Resilience and Growth in the Small States of the Pacific
Chapter

Chapter 16. Pacific Island Countries: Trade Integration in a Changing Global Economy

Author(s):
Hoe Khor, Roger Kronenberg, and Patrizia Tumbarello
Published Date:
August 2016
Share
  • ShareShare
Show Summary Details
Author(s)
Hong Chen, Lanieta Rauqeuqe, Shiu Raj Singh, Yiqun Wu and Yongzheng Yang 

Trade integration is widely considered to be essential for economic development and prosperity in Pacific island countries (PICs).1 This thinking is not without theoretical or empirical foundations. It is well known that domestic markets in PICs are too small to allow them to exploit economies of scale, a disadvantage exacerbated by their geographic remoteness from major global economic centers. Small size and remoteness reinforce each other, leading to high costs of production and trading, and hence contribute to the lower competitiveness of PIC exports. It is thought that, through trade integration, producers in PICs can effectively enlarge their markets and reduce those costs. Historically, a number of relatively small developing economies, such as Hong Kong Special Administrative Region, Mauritius, and Singapore, have managed to develop manufacturing industries that spearheaded industrialization and enabled significant increases in incomes per capita through trade integration with larger markets.

Two key questions PIC policymakers face is what needs to be done to facilitate trade integration with the rest of world and what industries can spearhead this process. Most observers doubt that PICs can or should try to replicate the use of manufacturing as the main platform for trade integration. Given the size and geographic disadvantages of these countries, their comparative advantage lies in industries other than manufacturing, and most likely in a combination of tourism, agriculture, fisheries, and minerals, depending on each country’s circumstances. All PICs seem to have potential in tourism, and, for those with abundant and fertile land, also in agriculture. PICs with large fishery resources can extract more economic benefits from these resources through better management and regional cooperation. Papua New Guinea and the Solomon Islands have significant commodity resources for production and export, while other PICs are engaged in onshore and offshore exploration. The export of labor services is another area with great potential for several of these countries. With small populations, PICs export labor services, which generate inbound remittances in a significant proportion to total output, without a large impact on the labor importing countries.

This chapter focuses on trade in goods and tourism in PICs and explores the potential for tourism to drive trade integration and inclusive economic growth. The analysis is made in the context of the eastward shift of global economic gravity, focusing on emerging Asia as an increasingly important source of demand for resource-based goods and services. This is particularly so for tourism and agricultural products, the latter both directly and indirectly (that is, by supplying the local tourism industry). It should be noted from the outset that the PICs’ traditional markets will remain important for a long time and should be further developed for deeper economic development beyond trade in goods and services. Nevertheless, PICs should continue to position themselves to tap into Asian markets for long-term benefits.

We begin by reviewing the current institutions for trade integration and trade performance in PICs, and examining the determinants of merchandise trade and tourism in PICs using gravity models. We then explore growth potential in tourism in the context of the shifting global economic gravity to Asia, and how a booming tourism industry in PICs can help revive agriculture and support more broad-based growth. We conclude with policy implications.

Trade Integration: the Institutions

The trade integration of PICs began late in the global context and has followed two tracks: intraregional and interregional integration.2Figure 16.1 summarizes the institutions that support trade integration in the region. The Pacific Island Countries Trade Agreement (PICTA) of 2001 is the key agreement promoting intraregional trade.3 In 2004, PICTA members agreed to extend the agreement to trade in services. Twelve of the 14 Pacific Islands Forum countries have signed PICTA and, apart from Micronesia, all 11 PICs that have signed PICTA have ratified the agreement. But only six have announced a readiness to implement it. Tariff reductions have long phase-in periods, up to 10 years for the least developed members and 13 years for the three PICs under the Compacts of Free Association with the United States (Marshall Islands, Micronesia, Palau). While PICTA has yet to be fully implemented, a subregional free trade agreement of the Melanesian Spearhead Group—effective in 1993 and comprising Fiji, New Caledonia’s Kanak and Socialist National Liberation Front, Papua New Guinea, the Solomon Islands, and Vanuatu—has moved ahead. Fiji, Papua New Guinea, and Vanuatu are trading duty free, with the exception of sugar, salt, and mackerel for Papua New Guinea. However, contingency measures have often been invoked to protect certain domestic industries (Duncan 2008). In 2012 the Melanesian Spearhead Group countries signed a memorandum of understanding on a scheme to facilitate the movement of skilled nationals within countries in the group for employment purposes.4

Figure 16.1Trade Agreements among Pacific Island Countries

Sources: Compiled by IMF staff, with information from ministries of foreign affairs.

Note: EPA = Economic Partnership Agreement; EU = European Union; MSG = Melanesian Spearhead Group; PACER = Pacific Agreement on Closer Economic Relations; PATCRA = Papua New Guinea and Australia Trade and Commercial Relations Agreement; PICTA = Pacific Island Countries Trade Agreement; SPARTECA = South Pacific Regional Trade and Economic Cooperation Agreement.

The late start and slow progress in intraregional trade integration are reflected in relatively low trade volumes among PICs (Figure 16.2).5 But the main reason for this, rather than trade policy, seems to be that PICs, while geographical neighbors with many common features, are not “natural” trade partners. Trade complementarity is low because of the similarity of export and import products among PICs, particularly for smaller PICs with fewer goods and services to offer. Moreover, PICs are distant from each other and scattered over an area about one-third of the Pacific. Fiji is farther from Papua New Guinea than from Australia, for example, and Palau is much closer to Asia than to most other PICs.

Figure 16.2Intraregional Trade

(Percent of GDP)

Source: IMF staff calculations.

PICs’ interregional trade integration has largely focused on the Australian and New Zealand markets given their geographic proximity and historical ties. The South Pacific Regional Trade and Economic Cooperation Agreement (SPARTECA) is the main accord promoting trade between PICs and Australia and New Zealand, providing nonreciprocal duty and quota free entry for PIC goods exports to both countries. Since its inception in 1981, SPARTECA appears to have played an important role in facilitating some exports from PICs to Australia and New Zealand, with Fiji’s textile, clothing, and footwear industry and Samoa’s auto wire harness assembly industry being the primary beneficiaries. However, SPARTECA’s effect has waned as preference margins on textile, clothing, and footwear have fallen with the removal of quotas and reductions in most-favored-nation tariffs on these goods. In Samoa, the decline of Australia’s automobile manufacturing industry in recent years has substantially reduced the demand for wire harnesses assembled there.6

The Pacific Agreement on Closer Economic Relations Plus (PACER Plus) is being negotiated to broaden PICs’ intraregional trade integration to include Australia and New Zealand on a reciprocal basis. Given that Australia and New Zealand already provide relatively easy access for most PIC goods, PACER Plus is likely to bring the largest benefits in three areas. The first is more secure access to the Australian and New Zealand agricultural markets by binding the two countries to more science-based quarantine restrictions. The second is development assistance from Australia and New Zealand to improve domestic supply capacity. And the third is temporary migration in the context of Mode 4 of the World Trade Organization General Agreement on Trade in Services (Braxton 2009). Both Australia and New Zealand already offer seasonal workers schemes, but PICs are seeking to expand these and to have them formalized in PACER Plus. At least in the short term, the focus on temporary migration as a driver of future growth is unsurprising given that some of the smaller PICs have limited capacity to export goods and services (even in tourism) to Australia and New Zealand (Box 16.1).

Another strand of PICs’ interregional trade integration has been their long-standing preferential trade agreement with the European Union (EU). As part of the African, Caribbean and Pacific group, PICs have benefited from the Lomé Convention and its successor, the Cotonou Agreement. Negotiations have moved slowly on a regional Economic Partnership Agreement to replace the trade-related sections of the Cotonou Agreement. Fiji and Papua New Guinea have signed an interim Economic Partnership Agreement with the EU to avoid interruptions of trade with Europe. This was done to protect Papua New Guinea’s market access for tuna exports to the EU, and Fiji’s sugar exports at preferential prices. For other PICs, exports to the EU are very limited and there appears to be little incentive for them to negotiate a regional Economic Partnership Agreement with the EU.

Box 16.1.Remittances and Pacific Seasonal Worker Programs

Remittances have long been a dependable source of foreign exchange and income for Pacific island countries (PICs). Limited domestic growth and special relationships with host countries have ensured continual emigration to developed economies, making PICs some of the most remittance-dependent countries in the world. During 2010-15 remittances averaged 20 percent of GDP in Tonga, 22 percent in Samoa, and 5 percent in Fiji. At the household level, remittances supplement incomes and help obtain education, health, and housing services, in addition to smoothing highly variable incomes and providing funds for investment and entrepreneurial activities without indebtedness. For many PIC households, remittances are an insurance against income loss and other financial difficulties. At the macro level, remittances have been critical for financing large trade deficits. In Fiji, remittances proved to be largely countercyclical amid political instability until recent years, which led to declines in official development assistance up to 2014.

Regional policymakers recognize that because of the remoteness, smallness, and resulting limited economic opportunities in PICs, greater access to labor markets in larger neighbors is important for regional economic integration, and is mutually beneficial for host and origin countries. While the three Compact of Free Association countries—the Marshall Islands, Micronesia, and Palau—have free access to the U.S. labor market, other Pacific island countries rely on seasonal worker programs in Australia and New Zealand for certain types of labor. These programs are quite attractive for Pacific workers skilled in agriculture and fisheries. New Zealand introduced the Recognised Seasonal Employer program in 2007. The current program includes workers from Fiji, Kiribati, Marshall Islands, Micronesia, Nauru, Palau, Papua New Guinea, Samoa, Solomon Islands, Tonga, Tuvalu, and Vanuatu.

Australia established the Pacific Seasonal Worker Pilot Scheme in 2008 and converted it into a permanent Seasonal Worker Programme in December 2011. The Seasonal Worker Programme offers seasonal labor to employers in the agricultural industry, and employers in selected locations in the accommodation industry who cannot meet their seasonal labor needs with local jobseekers. Participating countries include Fiji, Kiribati, Nauru, Papua New Guinea, Samoa, Solomon Islands, Timor-Leste, Tonga, Tuvalu, and Vanuatu.

Remittances from seasonal worker schemes are a beneficial link between migration and development. As with other remittance-dependent countries, efforts to reduce the cost of remitting have boosted flows. Seasonal worker programs will need to be further enhanced to sustain and increase remittances in the long term as an alternative or supplement to official development assistance.

Most PICs have been left out of economic integration in Asia. Papua New Guinea is the only PIC member of the Asia Pacific Economic Cooperation, and no PICs are members of the other groupings of Asian preferential trade agreements. This exclusion stems from their remoteness and smallness, which give them less influence and attraction to Asian countries. In the past, this arguably did not matter much as there was little trade between PICs and Asian countries (including trade in services such as tourism); Australia, New Zealand, North America, and Europe mattered more, especially given the trade preferences they offered. However, low trade integration with Asia will increasingly disadvantage PICs in a world of growing regionalism and given Asia’s emergence as a global center of economic activity.7 Deepening regional trade integration will be a particular challenge for PICs without mineral resources.

Trade Integration: the Performance

Trade performance has diverged between resource-rich and resource-poor PICs during 2006–10 and converged more recently. Export growth in the two resource-rich PICs, Papua New Guinea and the Solomon Islands, was very strong during 2011–14 (Figure 16.3), stemming from commodity booms.8 Micro and other PICs9 have also seen strong export growth since then after very slow growth in the previous decade. On the import side, growth largely mirrored export performance (Figure 16.4).

Figure 16.3Real Export Growth

(Percent)

Sources: IMF APDLISC database; and IMF staff calculations.

Note: PICs = Pacific island countries.

Figure 16.4Real Import Growth

(Percent)

Sources: IMF APDLISC database; and IMF staff calculations.

Note: PICs = Pacific island countries.

PICs have a high degree of trade openness because of their small size and hence need to trade to meet domestic demand. Openness has not changed significantly over the past two decades as measured by the trade-to-GDP ratio (Figure 16.5), except for Papua New Guinea and the Solomon Islands, whose openness has been boosted by commodity booms which lasted until 2014. Among non-resource-rich exporters, micro PICs tend to have higher openness, and other PICs lower openness relative to small states outside the region.

Figure 16.5Openness Comparisons, 1995–2014

[Goods and services trade; percent of GDP)

Source: IMF, World Economic Outlook database.

Note: PICs = Pacific island countries.

However, the high trade openness masks large trade deficits for most PICs, particularly among non-resource-rich exporters. In micro PICs, exports are often a small fraction of imports, but even in some of the larger ones (for example, Samoa and Tonga) trade deficits are large and compare unfavorably with small states outside the region (Figure 16.6).10 The sources for financing these deficits vary considerably, but in most PICs, including Fiji, Samoa, and Tonga, remittances and aid are major sources (Box 16.1). Most remittances in Kiribati and Tuvalu come from seafarers, but seafaring has been under pressure in recent years (Box 16.2). Income from industrial fishing access licenses is also important for some countries, especially for PICs with large exclusive economic zones, such as Kiribati, Papua New Guinea, and the Solomon Islands (Box 16.3).

Figure 16.6Trade Balance, Goods and Services, 2003–14

(Percent of GDP)

Sources: IMF APDLISC database; and IMF staff calculations.

Note: PICs = Pacific island countries.

PICs’ exports are highly concentrated, reflecting their narrow economic bases. Approximately two-thirds of merchandise exports are primary products, predominantly agricultural products and natural resources (Figure 16.7). In non-commodity-rich economies (Kiribati, Fiji, Marshall Islands, Micronesia, Palau, Samoa, Tonga, Tuvalu, and Vanuatu) agriculture alone accounts for over 60 percent of total merchandise exports, and even in resource-rich countries agricultural exports are larger than resource exports. Manufactured exports are significant in non-resource-rich countries, but they are very small in resource-rich ones.11 On the import side, agriculture (including food) is important, although this is more so in non-resource-rich countries, where it accounts for nearly 30 percent of total imports. Fuel imports are also important, again more so in non-resource-rich countries than in others. Agriculture and fuel combined account for nearly half of total imports in non-resource-rich PICs, and over one-third in resource-rich PICs. It is worth noting that resource-rich countries also import more machinery and equipment as a result of their relatively large investment in the resource sector.

Figure 16.7Trade Composition in Pacific Island Countries

(Percent)

Source: World Integrated Trade Statistics, recent year averages.

Note: PICs = Pacific island countries.

Box 16.2.Seafarers‘ Employment in Kiribati and Tuvalu

For Kiribati and Tuvalu, seafaring is an important source of employment and remittances, but both have clearly trended downward in recent years.

While remaining sizable, employment in seafaring fell sharply during the global financial crisis. As of October 2013 about 1,008 Kiribati and 112 Tuvalu seafarers had jobs on vessels, compared to 1,452 and 361, respectively, in 2006. Over the same period, seafarer remittances fell by 5 percent of GDP for Kiribati and 36 percent for Tuvalu. The depreciation of the U.S. dollar until 2013 had also significantly reduced the Australian dollar value of seafarer remittances (Figure 16.2.1).

Figure 16.2.1Seafarers’ Remittances: Kiribati and Tuvalu

Sources: Country authorities; and IMF staff calculations.

The recovery in world trade from the global financial crisis did not produce a corresponding recovery in seafarer employment for a number of structural reasons. The shipping industry still suffers from low profitability and overcapacity, and the increasing automation of ship operations has reduced demand for seafarers. Moreover, seafarers from Kiribati and Tuvalu have become less competitive compared with those from south and southeast Asian countries as transportation costs for Kiribati and Tuvalu seafarers traveling to their ships have become relatively high.

While the Kiribati Marine Training Center is considered one of the best vocational training institutes in the region, the Tuvalu Maritime Training Institute’s training programs are generally for traditional merchant vessels and have become inadequate for ships equipped with modern technology. And while the close connection with German merchant ships has traditionally provided stable employment and income, this concentration has reduced options to diversify into ships operated by other countries.

The difficulties facing seafarers in Kiribati and Tuvalu highlight the unique challenges facing small Pacific island countries, which have few options for earning foreign exchange other than maximizing ocean resources and manpower. Compared with some larger Pacific island countries, Kiribati and Tuvalu have little agricultural land, and their prospects for developing tourism are less favorable because they are poorly connected with the rest of the world. Both countries should therefore aim to diversify into seasonal worker programs in Australia and New Zealand, and make the most of ocean resources for job creation and government revenue (see Box 16.3).

This box was prepared by Xuefei Bai, Sergei Dodzin, and Jiangyan Yu.

Box 16.3.Fisheries in Pacific Island Countries

Pacific islands in the western and central Pacific Ocean hold one-third of the world’s tuna stocks. Industrial tuna fishing, mainly by international fishing fleets is a vital source of government revenue for several of these countries. In Kiribati, Micronesia, and Tuvalu, fishing license fees account for about 30 percent of total government revenue. Coastal fisheries help sustain rural livelihoods, and fish exports provide the largest source of foreign exchange earnings in several Pacific island countries (PICs), including Fiji, Kiribati, Samoa, and Tonga.

If properly managed, fishery resources can be an important source of sustainable economic growth and income. A key challenge, however, is how to extract more economic rents and value added from industrial fishing by ensuring the sustainability of ocean resources and improving regional cooperation in industrial fishing arrangements.

The Parties to the Nauru Agreement represents a major achievement in regional cooperation in managing PICs’ tuna resources. Under this initiative, Kiribati, the Marshall Islands, Micronesia, Nauru, Palau, Papua New Guinea, the Solomon Islands, and Tuvalu established a regional vessel register and subsequently a vessel monitoring system, enabling them to operate a so-called Vessel Day Scheme. By limiting the number of days for purse seine fishing by each member and allowing members to trade their approved days (quotas), the scheme has led to almost a tripling of license revenues, from US$70 million to US$230 million a year. To maintain and improve regional cooperation, PICs need to create greater scarcity of tuna resources by (1) expanding the scheme to nonmembers to increase leverage over international fishing fleets, (2) avoiding bilateral fishing agreements with countries outside the region, and (3) enforcing compliance with quota allocation by existing agreement members.

PICs have also made progress in the sustainable management of ocean resources. Pacific Islands Forum leaders have endorsed a regional Oceanscape Program to better manage ocean resources, recognizing that the current and potential economic benefits of these resources depend on their underlying health and environmental status. In supporting this initiative, the World Bank has provided a series of overlapping grants and credits to participating countries to enhance the natural capital embodied in the ocean resources.

Several PICs have made considerable effort to create jobs and increase local value added through downstream processing, but significant challenges remain. As with other manufacturing activities, small market size and remoteness, and inadequate infrastructure, technology, and capital, are major constraints. Many countries have tried to overcome them by initiating joint ventures with foreign companies. While some processing plants (such as those in Papua New Guinea) are doing reasonably well, many others, particularly in smaller PICs, are struggling for survival or have gone bankrupt. Countries need to examine their own circumstances to determine whether they should focus on maximizing fishing license fees or also venture into downstream activities, which may not be commercially viable or increase national economic welfare if processing plants cannot be run efficiently.

This box draws partly on material from a World Bank workshop on fisheries at the 2013 IMF/World Bank Annual Meetings.

PICs’ merchandise export destinations are quite concentrated on Australia and New Zealand. But the Australian and New Zealand markets have shrunk significantly, as have the North American12 and European markets in favor of emerging Asia, including China (Chapter 4). Turning to trade in services, inbound tourism has gained in importance for several PICs (Figure 16.8), and this is a bright spot in their trade integration with the rest of the world. The growth of tourist arrivals has averaged 6 percent a year since 2000—and it is not just Fiji that has done well. Tourist receipts now make an important contribution to several PIC economies (Figure 16.9). Australia, New Zealand, and the United States account for the bulk of tourist arrivals in most PICs, but the main sources of tourists for Palau have been the Asian economies, notably Japan, Korea, and Taiwan Province of China.13

Figure 16.8Annual Visitor Arrivals

Source: National tourism and statistics authorities.

Figure 16.9Tourism in the Pacific Islands

(Percent of GDP, 2005–14 average)

Sources: Country authorities; and IMF staff calculations.

Gravity Models For Pacific Goods and Tourism

To examine what drives the growth of goods trade and tourism in PICs, we employ two gravity models—one for goods exports and imports and the other for tourism—to aid our analysis. These are standard gravity models tailored to the circumstances in PICs. Our model for goods trade is based on a three-way error components model:14

where Tijt stands for trade flows from Pacific island country i to partner country j in period t (in the export equation Tij stands for exports, Xij, and in the import equation it stands for imports, Mij; Yit is the GDP of exporting country i and Yjt is the GDP of importing country j; Dj stands for distance between country i and country j; Fij is a dummy variable indicating if countries i and j are both signatories to the same preferential trade agreement (Fij = 1 if both countries are signatories and Fij = 0 if they are not); and Cij indicates if countries i and j share colonial ties, with binary values 1 and 0 indicating the existence and absence of such ties, respectively. αi and αj are export country and import country fixed effects; values for α (1 to 5) are the coefficients of corresponding variables, and μijt is the error term.

The equation is estimated using data for 1990–2012 covering six PICs as exporting countries15 and 100 countries as importing countries (Annex 16.1). The above three-way error component equation is estimated with two high-dimensional fixed effects estimators, such as the fixed effects least squares dummy variable estimator (FELSDV) and the two-stage FELSDV estimator (Annex 16.2 details the estimation methods and Annex 16.3 provides more information on the data). All estimated coefficients have the expected signs, and the equation has reasonably good explanatory power with R2 = 0.64 from the FELSDV estimator (Table 16.1).

Table 16.1Estimation Results of the Gravity Model for Pacific Island Countries’ Merchandise Exports
Independent VariableEstimated Coefficient
Exporting country real GDP (in logs)0.27
Importing country real GDP (in logs)0.44
Distance (in logs)−2.32
Preferential trade agreement (existence = 1)0.39
Colonial ties (existence = 1)1.31
Source: IMF staff calculations.Note: The dependent variable is real bilateral exports in logs; all coefficients are statistically significant at the 99 percent confidence level.
Source: IMF staff calculations.Note: The dependent variable is real bilateral exports in logs; all coefficients are statistically significant at the 99 percent confidence level.

A striking result from the regression is the very low elasticity of PIC exports with respect to their own GDP. The estimated coefficient indicates that, on average, with each percentage increase in the GDP of the exporting country, exports rise only by 0.27 percent. Thus, there is a tendency for export growth to lag behind output growth in PICs, indicating a strong inward orientation of economic activity. The elasticity of PIC exports with respect to importing country GDP is also low, but considerably higher than the elasticity with respect to PICs’ own GDP. This low elasticity may reflect the fact that PIC exports are primarily commodities, such as agricultural products and min-erals.16 For agricultural products, lack of product differentiation or processing could also be associated with low income elasticities of demand. Even with this low elasticity, income growth in importing countries should have helped PICs narrow their trade deficits. This is because in recent years their trading partners have grown much more rapidly (Papua New Guinea and the Solomon Islands being exceptions).

Remoteness is a major barrier to export growth in PICs. Based on the estimated coefficient on distance,17 for each percentage increase in distance to an export market, PICs’ exports decline by about 2.3 percent. Take Fiji and Tonga for illustrative purposes. Since Tonga’s distance to Australia (3,585 kilometers) is 11 percent greater than Fiji’s (3,224 kilometers), all else equal,18 Tonga’s exports to Australia would be 25 percent lower than Fiji’s.

The regression results also show that preferential trade agreements generally have a positive impact on bilateral trade. However, this impact varies across individual agreements, with the Melanesian Spearhead Group and SPARTECA showing positive effects, but the Lomé Convention, Cotonou Agreement, and the Economic Partnership Agreement having no significant effects. The impact of being a member of a preferential trade agreement on the bilateral trade of PICs is quite large. In the case of the Melanesian Spearhead Group, for instance, bilateral exports could be nearly 50 percent higher, while bilateral exports for SPARTECA could be 115 percent higher. It should be noted, however, that bilateral trade among most PICs is mostly very low, so even a large percentage increase in this trade would translate into only a small impact on overall trade.

Additionally, the welfare implications of the Melanesian Spearhead Group and SPARTECA can be very different despite both having a positive impact on exports. As a nonreciprocal agreement, SPARTECA essentially allows PICs to reap rents generated by tariffs and quotas imposed on Australia’s and New Zealand’s imports from non-PICs. Thus, there is little doubt that increased exports under SPARTECA translate into a welfare improvement for PICs. On the other hand, the Melanesian Spearhead Group Trade Agreement is reciprocal, and any increase in bilateral trade is a result of reciprocal tariff reductions among Melanesian Spearhead Group members. It is well known that such agreements can result in trade diversion as well as trade creation, and that the trade diversion effect is more likely to dominate the trade creation effect in free trade agreements that involve only small trading partners, a point emphasized by Duncan (2008).19

Colonial ties also seem to have a positive impact on PICs’ exports to their former colonial powers. On average, and other things being equal, a PIC exports 134 percent more to its former colonial power than to other countries. Such a positive impact on exports reflects the cultural (including language), political, and business ties that bind countries with their former colonial powers. However, it should be noted that the estimated effects of colonial ties are based on historical experience, and it is possible that such effects will diminish over time as trade preferences accorded to former colonies are gradually eroded or phased out.

Results from a similar import equation show that PICs have a tendency toward a deteriorating trade balance. The elasticity of their imports with respect to their GDP is estimated to be 0.91, substantially higher than the elasticity of exports to their GDP (Table 16.2). This suggests that as these economies expand, import growth tends to outpace export growth, and hence nontrade accounts (namely, services, income, and financial accounts) will need to generate a sufficient surplus to maintain the initial balance of payments position. This in turn will require either structural adjustment to increase goods and services exports and/or reduce imports unless external inflows (aid, remittances, and capital) continue to increase.

Table 16.2Estimation Results of the Gravity Model for Pacific Island Countries’ Merchandise Imports
Independent VariableEstimated Coefficient
Importing country real GDP (in logs)0.91
Exporting country real GDP (in logs)0.21
Distance (in logs)−1.74
Preferential trade agreement (existence = 1)0.36
Colonial ties (existence = 1)0.81
Source: IMF staff calculations.Note: The dependent variable is real bilateral imports in logs; all coefficients are statistically significant at the 99 percent confidence level.
Source: IMF staff calculations.Note: The dependent variable is real bilateral imports in logs; all coefficients are statistically significant at the 99 percent confidence level.

Other dependent variables have similar and expected impacts as those for the estimation of exports. The elasticity of imports with respect to trading partner GDP is positive but very low. This reflects the fact that, on average, PIC economies have been growing more slowly than their major trading partners and have not been able to absorb their exports at the same pace as the growth in their trading partners. This implies that PICs have become less important export destinations for their trading partners and increasingly marginalized in international trade. Another significant result is that distance seems to have a smaller negative impact on the imports of these countries than their exports. This may suggest that import consignments are larger and, therefore, their cost of shipping is lower. Preferential trade agreements have a magnitude of impact on the imports of PICs that is similar to that on their exports, but the colonial ties seem to be less important for imports than for exports. This result probably points to the wider range of options available for sourcing imports than diversifying export destinations.

The tourism equation has a structure similar to the goods model but with an expanded set of explanatory variables.

where Vijt stands for the number of tourist arrivals in PIC i from source country j in period t; Njt, is the population of source country j; YPCjt is the GDP per capita of source country j; Dij is the distance between PIC destination i and source country j; Lij is a dummy variable indicating if countries i and j share a common language (Lij = 1 if both countries share a common language and Lij = 0 if they do not); Si denotes land surface area of PIC destination i, and Uit is the share of urban population in PIC destination i, a proxy for domestic connectivity for tourist travel. βs are the coefficients of corresponding variables, and ɛijt is the error term. The tourism equation is estimated with the ordinary least squares estimator. Similar to the goods equations, the estimated coefficients are statistically highly significant and have the expected signs, as well as a relatively high level of overall explanatory power (R2 = 0.76).

The regression results highlight the importance of establishing tourism links with large and fast-growing source countries and increasing destination awareness (Table 16.3). For each percentage increase in source-country population, tourist arrivals rise by about 0.2 percent. This means that, all else being equal, only a small fraction of population growth translates into tourism growth in PICs. However, the same percentage increase from a large source country means a larger number of tourist arrivals than from a small source country. Given that this result is based on panel data, the low elasticity could also mean that the awareness of PICs as tourist destinations is lower in large source countries than in small ones. Such facts could hamper the long-term growth of tourism in PICs, but they also suggest that there is greater potential to attract tourists from large countries. On the other hand, distance has a smaller negative impact on tourism than on goods exports. For each percentage increase in distance from a destination country, tourist arrivals decline by 1.4 percent (compared with 2.3 percent for goods exports), reflecting distance-induced increases in travel costs. Clearly, this relationship may be nonlinear and only holds up to a certain distance.

Table 16.3Estimation Results of the Gravity Model for Pacific Island Countries’ Merchandise Imports
Independent VariableEstimated Coefficient
Log (source population)0.20
Log (source-country real GDP)0.95
Log (distance)−1.43
Common language (dummy = 1 if exists)1.48
Log (land surface of destination country)0.17
Degree of urbanization0.08
Constant8.67
Source: IMF staff calculations.Note: The dependent variable is tourist arrivals in logs; all coefficients are significant at the 99 percent confidence level. Destinations include Fiji, Palau, Samoa, Tonga, and Vanuatu. Source countries are listed in Annex Table 16.1.2.
Source: IMF staff calculations.Note: The dependent variable is tourist arrivals in logs; all coefficients are significant at the 99 percent confidence level. Destinations include Fiji, Palau, Samoa, Tonga, and Vanuatu. Source countries are listed in Annex Table 16.1.2.

Income is the most important force driving tourist arrivals from a source country. The results indicate an income elasticity of close to unity with respect to source-country income. This is more than twice that for the import-country income elasticity for goods exports. Nevertheless, the estimate indicates that, on average, tourism in PICs is not a luxury service.20 This seems to be consistent with anecdotal evidence that Australian tourists tend to go to North America, Europe, and Asia when they have more disposable income, whereas the Pacific is more likely to be regarded as a budget holiday destination. However, it is quite possible that the income elasticity varies among income groups and age cohorts and source countries. Information on such variations can be useful for tourism marketing and should be examined in country-specific research.

The results also indicate that a common language shared with a source country helps raise tourist arrivals. A PIC can expect 148 percent more tourist arrivals from a source country that shares its language than from a country that does not. Underlying the role of common languages could also be familiarity with destinations, and hence the availability of information about PICs in source-country languages could play an important role in attracting tourists. Training tour operators and local tourist guides to speak source-country languages could also help. Larger land surface in destination countries is also found to help attract more tourists, probably reflecting the higher capacity of larger countries to receive tourists and the greater diversity of their destinations. However, the role of land surface may also reflect the attraction of a greater range of tourism products and point to the possibility of the gains of collectively marketing a wider range of tourism products and attractions in PICs. Domestic connectivity, as measured by the degree of urbanization, also helps increase tourist arrivals, confirming the importance of general infrastructure for tourism development, although the estimated impact appears to be relatively small.

Taking Advantage of the Shifting Gravity

The results of the gravity model analysis suggest that tourism in PICs is likely to face more favorable conditions for growth than goods exports. The estimated demand elasticity of close to unity with respect to source-country income means that demand for PIC tourist arrivals can expand over time at a similar rate of income growth in source countries, and the smaller negative impact of distance on tourist arrivals helps moderate the disadvantage of remoteness. The common English language and political ties with traditional source countries such as Australia and New Zealand will remain a positive factor for tourist arrivals from these countries. Moreover, the wide use of English in PICs also helps minimize language barriers among Asian tourists, for whom English is the most common second language.21

However, it is the shifting global economic gravity that is likely to bring the greatest opportunities for tourism in PICs. Asia’s emergence as a global economic center has changed international tourism over the past two decades. Traditionally, Europe and the Americas dominated the global tourism market, both as sources and destinations for international tourism. According to United Nations World Tourism Organization statistics, while Europe remains the largest source of tourists globally, Asia and the Pacific has emerged as the second largest source, overtaking the United States. The region accounted for 23 percent of global tourist departures in 2012, an increase of 10 percentage points from 1990 (Figure 16.10). Tourists from China have increased particularly rapidly, with close to 100 million of its citizens traveling overseas in 2013, the largest country group in the world. The World Tourism Organization (2013) projects that world tourist arrivals will continue to grow robustly over the next two decades, at 3.3 percent per year and reaching 1.8 billion by 2030. International tourist arrivals in emerging markets are projected to grow twice as fast (4.4 percent) as in advanced economies (2.2 percent). Although there is no forecast by source country or region, departures from emerging markets are likely to outpace those from advanced economies, driven by higher populations and income growth.

Emerging markets in Asia could become a major source of tourists in PICs if the right conditions are created there, with China deserving special attention for its large population and rapid income growth. Since 1995, China’s travel departures have increased at an annual rate of 15½ percent. Song (2013) forecasts that travel departures could reach 345 million by 2020. Song’s implied 19½ percent average annual growth appears to be overly optimistic given the recent economic slowdown. But it is quite likely that the number of tourists from China will still grow rapidly over the medium term, and potentially by about 6.5 percent a year, matching the average forecast of China’s GDP growth (based on IMF October 2015 World Economic Outlook forecasts). However, household spending is likely to grow significantly faster than GDP as China rebalances growth toward greater reliance on domestic consumption, which would also lead to real exchange rate appreciation over time and therefore more affordable overseas travel.

Figure 16.10International Departures (Millions)

Sources: World Bank database; and UN World Tourism Organization.

Tourists from China have come in waves, and one appears to have just arrived in the Pacific. The first wave tended to concentrate on neighboring countries, particularly in northeast and southeast Asia. Over time and as the households grew richer, more tourists traveled to North America and Europe, and the next wave is likely to reach farther afield as nontraditional destinations are explored. Starting from an admittedly low base, tourist arrivals in the Pacific have surged over the past few years (Figure 16.11). The challenge for PICs is to sustain strong growth into the future. In this regard, Maldives provides a benchmark for PICs. Since 2005, Maldives has attracted tourists from China at an astonishing growth rate of 53 percent a year, surpassing 350,000 arrivals in 2014, according to Maldives government data. This represents more than a quarter of total tourist arrivals, with China now the country’s biggest source of tourist arrivals. High-end tourist facilities, diversification, and improvements in service, innovation, and marketing to cater to tourist demand have allowed Maldives to create a niche market that is now central to the success of its tourism industry.

Figure 16.11Visitors from China, 1995–2014

Source: National statistical agencies.

PICs need to create similar conditions to grow tourism from China and Asia in general. These include raising awareness of the Pacific as a tourism destination through marketing and other forms of information dissemination; more frequent and affordable flights; improved tourism infrastructure and services (for example, hotels and restaurants); and a greater variety and quality of tourist products. Obviously, the starting point varies considerably across PICs, and bottlenecks differ from country to country. Smaller and more remote PICs have a lower starting point and face tougher challenges. A small market makes it difficult to attract frequent and affordable flights and yet without such flights investment in tourist infrastructure will not be viable and tourist products will not develop. As such, an integrated approach to developing tourism with a concerted effort by public and private sectors may be warranted. In particular, the government needs to create the necessary conditions for domestic and foreign investment, including facilitating land leases for tourist infrastructure development. Scope also exists for intergovernmental or regional cooperation to overcome diseconomies of scale, such as through joint marketing and improved cooperation in aviation.

Securing a share of the Asian tourism boom is critical to the future of tourism in PICs. Figure 16.12 shows how tourist arrivals in Fiji could evolve with different degrees of success in attracting tourists just from China. Such rapid tourism growth, in turn, could significantly boost the agricultural sector in PICs, especially if synergies between tourism and agriculture can be exploited. PICs have long recognized the benefits of these synergies, but have been slow to exploit them. The linkage is particularly important for small states, as agricultural exporters face high transportation costs in selling their products to overseas markets, and tropical produce often faces more stringent sanitary and phytosanitary restrictions. By supplying the domestic tourism industry, producers could avoid the disadvantages of long distance and sanitary and phytosanitary restrictions. And those close to tourism sites could save significantly on domestic transportation costs, which are often high.

Figure 16.12Projection of Chinese Visitor Arrivals in Fiji

(Millions)

Sources: Fiji Bureau of Statistics; and authors’ projections.

The linkage between tourism and agriculture is important because it not only offers a way to reduce export costs and barriers, but also serves as a critical strategy to foster inclusive growth. Despite weak performance over the past decade, agriculture is by far the largest sector of most PIC economies and provides employment and income for more households than any other sector. Thus, linking agriculture to tourism can help revive agriculture and broaden the base for economic growth.

While there is little information about the current state of agricultural supply to the tourism industry, anecdotal evidence suggests significant potential. A study based on a “farm to table” project by the University of the South Pacific reports that 70 percent of food for the tourism industry in the Pacific is imported (Gibson 2013). The overall retention rate for tourist expenditure is about 44 percent in Fiji; that is, for each dollar a tourist spends, 56 cents leaks out of Fiji through spending on imported goods and services, a large portion of which is food and agricultural products. The Food and Agriculture Organization (2012) notes that with the exception of pork, virtually all meats are imported in Tonga and Samoa, particularly for high-end hotels and restaurants because local producers cannot supply these products in required volumes with consistent quality. Similarly, hotels and restaurants often rely on imported vegetables and fruits. Imports of food products are inevitable in PICs and help reduce costs given the undiversified production base and climate conditions in the region. Even so, there seems to be considerable scope to supply produce suitable for cultivation in the tropical climates of these countries. What has prevented the development of a supply chain to meet this local demand and consistently supply quality needs to be examined.

Once domestic producers can supply local hospitality industries with adequate volumes and consistent quality, local producers will be in a stronger position to export. The larger volumes and higher quality would effectively reduce the cost of agricultural exports and make PICs more competitive in overseas markets. In fact, overseas markets and domestic hospitality markets can be highly complementary. The seasonality of certain vegetables and fruits has been a major issue for local hotels and restaurants, but this is because small production volumes make it economically unviable to develop local storage facilities. Once production volumes reach critical mass, it will become more cost effective to develop such logistics to facilitate exports.

Agricultural development can also help enhance the tourism industry. As Rogers (2012) points out for Samoa and Tonga, agricultural systems are an integral part of the natural environment that provides the aesthetic context for a tourist destination. Thus, it is important to preserve the essential features of Pacific agricultural systems to ensure ecological sustainability and commercial value for tourism. Greater use of agricultural systems can enhance the tourist experience as well as increase local value added. However, care should be taken in developing tourism infrastructure to protect the agricultural environment. Similarly, agricultural development should minimize pollution and avoid damage to tourist attractions. Organic farming has often been advocated both as a way to produce higher-value-added products by product differentiation as well as to better preserve the natural environment. This requires a holistic approach to development planning and coordinated efforts between agriculture and tourism authorities.

Conclusions

PICs have made considerable effort—albeit with mixed success—to advance trade integration, both within and outside the region. Large preference margins offered by former colonial powers in earlier years appear to have helped boost certain exports, such as sugar; textile, clothing, and footwear products; and automobile parts. But as preference margins have fallen and Australia’s automobile manufacturing industry declines, PICs’ exports have suffered. Meanwhile, weak domestic supply capacity and rigorous quarantine requirements continue to hamper agricultural exports. PICs have increasingly turned to intraregional trade integration to boost export demand, but the lack of trade complementarity among PICs and the slow implementation of trade agreements mean that benefits may be limited, and their impact may not even be welfare improving because of trade diversion. Furthermore, the likely uneven distribution of trade expansion in favor of larger countries in the region may lead to tensions that continue to hinder trade liberalization.

Asia’s economic emergence globally further reinforces the rationale for unilateral trade liberalization in PICs. Australia and New Zealand will remain major sources of imports for these countries for a long time, but the rapidly growing importance of Asian imports has increased the chance of trade diversion from PACER Plus. Some of the major benefits for PICs from this are likely to be in the area of development assistance to improve their domestic supply capacity. This is especially so in agriculture and tourism, as well as in a scientifically based relaxation of quarantine restrictions on Pacific produce. Perhaps the largest benefits lie in an expanded and more institutionalized temporary migration scheme that would allow PICs to export labor services, especially the small countries with limited capacity, at least in the short to medium term, to export goods and services. All these measures should be included in the final PACER Plus agreement, but PICs should at the same time pursue unilateral liberalization to avoid trade diversion.

While continuing to expand trade, temporary migration schemes, and other forms of economic cooperation with traditional trading partners, PICs should make a greater effort to diversify trade into Asian markets. This will not be easy, as shown by the limited progress of non-resource-rich PICs in tapping the Chinese market. Apart from domestic supply constraints, this is partly because of ever-closer trade integration in Asia that has enabled southeast Asian countries to supply increasing quantities of tropical produce to northeast Asian countries. To improve competitiveness, PICs will need to make significant progress on two fronts. The first is to improve agricultural productivity. At the micro level, this requires, among other things, supporting infrastructure and services, such as extension services and efficient marketing arrangements, and land systems that provide secure access to land for productive purposes. At the macro level, PICs need to maintain exchange rates at appropriate levels through macroeconomic policies that keep inflation low and make greater use of aid and remittances for productive investment and minimizing the potential for Dutch disease. And second, PICs should negotiate a more level playing field in Asian countries, perhaps collectively, through freer market access and the extension of existing preferential access in some markets, such as China.

The prospects for diversification into Asian markets are much more promising in tourism, and here policies should focus on creating conditions for private business to thrive. PICs have comparative advantage in tourism, despite their remoteness from major global economic centers. As global economic momentum moves eastward, it presents PICs with an unprecedented opportunity to develop tourism. Surging outbound tourists from China and other Asian countries have already begun to benefit PICs, and this could be the beginning of a long boom. To seize the opportunity, PICs must market themselves proactively. Countries need to attract foreign investment and know-how in building and upgrading tourism infrastructure, and provide secure access to land leases. Regional cooperation in marketing and trade-related infrastructure could help overcome diseconomies of scale. Governments could help reduce barriers for entering the tourism market faced by local businesses, particularly small and medium-size ones; and help promote Pacific cultures as a tourist attraction, which would benefit local communities.

Strong tourism growth could also provide much-needed support to agriculture in PICs, and policies should actively support the establishment of agriculture-tourism linkages. But strong linkages need to be made between the two sectors; otherwise a tourism boom could increase pressure on the agriculture sector. In the past, some PICs may have experienced competition between agriculture and tourism, particularly in use of land and labor. Furthermore, strong exchange rates that are supported by tourism earnings may have had a negative impact on agricultural production and exports. Yet if the linkages are made, a tourism boom could generate demand for local food and other products, as there are considerable synergies to be exploited that would benefit both sectors. Government policies should therefore encourage initiatives to build supply chains to ensure a sufficient supply of food and agricultural products to the tourism industry with consistent quality, and unique Pacific agricultural systems should be integrated into local tourism products. Strong growth of both tourism and agriculture would provide a sound basis for inclusive growth.

Annex 16.1. List of Goods-Importing Countries and Tourist Source Countries
Annex Table 16.1.1Importers of Goods from Pacific Island Countries, 1992–2014(100 Countries in Total)
AlgeriaCôte d’IvoireIrelandNetherlandsSolomon Islands
AustraliaDenmarkIsraelNew ZealandSouth Africa
AustriaDominicaItalyNigeriaSpain
BahrainDominican RepublicJamaicaNorwaySri Lanka
BangladeshEcuadorJapanOmanSweden
BarbadosEgyptJordanPakistanSwitzerland
BelarusEstoniaKenyaPanamaTanzania
BelgiumFijiKiribatiPapua New GuineaThailand
BrazilFinlandKoreaPeruTonga
Brunei DarussalamFranceKuwaitPhilippinesTrinidad and Tobago
BulgariaGabonLao P. D. R.PolandTunisia
CambodiaGermanyLatviaPortugalTurkey
CanadaGreeceLebanonQatarTuvalu
ChileGuatemalaLithuaniaRomaniaUkraine
ChinaHaitiFormer Yugoslav Republic of MacedoniaRussiaUnited Kingdom
ColombiaHondurasMalaysiaSamoaUnited States
Costa RicaHungaryMaltaSaudi ArabiaVanuatu
CroatiaIcelandMauritiusSingaporeVenezuela
CyprusIndiaMongoliaSlovak RepublicVietnam
Czech RepublicIndonesiaMoroccoSloveniaYemen
Source: Authors’ compilation.
Source: Authors’ compilation.
Annex Table 16.1.2Major Source Countries of Tourist Arrivals, 2000–14
AustraliaJapanPhilippines
ChinaMalaysiaUnited Kingdom
IndiaNew ZealandUnited States
Source: Authors’ compilation.
Source: Authors’ compilation.
Annex 16.2. Model, Methodology, and Data Descriptions

Both goods and tourism analyses are based on a modified gravity model as follows:

where subscript represents individual PICs; j represents PICs’ trading partners in the analysis of goods and source country of tourist arrivals in the analysis of tourism; and t, year. T is trade flow in the analysis of goods and tourist arrivals in the analysis of tourism. T varies with i, j, and t. A0 is the constant term. Yit is individual PICs’ real GDP, which varies with i and t. Yjt is PICs’ trading partners’ real GDP in the analysis of goods and source countries’ real GDP per capita in the analysis of tourism. It varies with j and t. Dij is distance between capitals of pairwise trading partners. It is time-invariant, and varies with i and j. W is a matrix of other factors affecting trade flows and tourist arrivals. They will be defined in detail in models explaining trade flows and tourism arrivals, respectively. β and δ are parameters.

Note that among the above relationships between trade flow/tourism arrivals T and control factors, the linkage between T and PICs’ GDP Yit may be bidirectional, which may therefore lead to an endogeneity problem. An instrumental variables estimator should therefore be employed to yield unbiased estimates if endogeneity exists.

Modeling PICs’ Exports and Imports

Logarithmic transformation of equation (16.2.1) and taking into account other relevant factors X give an econometric model for modeling trade flows (exports and imports, respectively) for PICs:

where Tijt stands for trade flows of PICs, and we use X to denote exports and M imports; Yit is the GDP of export country i, and Yjt is the GDP of importing country j; Dij is the distance between capitals of pairwise trading countries; Fij is a dummy variable indicating if countries i and j are both signatories of the same preferential trade agreement (Fij = 1 if both countries are signatories, and Fij = 0 if they are not); and Cij indicates if countries i and j share colonial ties, with binary values 1 and 0 indicating existence and absence of such ties, respectively. F and C are the additional factors composing W in Equation (16.2.1), which are found significant in explaining PICs’ trade flows.22 Furthermore, ai and oj are the exporting country’s and importing country’s fixed effects, α is the vector of control factors’ coefficients, and μijt is the white noise error term.

Data for the trade flows (goods) cover six PICs and their 100 trading partners over 1990–2012.23Equation (16.2.2) is estimated with the fixed effects least squares dummy variables estimator (henceforth FELSDV, also referred to as the two high-dimensional fixed effects estimator) and two-stage FELSDV (henceforth TSFELSDV) to control for the endogenous effect of PICs’ real GDP if it exists, for exports and imports respectively. The null hypothesis of no fixed effects of exporting countries and importing countries is rejected at the 1 percent level with the p-value of zero for F critical statistic greater than the observed F statistic. This points to the necessary employment of the FELSDV estimator. In the TSFELSDV estimation, gross fixed capital formation (investment) is the external instrumental variable to control randomness of ln Yit in the first stage of estimation, which is found to be a strong instrument since the Wald test yields an F statistic of more than the threshold value of 10.

Annex Table 16.2.1Estimation Results of the Gravity Model for Pacific Island Countries’ Merchandise Exports and Imports
Dependent: In(exports)Dependent: ln(imports)
Independent VariableFELSDVTSFELSDVFELSDVTSFELSDV
PIC’s real GDP, lnYit0.27 (2.95)0.19 (1.62)0.81 (12.93)0.91 (11.23)
Trading partners’ real GDP, lnYjt,0.44 (10.78)0.44 (9.96)0.28 (9.69)0.29 (8.92)
Distance, lnDij−2.32 (-15.90)−2.36 (-15.47)−1.65 (-15.27)−1.74 (-15.17)
Preferential trade agreement, Fit0.39 (3.40)0.35 (2.67)0.40 (5.67)0.36 (4.43)
Colonial ties, Cit1.31 (8.89)1.28 (7.73)0.90 (7.73)0.81 (6.38)
Sample size3028258541043493
Instrumented variablelnYitlnY,
External instrumentsInvestmentInvestment
F-statistic for instruments’
significance-5554.57-5554.57
F(H,: αi = αj = 0)13.1542.38
R20.6400.6500.7630.768
Source: IMF staff calculations.Note: t-statistics are in brackets. FELSDV = fixed effects least squares dummy variable estimator; PIC = Pacific island country; TSFELSDV = two-stage FELSDV.
Source: IMF staff calculations.Note: t-statistics are in brackets. FELSDV = fixed effects least squares dummy variable estimator; PIC = Pacific island country; TSFELSDV = two-stage FELSDV.

Since there is no significant difference between the FELSDV and TSFELSDV estimates, the null hypothesis of exogeneity is not rejected; namely, ln Yit is not endogenous in models explaining PICs’ exports and imports. This leads to the conclusion that the FELSDV estimates are unbiased and consistent. Estimation results for goods exports and imports are summarized in Annex Table 16.2.1.

Modeling Tourist Arrivals

Similar to the model explaining merchandise exports and imports in PICs, an econometric model for modeling tourist arrivals in PICs is developed as follows:

where Vjt stands for the number of tourist arrivals in country i from country j; Nj, is the population of source country j; YPCjt is the GDP per capita of source country j; Dij is the distance between destination country i and source country j; Lij is a dummy variable indicating if countries i and j share a common language (Lij = 1 if both countries share a common language, and Lij = 0 if they do not); and Si stands for land surface area of country i. Uit is the share of urban population in destination country i, a proxy for domestic connectivity for tourist travel. βs are the coefficients of corresponding variables, and ɛijt is the white noise error term.

In equation (16.2.3), a destination country’s urban population ratio, Uit, which to a great extent reflects a country’s urbanization level, may be endogenous because development of the tourism industry may in turn speed up the country’s urbanization progress. If this is the case an instrumental variables estimator should be used.

Data for the analysis of tourist arrivals are based on a strongly balanced panel covering five PICs and nine source countries over 2000–14. Equation (16.2.3) is estimated with the ordinary least squares (henceforth OLS), two-stage least squares estimator (henceforth TSLS) and the FELSDV estimator. The TSFELSDV estimator was also tried, and it yielded the same result as the FELSDV: that Uit is not significant.

Estimation results for PICs’ tourist arrivals are summarized in Annex Table 16.2.2, which shows that the FELSDV estimator is more appropriate than the least squares (LS) estimators since the null hypothesis of no fixed effects is rejected at the 1 percent level. Although LS estimators provide more information about determinants of tourist arrivals in PICs, the FELSDV has a better goodness of fit, which means that heterogeneities of destination countries and source countries are more than the additional three factors identified in the LS estimates; namely, the destination country’s urban population ratio, the destination country’s land surface, and common language.

Annex Table 16.2.2Estimation Results of the Gravity Model for Tourist Arrivals, ln(tourist arrivals)
Independent VariableOLSTSLSFELSDV
Source country’s total population, lnNjt0.20 (3.41)0.22 (3.29)1.92 (1.82)
Source country’s per capita GDP, lnYPCjt0.95 (9.86)0.77 (6.30)1.38 (5.15)
Destination country’s urban population ratio, Uit0.08 (20.00)0.11 (9.87)−0.005 (-0.74)
Distance, lnDij−1.43 (-8.98)−1.68 (-8.60)−2.81 (-12.59)
Destination country’s land surface, lnSlt0.17 (5.83)0.21 (6.02)
Common language, Lij1.48 (10.04)1.60 (9.34)
Sample size273273273
Instrumented variableUtt-
External instruments-Destination country’s per capita GDP-
F-statistic for instruments’ significance45.62
Hausman F-statistic (p-value)16.38 (0.000)
F (H0,: αi, = αj = 0)65.62
Adjusted/centered R20.76010.68850.9424
Source: IMF staff calculations.Note: t-statistics are in parentheses. FELSDV = fixed effects least squares dummy variable estimator; OLS = ordinary least squares; TSLS = two-stage least squares estimator.
Source: IMF staff calculations.Note: t-statistics are in parentheses. FELSDV = fixed effects least squares dummy variable estimator; OLS = ordinary least squares; TSLS = two-stage least squares estimator.
Annex 16.3. Data Sources and Descriptions
Annex Table 16.3.1Series and Data Sources
AbbreviationSeriesSource of Data
XijtPICs’ exports (deflated by U.S. import price index, U.S. dollars)IMF
MijtPICs’ imports (deflated by U.S. export price index, U.S. dollars)IMF
PijPICs’ population (persons)IMF
PjtTrading partners’ population (persons)IMF
YitPICs’ real GDPIMF
YjtTrading partners’ real GDPIMF
DijDistance between capitals of pairwise trading partners (kilometers)CEPII - French Research Center in International Economics
FijTrade agreement between trading countries (binary series)Country authorities
CijExporter and importer share colonial ties (binary series)Issue Correlates of War Colonial History Data-Paul Hensel
VijtTourist arrivals in PICs (persons)Country authorities
UitPICs’ urban population ratio (percent)World Bank, World Development Indicators database
YPCjtSource countries’ real GDP per capita (U.S. dollars)IMF
NjtSource countries’ population (persons)World Bank
LijCommon language (binary series)InfoPlease database-Pearson Education Inc.
SiPICs’ land surface area (square kilometers)World Bank
Source: IMF staff compilation.Note: PICs = Pacific island countries.
Source: IMF staff compilation.Note: PICs = Pacific island countries.
Annex Table 16.3.2Summary Statistics: Exports from Pacific Island Countries (based on 3,028 observations)
PICs’ Exports (millions of U.S. dollars)PICs’

Population

(million

persons)
Trading

Partners’

Population

(million

persons)
PICs’ Real GDP (billions

of U.S.

dollars)
Trading Partners’

Real GDP

(billions of U.S. dollars)
Distance

(thousand

kilometers)
Trade

Agreement between Trading Countries
Exporter and

Importer

Share

Colonial Ties
Mean32.301.95124.752.801334.339.960.380.06
Standard deviation153.022.45290.243.092631.855.280.490.24
Maximum3,260.007.171,354.0415.0314,937.5619.391.001.00
Minimum0.100.100.000.210.010.750.000.00
Source: IMF staff calculations.Note: PICs = Pacific island countries.
Source: IMF staff calculations.Note: PICs = Pacific island countries.
Annex Table 16.3.3Summary Statistics: Imports of Pacific Island Countries (based on 4,104 observations)
PICs’ Imports (millions

of U.S.

dollars)
PICs’

Population

(million

persons)
Trading

Partners’

Population

(million

persons)
PICs’ Real GDP (billions of U.S.

dollars)
Trading Partners’

Real GDP

(billions of U.S. dollars)
Distance

(thousand

kilometers)
Trade

Agreement between Trading Countries
Exporter and

Importer

Share

Colonial Ties
Mean23.721.79112.612.651,106.5510.960.340.05
Standard deviation110.202.37270.733.012,321.524.970.470.22
Maximum2,856.607.171,354.0415.0314,937.5619.121.001.00
Minimum0.030.100.000.210.010.750.000.00
Source: IMF staff calculations.Note: PICs = Pacific island countries.
Source: IMF staff calculations.Note: PICs = Pacific island countries.
References

    ArmstrongHarvey W. and RobertRead.1998. “Trade and Growth in Small States: The Impact of Global Trade Liberalization.The World Economy21 (4): 56385.

    BoltonKingsley.2008. “English in Asia, Asian Englishes, and the Issue of Proficiency.English Today24 (02): 312.

    BraxtonNick.2009. PACER Plus and Its Alternatives: Which Way for Trade and Development in the Pacific?Auckland: Oxfam New Zealand.

    ChandSatish editor. 2005. Pacific Islands Regional Integration and Governance. Canberra: Asia Pacific Press.

    ChandSatish2010. “Shaping New Regionalism in the Pacific Islands: Back to the Future?Working Paper Series on Regional Economic Integration 61Asian Development BankManila.

    ChenHongLanietaRauquequeShiu RajSinghYiqunWu and YongzhengYang.2014. “Pacific Island Countries: In Search of a Trade Strategy.IMF Working Paper 14/158International Monetary FundWashington.

    DuncanRon.2008. Pacific Trade Issues. Pacific Studies Series. Manila: Asian Development Bank.

    EilatYair and LiranEinav.2004. “Determinants of International Tourism: A Three-Dimensional Panel Data Analysis.Applied Economics36 (12): 131527.

    Food and Agriculture Organization. 2012. “Report on a Scoping Mission in Samoa and Tonga: Agriculture and Tourism Linkages in Pacific Island Countries.Food and Agriculture Organization Sub-regional Office for the Pacific Islands.

    GibsonDawn.2013. “Farm to Table: Celebrating Local Produce and Cuisine.Unpublished. University of the South PacificSuva.

    PrasadBiman.2008. “A Quantitative Analysis of the Potential Benefit and Cost of a PACER-Plus Agreement.In Research Study on the Benefits Challenges and Ways Forward for PACER-Plus.Adelaide: Institute for International Trade, University of Adelaide.

    RogersTuifa’asisina S.2012. “Agriculture and Tourism Linkages in Pacific Island Countries.Food and Agriculture OrganizationRome.

    SongHaiyan.2013. “China Outbound Tourism: Trends and Prospects.Presentation for Hong Kong Polytechnic University Hong Kong Special Administrative Region.

    SongHaiyan and GangLi.2008. “Tourism Demand Modeling and Forecasting—A Review of Recent Research.Tourism Management29 (2): 203220.

    World Tourism Organization. 2013. Tourism towards 2030: Global Overview.Madrid: World Tourism Organization.

This chapter is based on Chen and others (2014).

In this chapter, PICs include Fiji, Kiribati, Marshall Islands, Micronesia, Palau, Papua New Guinea, Samoa, the Solomon Islands, Tonga, Tuvalu, and Vanuatu.

Chand (2010) looks at the history of regional integration in PICs, and Chand (2005) gives a comprehensive account of various aspects of regional integration and governance.

PICTA members include the 14 developing Pacific Islands Forum countries: Cook Islands, Fiji, Kiribati, Marshall Islands, Micronesia, Nauru, Niue, Palau, Papua New Guinea, Samoa, the Solomon Islands, Tonga, Tuvalu, and Vanuatu.

Temporary workers are typically employed one to three years, often extendable for another one to three years.

Fiji and Vanuatu are the only PICs that have a significant portion of their exports destined for other PICs. Some of Fiji’s manufactured goods, such as processed food and cement, are quite competitive in some other PICs, and a large portion of Vanuatu’s beef and kava exports goes to other PICs.

Before the global financial crisis, the wire harness plant in Samoa employed over 2,000 people and was the single largest private employer in the formal sector. Employment fell to less than 1,000 in 2012.

Armstrong and Read (1998) discuss the challenges facing small states in a globalizing world economy.

It is difficult to assess trade growth in PICs, because of the lack of export price statistics suitable for deflating the nominal values of their exports. In Figures 16.3 and 16.4, the nominal exports and imports of PICs are deflated using the U.S. import and export price indices, respectively.

Micro PICs include Kiribati, Marshall Islands, Micronesia, Palau, and Tuvalu. Other PICs include Fiji, Samoa, Tonga, and Vanuatu.

Chapter 4 finds that lower export openness in PICs has contributed to slow growth compared with other small states.

There may be classification issues with regard to the size of manufactured exports. Compared with non-resource-rich countries, the “others” category is very large and may include some manufactured products.

In this case, North America includes the United States and Canada.

Palau’s total tourist arrivals in recent years have increased rapidly, helped by more chartered flights from Asia.

Prasad (2008) examines the trade impact of PACER Plus using a gravity model. His analysis focuses on trade turnover rather than evaluating the impact on exports and imports separately.

These are Fiji, Papua New Guinea, Samoa, the Solomon Islands, Tonga, and Vanuatu.

We tried the real exchange rate as an explanatory variable in the equation, but it turned out to be statistically insignificant. Thus, the low income elasticity does not seem to reflect effects other than income.

Distance is measured in kilometers between the capital cities of the exporting and importing countries.

That is, even if Tonga had the same income level and its country size was the same as Fiji’s.

Put differently, if Melanesian Spearhead Group countries were dominant trading partners of each other, trade liberalization under the group’s trade agreement would have effects similar to unilateral liberalization, and, hence, the trade diversion effect would be limited.

Song and Li (2008) provide an extensive survey of estimated elasticities of demand for tourism. Most estimates surveyed are greater than one. Eilat and Einav (2004) suggest that income elasticities for tourism in high-income countries tend to be higher than in lower-income countries. Their estimate indicates that the elasticity for high-income countries ranges between 1.29 and 1.55, and between 0.41 and 1.48 for lower-income countries.

Bolton (2008) discusses the exponential increase in Asians knowing and speaking English in recent years.

We also considered the relative price index RELRERXj=RERXj/ΣkXαkRERkj, where j is importer, X is exporter, RER=CPIXCPIj/$X/$US$j/$US, and αk is the share of exports to country j from country k. The estimated coefficient of this relative price index turns out to be negative but highly insignificant.

Note that the two high-dimensional panel is not a balanced panel.

    Other Resources Citing This Publication