Stalled or stepwise fertility transition in Pacific Island Countries

Stalled or stepwise fertility transition in Pacific Island Countries Jean Louis RALLU INED, Paris XXXVI Chaire Quetelet Ralentissements, résistances et ruptures dans les transitions démographiques Session : Resistances and reversals of fertility in developing countries Louvain la Neuve, 24-26 novembre 2010 Pacific Island countries experienced different types of fertility transition. Polynesian countries witnessed rapid fertility decline from the 1960s while Micronesian countries presented various patterns, with earlier decline in Tuvalu and Kiribati than in the Marshall Islands or the Federated States of Micronesia. However, most Pacific island countries tend to show a stabilisation of TFR at levels between 3.5 and 4.5 that represents rather high level. This paper will present trends in TFR in the Pacific and possible causes of fertility transition interruptions. The causes vary according to countries although several countries share high emigration and a particular type of economy based on migration and remittances: the MIRAB (Bertram and Waters 1985). This type of islander migration has often been characterized as mass emigration resulting in nearly stable - or even declining population at some time in the past in the smallest countries. Thus, sustainable population size and economy have been a frequent concern for some of Pacific islands countries. We shall also compare the situation of independent Pacific island countries with French and US Territories in the region. 1.- Data We shall consider the period from 1960 because earlier estimates of fertility are not much reliable and emigration became significant from the 1960s in Pacific Island Countries (PICs). Fertility data for PICs are mostly indirect estimates based on the question of number of children ever born to women aged 15 and over in censuses. The method used is usually the Arriaga method, run with one census data or two censuses data. The OCM (Own Children Method) is also frequently used (Levin, Retherford 1986) and unpublished updates have been regularly done for recent censuses of several PICs 1. 1 We are grateful to Michael Levin for providing OCM results as personal communications. 1

UNDESA Population Prospects publish estimates for the largest PICs only. However, TFR are calculated for all PICs but they are not published, because of large uncertainty for smaller countries 2. American and French Territories have complete civil registration systems. American and French Territories publish yearly TFR based on civil registration or Health Department data and intercensal population estimates. Recently, Demographic and Health Surveys (DHS) were conducted in five countries : Marshall Islands, Solomon Islands, Tuvalu and Nauru (2007) and Papua New Guinea (1996 and 2006). Data for PICs used in this paper are: DHS data when available; OCM data; UN estimates; national estimates from census reports when they seem more reliable than UN estimates. For Territories, we use published data of national statistical offices. Migration data used in this paper are World Population Prospects estimates (UNDESA 2009). Data on marriage/union are derived from censuses. PICs censuses do not separate cohabitation from marriage. However, in some PICs marital life is frequently disrupted, mostly for cohabitations. Women who are alone after cohabitation can still report as single. This also applies to lone mothers who do not live with their children, in the frame of common adoption of children of single or divorced women, and eventually married women, by relatives. This would be most frequent for rural-urban migrant women and affect results of both the Arriaga and OCM methods. The impact is larger in the case of the Arriaga method that relies on reported ever-born children. The OCM distributes unmatched children proportionally by age of women however, most of unreported children of young women would be affected to older mothers. Contraceptive prevalence rates (CPR) are under estimated in PICs because data from health services do not include contraceptive distributed by NGOs or the private sector. CPR in most PICs are not consistent with fertility levels. Traditional methods are also widespread in the frame of strong religious (catholic and protestant) influence. However, there is little incentive from governments to promote contraception and frequently no budget is allowed to contraceptives. Contraceptives (including condoms) are provided by UNFPA, directly to health centres as well as to NGOs, as most of them do not have their own supply of contraceptives (Rallu, Robertson 2010). 2.- Fertility trends Polynesian countries experienced rapid fertility decline from the 1960s as concerns arose that rapid population growth may not be sustainable for long periods in limited environment of small Polynesian islands. 2 We are grateful to Patrick Gerland UNPD/UNDESA for providing data for all PICs as personal communication. 2

Figure 1 : Trends in TFR in some Pacific Island Countries from 1960 9 8 Marshall Is 7 Fed. States of Micronesia 6 Tonga 5 Samoa 4 3 2 1 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 2005-2010 However, the process of rapid fertility decline suddenly stalled in Tonga and Samoa from the mid or late 1970s (figure 1) (Rallu, Ahlburg 1999, Rallu, Ahlburg forthcoming). Then, TFR was nearly stable in Samoa until 1990 and a slow decline resumed thereafter. While TFR declined by 2.5 between 1960-1965 and 1975-1980, it remained just under 5 from 1975-1980 to 1985-1990 and was still around 4.5 in the early 2000s, or a decline of only 0.3 over 15 years. In Tonga, TFR declined by steps. Decline halted earlier (1970-1975) and at a higher level (5.5) than in Samoa, and remained stable around that level until 1980-1985. A new decline from 1985-1990 brought the TFR just above 4 in the late 1990s and it was followed by a new stabilization. The Federated States of Micronesia (FSM) and mostly the Marshall Islands are former TTPI 3. They delayed the onset of the fertility transition until respectively the early 1970s and mid 1980s. Then it was followed by a rapid decline, mostly in the Marshall Islands. TFR decline slowed down after 1985 in the FSM, but TFR nearly stabilized after 1990 in the Marshall Is. Three PICs are associated states with New Zealand: The Cook Is, Niue and Tokelau. They are the smallest PICs, with population of less than 15 000 in the Cook Is, and less than 1 500 in Niue and Tokelau. They show rather different transition from larger Polynesian countries. After a rapid fertility decline from 1960 to the late 1970s, the Cook Is did not experienced a stabilization of TFR like Tonga or Samoa, but only a slowing down of the fertility transition from the beginning of the 1980s (figure 2). The same applies to Niue with a time lag of about 10 years for the onset of fertility transition. Then, Niue fertility declined at a constant pace from 6.8 in 1970-1975 to 3.0 in 1995-2000. TFR decline continued to reach 2.8 in 2000-2005 but at a much slower pace. Since 1990, TFR is similar in Niue and the Cook Is. Tokelau shows a different story. It experienced rather earlier fertility transition and also an earlier slow 3 Trust Territory of the Pacific Islands under the United Nations and administered by the USA. 3

down in TFR decline that levelled-off just above 4 in 1975-1985. Then, it was followed by a strong reversal, and TFR reached above 5 in 1990-2000. TFR returned to 4.2 in 2000-2005. Figure 2 : Trends in TFR in Pacific Island Countries associated states of New Zealand, from 1960 9 8 7 Niue 6 Cook Is 5 4 Tokelau Niue Tokelau Cook Is 3 2 1 0 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 2005-2010 Two countries: Kiribati and Tuvalu experienced an active national family planning programme in the last decade of British rule (Booth 1994). Fertility decline was faster in Tuvalu than in Kiribati with TFR below 3.2 in the former against 4.5 in the latter in 1970-1975 (figure 3). Decline also lasted longer in Tuvalu, until the late 1970s against only until the first half of the 1970s in Kiribati. Then, TFR re-increased in Kiribati to reach just under 5 in 1980-1985, and was then hovering at or just below 4.5 until the end of the 20 th century. From the 2000s, Kiribati TFR started to decline more rapidly. Increase was much steeper in Tuvalu, bringing TFR from 2.8 in the late 1970s back to nearly 4 in 1990-1995, a level that has kept quasi stable thereafter. Nauru is a special case and less documented. The phosphate rich country had GDP per capita similar to Koweit, and the Nauruan diaspora was well established in financial centres in Melbourne and London. Mining activities brought migrants on temporary contracts, mostly from Kiribati and Tuvalu. Family reunification was only temporary and migrants left in the early 2000s when phosphates production was coming to an end. The 2007 DHS revealed a late and slow fertility transition, with TFR quasi stable around 4.4 in 1988-1997, slightly declining to 4.1 in 1998-2002 and more rapidly to 3.4 in 2003-2007, following a severe economic crisis from 2003 linked with the temporary end of phosphate mining, before new deeper layers of phosphates were found in the late 2000s. Thus, several PICs show slow (Cook Is, FSM) stalled or quasi stalled (RMI), step by step fertility transition (Samoa, Tonga) and even reversal or fertility trends (Kiribati, Tuvalu, Tokelau). Only Palau already completed fertility transition (figure 3). The only other Pacific populations to have reach replacement fertility are the Melanesians in New Caledonia, 4

Polynesians in French Polynesia and Wallis and Futuna as well as Fiji Indian population (although not ethnic Pacific islanders). Ethnic Fijians experienced a rather slow decline in TFR until 1999 when it stabilized around 3.3. Formerly, a previous plateau of TFR around 4 appeared for ethnic Fijians in the first half of the 1990. It seemed to follow a slight increase from the late 1980s, but this was based on indirect estimates from census data (Fiji Bureau of Statistics 1998) 4. Figure 3 : Trends in TFR in Kiribati, Tuvalu and Palau from 1960. 9 8 7 6 5 Kiribati 4 3 Tuvalu Palau 2 1 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 2005-2010 3.- Tentative explanation of interrupted or reversed fertility transitions What are the causes of such important and long lasting interruptions of fertility transition in most Pacific Islands countries in Micronesia and Polynesia? Proximate determinants: marriage and contraception The immediate causes of fertility decline or its reversal are changes in union formation trends and contraceptive prevalence. Given the lack of complete data on contraceptive prevalence (see above) it is not possible to address this issue. However, contraceptive prevalence is rather a mean to achieve a desired level of fertility. Ground causes have to be looked for in other areas. Data on unions do not show much impact on fertility levels. The proportion of single females slightly increased in Tonga as fertility slowly declined from 1986 to 1996 which is has expected. But, it is surprising to see that despite increases in the proportion single by 1.6 and 2.8 percentage points at 20-24 and 25-29, no decline in TFR is observed in 1995-2005. It is even more surprising that the 2001 census of Samoa shows proportions of single females 4 Other Melanesian countries: Papua New Guinea, Solomon Is and Vanuatu are experiencing slow fertility transition, with a period of reduced decline in the Solomon during the 1999 and early 2000s political crisis. 5

much lower than in the 2006 Tonga census: by 14 percentage points at ages 20-24 and 9 percentage points at ages 25-29 (and similarly higher proportions of currently married females), and TFR is only slightly higher in Samoa than in Tonga in the 2000s (4.5 against 4.2). This shows how the proportions of single or married females are loosely connected with fertility levels. Of course, differences in contraceptive behaviour can explain the different situations in Tonga and Samoa. But, the lack of reliable data on contraceptive prevalence in these two countries does not enable us to assess the impact of contraceptive prevalence. While the role of union formation and stability cannot be assessed, it does not seem to be a major cause of slow, stalled or reversed fertility transition. Moreover, trends in census figures do not reflect all of the phenomenon of unions; early (unstable) union formation and lack of union stability has always been common in Polynesia and more recently in Micronesia. Whenever, increases are observed in the proportion of single females at young ages by censuses, changes in sexual relations and behaviour in a context of low contraceptive use among teenagers as well as for many older women, including condoms (Rallu, Robertson 2010) have limited impact on fertility. Fertility and migration The comparison of trends in TFR and net migration rates 5 provides more interesting relations. Whenever data for five-year periods are not the best to analyze the timing of different phenomenon, the quasi simultaneity of trends is remarkable. In Tonga, in 1970-1985, when net migration is the most negative (on average -3.2%), fertility decline halts and TFR stabilizes at a rather high level: 5.5. From 1985, net migration becomes less negative (-1.8%) and TFR resumes a slow decline to just under 4.5 and remains above 4 until the 2000s (figure 4). In Samoa, fertility decline is rapid in the 1960s when net migration is only slightly negative, but TFR suddenly stabilizes at 4.5 when net migration becomes more negative. A small fertility decline resumes in the 1990s when emigration is reduced due to more difficult access to New Zealand and Australia following economic crisis and new migration laws in these countries. However, fertility decline is minor as emigration remains steady with net migration rate of nearly -2%. The opposite trends in TFR and net migration appear even more clearly in Tuvalu. Before 1985, while emigration is quasi nil in Tuvalu, TFR declines steadily. Once net migration becomes negative, TFR reverses trends and increases rapidly. Actually, increase in TFR started a little before emigration became apparent, but it follows the initiation of a special scheme to enter New Zealand for seasonal work in agriculture for Tuvaluans and I-Kiribati which put hopes that New Zealand would open to more Tuvaluan migration 6 like for former New Zealand colonies and actually, migration of Tuvaluans took momentum in the 1990s. 5 Net migration rates translate mostly emigration because there is little immigration in PICs. 6 The numerus clausus of this scheme as well as for the current Seasonal Employer Programme resulted in negligible impact on Kiribati population due to its size. 6

Figure 4 : TFR and net migration (%) in Polynesian countries, 1960-2005 Tonga Samoa Tuvalu Cook Is net migration Samoa TFR Samoa net migr Tonga TFR Tonga net migr Tuvalu TFR Tuvalu net migr Cook Is TFR Cook Is 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 Figure 5 : TFR and net migration (%) in Marshall Islands and Federated States of Micronesia, 1960-2005 10,0 8,0 6,0 4,0 2,0 Marshall Is net migr FSM TFR FSM net migr RMI TFR RMI FSM 0,0-2,0-4,0 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 In the Cook Islands, the timing is less clear: emigration was already high in the 1960s during the period of rapid fertility decline, and TFR decline did not stop, but only slowed when emigration peaked and resulted in population decline. A slow decline in fertility continued while net migration became less negative. Actually, emigration was always high from the 7

associated states of New Zealand. Cook Islanders, Niueans and Tokelauans have New Zealand citizenship and migration has become a kind of bi-residence for many of these islanders. For the two former, there is no strong relation between migration and fertility trends. Tokelau showed slowing down and reversing fertility transition at a time of reduced emigration which is contrary to other PICs. Reversal of fertility transition in Tokelau and slowing down of TFR decline in the Cook Is and Niue should rather be seen as sometimes late reactions to population declines that occurred in the 1970s. However, similar decline in Niue had no impact on fertility, except a quasi stabilization of TFR just under 3 in 1995-2005. The simultaneity of changes in trends in net migration and TFR appears to be the most striking in the Marshall Islands, reminding what happened in Tuvalu. As soon as net migration becomes highly negative in the 1990s, TFR stops declining and stabilizes between 4.5 and 4.3 (figure 5). In the Federated States of Micronesia, a first and short wave of emigration in the early 1970s (Ahlburg, Levin 1990) had seemingly no impact on a pretransitional TFR of 7. TFR started to decline from 1975 while emigration was negligible until the mid 1990s. Then, increased emigration seems to have temporarily slowed fertility decline in the 1990s. But, TFR would have continued to decline in the early 2000s. Some data put it at around 3.5, but this has to be confirmed 7. Why would there be a relation between trends in emigration and fertility? PICs are small countries. Migration is extremely high in comparison with island population and for some countries the diaspora is several folds the size of island populations. This is mostly the case for very small countries like Niue and Tokelau (1 500 inhabitants or less) and the Cook Islands (less than 15 000 residents) of which ethnic populations abroad are respectively 15, 5 and 3 times as large as in home countries. But, larger countries like Tonga (101 000) and Samoa (179 000) have much smaller diaspora, with about similar numbers of Tongan abroad as in Tonga and 1.5 Samoan abroad for each Samoan in Samoa. Limited resources and economies result in little wage opportunities and migration and remittances are the way to find jobs for young adults and ensure better living condition of parents and relatives left in the islands (Ahlburg 1996; Brown 2006). When migration is restricted by the immigration laws of Pacific rim countries, the poor future prospects lead to reduce fertility. But when there is hope that immigration will continue steady or increase, then, a larger number of migrants is a strategy to ensure higher remittances. This affects not only migration, but also fertility because second generation migrants send little or no remittances. Therefore, it is necessary to ensure a strong flow of migrants in the future to stabilize remittances. Although it is not possible to know what will be migration policies when children will reach working and migration ages, it seems that people react directly to the current context in migration. It is also true that higher emigration means smaller population growth in the future as it depletes future reproductive ages. Thus, it is logic to increase fertility when emigration increases. Given the size of emigration (between one third and one half of birth cohorts emigrate between ages 18 and 25 or 30), a high TFR is needed to avoid population decline in the future and actually, most PICs show fertility stabilizing or declining slower once TFR reaches under 4.5. There is also a kind of saying that 4 children are necessary: two who shall migrate and send remittances and two who shall take care of their elderly parents. Thus, stabilizing TFR at levels around 4 is intended to avoid population decline and ensure stable remittances flows. Interrupted fertility transition is thus part of a demographic and economic rational that aims to avoid the quasi disappearance for islands populations (Niue and Tokelau are striving to keep a population of around 1 500) and ensure sustainable development. 7 FSM boasted a 55% contraceptive prevalence rate that UNFPA Office for the Pacific considered highly unreliable. 8

The case of Kiribati However, there is another case of reversed fertility trends: Kiribati. A national family planning programme was developed in the British colony of the Gilberts and Ellice (Kiribati and Tuvalu after independence). Kiribati includes both Protestants and Catholics. Even before independence, Catholics became reluctant to further reduce their fertility because they feared to become a minority. Whenever there was no settlement emigration to rim countries from Kiribati 8, fertility re-increased from the mid 1970s and waivered just below 4.5 until the end of the 20 th century. TFR resumed decline in the early 2000s although the proportion of women in union did not change. An increase in contraceptive prevalence is likely during this period due to new activities in reproductive health lead by UNFPA in the frame of the MDGs; however, the Catholic Church again reacted negatively against the programme when there was evidence of its impact on fertility. The main relation between migration and fertility in Kiribati is due to separation of husbands and wives during seamen s contracts. Given the relatively small size of seamen migration and yearly holidays, it has limited impact and rather stable over time. Tuvalu, as a part of the Gilberts and Ellice, experienced the same family planning programme as Kiribati, but 99% of the population is protestant. Thus, we have to tuern again at migration to understand differences between Tuvaluan and Kiribati fertility trends. As a British colony, Tuvalu had no more migration outlet or special schemes to enter New Zealand than its neighbour, and Kiribati and Tuvalu migrants mostly consisted of seamen and miners to Nauru until the late 1980s. When New Zealand started its seasonal fruit workers programme, its relative impact on Tuvalu was much larger than for Kiribati as numerus clausus were not relative to the size of islands populations. Soon, it turned out that Tuvaluans had more facility to re-emigrate as economic migrants and become New Zealand residents. New Zealand authorities were also probably more lenient toward Tuvaluan than I-Kiribati overstayers because the population involved was much smaller and also because Tuvaluans are Polynesians, like Maori, Cook islanders and other former New Zealand population colonies, whereas I-Kiribati are Micronesians. Thus, Tuvalu has rapidly become an emigration country, with above 13% of its island-born population and 28% of its ethnic population living out of the country in 2006. This rapid increase in emigration has been accompanied with a steep increase in TFR in 1985-1995 and, since 1995 Tuvalu s TFR is close to 4, a level that is typical of Pacific emigration countries. Fijians Most of Fiji emigration consists of the Indian minority that experienced increased emigration following the 1987, 2000 and 2006 coups. However, the coups also raised awareness of the possibility of emigrating to rim countries for economic opportunities among ethnic Fijians. It is difficult to link trends in TFR with very small Fijian net migration. However, it is well acknowledged, from New Zealand and Australian data that Fijian emigration increased in the last decade of the 20 th century and more rapidly after 2000. Remittances to Fiji also increased steeply in the 2000s, but part of this increase is related to males working in security firms in the Gulf following the Afghan and Iraq wars. The role of other migrants remittances is limited and the size of Fiji economy makes it less reliant on remittances than for smaller 8 Mostly seamen migrated for temporary contracts on European and Asian merchant ships and as mine workers to Nauru. 9

countries 9. However, the stabilization of TFR in 1999-2004 (latest data available) reminds us of trends in other Pacific emigration countries in relation with increasing emigration. But, the level of Fijian fertility (3.3) is lower than for Polynesian and Micronesian emigration countries where it is frequently close to or above 4. Territories Economies that are not based on migration and remittances appear no to show similar reversal and even no interruption of fertility transition. Among French territories in the Pacific, French Polynesia and Wallis and Futuna consist mostly of Polynesian populations, and Wallis and Futuna experiences the same mass migration as Polynesian countries, with more natives living in New Caledonia than in their islands of birth. However, both have recently completed fertility transition, with rather constant TFR decline (figure 6). The situation looks different in American Samoa and Guam, with TFR decline suddenly stalling in the mid 1980s in the former and until 1995 in the latter. The cause of these trends in US territories is different from the ones prevailing in independent countries. More restrictive access to the US mainland, but not to American Samoa, was enforced against (Western) Samoans from the mid 1980s. It resulted in many Samoans in American Samoa unable to move to the US mainland. The population of American Samoa grew rapidly and these migrants had children in American Samoa. Thus, fertility decline halted in American Samoa and a slower decline resumed later, still due to Samoan migrants spending more time in the territory. For Guam, the temporary halt in TFR decline is mostly due to Asian (mostly Filipino) and Micronesian (mostly from FSM) migrants. Fertility decline resumed in Guam and accelerated in American Samoa from 2000. Figure 6 : Trends in TFR in American and French Territories in the Pacific from 1960 8,00 7,00 6,00 Wallis and Futuna 5,00 American Samoa Guam 4,00 French Polynesia American Samoa French Polynesia New Caledonia WF 3,00 Guam New Caledonia 2,00 1,00 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 2005-2010 9 Remittances in Fiji represent only 4.8% of GDP against 25% in Samoa and above 40% in Tonga (2007). 10

A former TTPI, Palau looks more as a counter example to emigration PICs and shows similarity with territories. Palau experienced a mix of emigration, with easy access to USA and US territories, and immigration by work contract, mostly males in construction which did not affect fertility. It did not witnessed stalling fertility, but just a slowing down when TFR reached 4 from 1975 when emigration increased. TFR reached replacement level in the last decade of the 20 th century (figure 3). Conclusion Interrupted and reversed fertility transition in Pacific Island Countries is a common feature in the region. Many countries (Tonga, Samoa, Marshall Islands, the Federated States of Micronesia) have had fertility stable or slowly declining between 4 and 4.5 for several decades, some in the frame of reversals after TFR had reached under 4 (Tuvalu, Tokelau). While the role of proximate determinants of fertility (union formation, stability and contraceptive prevalence) cannot be assessed due to data availability and quality, union formation does not seem to have major effects and low contraceptive prevalence is rather a mean of achieving a desired fertility level. Thus, the real causes of stalled, stepwise or reversed fertility transitions are related to other demographic, social and economic phenomenon. The simultaneous and opposite trends in emigration and TFR are typical of Polynesian and Micronesian demography. Mass migration necessitates a rather high level of fertility to avoid population decline and replacement of generations in the future. The relationship between migration and stable and relatively high fertility is also supported by an economic system based on remittances (MIRAB): constant outflows of migrants are necessary to ensure stable remittances inflows while there is still need of people to take care of their elderly parents. It seems also that there is an impact of the inability to migrate freely between islands and rim countries. Countries associated with New Zealand and US and French territories whose citizens are New Zealand, American or French citizens did not experience similar interruptions in fertility trends and the latter have completed fertility transition. However, small associated states of New Zealand, with highest emigration, have reacted lately to emigration and could not avoid population decline. The question is whether a similar relation between emigration and fertility exists in larger Asian emigration countries, for instance the Philippines, with support of religious influence groups. References: Ahlburg D.A., 1996. Remittances and the income distribution in Tonga. Population Research and Policy Review 15 (3): 391-400. Ahlburg, D., M. J. Levin, 1990, The North East passage : a study of Pacific Islander migration to American Samoa and the United States, National Centre for Development Studies, Research School of Pacific Studies, The Australian National University, 1990, 94 p Bertram G., Waters R. F, 1985, "The MIRAB Economy in South Pacific Microstates", Pacific Viewpoint, Vol. 26, n 3, Booth H., 1994, The demography of Kiribati: estimates from the 1985 census, Asia-Pacific population journal, nº 3, 1994, pp. 55-72 11

Brown R.P.C., 2006, At Home and Away: Expanding job opportunities for Pacific Islanders Through LabourMobility, The World Bank, 189 pp. Fiji Bureau of Statistics. 1998. 1996 Fiji Census of Population and Housing: Analytical Report, Part 1 Demographic Characteristics. Parliamentary Paper No 49. Parliament of Fiji, Suva Levin M.J., R.D. Retherford, 1986, Recent fertility trends in the Pacific Islands, Papers of the East West Population Institute, East West Center, Hawaii, 72p.. Rallu J. L., D. Ahlburg, 1999, Demography, in The Pacific Islands, Environment and Society, ed. M. Rapaport, Bess Press, Hawaii, 1999, pp.. 258-269. Rallu J.L., Ahlburg D.A., forthcoming, The Pacific, a slow or stalled demographic transition affecting development, in M. Rapaport ed., The Pacific Islands, Environment and Society, University of Hawaii Press. Rallu J.L., A. Sachs Robertson, 2010, The demographic window of opportunity in Pacific island countries : future prospects, in Pacific Perspectives 2009; Crises and Opportunities, UNESCAP, UN New York 2009, pp. 15-35. UNDESA, 2009, World Population Prospects, the 2008 Revision, United Nations, New York. Summary: Interrupted and reversed fertility transition is common in Polynesia and Micronesia. Many countries (Tonga, Samoa, Marshall Islands, the Federated States of Micronesia) have had TFR stable or slowly declining around 4 for several decades, some after a decline that brought them under this level (Tokelau, Tuvalu). Poor data on union formation and contraceptive prevalence do not bring much evidence of the role of proximate determinants and the latter is rather a mean to achieve a desired fertility level. The simultaneous and opposite trends in emigration and TFR are a typical Pacific phenomenon and certainly reveal causation. Mass migration necessitates a rather high level of fertility to avoid population decline. A continuous flow of migrants is also necessary to ensure sustainable level of remittances in the frame of an economic system based on remittances (MIRAB). 12