The Social ProTecTion indicaTor for The Pacific

This publication provides updates on Social Protection Indicators of 13 Pacific developing member countries of the Asian Development Bank. It presents an analysis of 2015 data on social protection programs in the countries, particularly the substantial progress in terms of expenditure and benefit size between 2009 and 2015. Spending favored the nonpoor over the poor. Yet, spending on women remained behind that on men, with the gap even widening. The Social Protection Index—now the Social Protection Indicator—was developed by the Asian Development Bank and its partners as the first comprehensive and quantitative measure of social protection systems in Asia and the Pacific.


Foreword
In 2015, the Sustainable Development Goals were adopted by United Nations member states to serve as a blueprint for achieving a better and more sustainable future for all. The provision of social protection was identified as one of the key elements of the first goal on ending poverty.  The report uses the Social Protection Indicator (SPI) for assessing the nature and effectiveness of social protection both within and across countries. The SPI shows the level of resources invested in social protection, the extent of coverage, benefit levels, and distribution of expenditure in terms of poverty and gender.

Social Protection Expenditure
The average social protection expenditure in the Pacific in 2015 was 6.0% of aggregate gross domestic product (GDP) and for each intended beneficiary as a share of GDP per capita, 5.3%. Aggregate expenditure varied from 13.8% of GDP in Timor-Leste to 1.1% of GDP in Papua New Guinea.
Social insurance refers to contributory schemes to help people respond to common risks such as illness, old age, and unemployment. Social insurance dominated spending on social protection in the region, with average spending of 3.4% of aggregate GDP and 3.0% of GDP per capita. Social insurance expenditure was largely taken up by old-age pensions (for persons aged 60 and above), equivalent to 2.8% of aggregate GDP. Pensions in the Pacific are administered through provident funds and offer support to those in formal employment. Expenditure on other social insurance programs, namely health assistance and other social insurance, was significantly less than that on pensions.
Social assistance spending averaged 2.1% of aggregate GDP and 1.9% of GDP per capita. Social assistance provides unrequited transfers to groups, such as the poor, who cannot qualify for insurance or would receive inadequate benefits from such a source.
The large difference in per capita social insurance expenditure contributed to the overall benefit gap between the poor and nonpoor. Social insurance is biased toward those in formal employment who are less likely to be poor. Social insurance spending on the nonpoor (2.4% of GDP per capita) was nearly five times as high as that on the poor (0.5% of GDP per capita). Social assistance also showed less per capita spending on poor (0.7% of GDP per capita) than on nonpoor beneficiaries (1.1% of GDP per capita). ALMPs showed a similar pattern: 0.1% of GDP per capita for the poor and 0.3% for the nonpoor.

Gender Dimension
Social protection spending in the Pacific favored men over women in 2015: 3.3% of GDP per capita on men and 2.0% on women. Social insurance expenditure on men, on average, exceeded that on women by 0.9 percentage points, as a higher proportion of men than women are in wage employment and therefore are more likely to be covered by social insurance. Social assistance spending also favored men, with a distributional gap of 0.3 percentage points, as did ALMPs, with a gap of 0.2 percentage points.

Trends in Social Protection over Time
Social protection made substantial progress across the region. Expenditure improved from 4.1% of aggregate GDP in 2009 to 6.0% in 2015. It also improved in per capita terms, from 3.3% of GDP to 5.3% over the period. Expenditure improved in all three social protection categories.
Social insurance expenditure increased from 2.1% of GDP per capita in 2009 to 3.0% in 2015. The majority of countries showed improvements. Social assistance expenditure increased from 1.1% of GDP per capita in 2009 to 1.9% in 2015. Per capita expenditure in ALMPs increased from 0.1% of GDP per capita to 0.5%.
Social protection coverage improved from 17.6% of intended beneficiaries in 2009 to 31.2% in 2015. Social assistance coverage increased by 8.2 percentage points, surpassing the gain in social insurance and ALMPs.
Social protection benefit size picked up across all three social protection categories. Social insurance benefits improved by 32.1 percentage points, from 112.5% of GDP per capita in 2009 to 144.6% in 2015.
Average social protection expenditure continued to favor the nonpoor in all three social protection categories in both 2009 and 2015. Furthermore, the gap in spending between the poor and nonpoor increased from 1.7 percentage points in 2009 to 2.5 percentage points in 2015.
Spending on women remained behind that on men in all three social protection categories, as the gap in spending widened from 0.9 percentage points in 2009 to 1.3 percentage points in 2015.
This report presents the analysis of comprehensive 2015 data on social protection programs in 13 Pacific countries. 1 It uses the Social Protection Indicator (SPI) for assessing social protection within and across countries.
This report continues the effort to systematically track and assess developments in social protection. The Asian Development Bank (ADB) and its development partners originally developed its precursor-the Social Protection Index-in 2005 as a tool to monitor social protection in Asia and the Pacific. This report updates the results of a 2012 analysis for the Pacific published in 2016 (ADB 2016).
Systematic and comprehensive monitoring and evaluation are essential for improving social protection systems to enhance their capacity to support the poor and vulnerable. The SPI approach offers uniform indicators that can help governments and other stakeholders assess the effectiveness of social protection systems and to track progress over time. At country level, it assesses social protection systems as well as progress. At cross-country level, it can help trace and assess patterns in social protection. This SPI monitoring helps track the progress of Target 1.3 of the Sustainable Development Goals (SDGs). 2 The analysis in this report is useful for two reasons. First, the report provides data on all types of government social protection initiatives, focusing on aggregate expenditure, coverage, benefit size, and poverty and gender dimensions. This enables one to assess social protection as a system by looking at social protection as a whole. The SPI thus complements existing national monitoring and evaluation tools, which tend to focus on individual programs rather than the overall system.
Second, the SPI analysis allows a focus on distinct areas of social protection activities. It provides disaggregated data on components of social protection 1 The 13 Pacific countries are the Cook Islands, the Federated States of Micronesia (FSM), Fiji, Kiribati, Nauru, Palau, Papua New Guinea, the Marshall Islands (RMI), Samoa, Solomon Islands, Timor-Leste, Tonga, and Vanuatu. 2 SDG target 1.3: Implement nationally appropriate social protection systems and measures for all, including floors, and by 2030 achieve substantial coverage of the poor and the vulnerable.

Overview and Methodology
systems and allows one to assess their relative significance. These components comprise three "categories"-social insurance, social assistance, and active labor market programs (ALMPs)-and include corresponding "programs" such as health insurance, pensions, and welfare assistance ( Table 1).
The analysis of 2015 data in this report follows the methodology used in the previous 2016 publication (using 2012 data). This report also presents comparisons of social protection indicators for 13 Pacific countries from 2009 to 2015.

What Is the Social Protection Indicator?
The SPI is an indicator for assessing social protection effectiveness within and across countries. Going beyond measuring social protection expenditure as a share of aggregate gross domestic product (GDP), the SPI assesses expenditure for each intended beneficiary as a share of GDP per capita. Intended beneficiary population groups are the key unit for assessing social protection effectiveness in this report-in other words, the share of the population that could qualify for benefits from a particular social protection category and program. The report distinguishes intended beneficiaries from actual beneficiaries, that is, those given support in practice. This distinction is captured by the coverage indicator, which measures the share of actual beneficiaries in the total intended beneficiary population. The SPI is a relative indicator against the GDP per capita of each country. 3 Table 2 presents the main intended beneficiary groups for each of the programs. A detailed explanation for defining and selecting the groups is in ADB (2011).

Methodology for Deriving the Indicators
Using 2015 data, this report draws on the methodology set out in ADB (2011), which provides a detailed explanation of defining and constructing the SPI and key data sources. The information required for the analysis falls into two general categories: basic statistics and indicators; and data on social protection expenditure and on the number of social protection beneficiaries. The following data sources were used: government statistics and reports, reports by international financial institutions and bilateral agencies, discussions and interviews with agencies responsible for social protection, and household survey data.
Throughout this report, most averages are unweighted. This represents a methodological choice, signifying that indicators for each country are treated as equal in weight to those in any other country.
For the 2015 Pacific update, a thorough and careful review of social protection measures has been undertaken across the Pacific region. With the availability of updated data for 2009 and 2012, the SPIs for these two years have therefore been revised. In particular, some important revisions to expenditure on social insurance have been made to 2009 and 2012 SPI data, notably for Kiribati, the Marshall Islands, Papua New Guinea, and Samoa. In addition, the cash-for-work ALMPs in Kiribati and the Marshall Islands have now been included in the SPI calculation; these two programs directly support incomes of poor outer-island households. The revised figures for 2009 and 2012, together with the 2015 data, where feasible, are presented in this report.
Results on the coverage of SPI show some countries exceeding 100% of intended beneficiaries. This simply means that in some countries, beneficiaries may receive more than one kind of social protection. For example, an elderly person may receive some form of pension while receiving health assistance benefits.
The analysis of social protection in the Pacific is also carried out by countries' income levels and regional groupings (Table 3). There are two country income groups based on World Bank classification: upper-middle-income countries (UMICs) and lower-middle-income countries (LMICs). 4 The regional groupings are, per ADB's classification, Micronesia, Melanesia, and Polynesia. These "regions" allow comparisons with other countries in the Pacific region that broadly share similar geographic and resource endowments. Four of the Micronesian countries of the North Pacific are primarily comprised of atolls, and/or small reef or volcanic islands-Kiribati, the Federated States of Micronesia (FSM), the Marshall Islands, and Palau-while Nauru is a single raised limestone island. These countries have small scattered populations-ranging in size from around 10,000 in Nauru to 100,000 in the FSM and Kiribati-and, apart from fishing in their exclusive economic zones, are generally resource poor.  The Melanesian countries are comprised primarily of larger volcanic islands, but still include small isolated populations due to the mountainous geography and the multitude of islands in their respective archipelagoes. These countries are much more resource-rich, particularly Fiji, Papua New Guinea, and Timor-Leste. Although less well endowed, Solomon Islands has exploitable mineral deposits, and Vanuatu has considerable agricultural potential. The populations in Melanesia are much larger, ranging from around 300,000 in Vanuatu to over 7 million in Papua New Guinea.
The three Polynesian countries-the Cook Islands, Samoa, and Tonga-comprise a mixture of widely scattered small to medium-sized volcanic islands and atolls. They have few natural resources and small and dispersed populations (although in Samoa, the population is concentrated on two islands).

Structure of the Book
Chapter 2 discusses the 2015 results on the overall social protection expenditure as a share of aggregate GDP and of GDP per capita for each intended beneficiary, while Chapter 3 discusses expenditure by the main categories and programs of social protection. Chapter 4 discusses the coverage and benefit size of social protection benefits, and Chapter 5 deals with the distributional dimensions of social protection along poverty and gender lines. Chapter 6 discusses the progress of social protection from 2009 to 2015. Chapter 7 provides the conclusions.
• In 2015, overall social protection expenditure in 13 Pacific developing member countries averaged 6.0% of aggregate GDP. • Such expenditure varied from 13.8% of GDP in Timor-Leste to 1.1% of GDP in Papua New Guinea. • Expenditure for each intended beneficiary as a share of GDP per capita-the SPI-averaged 5.3%. • The SPI ranged from 12.6% (Timor-Leste) to 0.8% of GDP (Papua New Guinea). Figure 1 illustrates social protection expenditure overall as a share of aggregate GDP and for each intended beneficiary as a share of GDP per capita-the SPI. 5 The average share of social protection expenditure in aggregate GDP was 6.0%, and expenditure in five countries was above this average. Expenditure for each intended beneficiary was 5.3% of GDP per capita, with five countries exhibiting results above the average. Spending on Social Protection-General Results those receipts are allocated to a sovereign wealth fund, providing the country with the ability to spend more than its non-oil GDP would normally permit. For its part, Kiribati has had a major boost to its government budget and gross national income (GNI) from very high levels of fish license revenues since 2010. 6 From 2010 to 2015, its GDP rose by an estimated 28%, while GNI rose by almost 75%. This enabled total budget expenditure to increase by around one-third, including a 30% increase in the allocation to the cash-for-work program for the rural poor.
Palau, which is one of the most affluent countries in the Pacific (with a per capita GDP of $16,273.14), spent 11.0% of GDP on social protection. Other countries with high social protection expenditure, such as the Marshall Islands and the FSM, have Compacts of Free Association with the United States (US), which provide significant budget and development support, enabling their social security administrations to provide relatively generous social insurance, in turn boosting their per capita social protection expenditure.   In the UMIC group (the Cook Islands, Fiji, the Marshall Islands, Nauru, and Palau), social protection expenditure came to 7.3% of aggregate GDP and 5.9% as a share of GDP per capita (Figure 2), both higher than the Pacific's averages.
In the LMIC group (the FSM, Kiribati, Papua New Guinea, Samoa, Solomon Islands, Timor-Leste, Tonga, and Vanuatu), however, the averages were at 5.2% of GDP and 4.9% of GDP per capita, which were both lower than the overall averages for the Pacific (Appendix , Table A.1 and Table A.5).
By region, in Micronesia (the FSM, Kiribati, the Marshall Islands, Nauru, and Palau), the average expenditure in terms of aggregate GDP (9.2%) and GDP per capita (7.9%) were higher than the averages for the Pacific (Figure 2 GDP. This chapter presents social protection expenditure by the three categoriessocial insurance, social assistance, and ALMPs-as well as their associated programs (see Table 1).
Across the Pacific, social insurance programs have taken the biggest share of expenditure both in terms of aggregate GDP and GDP per capita for each intended beneficiary (Figure 3). More than half the Pacific's expenditure went on social insurance, at an average of 3.4% of aggregate GDP, while social assistance expenditure came to 2.1% of aggregate GDP. Spending on ALMPs was minimal, at 0.5% of aggregate GDP.
Social protection expenditure for each intended beneficiary averaged 5.3% of GDP per capita. The equivalent figures were 3.0% for social insurance, 1.9% for social assistance, and 0.5% for ALMPs.
Social insurance expenditure was dominated by pension programs. Much social assistance expenditure was taken up by welfare assistance and assistance to the elderly. In ALMPs, spending on food and cash for work was greater than that on skills development and training.

Social Protection Spending by Category and Program Social Insurance
Social insurance refers to contributory schemes to help people respond to common risks such as illness, old age, and unemployment. It includes pensions; health insurance; and other social insurance programs such as unemployment and disability insurance, maternity benefits, and benefits from provident funds.
Social insurance spending in four countries significantly exceeded the regional 3.4% average of aggregate GDP: the Marshall Islands (10.7%), Palau (10.3%), the FSM (6.2%), and Kiribati (6.1%) (Appendix, Social insurance expenditure mainly accrued from the pension programs covering both public and private employees. Health insurance and other social insurance expenditure was significantly less than those for pensions (Appendix ,  Table A.2).
For both income groups, spending on social insurance was higher than that on social assistance and ALMPs (Figure 4). Social insurance expenditure in the UMIC group at 5.0% of GDP was twice as high as that in the LMIC group. The shares of the Marshall Islands (10.7%) and Palau (10.3%) drive up the average social insurance spending in the UMIC group. In the LMIC group, Kiribati (6.1%) and the FSM (6.2%) showed higher than average shares.
The two income groups exhibited a similar pattern on per capita expenditure, with social insurance spending the highest among the three categories, averaging 4.0% of GDP per capita in the UMICs and 2.3% in the LMICs. The Marshall Islands (9.0%) and Palau (7.7%) posted above-average shares of social insurance expenditure in the UMIC group. The remaining three countries in this group  In the LMIC group, above-average per capita social insurance expenditure was seen in Kiribati (5.9% of GDP), the FSM (5.2%), and Solomon Islands (2.7%). Samoa spent 2.3% of GDP per capita, but the remaining LMIC countries-Papua New Guinea, Timor-Leste, Tonga, and Vanuatu-spent less than 1.0%.
By region, social insurance expenditure was high in Micronesia, at 6.7% of aggregate GDP and 5.6% of GDP per capita ( Figure 4). Of the five member countries in this region, only Nauru spent less than 1.0% of aggregate GDP on social insurance. The Marshall Islands and Palau spent more than 10.0% of aggregate GDP and the remaining two countries, Kiribati and the FSM, spent about 6.0% of GDP. On per capita expenditure, a similar pattern holds, with the Marshall Islands having the highest social insurance expenditure and Nauru the least.

Pensions and Health Insurance
Social insurance comprises pensions, health insurance, and other social insurance. Pensions include contributory benefits for older persons aged 60 years and above. Health insurance refers to contributory programs that enable access to health care. Other social insurance programs include passive labor market programs such as unemployment and disability insurance, maternity benefits, and benefits from provident funds.
Pensions accounted for the largest share (2.8% of aggregate GDP) of social insurance expenditure. Health insurance (at 0.1%) and other social insurance programs (at 0.5%) took small shares (Appendix, Table A.2). All the 13 Pacific developing member countries provided pension insurance, yet health insurance was present in only two: Palau and Tonga.
Pensions in the Pacific primarily provide retirement support for those in formal employment. In most countries, they are provided through national provident funds offering contributory benefits mainly for those in formal employment (Box 1). Those in informal employment, or in mixed cash/subsistence farming or fishing, have often been excluded or otherwise unable to participate in these schemes. One possible reason is either that the law does not allow for their participation or because of poverty.
Box 1: Provident funds-The core of Social insurance in the Pacific Ten of the 13 Pacific developing member countries have national provident funds as their primary source of social insurance. The other three Compact countries (the Federated States of Micronesia, the Marshall Islands, and Palau) of the North Pacific have social security administrations. These provident funds are all similar in structure and purpose and were mostly established around the time that each country became independent. In Tonga, the Retirement Fund Board evolved from the public service pension scheme in 1999 and was followed by the National Retirement Benefits Fund targeting the private sector in 2012. The provident funds are all established by statute, and their general purpose is to ensure retirement benefits for both private and public sector employees.
The funds are compulsory savings schemes that cover those in formal employment. In some countries, the self-employed, informal employees, or migrant workers are eligible to become voluntary contributors. All employees employed under a contract of service must contribute a certain percentage of their income to the fund and this is matched by contributions from the employer. Although contribution levels vary by country, contributions are often 10% from both employer and employee.
Benefits can be drawn from the funds in a variety of ways. The ideal situation is for members to contribute throughout their working life and to build a lump sum for their eventual retirement. In some countries, this will then be withdrawn as a lump sum; in others, the member may purchase an annuity. Many funds have now made it possible for members to use their provident fund balances as collateral for housing, education, or other personal loans. Following Tropical Cyclone Winston in 2016, the Fiji provident fund allowed affected members to draw up to $1,000 from their fund account to assist with cyclone recovery. The relaxation of the rules on early access has, however, diluted the retirement purpose of many funds, such that at retirement many members have only small balances after loans have been paid off.
As provident funds focus on those in formal employment, their coverage in practice tends to be very narrow. The provident fund benefits, however, can be very high when lump-sum benefits are taken at retirement.
The principal challenges to expanding provident fund coverage to informal workers relate mainly to administrative capacity and affordability. It could be costly in financial and administrative terms to manage many small and irregular contributions and withdrawals. The amounts collected would likely be small and the higher costs would then have to be covered by lower returns to all members. Furthermore, due to the large share of informal sector employment, many informal workers cannot afford to contribute regularly.

Social Assistance
Social assistance provides unrequited transfers to groups, such as the poor, who cannot qualify for insurance or would receive inadequate benefits from such a source. Its main policy programs are welfare assistance, such as cash or in-kind transfers, child welfare, assistance to the elderly, health assistance (tax-funded benefits), and disability benefits.
Social assistance expenditure is the second largest in terms both of aggregate GDP and of GDP per capita for each intended beneficiary. In 2015, the Pacific's average social assistance expenditure was equivalent to 2.1% of aggregate GDP (Appendix , Table A.1). The highest spending was in Timor-Leste with 12.9%, followed by Nauru with 4.4%, and the Cook Islands with 3.9%. Expenditure in the other 10 countries was below the Pacific's average for this category.
On a per capita basis, the Pacific's average for social assistance was 1.9% of GDP (Appendix, Table A.5). Again, Timor-Leste had the highest figure, with 11.8%. Two-thirds of this expenditure stemmed from social pension/welfare payments to veterans of the struggle for national liberation (Box 2). Similarly for Nauru, social assistance of 4.2% of GDP per capita was above the regional average, reflecting the school feeding program introduced in 2014 and the higher levels of elderly pension payments paid in 2015. 7 Yet Nauru had low per capita social insurance expenditure (0.1% of GDP).
Conversely, social assistance in three North Pacific countries of the FSM, the Marshall Islands, and Palau was very small, but these countries provided relatively generous social insurance. Tonga's social protection expenditure was 0.3% of GDP per capita because, with a strong and homogenous society, the country has relied on traditional, custom-based social protection systems rather than government interventions.
Social assistance expenditure averaged 2.1% of GDP for both income groups ( Figure 4). In the UMIC group, the Cook Islands (3.9%) and Nauru (4.4%) exhibited above the average shares, and the remaining countries under this income group spent less than 1.0% of GDP on social assistance. For the LMIC group, Timor-Leste provided the largest share of expenditure with 12.9% of GDP. The remaining countries in this group spent at most 1.0% of GDP on social assistance. Finally, in Solomon Islands, social assistance spending was negligible, and in Papua New Guinea, there are no social assistance programs of any significance.
In terms of per capita expenditure, a small difference was registered in the two income groups, with 1.8% of GDP per capita for the UMIC group and 1.9% for the LMIC group ( Figure 4). Two countries in this group, the Cook Islands (3.0%) and Nauru (4.2%) had above-average per capita expenditure on social assistance. The three remaining countries (Fiji, the Marshall Islands, and Palau) spent less than 1.0% of GDP per capita. In the LMIC group, as in the overall social protection, Timor-Leste was the only country that exhibited a significant amount of per capita expenditure at 11.8% of GDP (Appendix, remaining countries in this income group spent at most 1.0% of GDP per capita on social assistance. Regionally, expenditure on social assistance in Melanesia and Polynesia dominated social protection expenditure ( Figure 4). In Melanesia, social assistance spending was 2.8% of aggregate GDP and 2.5% of GDP per capita. The social assistance program in Timor-Leste was the driver for this relatively high expenditure. The programs of social assistance are now discussed in more detail.

Welfare Assistance and Assistance to the Elderly
Welfare assistance includes conditional and unconditional cash and food/ in-kind transfers for various groups, while assistance to the elderly refers to noncontributory transfers or social pensions targeted specifically at persons aged 60 and above.
The dominant programs in social assistance are welfare assistance (with expenditure of 1.1% of GDP) and assistance to the elderly (0.7% of GDP) (Appendix , Table A.2).
In seven Pacific countries (the Cook Islands, Fiji, Kiribati, Nauru, Samoa, Timor-Leste, and Tonga), benefits were available to all the elderly who reached the national qualifying age. 8 In most Pacific countries, women outnumber men among the elderly; women therefore benefit more than men from transfers to the elderly, especially where these benefits are universal. This is important, because far fewer females are covered under social insurance.

Child Welfare
Child welfare refers to assistance solely targeted at children, such as school feeding programs, child allowances, and child grants. Child welfare is present in most countries, although the expenditure provided for this type of program was only 0.2% of GDP in 2015 (Appendix , Table A.2). Child welfare programs, including school feeding, are generally universally available to all children in those countries that provide such benefits. These countries maintain an adequate level of gender equity in child welfare benefits, at least at primary level, where there are high rates of enrollment for boys and girls.

Health Assistance and Disability Assistance
Health assistance includes tax-funded benefits that enable individuals to use health services, but spending was the lowest across all social assistance programs. Health assistance was present in four countries (Samoa, Timor-Leste, Tonga, and Vanuatu), but the combined expenditure in these countries was negligible, at 0.01% of GDP (Appendix, Table A.2).
Family health associations in several Pacific developing member countries exemplify health assistance and primarily provide reproductive health services for young women. They are funded through the International Planned Parenthood Federation, filling a gap that is not well served by national medical services.
Disability assistance includes cash benefits for persons with disabilities. Disability assistance amounted to 0.1% of GDP, and programs were present in only two countries: Nauru and Timor-Leste. The allocations in these countries were still small compared with other social assistance programs. In Timor-Leste, some 7,400 beneficiaries (0.6% of the total population) received disability benefits (Appendix , Table A.4). In Nauru, only 158 beneficiaries (1.4% of the total population) received weekly cash benefits (Nauru Country Report, 2017). The FSM provided special education services for children with disabilities.
The increasing incidence of noncommunicable diseases (NCDs) and the complications that arise from serious cases are leading to a rising number of people living with a disability. 9 According to World Health Organization (WHO) surveys, about 40% of the populations of the Pacific islands have been diagnosed with an NCD, and NCDs account for more than 70% of all deaths, a large share of which are premature and preventable. 10 The primary health concern now for many Pacific island countries is the increasing earlier onset of NCDs, especially those relating to lifestyle and diets. Consequently, increasing rates of disabilities among Pacific populations are inevitable in the coming years.
At present, persons with disabilities account for an estimated 0.4% of all actual social protection beneficiaries in the region (Appendix , Table A.6). This figure is in line with the 4.1% of people in Pacific island countries (excluding Papua New Guinea, which has no data) who are estimated to be living with disabilities. 11 Although relatively small in absolute numbers, the incidence of disabilities will increase most among the middle-aged and elderly, who are more vulnerable to NCD complications. Care for these age groups can be time-consuming and expensive. The need for and costs of social protection for carers and the caredfor will become an increasing issue for governments of Pacific island countries, either directly in health costs or through social protection benefits and services.

Active Labor Market Programs
ALMPs comprise cash-and food-for-work programs as well as skills development and training programs. They mainly provide limited income support to lowincome households, including direct support to outer-island copra cutters or assistance in getting seasonal work in Australia and New Zealand.
Only five countries in the Pacific have ALMPs and spend significant amounts on them (Kiribati, the Marshall Islands, Solomon Islands, Timor-Leste, and Tonga). Another five spend negligible amounts on these programs overall (Fiji, Palau, Papua New Guinea, Samoa, and Vanuatu). Three countries-the Cook Islands, the FSM, and Nauru-do not have ALMPs (Appendix , Table A.1). ALMPs therefore show tiny expenditure across the Pacific, averaging 0.5% of GDP in 2015, both aggregate and per capita ( Figure 4).
In Kiribati and the Marshall Islands, the main ALMPs provided direct income support to poor outer-island copra cutters through cash-for-work programs (Box 3); Solomon Islands and Timor-Leste provided cash for work for unemployed youth, and Vanuatu had a small technical/vocational training program targeted at unemployed rural youth. In Kiribati and Solomon Islands, the ALMPs were larger than their per capita expenditure on social assistance. Tonga spent 0.1% of GDP, offering assistance for those engaged in the Australian Seasonal Worker Programme and the New Zealand Recognized Seasonal Employer scheme.
Expenditure on ALMPs averaged less than 1.0% of aggregate GDP for both income groups: 0.6% in the LMIC group and 0.2% in the UMIC group ( Figure 4). Per capita, the average for the UMIC group was 0.2% of GDP and for the LMIC group, 0.6% of GDP.
Skills development and training programs target the unemployed and underemployed and seek to enable them to take advantage of employment opportunities and improve employment quality. These programs are available in seven countries of the Pacific (Fiji, Kiribati, Palau, Papua New Guinea, Samoa, Solomon Islands, and Tonga), but their combined expenditure was only 0.1% of aggregate GDP in 2015 (Appendix, Table A.2).

Box 3: cash-for-Work Programs in Kiribati and the Marshall islands
The copra cash-for-work programs in Kiribati and the Marshall Islands provide income support for poor outer-island households whose economic opportunities are severely limited by isolation from markets and lack of employment opportunities. In both countries, the government provides a price bonus for each ton of copra produced. The cash is paid directly to the copra cutters when the copra is collected by visiting interisland vessels. The programs are self-selecting based on the "no copra, no payment" principle and are therefore targeted and directly linked to work undertaken.
The cash-for-work payment for copra production in the Marshall Islands represents about one-quarter of total agricultural income and about 10% of total outer-island income. It is thus an important income supplement for the rural poor.
Source: Social Protection Indicator country reports, 2017.
Social protection coverage indicates how many of the intended beneficiaries of social protection receive the benefits. The size of those benefits indicates the amount that actual beneficiaries receive.

Coverage
• Social protection in the Pacific covered only 31.2% of intended beneficiaries in 2015 ( Figure 5), meaning that nearly two-thirds of eligible persons were left without support. • Social insurance covered 8.7% of intended beneficiaries. Palau has nearuniversal social insurance coverage (84.2% of intended beneficiaries). • Social assistance had the widest coverage of the three categories, at 20.0%. • Active labor market programs (ALMPs) had narrow coverage of 2.4%.
Differences in social protection coverage are stark: Palau had the highest rate, at 96.9% of intended beneficiaries, and Papua New Guinea the lowest, at a mere 0.3% (Appendix, Table A.7). Coverage was wider in the UMIC group, with 43.7% of intended beneficiaries, than in the LMIC group, with 23.8% ( Figure 6), due to the high rate of coverage in Palau. Coverage in the other four countries in this income group (the Cook Islands, Fiji, the Marshall Islands, and Nauru) averaged 30.4% of intended beneficiaries. In the LMIC group, Samoa (45.9%), Timor-Leste (46.4%), and Vanuatu (66.4%) exhibited wider coverage.
By region, the widest coverage was in Micronesia, with 34.8% of intended beneficiaries (again reflecting Palau), followed by Polynesia (29.4%), and Melanesia (28.6%) ( Figure 6). Excluding Palau from the computation would give the remaining four countries (the FSM, Kiribati, the Marshall Islands, and Nauru) 19.3% coverage, to become the lowest of the three regions.   In Melanesia, two countries posted coverage significantly below the region's average: Papua New Guinea at only 0.3% and Solomon Islands at 2.7%. In Polynesia, coverage in the Cook Islands was 34.6% and in Samoa 45.9%. Tonga pulled down the regional average, covering only 7.6% of intended beneficiaries.

Social Insurance
Social insurance coverage of intended beneficiaries comes second to social assistance ( Figure 5), at less than half the rate (8.7% against 20.0%), even though social insurance (mainly pensions) had the highest expenditure in the Pacific (Appendix , Table A.3).
Social insurance coverage was substantial only in Palau, at 84.2% (Appendix ,  Table A.7). Palau's coverage comes from the provision of two health insurance schemes: National Health Insurance and the Medical Savings Account, which cover almost the entire population. Palau heavily skews the Pacific average upward and is the only country to exceed the 8.7% Pacific average. Excluding Palau from the computation, the 12-country average falls to 2.7%. This narrow coverage stems from the contributory nature of social insurance and the link with formal employment. Four countries (Fiji, the FSM, Kiribati, and the Marshall Islands) had social insurance coverage exceeding 2.7%, while five countries (Nauru, Papua New Guinea, Timor-Leste, Tonga, and Vanuatu) had tiny coverage of less than 1.0% (Appendix , Table A.7).
There is a wide difference in social insurance coverage between the UMIC and LMIC groups. UMIC coverage was 19.9% ( Figure 6; Appendix, Table A.7) (but excluding Palau, only 3.8%). With only one pension insurance program, Nauru had the lowest coverage at 0.3%. For the LMIC group, coverage was 1.8%, varying from 5.3% in the FSM to 0.1% in Timor-Leste.
Mainly because of Palau, coverage in Micronesia (20.3% of intended beneficiaries) was far above that in Melanesia (1.6%) and Polynesia (1.3%) ( Figure 6). And even when Palau is excluded from the computation for Micronesia, coverage was still higher than in the other two regions (4.3% of intended beneficiaries) (Appendix ,  Table A.7).

Social Assistance
Among the three social protection categories, social assistance had the widest coverage in 2015, reaching 20.0% of intended beneficiaries ( Figure  In Samoa, social assistance covered 42.6% of intended beneficiaries, mainly owing to the health assistance program. In Timor-Leste, beneficiaries of the school feeding program, which reached about 343,000 in 2015, drove the 38.6% figure. At the other end of the scale, in Solomon Islands, only 0.5% of intended beneficiaries were reached (Appendix, Table A.7).
There was not much difference in coverage between the UMIC and LMIC groups ( Figure 6). In the UMIC group, social assistance coverage was 20.9%. In this income group, Nauru had the widest coverage of 36.8% of intended beneficiaries, mainly through the provision of the school feeding program. The Marshall Islands, with only one program on special education, had the least coverage of 1.3% in this group. In the LMIC group, social assistance coverage was 19.4%.

Active Labor Market Programs
ALMPs had the least coverage among the three categories, at only 2.4% of intended beneficiaries in the Pacific in 2015 ( Figure 5). The Marshall Islands had the widest coverage at 13.9%, due to its income support program for copra farmers. Coverage in the two income groups showed some difference: 2.9% in the UMIC group against 2.2% in the LMIC group ( Figure 6). ALMPs have very limited coverage in all three regions (Appendix, In most countries, the size of benefits in 2015 received by each actual beneficiary was less than GDP per capita. It is possible, however, for the benefit size to exceed GDP per capita (see Appendix, Table A.8) where (lump-sum) benefit levels are particularly high and the number of beneficiaries is relatively low, such as with social insurance benefits paid out under social security arrangements and provident funds in some countries. This is notable in countries with relatively low GDP per capita and low rates of formal employment.
Social protection benefits in the Pacific averaged 54.1% of GDP per capita in 2015 ( Figure 7), with 144.6% for social insurance, 16.4% for social assistance, and 17.9% for ALMPs.
Countries displayed a wide variation in social protection benefits, ranging from 251.1% of GDP per capita in Papua New Guinea to 2.1% of GDP per capita in Vanuatu (Appendix,  Generous benefits in the Pacific were provided through pensions or through the lump-sum payments from provident funds on retirement. Many retiring employees prefer to take their provident fund benefits as lump sums, rather than as annuities or pensions. A long-serving public servant can often receive a substantial lump-sum amount, often far above GDP per capita.
These lump sums are frequently spent rather than saved and invested to provide an income in retirement. This is due partly to the lack of investment opportunities in the generally undeveloped financial markets of LMICs, and the very low interest rates on standard savings accounts. It is also due partly to the Pacific tradition of sharing wealth and resources with family, especially when related to family events such as weddings and funerals, and in contributions to the church.
The two income groups show big differences, both as a share of GDP per capita and in monetary terms: 46.2% ($2,871.72) for the UMIC group and 206.1% ($4,815.81) for the LMIC group (Table 4).
The social insurance benefits in Melanesia were notably higher than the benefits in Micronesia and Polynesia ( Figure 8). Four out of the five countries in Melanesia had benefits of more than 100% of GDP per capita (Appendix , Table A.8), as did two countries in Micronesia (Kiribati and the Marshall Islands). In monetary terms, Polynesia was ahead of the other two regions (Table 4).

Social Assistance
The average social assistance benefit was 16.4% of GDP per capita, or $627.91 in monetary terms. Both the Pacific average and benefits for countries were far smaller than social insurance benefits (Figure 7).
In the Marshall Islands, the benefits were 62.3% of GDP per capita, the highest in the Pacific, mainly associated with the special education program. Benefits in the FSM at 48.8% of GDP per capita were also on the higher end. Benefits in Timor-Leste (30.7% of GDP per capita) and Kiribati (28.2% of GDP per capita) were also higher than the Pacific average. However, their monetary values ($348.18 in Timor-Leste and $415.34 in Kiribati) were lower than the Pacific average of $627.91 (Table 4).
Social assistance benefits in the Cook Islands were only 9.2% of GDP per capita, but the monetary value was the third highest in the Pacific, at $1,533.09, because of the country's high GDP per capita. In Vanuatu, the benefit was only 0.7% of its GDP per capita, equivalent to only $19.05.
There was little difference between the two income groups in terms of share of GDP per capita. UMIC countries spent 17.9% of GDP per capita and LMICs 15.5% ( Figure 8). However, in monetary terms, the benefits were much higher in the UMIC group ($1,071.01) than the LMIC group ($350.98) ( Table 4).
The regional pattern reveals more discernible differences. Benefits in Micronesia of 31.0% of GDP per capita were above the Pacific average ( Figure 8). Melanesia and Polynesia provided similar benefits (7.4% and 7.1% of GDP per capita), both lower than the Pacific average (Appendix, Active Labor Market Programs The regional average in ALMPs was 17.9% of GDP per capita (Figure 7). The benefit was $391.28 in monetary terms, significantly lower than the monetary benefits in social insurance and social assistance (Table 4). Kiribati had the highest benefit in the Pacific as a share of GDP per capita, at 128.3%, mostly stemming from its cash-and food-for-work program and its support for outward movement of labor under the New Zealand Recognized Seasonal Employer scheme. The benefit in Solomon Islands was 48.5% of GDP per capita, mainly attributable to the country's skills development and training program and support for domestic unemployed youth. In monetary terms, the benefit in Kiribati was the highest at $1,889.63 (Table 4).
There was a wide variation between the two income groups: 26.7% of GDP per capita for the LMIC group, equivalent to $438.58, and 4.0% for the UMIC group, equivalent to $315.58 (Figure 8, Table 4). Among the regional groups, the highest benefits were in Micronesia at 28.0% of GDP per capita ( Figure 8). In Melanesia, benefits came to 17.5% of GDP per capita or $360.60 per beneficiary. Benefits in Polynesia of 1.9% of GDP per capita or $74.89 were significantly below the Pacific average.
This chapter discusses the poverty and gender dimensions of the Social Protection Indicator or of the SPI. The former provides a measure of the distribution of social protection expenditure as a share of GDP per capita between poor and nonpoor beneficiaries. The latter does the same thing for the distribution between women and men.

Distribution of Expenditure between the Poor and Nonpoor
• Social protection spending in the Pacific tends to favor the nonpoor over the poor. • Social insurance spending on the nonpoor in 2015 (2.4% of GDP per capita) was nearly five times as high as that on the poor (0.5% of GDP per capita). Nearly all countries spent more on the nonpoor. • The gap in social assistance spending was less pronounced than that in social insurance, but the nonpoor still received more than the poor (1.1% versus 0.7% of GDP per capita) in the Pacific. Most countries spent more on the nonpoor. • In ALMPs, the difference in spending was very small, at 0.2 percentage points in favor of the nonpoor. All but two countries spent more on the nonpoor.
In 2015, social protection expenditure for the poor amounted to 1.4% of GDP per capita, against 3.9% for the nonpoor (Figure 9). For all social protection categories, expenditure for the nonpoor was higher than for the poor. Only in one country (Fiji) did spending favor the poor over the nonpoor, and only by 0.1 percentage points (1.7% versus 1.6% of GDP per capita) ( Figure 10).
In 2015, social insurance expenditure for poor beneficiaries amounted to 0.5% of GDP per capita, versus 2.4% for nonpoor beneficiaries. This is largely because contributory social insurance programs are tied to formal employment, but the Poverty and Gender Dimensions of Social Protection Spending  poor are less likely to be formally employed (the majority of them are in the informal economy). Even when voluntary contribution options exist, the poor in the informal economy are often unable to contribute to insurance schemes because they have low incomes and may not see the benefit of paying insurance contributions (Handayani 2016).
Social assistance showed a smaller gap in spending on poor beneficiaries (at 0.7% of GDP per capita) and nonpoor beneficiaries (at 1.1%). For ALMPs, expenditure spent on the poor was at 0.1% of GDP per capita, lower than 0.3% for the nonpoor.
In all countries except Fiji, per capita social protection expenditure that went to the poor was lower than that to the nonpoor ( Figure 10). In Kiribati, the gap was widest: the poor received 1.6% of GDP per capita while the nonpoor received 8.8%, reflecting relatively heavy social insurance spending.
The two income groups also recorded social protection spending differences ( Figure 11): 1.6% of GDP per capita in the UMICs and 1.3% in the LMICs for the poor, but 4.4% in the UMICs and 3.6% in the LMICs for the nonpoor. Regionally, Micronesia exhibited significant differences along the poor/nonpoor axis for all social protection categories ( Figure 11). In this region, per capita expenditure on the poor averaged 1.6% of GDP per capita-4.7 percentage points lower than expenditure on the nonpoor. For social assistance and ALMPs, there was not much variation in expenditure. In Melanesia, the gap between the poor and nonpoor was much less, at 0.9 percentage points (1.7% versus 2.6%). In Polynesia, the three countries spent, on average, 0.6% of GDP per capita on the poor and 2.1% on the nonpoor.

Social Insurance
There were wide differences in per capita social insurance expenditure between the poor and nonpoor, with spending across nearly all countries favoring the nonpoor ( Figure 12). Fiji was the only country where social insurance expenditure for the poor and nonpoor was equal, at 1.4% of GDP per capita.
A substantial gap was observed in the FSM, Kiribati, the Marshall Islands, and Palau. Expenditure in the Marshall Islands had the widest difference (1.6% versus 7.4%). This was because social insurance programs there mainly  cover those in formal employment, while most of the poor remain outside their coverage.
In the UMIC group, per capita social insurance expenditure was 0.8% of GDP on the poor and 3.1% on the nonpoor (Figure 11). The LMIC group showed a smaller gap, between 0.3% and 2.0%.
By region, the poor/nonpoor difference was most pronounced in Micronesia, at 3.9 percentage points. Poor beneficiaries there received 0.8% of GDP per capita and the nonpoor 4.7% (Figure 11). Four out of Micronesia's five countries showed significant differences. Palau, where all the social insurance programs are tied to employment, had the highest distributional gap of 5.5 percentage points. The differences were less in Melanesia (0.7 percentage points) and Polynesia (0.9 percentage points) ( Figure 11).

Social Assistance
Expenditure for the poor was 0.7% of GDP per capita and for the nonpoor 1.1% (Figure 9). Spending in Fiji showed a 0.1-percentage-point preference for the poor, and, in the Marshall Islands, exhibited an equal distribution ( Figure 13). The other countries spent more on the nonpoor than the poor.
The widest distributional gap-1.2 percentage points-was in the Cook Islands, where expenditure on the poor was 0.9% of GDP per capita, against 2.1% for the nonpoor. This gap is because the assistance to the elderly program-the biggest in the Cook Islands-covers only a small number of poor beneficiaries. Timor-Leste, with the highest per capita expenditure on social assistance, showed a 1-percentage-point difference between poor (5.4% of GDP) and nonpoor (6.4%) beneficiaries.
Some social assistance programs in the Pacific are universal, and therefore spending on the nonpoor does not necessarily indicate leakage. Very few social assistance programs in the region are actually poverty targeted. Those that are include the Bolsa da Mãe program in Timor-Leste, the Poverty Benefit Scheme in Fiji, and the destitute allowance program in the Cook Islands. Yet despite these poverty-targeted programs, the nonpoor still received more than the poor in the Cook Islands and Timor-Leste. In Fiji, the poor had a slight advantage (0.1 percentage points) over the nonpoor.
The income groups show narrower distributional gaps, each of 0.4 percentage points. Spending received by the poor in the UMIC group was 0.7% of GDP per capita, and by the nonpoor 1.1% (Figure 11). In the LMIC group, spending on the poor was 0.8% of GDP per capita and on the nonpoor, 1.2%.
In all three regions, social assistance expenditure for the nonpoor was higher than for the poor, although the gap was not as pronounced as in social insurance. In Micronesia, expenditure on the poor averaged 0.5% of GDP per capita, less than the 0.9% on the nonpoor ( Figure 11). In Melanesia, spending was higher and the difference smaller than in the other two regions, as 1.2% of GDP per capita was spent on the poor and 1.4% on the nonpoor. In Polynesia, the gap was widest at 0.7 percentage points-0.4% versus 1.1%.

Active Labor Market Programs
At such low spending rates for ALMPs-0.3% of GDP per capita on the nonpoor and 0.1% on the poor (Figure 9)-the poor/nonpoor difference was inevitably inconsequential, at 0.2 percentage points (Appendix,  capita. Its cash-for-work program, funded by the Government of the Marshall Islands for copra production, accounts for about one-quarter of total agricultural income and about 10% of total outer-island income. It is thus an important income supplement for the rural poor.
The difference in expenditure was pronounced in Kiribati, with 0.8% of GDP per capita on the poor and 2.8% on the nonpoor. Kiribati's major ALMPs offer income support for copra farmers, but this fails to provide higher per capita expenditure for poor beneficiaries. Timor-Leste has an ALMPs that targets cash for work for rural youth. It spent 0.2% of GDP per capita on the poor and 0.5% of GDP per capita on the nonpoor.
The UMIC group had no poor/nonpoor difference in per capita expenditure ( Figure 11). The difference of 0.3 percentage points in the LMIC group was mainly driven by Kiribati.
By region, Micronesia spent 0.2% of GDP per capita on the poor and 0.6% on the nonpoor (Figure 11), largely due to spending on the nonpoor in Kiribati (Appendix , Table A.9). Melanesia spent only 0.1 percentage points more on the nonpoor. Spending on both the poor and nonpoor in Polynesia was negligible.

Distribution of Expenditure by Gender
• Social protection in the Pacific favored men over women, at 3.3% of GDP per capita versus 2.0%. • Social insurance expenditure on men, on average, exceeded that on women by 0.9 percentage points. • The difference in social assistance spending was 0.7 percentage points, but spending in four countries favored women and was distributed equally in another five (see Figure 18). • There was a small difference of 0.2 percentage points in ALMPs benefits, on average, between men and women. Still, expenditure favored men in three countries and was equally shared in one country.
Overall per capita social protection spending on male beneficiaries was higher than that for female beneficiaries in 2015, and in all three categories ( Figure 14). The widest gap was in social insurance, at 1.1% versus 1.9%. In social assistance, the gap was 0.8% against 1.1%, and in ALMPs, 0.1% versus 0.3%.
The unequal distribution of social protection expenditure needs to be considered in the context of overall gender inequalities in the region. The basic-needs poverty rates for men and women are estimated to be similar in the Pacific. This, however, can hide inequalities between men and women within households themselves (UNDP 2014, National Poverty Reports). The gender gap is driven by the situation in the labor market. A higher proportion of men than women are in wage employment and a higher proportion of women are confined to domestic duties as carers for children and the elderly. Women without their own source of income are generally reliant on their husbands, or male breadwinners, to provide the cash needed for the daily needs of their families. Sociocultural circumstances further perpetuate these inequalities in the home. In this context, it is especially important to provide greater targeted social protection support for women. Overall per capita social protection expenditure is higher for male than female beneficiaries in all Pacific countries except the Cook Islands and Nauru (where expenditure was higher for women than men) and Tonga (where spending was distributed equally) ( Figure 15).
Social protection expenditure was higher for male than female beneficiaries in both income groups in all three categories ( Figure 16). In the UMIC group, the average figure for females was 2.3% and for males, 3.6%. In the LMIC group, the corresponding figures were 1.8% (for females) and 3.2% (for males), a similar percentage point difference for both groups.
By region, overall expenditure for women was 3.0% of GDP per capita for Micronesia, 1.5% for Melanesia, and 1.3% for Polynesia (Figure 16), against 5.0% (Micronesia), 2.8% (Melanesia), and 1.4% (Polynesia) for men. This suggests that nearly one-third of social protection in Micronesia and Melanesia was received by women, and about two-thirds by men. Polynesia has a greater gender balance, largely due to the higher proportion of women in formal employment than in Micronesia and Melanesia. Furthermore, a higher proportion of elderly, and particularly elderly women, receive benefits in Polynesian populations.

Social Insurance
Expenditure on female beneficiaries was 1.1% of GDP per capita and on male beneficiaries 1.9% ( Figure 14). The gap was especially pronounced in the Marshall Islands ( Figure 17), with 2.3% for female beneficiaries and 6.7% for males. In two countries-the Cook Islands and Tonga-social insurance spending was virtually equal (Appendix , Table A.10).
There is also a notable gender difference in the income groups. Spending on women averaged 1.3% of GDP per capita in the UMIC group and 0.9% in the LMIC group (Figure 16), against 2.6% (UMIC group) and 1.5% (LMIC group) for men.
All three regions favored men over women. The widest gap was in Micronesia, at 3.6% versus 2.0% ( Figure 16). The discrepancy was substantial in four of the five countries in this region, especially in the Marshall Islands (Appendix , Table A.10).
In Melanesia, spending on men was double than that on women (1.0% versus 0.5%). In three countries of this region-Fiji, Solomon Islands, and Vanuatuspending on men was nearly twice as high. The results were almost equal in Polynesia, with spending on men only slightly higher than that on women-0.6% versus 0.5%.

Social Assistance
Against a backdrop of spending of 0.8% of GDP per capita for women and 1.1% for men (Figure 14), three countries-the Marshall Islands, Palau, and Timor-Leste-conspicuously favored men over women. The widest gap was in Timor-Leste, at 4.1% to 7.7% (Figure 18), despite the country's Bolsa da Mãe (Mother's Allowance) program targeted at women, due to the high payments to predominantly male war veterans.
In four countries (the Cook Islands, Fiji, Kiribati, and Nauru), spending on women exceeded that on men, though only marginally in the Cook Islands and Fiji. Both countries have programs that target women: maternity benefits in the Cook Islands and the Care and Protection program in Fiji. In Nauru, the birth claim benefit was targeted at women, and women made up almost two-thirds of the beneficiaries of the old-age pension. In another five countries (the FSM, Samoa, Solomon Islands, Tonga, and Vanuatu), social assistance expenditure was virtually equal for men and women. Spending between the income groups differed somewhat: in the UMIC group, that for women and men averaged 0.9% of GDP per capita, while in the LMIC group, there was a 0.4 percentage point difference in favor of male beneficiaries (1.2% versus 0.8%) ( Figure 16).
Regionally, spending in Micronesia and Polynesia exhibited a very slight bias toward women (0.8% versus 0.7%). In Melanesia, spending on men (1.6% of GDP per capita) was 0.7 percentage points higher than on women (0.9%), mainly due to Timor-Leste (Appendix , Table A.10).

Active Labor Market Programs
Across the Pacific, ALMPs spending as a share of GDP per capita favored men (0.3% of GDP versus 0.1%) (Figure 14), reflecting higher spending on men in three of the four countries with data and spending nonnegligible amounts (Appendix, Only the LMIC group showed much difference in per capita expenditure between female and male beneficiaries, at 0.1% (for women) and 0.5% (for men) (Figure 16). In the UMIC group, spending was 0.1% for both genders.
Micronesia and Melanesia spent more on men than women ( Figure 16). Micronesia was more unequal, at 0.6% versus 0.2%, largely reflecting the gap in Kiribati (Appendix,

Social Protection Expenditure
• Social protection expenditure in the Pacific rose from 4.1% of aggregate GDP in 2009 to 6.0% in 2015 ( Figure 19). It also improved as a share of GDP per capita, from 3.3% in 2009 to 5.3% in 2015 ( Figure 20). • Expenditure generally improved in all social protection categories, in both aggregate and per capita terms.
The two main social protection categories saw higher spending as a share of aggregate GDP from 2009 to 2015: on social insurance from 2.6% to 3.4%, and on social assistance from 1.5% to 2.1% ( Figure 19).
Both categories also lifted their expenditure as a share of GDP per capita for each intended beneficiary (Figure 20): on social insurance from 2.1% to 3.0%, and on social assistance from 1.1% to 1.9%.
By income group, the share of expenditure for the UMICs grew from 6.0% in 2009 to 7.3% of aggregate GDP in 2015, and for the LMICs from 2.7% to 5.2% ( Figure 21). Per capita expenditure also increased for both groups: from 4.8% of GDP per capita in 2009 to 5.9% in 2015 for the UMIC group, and from 2.0% to 4.9% for the LMIC group ( Figure 22). Spending in the UMICs was boosted by the relatively generous social insurance provided in the Marshall Islands and Palau; and that in the LMICs by the relatively high social insurance in Kiribati and the FSM, and the high social assistance in Timor-Leste (the highest of all Pacific island countries).  By region, social protection expenditure in Micronesia increased from 5.3% of aggregate GDP in 2009 to 9.2% to 2015 (Figure 21), for gains in both main categories. Expenditure in Micronesia was dominated by social insurance, and a 7.3 percentage point increase in social insurance spending in the Marshall Islands was especially notable. Progress was also driven by gains in social assistance, particularly in Nauru.

Changes in Social
In the five Melanesian countries, the gain was smaller, with expenditure rising from 3.6% of aggregate GDP to 4.6% over the period. Average social protection expenditure was boosted by advances in social assistance expenditure in Timor-Leste and social insurance in Solomon Islands.  This expenditure pattern was similar for GDP per capita. The biggest improvement was in Micronesia, from 4.4% in 2009 to 7.9% in 2015 ( Figure 22). Melanesia increased spending from 2.7% to 4.3% over the period. Polynesia increased its outlays from 2.1% in 2012 to 2.6% in 2015. Figure 23 shows per capita social protection spending by country for the three years.

Social Insurance
Social insurance expenditure in the Pacific increased from 2.1% of GDP per capita in 2009 to 3.0% in 2015. Of the nine countries with comparable data from 2009 to 2015, only Fiji and Nauru showed decreases, while most showed improvements (Appendix , Table A.5). Kiribati, which was included in the regional compilation only in 2012, showed a decline in its social insurance per capita expenditure from 7.1% of GDP per capita in 2012 to 5.9% in 2015.
The declines in the Kiribati and Fiji per capita social insurance expenditure were caused by increases of almost 25% in beneficiaries from 2009 to 2015, 13 while social insurance expenditure in each country remained almost unchanged.
The Marshall Islands recorded the largest increase in social insurance over the period, from 3.0% of GDP per capita to 9%. This was due to its significant withdrawals from its provident fund. Solomon Islands also recorded a large 13 Information from Kiribati Provident Fund.  relative increase, from 1.0% to 2.7%, reflecting a doubling in social insurance expenditure (with an increase of around 50% in the number of beneficiaries). 14 Per capita expenditure improved in both income groups: slightly in the UMICs (from 3.9% to 4.0%), but more strongly in the LMICs (from 0.7% to 2.3%) ( Figure 24).

Social Assistance
Social assistance expenditure increased from 1.   All three regions showed some improvement ( Figure 25). The largest was in Melanesia (from 1.2% to 2.5%, in view of Timor-Leste). Micronesia increased its per capita spending from 0.9% to 1.4% (Nauru) and Polynesia from 1.1% to 1.4% (the Cook Islands) (Appendix, Table A.5).

Social Protection Coverage
• Coverage improved sharply, from 17.6% of intended beneficiaries in 2009 to 31.2% in 2015. • Progress in social assistance coverage surpassed that in social insurance.
• Nine countries expanded coverage in both social insurance and social assistance programs.  Overall social protection coverage rose steadily from 2009 to 2015 (Figure 26), mirroring coverage gains in 10 countries. Yet despite coverage progress, benefits still failed to reach more than two-thirds of intended beneficiaries in 2015.
The overall improving trend is grounded in progress in both main categories. The largest extension in coverage of 8.2 percentage points was observed in social assistance, from 11.8% of intended beneficiaries to 20.0% in 2015. It was followed by social insurance, with its 4.9 percentage points.
Both income groups increased coverage from 2009 to 2015 ( Figure 27): the UMICs from 18.5% to 43.7% (with large increases in social insurance and especially in social assistance), and the LMICs from 16.9% to 23.3% (with both main categories recording progress).
Regionally, social protection coverage improved in Micronesia (from 20.9% to 34.8%, on social insurance and social assistance) and particularly in Melanesia (from 11.2% to 28.6%, largely on social assistance). In Polynesia, coverage increased only slightly (from 28.0% in 2012 to 29.4% in 2015) ( Figure 28). (Only Samoa was included in the 2009 data for this region.)

Social insurance
Social insurance coverage more than doubled, from 3.   Figure 27). Palau was the main contributor. The LMICs saw a much smaller increase, from 0.7% to 1.8% over the period.
The widest progress was in Micronesia, from 8.1% in 2009 to 20.3 in 2015 ( Figure 28), again mainly driven by Palau. Coverage in the Melanesia declined slightly, from 1.7% to 1.6%, mainly due to decreased coverage in Fiji (Appendix, Driving the results were the extension of coverage in Fiji (from 4.0% to 20.5%), Nauru (from 5.0% to 36.8%), and Vanuatu (from 3.6% to 64.5%) (Appendix ,  Table A

Social Protection Benefit Size
• Social insurance benefits in the Pacific improved by 32.1 percentage points, and social assistance benefits by 2.8 percentage points, for each actual beneficiary as a share of GDP per capita. • Five countries improved benefits in social insurance, and four in social assistance programs. • Benefits declined in eight countries in both social insurance and social assistance programs.
The average social protection benefit declined across the region from 2009 to 2015, from 81.3% of GDP per capita (nine countries), to 69.6% with the inclusion of an additional four countries in 2012, and further to 54.1% in 2015 ( Figure 29). This decline can be explained by a substantial decrease in social insurance benefits in eight countries.  Benefits in both income groups decreased from 2009 to 2015 ( Figure 30). The decline was more substantial in the LMICs, where benefits fell from 124.0% of GDP per capita to 76.7%. This was due to a decrease in social insurance benefits in the Cook Islands, Fiji, and Palau and in social assistance benefits in Fiji (Appendix , Table A.8). The decrease in the LMICs was 10.2 percentage points, from 28.0% of GDP per capita to 17.8%. This was mainly driven by a decline in social insurance benefits in five countries.
By region, benefit size in Micronesia increased substantially, from 19.4% of GDP per capita to 50.6% (Figure 31). This was due to progress in social insurance and social assistance benefits in the Marshall Islands and social assistance in the FSM (Appendix,

Social Assistance
The improvement in social assistance benefits was more modest than that in social insurance. Benefits increased from 13.

Benefit Size and Coverage
Countries with the highest benefits, such as the FSM, Kiribati, Papua New Guinea, and Solomon Islands, had proportionally lower coverage (Appendix, A high size of benefits combined with a low coverage implies that a small number of beneficiaries receive a disproportionate share of the benefits. Where the high size of benefits is biased toward social insurance, it further suggests that the beneficiaries are unlikely to be among the poorest, because social insurance is generally available to those in formal employment ( Figure 32).   For example, social insurance benefits in Papua New Guinea are the most generous in the region, at 437.1% of GDP per capita. However, social insurance covers only 0.2% of the country's eligible population. Social insurance in Papua New Guinea is narrowly targeted at those in formal employment and there are no social assistance programs of any significance.
Conversely, where there is high coverage and low benefit size, the implication is that while there may be a significant number of beneficiaries, benefit levels are quite low and thus may be of little real assistance in alleviating hardship or poverty. Thus, a low social assistance benefit is associated with a high coverage in Palau, Samoa, and Vanuatu. For example, social assistance in Vanuatu covers 64.5% of the eligible population, offering each beneficiary small benefits, equivalent to 0.7% of GDP per capita (Appendix,

Distribution of Expenditure between the Poor and Nonpoor
• The average expenditure on the poor was lower than on the nonpoor in all three social protection categories in both 2009 and 2015, with a worsening trend over the period. • Five countries increased their spending on the poor in social insurance, and four in social assistance programs.
Average social protection expenditure in the region has increasingly favored the nonpoor since 2009. The spending on the nonpoor in 2009 (2.5% of GDP per capita) was 1.7 percentage points higher than that on the poor (0.8%), but by 2015, the gap has widened to 2.7 percentage points (3.9% versus 1.2%) (Figure 33). The per capita expenditure on the poor was lower than on the nonpoor in all three social protection categories, but especially in social insurance, as this category's high benefits went almost exclusively to the nonpoor in formal employment.
Still, per capita spending on the poor increased over the period from 0.8% of GDP per capita to 1.2%. Nine countries increased their spending on the poor during this period (Appendix , Table A.9), notably Timor-Leste, Nauru, and the Marshall Islands, reflecting new social assistance measures in Timor-Leste and Nauru, and an expansion in social insurance spending in the Marshall Islands. Spending on the nonpoor increased from 2.5% of GDP per capita to 3.9% in nine countries.
Spending in all regions continued to favor the nonpoor in 2015, with varying trends: the gap increasing in Micronesia in favor of the nonpoor, staying the same in Melanesia but decreasing in Polynesia ( Figure 35). All three regions increased spending on the poor and nonpoor: from 0.7% of GDP per capita to 1.6% in Micronesia, from 0.9% to 1.7% in Melanesia, and from 0.4% to 0.6% in Polynesia.

Social Insurance
Average social insurance expenditure on the poor was lower than on the nonpoor in both 2009 and 2015, with the gap widening from 1.4 to 2.1 percentage points. Countries' social insurance spending favored the nonpoor in both years (Appendix , Table A.9), Fiji aside, where expenditure became the same (1.4% of GDP per capita) in 2015, as spending on the nonpoor declined. Spending on the poor increased in five countries, decreased in three, and remained the same in another five. The Marshall Islands led the way, lifting pro-poor spending from 0.5% to 1.6% (Appendix , Table A.9). Spending on the nonpoor increased in seven countries, decreased in four, and remained the same in two. Most notably, spending increased in the Marshall Islands by 5 percentage points and in Solomon Islands by 1.3 percentage points.
Spending in both income groups continued to favor the nonpoor from 2009 to 2015, though the gap in the UMICs decreased from 2.7 to 2.3 percentage points, while in the LMICs it increased from 0.6 to 1.7 percentage points ( Figure 34). Two countries increased per capita expenditure on the poor between 2009 and 2015: Timor-Leste (from 2.0% to 5.4% of GDP per capita), and Nauru (from 0.3% to 1.6%). Three countries decreased it by 0.1 percentage points (Fiji, the Marshall Islands, and Palau). Six countries increased their spending on the nonpoor (notably Timor-Leste and Nauru), three decreased their spending, and another three retained the same level.
The improvement in expenditure on the poor in Timor-Leste was due to the increase in pensions to war veterans, of whom around half were poor, and higher expenditure on the poverty-targeted Bolsa da Mãe program. From 2009 to 2015, such pensions increased by almost two-thirds, to $129.45 million, while spending on the Bolsa da Mãe program rose by almost 400%, to $8.37 million. The increase in Nauru was a result of the introduction of the school feeding program in 2014 and the 50% improvement in poverty targeting for school-age children. 15 Social assistance spending in both income groups favored the nonpoor over the poor in both years (Appendix , Table A.9). The gap increased from 0.2 to 0.4 percentage points in the UMICs, but slightly narrowed, from 0.4 to 0.3 percentage points, in the LMICs.
In all regions, expenditure favored the nonpoor in 2009 and 2015, with the gap increasing in Micronesia from 3.1 to 4.7 percentage points, but narrowing in Polynesia from 2.0 to 1.5 percentage points, mainly due to reduced spending on the nonpoor (Figure 35).  , Table A.9). It, however, decreased by the same amount in Micronesia from 0.3 to 0.2 percentage points and in Polynesia from 0.8 to 0.7 percentage points. All three regions increased spending on the poor (see Appendix, Table A.9).

Distribution of Expenditure by Gender
• Spending on women in the Pacific is still behind that on men in all three social protection categories, and the gap is widening. • Spending on women improved in 10 countries and decreased in three.  Spending on women improved in 10 countries and decreased in three (Appendix ,  Table A.10), with the largest increases in Timor-Leste (from 2.2% to 4.2%) and Nauru (from 0.4% to 2.3%), and the largest decrease in the FSM of 0.7 percentage points. Nine countries increased spending on men, with Timor-Leste the most (from 3.0% to 8.4%) and four decreased it, with Nauru the most (from 5.0% to 2.0%). Social insurance spending for women and men increased. That for women rose from 0.7% of GDP per capita in 2009 to 1.1% in 2015. Spending for women increased in seven countries, remained the same in two, and decreased in four. The countries that registered the largest increase were the Marshall Islands (by 1.0 percentage point) and Palau (by 0.9 percentage points). The FSM decreased its spending on women by 0.7 percentage points from 2012 to 2015. These three countries have generous social insurance systems and provide survivor benefits that often accrue to the wives of deceased members.

Social Assistance
The overall per capita social assistance expenditure on women was lower than Four countries increased per capita expenditure for women during the reference years. Five countries retained their spending and two countries (Palau and Samoa) slightly decreased spending on women. Nauru and Timor-Leste increased per capita expenditure for women from 0.3% of GDP per capita to 2.3%, and Timor-Leste from 2.1% to 4.1%. The increase in Nauru's expenditure for women came from the introduction of the school feeding program in 2014. The program is universal for all children and thus had an equitable split between males and females. The expenditure for females in Timor-Leste was boosted by the substantial increase in funding for the Bolsa da Mãe program that targeted poor women and children. On men, five countries increased their spending, four decreased it, and three kept the same rate. The most significant increases were in Timor-Leste (4.8 percentage points) and Nauru (1.6 percentage points).
Social assistance in the UMICs favored men over women in 2009, but spending was equally distributed by 2015 (Appendix , Table A.10). The LMICs spent more on men than women in both years, with the gap increasing from 2009 to 2015 from 0.2 to 0.4 percentage points.
By region, in 2009, Micronesia spent twice the amount on men than on women, but spending on women came to exceed that on men by 0.1 percentage points by 2015 (Appendix, Table A.10). This was partly due to the increased spending on women in Nauru. In Melanesia, spending on men was higher in both years. The gap between women and men in Micronesia increased by 0.5 percentage points, from 0.2 to 0.7 percentage points in favor of men. Spending on women in Polynesia was higher than that on men by 0.1 percentage points in both 2012 and 2015.
Social protection expenditure in the Pacific is increasing, but more is needed Overall social protection expenditure in 13 Pacific countries averaged 6.0% of GDP in 2015. There has been significant progress in the region as expenditure increased by nearly 2 percentage points, from 4.1% of GDP in 2009. This translated into an increase in the per capita social protection expenditure for each intended beneficiary from 3.3% GDP per capita in 2009 to 5.3% in 2015.
Financing was possible due to oil revenues in Timor-Leste and fish licenses in Kiribati as well as increased financial assistance from the United States to the FSM, the Marshall Islands, and Palau. It is also a reflection of governments' commitment to strengthen social protection.

Social insurance dominates other social protection categories
Social insurance dominates spending on social protection in the Pacific, with an average spending of 3.4% of GDP. Social assistance spending was smaller, at 2.1% of GDP. Spending on ALMPs was the lowest in the region, at 0.5% of GDP. All three social protection categories reported improvements over the period 2009-2015. Especially notable was progress in social insurance in the Marshall Islands and Solomon Islands and in social assistance in Nauru and Timor-Leste. The number of programs and level of expenditure remained very small for ALMPs, however.
Coverage is improving but needs to be extended further Social protection in the Pacific covered 31.2% of intended beneficiaries in 2015, nearly doubling from 2009. Social assistance achieved the widest Conclusions coverage, reaching 20.0% of intended beneficiaries. These results demonstrate important improvements but also that around two-thirds of eligible persons in the Pacific remain unserved by social protection. Social insurance programs, while found in all the countries, remain very narrowly based on those in formal employment. Social assistance programs reach only a fifth of intended beneficiaries.
Social insurance coverage should be broadened to reach those in the informal economy, including those working overseas. By instituting innovative financing and institutional arrangements, it is possible to enable these persons with an easier and more flexible access to the various national provident funds and social security administrations. For example, Fiji allows existing members and those not covered by the regular provident fund scheme to voluntarily contribute up to 12% of their gross income as additional contributions.
It is important to expand coverage of social assistance programs to better reach vulnerable groups, such as children, older persons, and persons with disabilities, as explained in more detail below.

Benefits remain limited for the poor majority
The average social protection benefit received by each beneficiary in the Pacific was 54.1% of GDP per capita. Social insurance benefits were the largest, with average benefits at 144.6% of GDP per capita ($4,068.08), substantially exceeding benefits for social assistance (16.4% of GDP per capita; $627.91) and ALMPs (17.9% of GDP per capita; $391.28). The generous social insurance benefits were provided through pensions or through the lump-sum payments from provident funds upon retirement.
Social insurance that targets workers in formal employment tends to provide generous benefits to a small population subset. Social assistance in the region has relatively high coverage, but offers small benefits. Most Pacific island countries need to substantially build financial and institutional capacity to attain an optimal combination of benefits and coverage that would have a real impact to those who are most in need.

Active labor market programs remain underdeveloped
In 2015, average expenditure for ALMPs was a mere 0.5% of aggregate GDP and 0.5% of GDP per capita. Three countries (the Cook Islands, the FSM, and Nauru) do not have ALMPs. ALMPs had the lowest coverage among the three social protection categories-only 2.4% of intended beneficiaries. Further investment in ALMPs is clearly needed to help address the needs of the unemployed and underemployed through skills development. It can also help support the poorest segments by involving them in food-and cash-for-work programs and providing them with immediate relief.

Social assistance does not fully support vulnerable groups
Spending on each of the social assistance programs is inadequate to address poverty and vulnerability. Welfare assistance and assistance to the elderly were the dominant social assistance programs, but their expenditure was only 1.1% and 0.7% of GDP. Child welfare programs were present in most countries, but their expenditure was just 0.2% of GDP. Disability assistance was equivalent to 0.1% of GDP and was present only in two countries. Health assistance was present in four countries, but the combined expenditure was negligible at 0.01% of GDP.
Financial considerations are critical when determining policy design, and it may be assumed that any social protection measures targeted at the elderly and those living with a disability are likely to experience rapidly increasing demands from beneficiaries.

Child welfare assistance needs to be expanded
High levels of basic-needs hardship and poverty persist. Poverty in childhood can have detrimental impacts lasting long into adult life, perpetuating the cycle of poverty and vulnerability.
The challenges for policy-makers are to identify appropriate eligibility and targeting arrangements (e.g., targeted or universal), the amount of the benefits, whether cash or in kind, the delivery mechanism, the manner in which the delivery is to be monitored, and its impact evaluated. While a universal program will benefit the well-off and the poor alike, it will minimize the administration costs and eliminate issues related to targeting. This mechanism may be feasible in the long run especially as the proportion of children in the population is declining in most Pacific countries.

Progress in pro-poor spending is slow
The average expenditure on the poor in 2015 was lower than that on the nonpoor in all three social protection categories, and the gap in spending between the poor and nonpoor increased from 1.7 percentage points in 2009 to 2.5 percentage points in 2015. Few countries showed a progressive trend from 2009 to 2015.
Social insurance coverage should be expanded to enable people working in the informal economy to gain easier and more flexible access to the various national provident funds and social security administrations. Governments should give more attention to improving the targeting of social assistance benefits.
Social protection needs to be more gender sensitive Average social protection expenditure on women in the Pacific in 2015 was lower than that on men in all three social protection categories, with the gap in spending widening from 0.9 percentage points in 2009 to 1.3 percentage points in 2015, although a few countries showed a progressive trend in some programs.
Governments need to make social protection more inclusive and gender sensitive. Women in the Pacific region are largely concentrated in the informal economy and therefore have fewer economic resources. Social protection can address female poverty and vulnerability if programs are designed and implemented to address the gender-specific drivers of poverty and vulnerability.

Social protection policy frameworks need to be strengthened
Strengthening social protection requires not only financial commitment but also efforts to develop broad-based and comprehensive social protection systems. Governments should develop comprehensive policy frameworks for coordinated provision of social protection to the poor and vulnerable. The Cook Islands, Fiji, and Timor-Leste are among the few that have broad-based and reasonably comprehensive social protection systems that target the elderly, vulnerable children, and those living with a disability.
An increasing number of ad hoc and uncoordinated social protection measures across the region have given rise to inconsistencies in targeting beneficiaries and in the levels of benefits provided. Greater policy coordination would help ensure that increasing levels of social protection expenditure benefit the poor, women, and the vulnerable.