Chapter 4. Systematising social protection in Cambodia

This chapter proposes a scaling-up of social assistance and a consolidation of social insurance arrangements in line with Cambodia’s 2017 Social Protection Policy Framework (SPPF). While domestic resource mobilisation policies are proving highly successful, the sustainable implementation of the SPPF will require a financing strategy. Finally, this chapter recommends measures to improve coherence across the social protection system and the expansion of tools for implementing, monitoring and evaluating social protection provision.

This chapter proposes policies to support the Royal Government of Cambodia’s plans for establishing of a social protection system. These recommendations are based on the evidence presented in the first three chapters of the review and are aligned to the policies and priorities outlined by the Social Protection Policy Framework (SPPF) published in 2017.

Chapter 1 identifies considerable livelihoods challenges lie ahead for Cambodia in spite of its impressive progress in reducing poverty and developing its economy. The decline in poverty is slowing and broader measures of deprivation have not fallen as far as monetary policy. Children are especially at risk: malnutrition among children under five years old is a problem across all income groups, while the low proportion of children who finish secondary school and the large number of children who are in the labour market are also a major concern. Informal employment is growing rapidly and many workers are in low-paid or vulnerable employment. Climate change is emerging as a major threat to homes and livelihoods and Cambodia’s demographics will become less favourable over time.

As Chapter 2 recognises, Cambodia has the building blocks for a social protection system that can respond to these challenges. A legal framework is in place, the SPPF is the second policy document for the sector as a whole and numerous schemes exist for each one of the system’s three pillars – social assistance, social insurance and labour market programmes. Health Equity Funds are leading the way towards Universal Health Coverage in Cambodia, with the poorest households among the first to benefit. The IDPoor targeting system is both transparent and effective at identifying those most at need. However, not all risks are covered by existing social protection instruments and Cambodia’s rapidly growing urban areas are poorly served by anti-poverty schemes.

Moreover, as Chapter 3 shows, existing social protection schemes are limited in terms of coverage and benefit levels; as a result, they have little impact on poverty. Although Cambodia’s public expenditure on social protection is increasing, its spending remains low by regional standards and the vast majority of these resources are used to finance pension benefits paid to retired civil servants and military veterans through arrangements that are still tax-financed rather than contributory. As a result, social spending is neither pro-poor nor sufficient to offset the impact of taxes on household consumption among poor and vulnerable households, with the overall effect that the fiscal system exacerbates poverty.

This chapter identifies specific interventions that can underpin the development of Cambodia’s social protection system, as envisaged by the SPPF. These proposed interventions will operate at a policy, programme or administrative level and will adopt a four-pillar approach, based on creating, expanding, reforming and sustaining the components of a new social protection system within a single, coherent and overarching framework.

Achieve a strategic and systematic expansion of social assistance

Chapter 1 indicated various factors which will not only limit further gains in reducing poverty but will also jeopardise the gains that have been achieved. It shows that a significant proportion of the poor are chronic poor, meaning that they have not benefited from the rapid growth of Cambodia’s economy in recent years and are not likely to do so in the future unless the nature of this growth changes. At the same time, the characteristics and location of poverty are evolving and there is frequent movement in and out of poverty. It is also important to note that overall living standards have not improved as fast as incomes: the multi-dimensional poverty rate is far higher than income poverty.

This section proposes different policy options for enhancing the impact of social assistance. It does not examine the implementation issues associated with these proposals, since these form the basis of a separate analysis of the capacity to implement the SPPF’s social assistance recommendations, as carried out by the European Union Social Protection Systems Programme (Box 4.1).

Box 4.1. Developing the capacity to expand social assistance

The European Union Social Protection Systems Programme (EU-SPS) has analysed the RGC’s capacity to implement the SPPF objectives related to social assistance (THL, forthcoming). This research ran in parallel to this review and its findings are an essential complement.

This report reviews the governance and implementation landscape of the current social assistance system and relates this to broader public sector reforms, most notably the decentralisation process. It also identifies the key actors involved in the co-ordination and implementation of social assistance programmes at central and subnational levels and assesses their capacity to enact reforms proposed by the SPPF. Capacity in this case refers to institutions’ formal mandate as well as the scope and effectiveness of their administrative structures.

The study suggests a number of guiding principles on the future governance structure for the social assistance schemes, the development of appropriate capacity to implement the system envisaged by the SPPF and the need for sequencing reforms to overcome current capacity constraints. It also analyses appropriate structures for the national-level co-ordination of social assistance, the co-ordination of implementation and the implementation activities themselves.

Scaling up social assistance with a focus on the young and very young

Existing social assistance programmes – including programmes implemented by the RGC and development partners respectively – address the needs outlined in Chapter 1 but are highly fragmented and do not operate at sufficient scale to address these problems at a national level. Nor are they linked to other social protection programmes.

There is a clear case for these programmes to be scaled up, not only to improve the living conditions of children but also to address broader concerns around the lack of skills among the workforce and long-term issues around competiveness. The SPPF maintains the current emphasis on programmes that support human development. In a context where caregiving responsibilities are an important contributor to poverty, scholarship programmes promote school attendance by offsetting the direct and indirect costs of educating children.

However, these schemes only operate for certain grades, do not operate country-wide, are implemented by different agencies (domestic and international) and are reliant (in part) on financial support from development partners, which raises doubts as to their sustainability and capacity to expand.

Moreover, there is a clear case for starting social protection interventions earlier, as demonstrated by the high levels of malnutrition among children under five years old. This review strongly supports the SPPF’s commitment to the gradual implementation of a nutrition promotion programme for children and pregnant women at a national level which is complemented by a policy for free counselling and medical check-ups for pregnant women. Given that malnutrition is a problem that affects a large proportion of children across the income distribution, consideration should be given to making this a universal programme.

There is a clear investment case for introducing a programme for children and pregnant women. Early childhood interventions not only strengthen cognitive development at an age when such programmes yield the greatest long-term gains (Heckman, 2008), but also reinforce the impact of scholarships and school feeding programmes when children are ready for school.

The Identification of Poor Households Programme (IDPoor) database (discussed in greater detail below) could provide a mechanism for ensuring that beneficiaries of a programme aimed at one stage of the lifecycle are automatically enrolled in a social protection programme later in life, including labour-market initiatives beyond education. In practice, existing social protection schemes do not operate at sufficient scale to ensure this continuity of coverage, with the possible exception of the Health Equity Funds (HEF).

Shifting from emergency response to developmental programmes

The need for better co-ordination between emergency response measures and other social assistance programmes (as proposed by the SPPF) is an important aspect of systematisation. It is important that the national disaster management strategy identify areas that are regular recipients of emergency assistance so that permanent social protection programmes be established in these areas. This shift will not only improve the speed at which communities in these areas who are affected by a natural disaster can be reached but will also allow these communities to increase their resilience ex ante, for example by climate-proofing their land and infrastructure or diversifying their activities. In so doing, social protection not only safeguards livelihoods but, at an aggregate level, promotes greater long-term productivity in the agricultural sector.

Placing greater emphasis on social assistance programmes aimed at improving human development rather than those that provide temporary support (typically food aid) will also yield longer-term benefits for recipients. However, better aligning the emergency response with human development social assistance programmes faces an institutional constraint, namely the fact that the Ministry of Economy and Finance (MoEF) directly manages the budget for emergency response. Such an arrangement reflects the challenge that line ministries experience in budgeting for national disasters, whose impact, location and duration is impossible to know in advance. Targeting especially at-risk areas with conventional social assistance programmes is a useful means of reducing the proportion of the emergency relief budget for which the MoEF is responsible.

Better integrating public works programmes to social protection and development plans

Public works programmes (PWPs) have been an important poverty-reduction instrument in Cambodia since 1993, with the result that the country has generated significant experience and expertise in implementing such programmes. However, PWPs have relied overwhelmingly on support from the World Food Programme and the Asian Development Bank and do not feature in the SPPF. This exclusion likely reflects two factors: the low rate of unemployment and the fact that PWPs are costly relative to other interventions. However, PWPs can adapt from being primarily poverty-alleviation instruments to interventions which simultaneously develop economic assets and provide training and work experience to low-skilled workers. Although this shift would increase the cost of such programmes, it would also greatly increase their broader economic benefits.

As an example of how this might work, Cambodia suffers from severe infrastructure deficits, which are particularly pronounced in rural areas and thus a major constraint on the competitiveness of the agricultural sector. PWPs can help not only meet these infrastructure needs (for example by building rural roads and improving irrigation) but can also maintain new infrastructure assets and thus greatly increase their lifespan. At the same time, the programmes can target specific groups at risk of under-employment or unproductive work (such as youth) and emphasise the training provided to participants. In so doing they can increase social protection among the vulnerable labour force, which is currently reliant on technical vocational education and training (TVET) (whose coverage is limited).

This might also be a means of expanding access to social protection into urban areas, which are currently underserved in this regard – a situation that cannot be sustained in a context of rapid urbanisation and thus an increase in the absolute number of urban poor. Expanding and adapting the IDPoor targeting mechanism (as mentioned below) to urban contexts is an important means of understanding the characteristics and needs of the urban poor and vulnerable. This knowledge should be used to ensure that the social protection programmes implemented in urban areas do not simply mimic the design of rural interventions.

The SPPF requires a long-term financing plan

Scaling up social assistance in the ways that are outlined above would allow it to make a significant impact in reducing poverty. However, it would also require a significant increase in the resources allocated to social protection, which would have to be sustained over a prolonged period. The SPPF proposes a gradual increase to social assistance that should not represent an unmanageable burden on public finances. The macroeconomic situation is favourable to such a scaling up: the RGC has greater fiscal resources at its disposal to increase the scope of social protection thanks to progress in increasing revenue generation at an even faster rate than that at which the economy has grown.

However, as Chapter 3 shows, social protection spending is low by regional standards and tax-funded pension payments to retired civil servants and military veterans are eliminating any fiscal space that might be opening up for social protection. A detailed plan for sustainably financing the expansion of social protection is an important complement to the SPPF.

Tax revenues as a proportion of gross domestic product (GDP) have risen from 9.7% in 2009 to an estimated 14.3% of GDP in 2015 (Figure 4.1) – a performance driven by the RGC’s success in enhancing tax compliance, particularly with regard to encouraging businesses to register formally. This increase has helped the government to consolidate public finances by reducing the fiscal deficit rather than increase spending, in keeping with a counter-cyclical approach to public spending.

Figure 4.2 shows the revenues generated by different tax instruments from 2010 to 2015. Indirect and trade taxes provide the bulk of trade revenues but there is evidence of an upwards trend in receipts from taxes on profits and salaries (Cambodia doesn’t have a personal income tax per se). This shift towards more progressive tax instruments complements the proposed increase in social protection spending both in terms of reducing poverty and improving the distributional impact of the fiscal framework.

These increases in revenues have coincided with stable spending levels. As shown in Figure 4.3, total spending in 2015 was 21% of GDP, which is the same level as in 2009 (though it did increase in the intervening years). The combined effect of increased revenues and steady spending has been to reduce the fiscal deficit from 9.5% of GDP in 2009 to 4.5% of GDP in 2015. At the same time, the debt level remains low (at 33% of GDP). However, wage increases for civil servants are expected to increase expenditure from 2016 onwards and are likely to increase the fiscal deficit.

While the SPPF identifies the financing challenges that confront different social insurance instruments, it does not contain a long-term financing strategy for expanding the social protection system as a whole. Given that the SPPF recommends the introduction and scaling up of a number of different programmes, it is important to understand the overall financing requirement and devise an implementation strategy that reflects the priorities for the sector through a costed action plan, containing different options for scaling up according to different future scenarios (for example, related to economic growth).

It is also necessary to forecast how the RGC’s resource envelope is likely to grow over time and how much of the spending envelope is likely to be allocated to social protection. Integrating social protection within the RGC’s broader development plans is crucial in this regard. Moreover, it is also essential to understand the impact of the tax system in terms of both redistribution and the reduction in consumption: increasing taxes to finance social protection risks exacerbating poverty if poor individuals are net payers to the fiscus rather than net beneficiaries.

Box 4.2. The level and type of development assistance is changing

Development partners have played a major role in funding social protection programmes in Cambodia. The World Food Programme, Asian Development Bank, United Nations Children’s Fund (UNICEF), Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), European Union (EU) and others fund the pillars of Cambodia’s social protection system: scholarships, TVET and HEF. They also finance relatively small-scale pilots, providing for antenatal and early childhood benefits and home-grown school feeding programmes.

However, Cambodia’s graduation to lower-middle income status is likely to change the dynamics. Already, the level of official development assistance (ODA) is not keeping up with growth in the economy, falling from 4.1% of GDP in 2010 to 1.9% of GDP in 2015 (National Bank of Cambodia, 2016). Moreover, the composition of Cambodia’s ODA is changing (Figure 4.4). Data from the OECD Development Assistance Committee shows concessional lending for economic and infrastructure services has grown strongly and now accounts for the majority of ODA, while assistance for social infrastructure and services in 2015 was at approximately the same level in absolute terms as in 2010 and has thus fallen significantly as a proportion of GDP.

This change in the structure of ODA is typical as countries’ income level rises. A corollary of this change is that the terms of assistance become less favourable for the government, since the investments supported by ODA are expected to generate an economic return.

Integrating social insurance arrangements for efficiency and equity

Cambodia’s social insurance programmes are in their infancy and, at present, are highly fragmented. Given that the population will start to age rapidly within the next 30 years (as discussed in Chapter 1), it is important that the SPPF’s recommendations for developing social insurance are acted upon quickly while preserving the overall objective of deepening systematisation.

As Chapter 2 shows, access to social security relies on individuals working in the formal sector even though informal employment is increasing rapidly and accounts for the majority of the workforce. Moreover, the extent and type of coverage depends on whether employees are in the private or public sector: public sector workers have access to pension arrangements but not yet health insurance, while private-sector workers do not have access to a pension arrangement but are covered by employment injury insurance and health insurance through the National Social Security Fund.

This fragmentation is reflected in the institutional structure for social insurance. Three principal arrangements exist, each established by different legislation with a separate governance structure: the National Social Security Fund (NSSF), the National Social Security Fund for Civil Servants (NSSF-C) and the National Fund for Veterans (NFV). Financing arrangements also differ between the three. The NSSF and NSSF-C are intended to be contributory arrangements while the NFV is run on a non-contributory basis, meaning pension payments are financed by general tax revenues. In reality, the NSSF-C is also non-contributory: the 2011 decree stipulating a total contribution rate of 24% of a civil servant’s salary has yet to be implemented.

As a result, pension payments for retired civil servants are financed through general taxation via the budget of the Ministry of Social Affairs, Veterans and Youth Rehabilitation (MoSVY). It accounts for a large (and rapidly growing) proportion of spending by the MoSVY. Moreover, the government balance sheet does not reflect the long-term liability that is being generated by the NSSF-C through the commitment to provide a pension to current civil servants when they retire. This implicit debt is growing as civil service wages increase. In actuarial terms, this arrangement is insolvent for as long as this debt is not acknowledged and matched.

Notwithstanding the urgency of enforcing contribution to the NSSF-C (the NFV arrangement is expected to remain non-contributory) and recognising both funds’ respective liabilities, the present situation is an opportunity for the RGC to devise an integrated framework for social insurance. This framework would ideally be based on a single pension arrangement (providing a comprehensive range of benefits) for all formal-sector workers. A similar arrangement for health insurance would also be optimal, as is discussed below.

Consistent with this, the RGC is currently considering a general pension scheme, in which private-sector workers and new civil servants will be enrolled. Under this dispensation, the NSSF-C would be closed to new members and its debt would be ring-fenced so that it is not transferred to the private sector. The NSSF-C would still become a contributory scheme but would slowly wind down.

From a systematic perspective, a single statutory pension arrangement provides numerous advantages at an institutional, administrative and policy level. Integrating mandatory pension arrangements for private-sector and new public-sector workers into a single statutory scheme would maximise potential economies of scale in terms of administration and governance. It would also guarantee fairness across these arrangements and minimise any distortions of the labour market associated with individuals opting to work in one sector due to better social security arrangements (at a given salary level).

The SPPF envisages that the mandatory pension system will comprise public and private sector arrangements which will share information systems and will “be consistent” with each other to ensure portability of benefits (and thus facilitate staff movement between the public and private sectors). However, there is a strong case for merging the public and private sector arrangements. It has not yet been decided whether the public and private sector funds will be designed as funded or unfunded arrangements but in either case the rationale for discrete arrangements is weak if the intention is to ensure consistency between the two.

If a funded defined contribution arrangement is selected for both arrangements, the system is better understood as one comprising individual accounts for each contributor rather than stand-alone funds for the public and private sectors. Achieving consistency in such a context is straightforward: the contribution rate, investment rules, governance arrangements and pay-out mechanisms should be the same for workers in the public and private sectors. If an unfunded arrangement (either defined benefit or defined contribution) is selected, ensuring consistency between discrete public- and private-sector arrangements would be much harder because the risk pools are different.

For example, in most countries the contribution density and risk profile differ between public and private sector workers (often in favour of the former). While it might be possible to offer the same level of benefits to members of these different arrangements under such circumstances, other parameters would have to differ in order to ensure they are both sustainable, such as the contribution rate or retirement age. This undermines consistency between the two arrangements, complicates portability and raises major equity concerns. This issue would fall away if a general pension scheme is established.

A detailed discussion of the appropriate design for the new pension system is beyond the scope of this paper. In broad terms, the shallowness of Cambodia’s financial markets is a major obstacle to generating the investment returns on which funded arrangements depend. The country’s demographics are supportive of a pay-as-you-go arrangement, which (if a defined benefit design is chosen) can guarantee an adequate level of benefits. However, it is important that the system does not make excessive promises to today’s workforce that will impose an unfair or unmanageable burden on future generations in the form of higher contribution rates to finance pension payments. Whichever system is chosen, it is essential that the legacy costs of the current system (the non-contributory NSSF-C and NFV arrangements) are properly accounted for.

Box 4.3. Placing the garment sector at the heart of the social protection system

According to the International Labour Organization, registered garment and footwear factories employed 610 000 workers in 2016, which is roughly a third of the total enterprise workforce (as calculated by the 2011 Enterprise Survey) (ILO, 2017; NIS and MoP, 2011). The total annual salary bill for the sector was USD 1.4 billion in 2016. The RGC’s strategy for including these workers within tax and social security arrangements will have major implications for public finances and the viability of social security arrangements in the future.

The RGC has focused on garment and footwear factories for implementing both the Employment Injury Insurance (EII) and the social health insurance (SHI), which has led to relatively quick increases in coverage of both arrangements through the NSSF. This success indicates that it would also be relatively straightforward administratively to implement a pension scheme for these workers.

However, the same workers are deliberately excluded from paying tax: the minimum salary tax threshold of USD 250 per month in 2016 was well above the minimum wage (USD 140 per month) and the average salary in the sector (USD 222 per month). Fringe benefits are also exempt from taxation.

The exclusion of garment workers from the tax system lowers the cost of employment, which is an important aspect of Cambodia’s export competitiveness. However, it also imposes a constraint on domestic resource mobilisation and could have implications for the distributional impact of the tax system if it requires additional resources to be generated from indirect taxes rather than the graduated (and thus progressive) salary tax.

Moreover, it should be noted that the rate of pension contributions (when these are introduced) is likely to be considerably higher than the 5% marginal income tax rate for workers in the lower band (which has an upper threshold of USD 375 in 2016). Such contributions, whether made by the employer or employee, have a similar impact on the cost of employment as taxes on income.

It is necessary to understand the likely interaction between the mandatory arrangements and an eventual social pension. The value of the social pension benefit should be set at a level which protects recipients from poverty but which is not sufficiently high to deter participation in the mandatory scheme. This would especially important in the case of the voluntary fund for informal workers currently envisaged by the SPPF. If the value of the contributory pension is not significantly higher than that of the social pension then potential contributors to the former will have no reason to enrol and the cost of protecting them from poverty in old age will fall predominantly on the state – specifically, the employers and employees of the day who pay their taxes. Given the current level of social benefits, however, this problem is unlikely to arise.

The SPPF envisages separate processes for reforming or establishing the civil service pension scheme, the private-sector pension scheme, a voluntary fund for higher earners, a fund for informal workers and the social pension. To ensure coherence at an institutional and policy level across the retirement space, the process for designing different components of the retirement landscape needs to happen at once and include all role players, including social partners. This might prolong the process but retirement systems have a lifespan and an impact (for better or worse) that cross multiple generations; it is worth tolerating delays in the short term to design a system that is fair and sustainable in the long term.

Universal health coverage can be achieved through greater systematisation

The SPPF traces an ambitious path to universal health coverage (UHC) for Cambodia, building on existing and recently developed health financing and health system innovations. HEF have improved financial access to health services, especially for the poorest 20% of the population. The recently approved Health Equity and Quality Improvement project, funded by donors and the RGC, will further strengthen the operational environment of HEF. At the same time, recent progress in rolling out social health insurance through the NSSF has increased coverage among formal sector workers and built up capacity required to manage a health insurance mechanism.

Alongside the demand-side reforms, improving the supply of health services is crucial. Here, the RGC can build on the service delivery grants (SDGs) and the monitoring and evaluation mechanisms built into them - although more evidence is still needed on the effectiveness and efficiency of the SDG model. The challenge of upgrading public health service provision is noted by the SPPF and is seen as a necessary step in increasing effective coverage. The accreditation body envisaged by the SPPF, which will ensure quality in both public and private health facilities, is also an important element of the overall UHC strategy.

A systemic approach to achieving UHC will be developed on three fronts:

Further extending the formal sector coverage (coverage of all private sector formal salaried workers and rolling out the health insurance system to civil servants)
Further expanding population and service coverage through HEF as a targeted measure for the poor and near poor and consolidating financing for HEF
Reaching the “missing middle” – the “non-poor” informal sector workers – through a comprehensive approach as discussed below.

While many countries are following this path to UHC, there are many pitfalls to avoid. One of the major challenges lies in reaching the “missing middle”, often defined as “non-poor” informal-sector workers who are expected to enrol through direct contributions. It has been shown in many countries, including Cambodia in the case of community-based health insurance (CBHI), that voluntary (or de facto voluntary) contributory health insurance mechanisms tend to fail in encouraging enrolment. They often struggle with adverse selection; in Indonesia, for example, the claims ratio for the voluntary, non-poor informal sector is around 600%, indicating that enrolment attracts mainly high-risk (and therefore relatively costly) individuals (Tandon et al., 2016).

In the context of Cambodia, where 55% of the population is either poor or vulnerable and informal employment predominates, the level of contributions for non-salaried workers will confront challenges related to affordability and regularity. At the same time, the non-poor informal sector with capacity to pay into a contributory system is a relatively small cohort; a separate pool designed for this population segment will face great challenges with regard to sufficiency of revenues and adverse selection.

There are three principal mechanisms for reaching the “missing middle”:

Extension of health coverage through formalisation policies.

Formalisation of the workforce is a key means of ensuring the enhancement and financial sustainability of health coverage. Formalisation will increase coverage by (gradually) squeezing out the “missing middle”. As the workforce formalises, the effect of more people being under a clear mandate to enrol in social protection schemes and the new revenue streams this entails, will increase health coverage and ensure its sustainable funding.

Creating cross-subsidies between different population groups.

The SPPF establishes principles for moving towards a single health insurance operator. However, there is still some uncertainty about the extent of cross-subsidisation between the different schemes with the single operator as well as if and how much of the funds will be pooled together. Pooling funds together from different revenue streams would allow cross-subsidies between population groups and would open options to enrol the vulnerable through contribution subsidies.

Higher funding through budget transfers.

The size of the poor and vulnerable population means that inclusion under any coverage mechanism cannot rely on large, direct contributions from individuals. Increasing coverage to the “missing middle” will require their contributions to be subsidised in some way. While this can, to some extent, be achieved through cross-subsidies from contributions of formal sector workers, it will also require matching contributions from the government, financed by general tax revenues. The envisaged expansion of HEF already relies on public funding.

In addition to the extension of coverage, focus needs to be given to the health benefit package, which will determine the “depth” of coverage. The SPPF proposes to study the feasibility of a unified package for all social health insurance schemes. Currently there are several packages for different schemes; unifying these packages and making informed decisions on inclusion and exclusion of services and medicine based on cost-effectiveness and equity criteria is essential for establishing clear entitlements.

Chile, for example, has established a guaranteed package of services based on clear guidelines related to both cost-effectiveness and equity criteria. This package, part of the AUGE reform, is accompanied by guaranteed access to services and guaranteed minimum waiting times (Missoni and Solimano, 2010). However, in the Cambodian context, the supply-side gaps are still important; effective coverage, especially in rural areas, will not be defined by the benefit package but by the availability and quality of health services. The upgrading of health service provision is thus a necessary condition for moving towards UHC.

A well-designed provider-payment mechanism and purchasing strategy are also necessary to ensure that incentives for providers are aligned with the objectives behind the UHC policy and that resources are used efficiently. In Cambodia, HEF and the service delivery grants have already laid the ground for purchasing mechanisms that focus on incentivising efficiency and improved quality. This needs to be further developed and brought to a system-wide level. It is also important to note that the SPPF emphasises the need to include private-sector service providers in the delivery of the package. This will be crucial: as Chapter 2 shows, private health providers are the primary choice for individuals seeking healthcare even if they are covered by HEF. The provider payment design will thus need to incentivise the engagement of the private sector.

Finally, institutional integration and setting up an overarching governance system for health coverage are both essential. The health sector will need to work towards institutional integration of the different insurance mechanisms in creating a single operator as intended by the SPPF. Managing this transition will not happen in isolation; it needs to be aligned with the broader social insurance integration discussed above. In practice, this would require that the single operator for the integrated health insurance system has a clear mandate and governance structure which are defined as part of the broader integration of social insurance and social protection processes to avoid overlaps and ambiguities.

Building a social protection information system based on IDPoor and CSES

Cambodia’s arrangements are currently fragmented, with numerous individual programmes operating in different contexts and locations. Little information is available regarding who is covered by these programmes, what impact these programmes have and how much they cost. This limits the transparency of the social protection sector, reduces the efficiency of programmes and will undermine efforts to create a social protection system. Nonetheless, Cambodia is well-placed to improve this situation through the IDPoor database and the Cambodia Socio-Economic Survey (CSES).

The IDPoor database is a reliable and well-established social registry that should be strengthened in support of a new social protection system. It is used by both by the RGC and development partners to assess eligibility for the majority of social protection programmes; its accuracy is reflected by the progressive beneficiary incidence of programmes such as HEF. It is also a well-known and trusted system, recognised throughout rural Cambodia due to its transparent implementation and grievance procedures at the community level. The willingness of local communities to implement and provide feedback to the system makes it unique and an example to other countries.

The SPPF envisages that the IDPoor database will evolve from a social registry to become the basis for a single registry for the poor and vulnerable which can optimise the efficiency of social protection programmes. There are a variety of steps that need to be taken before this can occur, in terms of the scope, timing and coverage of the database. However, in addressing these issues, the RGC can ensure that the IDPoor is a means of improving not only the targeting of social protection policies but also their design and impact.

The first step in expanding the role of the IDPoor would be to include information on social protection programmes, covering enrolment and benefits. It would thus become a nexus of information for the RGC to implement and monitor its social protection system. In its current form, the IDPoor questionnaire contains a scoring section – the proxy means test – as well as a basic non-scoring section capturing special circumstances causing a reduction in living standards such as illnesses, theft or reduced crop production. This section could be enhanced to cover more dimensions of living standards. However, the balance between the benefit of additional information and burden of fielding a longer questionnaire needs to be weighed carefully.

Box 4.4. Single vs social registries

When discussing the information base for social protection systems it is important to clarify between two types of registries often mentioned: social and single registries. These two types of registry differ in their functions, characteristics and feasibility.

Single registries gather information from various types of social protection programmes and can be used as a monitoring tool. They can also be merged with various databases, including programme-specific Management Information System (MIS) databases, as well as government databases on taxation or civil registration. Single registries typically include information on the number and characteristics of beneficiaries across social protection programmes and information on transfers received (amount, frequency, type). Single registries can thus facilitate the application and registration process for individuals and households.

Social registries, also known as unified targeting databases, provide information on households on the basis of which their eligibility for poverty-targeted social programmes can be established. They are thus more limited in their scope than single registries, their primary function being the targeting of social protection beneficiaries.

Social registries play an important role in climate-responsive social protection programmes by informing the authorities which households are likely to be most vulnerable in the event of a climate-related shock. This in turn ensures that emergency assistance can be directed to those most at need, sometimes before the shock has even occurred if early warning systems are in place.

Source: Chirchir and Farooq (2016), “Single Registries and Social Registries: clarifying the terminological confusion”, Pathways’ Perspectives, Issue No. 23.

This dilemma can be resolved in part by linking IDPoor to other databases, such as civil registration or social insurance data. This would allow policy makers to crosscheck information across various sources, thereby facilitating application processes and increasing the accuracy of information.

The IDPoor database needs to be updated more frequently to ensure data quality and accurate targeting. With survey waves carried out, on average, every three years, the system is at risk from errors of exclusion and inclusion, especially given the extensive transitioning in and out of poverty and the high levels of internal migration. Additionally, the SPPF specifically targets vulnerable groups such as pregnant women and children under five, who would need to be accurately registered to become beneficiaries. Increasing the frequency of survey waves will make IDPoor a more reliable source of information on the current poverty status of households.

On-demand registration can mitigate low survey frequency. The incorporation within the IDPoor database of households designated as poor through HEF’s post-identification mechanism is a compromise to improve accuracy until an on-demand identification process can be implemented nationwide. The current post-identification mechanism does not match the quality of the IDPoor process due to its simpler and non-participatory procedure. Additionally, the Ministry of Planning (MoP) is currently evaluating various options aimed at allowing greater flexibility in on-demand identification, including electronic data collection through commune councils at district level.

IDPoor’s focus on rural areas has restricted access to social programmes; achieving national coverage of the database is an integral part of scaling up social protection itself. With the discrepancy between rural and urban poverty rates decreasing significantly in recent years, the extension of the IDPoor system to urban areas, piloted in 2016, would be an important step towards improving access to social protection for the urban poor. Rolling out the IDPoor database across the entire country will not only improve social protection policies but can also support decentralisation reforms by providing information to local governments to adjust policies to the needs and characteristics of their vulnerable populations.

At the same time as the IDPoor system is transitioning towards a single registry, the CSES could become a high-level monitoring and evaluation tool for social protection. The CSES is nationally representative and conducted annually but its value for social protection planners is so far limited. The survey currently only provides household level data on uptake of broad categories of social protection programmes and the possession of an Equity Card for accessing HEF. Including a more detailed module to capture information on the recipients of social protection would allow for regular, nationally representative stock-taking.

Adding individual and programme-specific data would allow for assessment of targeting mechanisms, uptake rates and adequacy of specific benefits. This would be particularly useful given the breadth of information on individuals and households contained in the CSES, which would allow for detailed analysis of social protection in relation to individual characteristics, such as employment, assets, consumption and education. Over time, the CSES should become an important monitoring and evaluation tool for the social protection system, though it would still need to be complemented by individual programme evaluations.

Processing and delivery mechanisms

Uptake of social protection programmes globally is often hindered by potential beneficiaries not being aware of the benefits for which they might be eligible or lacking information about or access to programmes. Cambodia is particularly prone because it has very few national social protection programmes and the delivery mechanism for even the largest programmes, such as HEF, is fragmented (ILO, 2012). Improving awareness and providing counselling on eligibility could enhance the performance of the social protection system.

Portability or flexible access to social protection transfers is an important delivery issue. Households identified through IDPoor as eligible for certain benefits may not be able to access them outside their locality. This is a particular challenge in Cambodia, given the high levels of internal migration and the lack of national schemes. Alleviating this requires integration of delivery mechanisms, in particular payment mechanisms, as seen in Mexico, Brazil and Colombia (Box 4.5).

Box 4.5. Electronic benefits payment in Mexico, Brazil and Colombia

New payment mechanisms can multiply the livelihood and empowerment impact of social protection programmes while improving delivery performance, cost-efficiency and value for money. Governments have two options for benefit transfers: the traditional “pull” mechanism, whereby beneficiaries collect their benefits from fixed pay points; or “push” systems, which deliver funds directly to beneficiary bank accounts, smart cards or mobile phones. While most benefit payment systems globally employ “pull” approaches, many countries are piloting or have transitioned to “push” systems.

Mexico’s cash transfer programme, Prospera (formerly known as Oportunidades and Progresa at different points in time), established in 1997, employs an electronic payment system to deliver benefits to bank accounts. The system lowers government delivery expenses and private beneficiary costs, while improving delivery reliability. Moreover, the system deepens the programme’s developmental impact by promoting and strengthening beneficiary access to financial services. A recent study demonstrates that the payment mechanism changes beneficiary behaviour; they reduce their reliance on informal lending sources and are better able to cope with shocks using their savings rather than suffering a decline in consumption or resorting to often usurious loans (Masino and Niño-Zarazúa, 2014).

Brazil’s Bolsa Familia cash transfer programme is another successful example of electronic payment. The broad footprint of Caixa, Brazil’s state-run rural development bank, enabled a shift to bank account payments, lowering costs to beneficiaries by 5.8% and deepening beneficiary access to financial services. However, not all countries in Latin America have immediately succeeded in lowering costs using financially inclusive mechanisms. Colombia’s shift from cash payments to account-based transfers actually increased the delivery cost from USD 5.20 to USD 6.24 (Bold, Porteous and Rotman, 2012). In the absence of cost-effective banking infrastructure, a dedicated manual cash payment system may provide easier and lower-cost access for beneficiaries. However, it comes with the potential disadvantages of increased fiduciary risks, possible leakages, time-consuming verification processes and the foregone opportunities to provide financial, communications and information services that can enhance livelihood development.

A one-stop-shop embedded in government institutions that delivers multiple services (from information dissemination to enrolment and service provision) can realise efficiency gains for providers and enhance programme uptake by simplifying the process for beneficiaries. As such, a single window service facilitates access for poor households by reducing travel and time costs associated with accessing social protection services. It would also contribute to the extension of programme coverage, improve transparency and facilitate information-sharing across government institutions. By fostering horizontal and vertical integration, these services can strengthen sub-national authorities.

A Social Service Delivery Mechanism was piloted by the ILO until 2016 as a means to unify registration and benefit payments. This pilot revealed a need for decentralised case management for social protection programmes that can be performed on a rolling basis.

References

Bold, C., D. Porteous and S. Rotman (2012), “Social cash transfers and financial inclusion: Evidence from four countries”, CGAP Focus Notes, No. 77, Consultative Group to Assist the Poor, Washington, D.C., cgap.org/sites/default/files/Focus-Note-Social-Cash-Transfers-and-Financial-Inclusion-Evidence-from-Four-Countries-Feb-2012.pdf.

Chirchir, R. and S. Farooq (2016), “Single Registries and Social Registries: clarifying the terminological confusion”, Pathways’ Perspectives, Issue No. 23, November, Development Pathways, London.

Heckman, J. (2008), “The case for investing in disadvantaged young children”, in Big ideas for children: Investing in our nation’s future, First Focus, Washington, DC, pp. 49-58.

ILO (2017), “Recent trade policy developments and possible implications for Cambodia’s garment and footwear sector”, Cambodian Garment and Footwear Sector Bulletin, Iss. 5, International Labour Organization, Geneva, ilo.org/asia/publications/WCMS_541288/lang--en/index.htm.

ILO (2012), Cambodia: Toward Integrated Employment and Social Protection Policies, EU/ILO Project on Improving Social Protection and Promoting Employment, International Labour Organization, Geneva, ilo.org/wcmsp5/groups/public/---europe/---ro-geneva/---ilo-brussels/documents/publication/wcms_196019.pdf.

Masino, S. and M. Niño-Zarazúa (2014), “Social service delivery and access to financial innovation: The impact of Oportunidades’ electronic payment system in Mexico”, UNU-WIDER Working Paper 2014/034, United Nations University-World Institute for Development Economics Research, Helsinki, wider.unu.edu/publication/social-service-delivery-and-access-financial-innovation.

Missoni, E. and G. Solimano (2010), “Towards Universal Health Coverage: the Chilean experience”, World Health Report 2010 Background Paper, World Health Organization, Geneva, who.int/healthsystems/topics/financing/healthreport/4Chile.pdf.

National Bank of Cambodia (2016), “Cambodia Balance of Payments Statistics Bulletin – Series No.54, Fourth Quarter 2016”, National Bank of Cambodia, Royal Government of Cambodia, Phnom Penh, nbc.org.kh/download_files/economic_research/bop_eng/BOP-Bulletin-No-54.pdf.

NIS and MoP (2011), “Pilot Survey 2010 of 2011 Economic Census of Cambodia”, National Institute of Statistics, Ministry of Planning, Royal Government of Cambodia, Phnom Penh, nada-nis.gov.kh/index.php/catalog/28/related_materials.

OECD (2017), OECD.Stat (database), stats.oecd.org, (accessed May 2017).

Tandon, A. et al. (2016), Indonesia - Health financing system assessment: spend more, right, and better, World Bank Group, Washington, DC, documents.worldbank.org/curated/en/453091479269158106/Indonesia-Health-financing-system-assessment-spend-more-right-and-better.

THL (forthcoming), The Capacity Assessment for the Coordination and Implementation of Social Assistance Programs in Cambodia, The National Institute for Health and Welfare [THL], Helsinki.