting (Lobban and Schefter, 1997).

Problem and Scope

Because of these often irreversible socio-environmental impacts, critics are reassessing insular tourism's sustainability (Patullo, 1996). Some have argued for more comprehensive, integrated and participatory tourism planning (Briguglio and others, 1996a); Innskeep, 1994) while others have emphasized the need for greater destination control over visitor demand (Butler, 1991) and the high-volume imperatives of heavily capitalized multinational air, cruise and hotel interests (McElroy and de Albuquerque, 1996).

This research focuses on improving tourism measurement for strategic planning. Because tourism represents a series of products/services consumed in stages often from fragmented suppliers, there is no distinct tourism output that is easily defined or modeled in traditional economic analysis (Ashworth, 1995). A fortiori there are no accurate and agreed-upon measures of tourism's social and environmental externalities. In addition, there is no standard integrated measure-economic, social, environmental-of overall tourism scale and penetration. Developing such a comprehensive measure is important particularly for small islands when the size of an industry, often considered environmentally benign, has become a major source of conflict and concern (Stabler, 1997). Such an indicator would enable policy makers to assess the overall scale of island tourism and, in the comparative macro context of other island destinations, provide an early warning signal for anticipating nonsustainable levels of visitation.

This study has five parts: (1) a brief review of the literature on tourism impact measurement, (2) the construction of an index of overall tourism penetration, (3) an application of the index to a sample of 35 islands, (4) a cluster analysis to corroborate the findings of the index and (5) a conclusion highlighting the policy implications of the index for sustainable tourism planning in island microstates.

Literature

Despite a long and ongoing history of attempts to conceptually and empirically define tourism carrying capacity, little advance has been made (Johnson and Thomas, 1996). Most efforts have focused on qualitative definitions of saturation: infrastructure breakdown, subsistence disruption, congestion, realty inflation, rising crime and host hostility (Gray, 1974), declining cultural values and visitor satisfaction, and the replacement of lost natural amenities with man-made attractions (Butler 1980). Limited guidance is available, however, for empirically identifying the pace and scale of tourism development responsible for these changes.

According to a recent review (McElroy and de Albuquerque, 1998), actual measures of tourism penetration have been relatively crude and uni-dimensional. They have focused primarily on economic impact such as contribution to GDP and employment, spending per capita, hotel occupancy rate and so on (Mathieson and Wall, 1982), but such indicators are limited because they ignore tourism's important social and environmental effects. This same partial quality is characteristic of the more common measures of social and environmental pressure. It applies to both simple impact indicators like the Tourism Intensity Rate (visitors per host population and land area, Harrison, 1992) and the Defert Index (rooms or beds per host population, Oppermann and Chon, 1997) to the more complex measures like the CTO's (1993) Tourism Penetration and Density Ratios (visitors x average stay per population or land area x 365).

Early attempts by Bryden and Faber (1971) and Doxey (1976) to combine indicators into composite measures of overall impact were relatively rudimentary and arbitrary in construction (McElroy and de Albuquerque, 1998). The authors' early efforts (de Albuquerque and McElroy, 1992) to develop a simple tourism penetration model fashioned after Butler's (1980) resort cycle clustered small Caribbean islands according to three stages primarily determined by descriptively interpreting a dozen tourism indicators. Follow-up studies (1993, 1994) including Pacific islands suffered the same limitations. These shortcomings led to the recent construction of the Tourism Penetration Index successfully applied to a homogeneous sample of 20 small Caribbean islands (McElroy and de Albuquerque, 1998). The present study briefly reviews the Index construction and extends the application to a more heterogeneous mix of 35 small islands that include Atlantic, Caribbean, Mediterranean, Pacific and Indian Ocean destinations.

The Tourism Penetration Index (TPI)

A simple three-variable index was constructed duplicating the Caribbean (1998) study. To

ensure wide applicability and ease of interpretation, the TPI was derived from the same three variables: (1) visitor spending per capita to measure economic penetration, (2) average daily visitor density per 1,000 population to measure social penetration (visitors x average stay per population x 365), and (3) number of hotel rooms per square kilometer of land area to measure environmental pressure. Despite their uniform estimation and widespread availability, the three annual measures suffer at least three limitations. They do not capture the seasonal concentration of visitor flows characteristic of tourism; nor do they capture shoreline concentration so common in mountainous islands catering to mass sun-lust tourism; and they fail to measure the length of a destination's experience with and/or adaptation to tourism over time.

To operationalize the three variables, a sample of specifically small islands was selected not only because tourism has policy significance in such microstates with few development alternatives, but also because the delicate ecologies and social structures of such destinations are particularly vulnerable to the scale and style of mass tourism. Two commonly employed measures were used to define size, i.e. less than one million inhabitants (Kakazu, 1994) and less than 10,000 square kilometers in area (Beller and others, 1990). From an initial list of 42 islands a sample of 35 destinations was selected based primarily on whether they med both small size criteria and whether sufficient annual published data were available.

To ensure uniformity, all tourism data were taken from a standard and accessible source (WTO, 1997) and included: population, area, stayover (tourist) and one-day visitors, average length of visitor stay, total visitor spending, and total hotel rooms. Excluded destinations that met the demographic but not the area criterion included Bahamas in the Caribbean and the Pacific islands of American Samoa, Fiji, New Caledonia, Solomon Islands, and Vanuatu. The British Virgin islands was excluded because our previous work (1998) suggested the land-based indicators of the TPI tend to seriously overstate the impact of the dominantly sea-based character of its yachting tourism style. To rule out potential regional cross-sectional distortions in aggregate tourism performance, the year 1993 was chosen as sufficiently beyond the 1991 recession and Gulf War disruption but prior to destructive 1994-96 hurricane seasons in the Caribbean.

The sample island include 18 from the Caribbean, nine from the Pacific, four from the Indian Ocean (Comoros, Maldives, Reunion and Seychelles), Cyprus and Malta from the Mediterranean, and Bermuda and Cape Verde in the Atlantic. Table 1 provides he background data for the 35 destinations. Table 2 presents calculations of the three basic variables for each

(Table 1 about here)

island, the standardized indices based on these variables, and the resulting TPI scores and rankings from most to least penetrated. The standardized indices were calculated by taking the value of



each variable for each destination, subtracting the minimum value of that variable, and dividing the



result by the maximum value of the variable minus its minimum according to the formula:



(X - Xmin)/(Xmax - Xmin)



The resulting standardized values or indices are presented in the right panel of Table Two.

(Table 2 about here)

The TPI scores were determined with the following assumptions. Because different weights for each impact variable could not be established a priori either from the theory or the literature, it was assumed that each type of separate impact--economic, social, environmental--was as important as the other two in contributing to overall tourism penetration. This assumption yielded the formulation of an unweighted (equivalent impact) average Tourism Penetration Index whereby the three standardized values or indices were simply averaged. As additional experiments, two differentially weighted TPIs were also constructed assigning (in the first) visitor spending per capita twice the weight of the other two measures, and (in the second) similarly for both visitor spending and visitor density. These were justified on the basis that economic and social impacts were the more direct and accurate measures of penetration among the three indicators. The results (not shown) were almost identical to the equivalent weight (unweighted) formulation and left the basic island groupings intact with only a few minor one-step changes in destination position. These approaches, however, were rejected in preference to the more conservative unweighted formulation for two reasons: (1) to avoid potential bias arising from regional differences in the degree of unmeasured subsistence activity in the case of the per capita spending variable, and (2) in the case of the visitor density variable, to rule out any distortion due to dualistic development in archipelagic states which comprise half the sample.

Results

Although quite simple and rudimentary, the TPI scores yield results that broadly confirm what is expected from historical observation. The 35 islands are loosely ranked from most penetrated (St. Maarten) to least penetrated (Kiribati). Generally, the more traditional, developed, and accessible Caribbean and Mediterranean destinations populate the top half of the rankings while the more isolated, less visible, and recently emerging Pacific and Indian destinations populate the bottom half.

Although it is difficult to precisely determine island groupings based on decrete levels of penetration, as revealed by the TPI, the most penetrated islands are clearly distinct from the least. The former include a subgroup of seven high-density destinations characterized by per

capita visitor spending significantly over $8,000 and an average daily visitor density above 150 tourists per 1,000 residents. This is equivalent to a 15 percent increase in the daily population. Such islands include four mature Caribbean resort areas (Aruba, Caymans, St. Maarten, U.S.V.I.) and Bermuda, plus Malta, a favorite British Mediterranean resort, and the Northern Marianas, a popular honeymoon getaway for the Japanese. According to previous research (McElroy and others 1993), such affluent islands are characterized by crowding, short average stays, declines in repeat visitation and visitor satisfaction, large-scale facilities, and an increasing propensity to replace lost natural amenities with man-made attractions (shopping, gambling, water sports). These are also the destinations most frequently cited for environmental spillovers induced by tourism development (Jenner and Smith 1993).

The group of least penetrated islands that populate the bottom end of the TPI scale includes 13 low-density destinations typified by visitor spending less than $400 per capita and visitor and room densities less than 15 per 1,000 and 5 per kmē respectively. This last contrasts sharply with the average 50 room per km2 for the most penetrated islands. They include two sub-groups comprising seven new entrants at the bottom of the scale - five Pacific destinations plus Capte Verde and Comoros - and six more developed destinations comprising four Windward islands in the Caribbean and two outposts: French Polynesian in the Pacific and Reunion in the Indian Ocean. Our past research (1992, 1993, 1994) suggests such destinations are generally characterized by relatively pristine natural assets, small-scale facilities and infrastructure, long average stays and limited international visibility. These islands have the greatest potential for developing into ecotourism destinations with a sustainable visitation style based on the low-density consumption of unusual natural and cultural assets. In fact, both Dominica and Western Samoa have recently achieved international recognition because of their spectacular beauty and cultural uniqueness respectively.

Although it is difficult to define precisely the boundaries between the intermediate destinations and these most and least penetrated groupings, the TPI scores plus past historical observation suggest two subgroups: the more and the less penetrated. The more penetrated intermediate islands include nine destinations averaging visitor spending above $3,500 per capita and over 80 daily visitors per 1,000 population. They include six popular Caribbean resort areas, two Pacific islands (Guam and Cook Islands), and Cyprus. Past analyses suggest they are usually characterized by rapid arrival growth and hotel room construction. This growth in recent decades has been characteristic of Cyprus in the Mediterranean (Pearce 1987), of Antigua, St. Kitts-Nevis and Turks and Caicos in the Caribbean (CTO 1993), and of Guam in the Pacific (WTO 1993), the offshore golf getaway for the Japanese. This is less true of Montserrat whose retirement tourism began earlier (1960s) and has been virtually destroyed by volcanic eruptions since 1995. Barbados remains a more mature tourist area with a balanced agro-industrial economy while a well-managed marine park has established Bonaire as a popular international diving destination. In this more-penetrated subset of intermediate islands, Anguilla, Antigua and Guam appear the most likely, without strong domestic policy to limit visitor densities, to cross the threshold into high-density status in the coming years.

The less penetrated intermediate islands include six destinations averaging roughly $1,300 in per capita visitor spending, 40 daily visitors per 1,000 population, and 10 rooms per square kilometer. They include four Caribbean islands with relatively diversified economies - Curacao, Martinique, St. Lucia, St. Kitts/Nevis - plus the Maldives and Seychelles in the Indian Ocean. Although they share with the more developed intermediate islands the migration of land, labor and capital from agriculture and fishing to tourism as well as increasing resource-use conflicts, the pace of change is slower affording more opportunity for anticipatory planning. Because of this they may have a greater chance to establish sustainable tourism. In fact, both Maldives and Seychelles are noted for long-range planning and enforcing strict controls on the level of visitation to minimize intrusive socio-environmental impacts (Innskeep, 1994).

Cluster Analysis

Because of the potentially arbitrary nature of grouping islands according to the TPI scores alone, a cluster analysis was performed on the three standardized indicator values in the unweighted (equivalent weights) formulation to independently classify levels of tourism penetration. A standard software program (NTSYS-pc 1993) was used to calculate the unweighted average distance between each pair of islands across the three indicators. This produced a matrix of pairwise average distances. On the basis of this matrix, the program grouped islands into similar clusters as shown in Figure One.

These results broadly confirm groupings determined with the TPI scores, although some differences appear particularly at the lower end of the scale. There are only two primary clusters: (1) the seven most penetrated destinations with the highest TPI scores, and (2) the rest of the sample. Within the second cluster there are three sub-clusters that correspond roughly to the more, less and least penetrated groups from the TPI analysis. The cluster analysis identifies the same seven most penetrated destinations with the exception that Guam replaces Malta. The latter is treated as an outlier in a class by itself largely because of its unique (to the sample) tourism style: high-volume, low-cost package tourism to high-density (bargain) hotel and second-home (retirement) accommodations (Briguglio and others, 1996b). The cluster abalysis also identifies the same six islands at the bottom of the TPI scale and treats Maldives as an outlier in the less penetrated intermediate subgroup, again largely because of its relatively heavy built environment.

The major difference between the two analyses is in defining the boundaries between the intermediate and least penetrated destinations. The cluster analysis tends to compress the more penetrated intermediate group and classify the TPI's less penetrated intermediate islands with the least penetrated group. These differences are not surprising given the different emphases of the two approaches. Whereas the TPI groups destinations on the total level (score) of penetration alone, the cluster analysis groups them according to their similarities along all three dimensions of penetration: economic, social and environmental. On the other hand, the individual positions of each island in the clusters roughly parallel the TPI rankings with the few exceptions noted.

Policy Implications

In the postwar tourist restructuring of island microstates, evidence is mounting that the older more successful destinations have, in some cases, sacrificed socio-environmental stability for rapid development. Part of the problem has been the absence of an early warning system, i.e. a comprehensive, integrated measure of tourism's pervasive impact on insular society.

Without an early warning system in place, island decision-makers rush to embrace the

varied economic benefits of tourism but fail to anticipate the destructive intrusions

of mass tourism practice: in particular, how the interplay of inadequate facility,

infrastructure, and amenity planning and management, with the high-volume propen-

sity of capital-intensive travel interests, tends to cumulatively overrun the declicate

insular carrying capacity. (McElroy and de Albuquerque, 1998: 164)



The Tourism Penetration Index is a preliminary response. Despite its simple construction, preliminary results suggest it does relatively accurately score the degree of tourism penetration across the 35 islands examined. The TPI also allowed the classification of destinations into four groupings based on degree of penetration and, in conjunction with the cluster analysis, tended to perform best distinguishing the most from the least penetrated destinations. In particular for the most penetrated islands, the TPI also highlights the need to reassess the past practice of mass tourism promotion in fragile island ecosystems.

The Index is not without its limitations, however, in addition to its failure to capture important seasonal and geographic visitor concentrations. It is excessively aggregative by definition and cannot account for island-specific variations in tourism style that may play a crucial role in determining degrees of concentration. Because of its cross-sectional construction at one point in time, it also cannot measure the duration of tourism development and a fortiori a host society's historical adaptation or antagonism to the industry. It also does not unambiguously discriminate destinations into discretely different degrees of penetration as the cluster analysis suggested. While this may be an accurate reflection of reality, it may also derive from the small number and particular choice of variables employed and the sample of islands tested. Further refinements using better and additional variables and perhaps a more sophisticated but defensible weighting scheme may improve the TPI's performance.





























References

Ashworth, G.J. "Environmental Quality and Tourism and the Environment." In Sustainable

Tourism Development, eds. H. Coccossis and P. Nijkamp, 49-63. Aldershot, UK:

Avebury, 1995.



Beller, W., P. d'Ayala and P. Hein, eds. Sustainable Development and Environmental Manage-

ment of Small Islands. Paris: UNESCO-Parthenon, 1990.



Briguglio, L., B. Archer, J. Jafari, and G. Wall, eds. Sustainable Tourism in Islands and Small States: Issues and Policies. London, England: Pinter, 1996a.



Briguglio, L., R. Butler, D. Harrison, and W.L. Filko, eds. Sustainable Tourism in Small Island

States: Case Studies. London: Pinter, 1996b.



Bryden, J. and M. Faber. "Multiplying the Tourist Multiplier." Social and Economic Studies.

20 (1971): 61-82.



Butler, R.W. "Tourism, Environment, and Sustainable Development." Environmental Conserva-

tion 18(3,)( 1991): 201-209.



Butler, R.W. "The Concept of the Tourist Area Cycle of Evolution: Implications for Management of Resources." Canadian Geographer 24 (Spring 1980): 5-12.



Caribbean Tourism Organization. Caribbean Tourism Statistical Report 1993. St. Michael,

Barbados: CTO, 1994.



De Albuquerque, K. and J. McElroy. "Caribbean Small Island Tourism Styles and Sustainable Strategies. Environmental Management 16(5) (1992):619-632.



Doxey, G.V. "When Enough's Enough. The Natives are Restless in Old Niagara. Heritage

Canada 2(2) (1976): 26-27.



Gray, H.P. "Towards an Economic Analysis of Tourism Policy." Social and Economic Studies 23(1974):386-98.



Hall, C.M. and S.J. Page, eds. Tourism in the Pacific: Issues and Cases. London: International

Thomson Business Press, 1996.



Harrison, D. ed. Tourism and the Less Developed Countries. West Sussex, UK: Wiley, 1992.



Innskeep, E. National and Regional Tourism Planning: Methodologies and Case Studies.

London: Routlege, 1994.



Jenner, P. and C. Smith. The Tourism Industry and the Environment. London: Economist Intelligence Unit, 1992.



Johnson, P. and B. Thomas. "Tourism Capacity: A Critique." In Sustainable Tourism in Islands and Small States, eds. L. Briguglio and others, 118-36. London: Pinter, 1996a.



Kakazu, H. Sustainabale Development of Small Island Economies. Boulder, CO: Westview

Press, 1994.



Lobban, C.S. and M. Schefter. Tropical Pacific Island Environments. Mangilao, Guam:

University of Guam Press, 1997.



Lanfant, M.F., J.B. Allcock, and E.M. Bruner. International Tourism: Identity and Change. London: Sage Publications, 1995.



Mathieson, A. and G. Wall. Tourism: Economic, Physical and Social Impacts. London: Longman, 1982.



McElroy, J. and K. de Albuquerque. "Tourism Penetration Index in Small Caribbean Islands." Annals of Tourism Research 25 (1, 1998): 145-168.



McElroy, J. and K. de Albuquerque. ""Sustainable Alternatives to Insular Mass Tourism:

Recent Theory and Practice." In Sustainable Tourism in Islands and Small States: Issues

and Policies, L. Briguglio and others, eds., 47-60. London: Pinter, 1996a.



McElroy, J. L. and K. de Albuquerque. "Island Tourist Profiles Across the Destination Life

Cycle." Insula 3(1) (April 1994): 12-18.



McElroy, J., B. Potter and E. Towle. "Challenges for Sustainable Development in Small Caribbean islands." In Sustainable Development and Environmental Management

of Small Islands, eds. W. Beller, P. d'Ayala and P. Hein. Paris: UNESCO, 1990:

299-316.



McElroy, J., K. de Albuquerque, and A. Dioguardi. "Applying the Tourist Destination Life Cycle Model to Small Caribbean and Pacific Islands." World Travel and Tourism Review 3:236-44.



Opermann, M. and K-S Chon. Tourism in Developing Countries. London: International

Thomson Business Press, 1997.



Pattullo, P. Last Resorts: The Cost of Tourism in the Caribbean. London, Cassell, 1996.



Pearce, D.G. "Mediterranean Charter: A Comparative Perspective." Tourism Management 8(Winter 1987):291-305.



Priestley, G.K., J.A. Edwards, and H. Coccossis. Sustainable Tourism? European Experiences. Oxon, UK: CAB International, 1996.



Rohlf, J. Numerical Taxonomy and Multivariate Analysis Systems (NTSYS-pc). Setauket,

NY: Exeter Software, Version 1.8 1993.

Stabler, M. J., ed. Tourism and Sustainability: Principles to Practice. New York: CAB

International, 1997.



Turner, L. and J. Ash The Golden Hordes: International Tourism and the Pleasure Periphery. New York: St. Martin's Press, 1976.



Vellas, F. and L. Becherel. International Tourism: An Economic Perspective. New York: St.

Martin's Press, 1995.



Wilkinson, P. "Strategies for Tourism in Island Microstates." Annals of Tourism Research 16(Summer 1989):153-77



Williams, A.M. and G. Shaw. Tourism and Economic Development: European Experiences. London: Belhaven Press, 1991.



World Tourism Organization. Compendium of Tourism Statistics 1991-1995 (17th Edition). Madrid, Spain: WTO, 1997.