Assessment and recommendations
Recent years have seen a heightened recognition of the potential role of social protection in the development process. Social protection now constitutes an essential component of the global agenda for sustainable development and it occupies a large place in several regional and national commitments. To a large extent, this infatuation for social protection has been fuelled by the recognition of social protection as a human right under international human rights law, as well as overwhelming evidence that investing in social protection is crucial for tackling poverty and vulnerability, and for improving job quality. Making a strong investment case for social protection during budget discussions can remain a difficult task, however. Not only the economic impact of social protection investments, beyond cash transfers, remains insufficiently documented empirically, but contrasting views still exist about the contribution of social protection to growth and equity. Some, and there are many, might see social protection investments as drivers of overall economic growth and inequality reduction. Others, in contrast, might emphasis the possible adverse effects of social protection on growth through tax distortions and changes in labour allocation and precautionary savings. Clearly, this shows the need to better document empirically the economic impact of social protection programmes and to build a more solid economic case for investing in such programmes.
This study investigates the rationale for social protection from an inclusive economic perspective and asks: Can social protection investments be an engine for inclusive growth? The study begins by laying out a methodological framework, which draws on the OECD concept of inclusive growth, disentangles social protection into social assistance and social insurance, identifies the transmission channels from social protection investments to inclusive growth, and proposes a practical way to measure empirically the impact of social protection on growth across different income groups. It then presents recent and new evidence on the impact of social protection on the micro-level drivers of inclusive growth through different stages of life. The empirical analysis is undertaken for countries at different stages of development and separately for social assistance and social insurance programmes.
This study has two main objectives. First, to contribute to fill-in important knowledge gaps as regards the impact of different types of social protection programmes on inclusive growth. Second, to create more solid economic arguments for investing in social protection that can feed budget discussions and social dialogue.
The study is intended primarily for the use of development practitioners, both national policy makers and social partners, as well as international and bilateral development partners. It draws on an in depth review of the theoretical and empirical literature, enriched by 11 new impact evaluations of social protection programmes implemented in Brazil, Ghana, Germany and Indonesia. The rationale for choosing these countries is threefold. First, their diversity in terms of development stages and geographical location. Second, the existence of well-enough established social protection systems for which an evaluation exercise could bring enough value-added from a global learning perspective. Third, the availability of recent and adequate data for conducting rigorous quantitative impact evaluations of the main national social protection programmes.
This study also acknowledges some of its limitations and calls for caution in the interpretation and the generalization of its results, due to important challenges related with the availability and quality of suitable data to measure and understand growth impacts. First, the health related drivers of inclusive growth are not included in the study because of the difficulty to adequately measure health outcomes in non-specialised household surveys. Second, while new evidence from 11 impact evaluations is provided in this report, these are based on quasi-experimental approaches and not on randomized controlled trials. Third, although design and implementation issues, as well as the level of social protection benefits, are likely to play a critical role on the observed outcomes, these are not well captured in the analysis that rely mostly on quantitative methods.
A conceptual framework to measure the impact of social protection on inclusive growth
Inclusive growth is defined by the OECD as improvement of living standards and shared prosperity across all social groups. The concept of inclusive growth has gained recognition in development circles because it has broadened the discourse beyond a focus on the extreme poor, and increasingly shifted policy focus from poverty reduction to determining how growth can be made more equitable and more inclusive. The conceptual framework developed for this study refers to the OECD definition of inclusive growth, thus recognising the importance to look at redistributive issues when assessing the economic impact of social protection investments.
Social protection may affect inclusive growth through several transmission channels
Social protection refers to policies that aim to prevent and reduce poverty, vulnerability and social exclusion throughout the life cycle. Accordingly, social protection systems often provide benefits to individuals or households in order to guarantee income security and access to health care throughout different stages of life. Besides its impact on poverty and vulnerability, social protection may also influence the quality of growth. The framework developed in this study identifies three main transmission channels through which social protection may affect inclusive growth. First, social protection can help lift credit constraints by facilitating access to bank loans and extend credit to low-income households. Second, social protection can help households cope with risks and protect their consumption and assets against adverse shocks, which leads to a more efficient use of resources. Third, social protection can also affect the allocation of resources and time use in the household, which in turn have implications for income growth.
The transmission channels may operate at the micro, meso and macro levels
One way social protection can influence inclusive growth is through its direct impact on individuals and households. At such individual and household (micro) level, a pure growth effect may be expected by: (i) enabling households to accumulate productive assets, (ii) preventing the loss of productive capital after a shock; (iii) enabling innovation and entrepreneurship, (iv) affecting labour market participation and savings and (v) supporting investments in human capital. While most of these factors are expected to have a positive impact on growth, the positive growth effect may be moderated by a possible negative growth effect of social protection induced by a decline in labour force allocation and savings, creating dependency and adverse incentives to work and save.
Such a growth effect induced by social protection investments may further interact with an effect on inequality. Social protection, especially social assistance, can indeed contribute to make the positive growth effect equalising through two main complementary paths. First, by guaranteeing a minimum level of economic and social wellbeing, serving not only as safety nets for low-income and vulnerable households and individuals to mitigate the risk of poverty, but also as spring boards that enable social mobility and help close inequality gaps. Second, by enabling equal access to opportunities, thus overcoming the savings and credit constraints among less wealthy households that can prevent human capital investments and the disruption of the cycle of inter-generational poverty.
Besides the more direct effect of social protection on inclusive growth that operates at the micro level, social protection might also affect growth and inequality outcomes at community (meso) and national (macro) levels. As regards the growth effect, at meso level, social protection investments can generate multiplier growth effects from increased local consumption and production and enable the accumulation of productive community assets. At macro level, social protection can have significant and broad growth enhancing effects on the economy by increasing aggregate household productivity, stimulating aggregate demand and thus increasing employment, in particular through counter-cyclical spending during economic downturns, and raising consumption and income tax revenues. In addition, indirect effects such as facilitating economic reforms, building human capital, enhancing social cohesion and influencing fertility can further help spur growth.
As regards the inequality effect, social protection may affect the level of inequality at meso or macro level by contributing to the provision of equal access to opportunities. Ultimately, however, such redistributive effect at meso and macro levels are likely to depend on the level of coverage, the generosity of the benefits, and the type of the programme, in particular whether it is targeted to vulnerable groups as with social assistance.
The measurement framework proposed in this study focuses on the micro- determinants of inclusive growth for which a theoretical link exists with social protection and which can be measured in non-specialised household surveys.
The conceptual framework developed in this report shows that social protection investments may affect growth and inequality through a multiplicity of effects at micro, meso and macro level. Measuring these effects is often a challenge, however. Key measurement challenges include the heterogeneity of social protection investments, the multiplicity of possible effects that may cancel each other out, the presence of endogeneity, and, for the macro effects, the scarcity of internationally comparable data on social protection investments broken down by types of programmes. For all these reasons, this report adopts a careful approach to measure the impact of social protection investments. It focuses on the micro determinants of inclusive growth for which a theoretical link exists with social protection investments and which can be measured through non-specialised household surveys. It thus looks at the more direct effects of social protection investments that can be measured in most household budget surveys.
The resultant measurement framework then identifies a number of micro determinants of inclusive growth around different stages of life – the so-called outcome variables – that can be observed with reasonably good household survey data and which are, at least in theory, likely to be influenced positively or negatively by social protection investments. The outcomes of interest for which a strong theoretical justification exists and that can usually be measured empirically typically refer to education outcomes, early pregnancy, fertility, child labour, employment outcomes, migration, consumption, and savings.
According to theoretical expectations, many micro-level effects of social protection on inclusive growth shall be positive. Other effects would a priori be unclear or negative.
Social protection may affect the micro-determinants of inclusive growth in different ways. Many of the expected micro-level effects of social protection on inclusive growth are positive. Social protection is likely to support consumption, to improve educational outcomes in financially constrained households, and to foster innovation and investments among the poor. Social protection is also expected to reduce fertility, which may affect positively inclusive growth in low-income countries where high fertility prevails. Yet, some of the effects of social protection on inclusive growth are a priori unclear or negative. Social protection can have indeed opposite – and thus a priori undetermined – effects on labour supply and migration, and is expected to alter savings patterns.
Evidence on the micro-level impact of social assistance on inclusive growth
Social assistance programmes are a key component of social protection investments that are expected to affect economic growth and equity due to their targeted benefits to the poor and their non-contributory nature. Yet, the extent to which social assistance impact on inclusive growth remains ultimately an empirical question. Recent empirical studies and new impact evaluations undertaken for five social assistance schemes implemented in Brazil, Germany, Ghana and Indonesia are analysed for different stages of life and, whenever possible, different household income deciles. Findings show that overall, social assistance seems to have a positive impact on inclusive growth mostly through its positive impact on children and youth outcomes.
Social assistance tends to spur inclusive growth largely by improving children and youth education outcomes among low-income households
Early on in the life stages, social assistance is expected to spur inclusive growth through its effect on human capital. The effect may be particularly strong among poorer children and youth given the targeted nature of social assistance.
Empirical findings seem to support the theoretical expectations as regards the impact of social assistance on education outcomes among children and youth. For targeted cash transfers, there is solid evidence that they spur investments in child schooling, and even more so when they are conditional on school attendance. New evidence from Brazil, Ghana and Indonesia also show that the strongest effect on school attendance is found for children in poor households. Another education outcome analysed in the Brazilian and Indonesian impact evaluation studies is school attainment. These studies find a positive impact on school attainment of children and youth in the poorest income group. Similar findings arise from scholarship programmes for low-income families, which tend to have positive impacts on school attainment, especially among the poorest students. In contrast to targeted transfers, Universal Child benefits appear to have no or limited aggregate effects on children’s education. New evidence from Germany is in line with previous results. This suggests that cash transfers may mainly influence the education outcomes of children and youth from disadvantaged families who may be financially constrained, while they have little effect on better-off families for which the income effect may be too small given the relatively low level of the child benefits as a proportion of family income.
Besides education outcomes, cash transfers may also reduce the burden on children to contribute to the household income, and thereby reduce child labour both within and outside households. Transfer programmes linked to or conditional on children attending school are likely to have an even stronger effect on child labour. Empirical findings show that cash transfers can decrease child labour in some cases (mostly in Latin America) but not in others (in sub-Saharan Africa), confirming that poverty may not be the sole driver of child labour. As economic opportunities increases, so can do the demand for child labour.
An additional possible effect of cash transfers on children and youth is early pregnancy. Delaying childbearing is an important factor to improve educational and health outcomes for young women, and in the longer run break intergenerational transmission of poverty. Additional income from cash transfers can reduce young women’s financial dependence on others and delay decisions on marriage and childbearing. Conditional cash transfers can also have an indirect effect on early pregnancy through its positive effect on educational attainment. Yet, empirical findings show that CCT programmes do not automatically decrease early pregnancy and that the type of conditions tied to the programme matter for the effect on early pregnancy.
The inclusive growth effect of social assistance is less apparent for the working age and elderly population
Social assistance can play an important role in ensuring income security for disadvantaged women and men of working age and the elderly, and thereby affect their behaviours in a way that can spur inclusive growth. During working age, social assistance programmes can increase consumption, affect labour and employment outcomes such as participation and intensity, but also other outcomes of the working age population such as fertility rates and entrepreneurship. During old age, social pension may impact consumption and saving patterns.
Empirical evidence shows that the impact of CCT on employment and entrepreneurship is mixed. Modest transfers do not seem to have strong impacts on employment outcomes, and when a significant impact is found, the effect may be negative or positive. CCT programmes tend also to have either a positive or no effect on investments in small businesses. New evidence for Brazil and Indonesia broken down by income groups further indicate that CCT income raises business investments only among poorer households and has no impact on investments in larger formal businesses whatever the income group.
Scholarships for the poor can also have positive spill-over effects on household consumption and investment, although there is still limited evidence. New evidence using student scholarship programme data for Indonesia show a positive impact of the programme on self-employment and consumption. Although much less documented empirically, there is also some evidence that social pension can boost consumption and investments, including investments in human capital of younger members.
Cash transfer programmes may also have effects on other household and individual outcomes, including fertility. Programmes that provide a regular cash transfers per child can encourage households to increase the size of the household to increase the amount of transfer. Concerns that cash transfers (especially unconditional) may increase fertility rates and have negative effects on population control programmes have been put forward in policy discussions in low-income countries where fertility rates tend to be high. However, existing and new empirical evidence produced in this report does not give much support to these concerns. If anything, CCT programmes do seem to reduce fertility.
Evidence on the micro-level impact of social insurance on inclusive growth
The primary objective of social insurance programmes are not to support growth and equity but, in a more pragmatic way, to protect insured persons and their dependents against a number of life contingencies through contributory mechanisms. That said social insurance may influence inclusive growth through its direct impact on a number of micro-economic channels. Findings based on recent empirical evidence enriched with six new impact evaluation of social insurance schemes implemented in Brazil, Germany, and Indonesia suggest that while the overall micro effect of social insurance on growth and inequality is more ambiguous than for social assistance, the most straightforward way social insurance may spur economic growth is by increasing consumption and, to a lower extent, by reducing fertility. The evidence base on the impact of social insurance on other outcomes remains limited, however, and further research is needed on this.
The inclusive growth impact of social insurance for children and youth is not as obvious as that of social assistance
Findings as regards the inclusive growth impact of social insurance related to children and youth outcomes are often mixed and vary across countries both within developed and developing economies. New empirical evidence produced for this report on Brazil and Indonesia confirms the mixed effects of social insurance on education. While social insurance appears to have very limited impact in Brazil, it significantly boosts educational outcomes in Indonesia, especially among less wealthy families.
Beyond educational outcomes, other important children and youth outcomes are child labour and early pregnancy, which are known to have adverse effects on inclusive growth. Few empirical studies have analysed the potential effects of social insurance benefits on such outcomes. Most existing studies have focused on social assistance programmes – social pensions and other cash transfer programme – and find mixed effect of social transfers on participation and time spent in child labour. New empirical evidence produced for this report reveals that, in the case of Brazil, old age contributory pensions do not affect the occurrence of early pregnancies but are positively associated with child labour among poorer households.
Among the working age and the elderly, social insurance tends to support inclusive growth mostly through a positive effect on consumption and a small negative effect on fertility …
Most of the evidence on the inclusive growth effect of social insurance programmes among the working age and the elderly comes from their positive impact on consumption. Although the empirical literature often provides mixed results, a number of studies supports the theoretical hypothesis that social insurance spurs consumption. Social insurance also tends to have a small negative impact on fertility, which, in the context of developing countries where high fertility prevails, may spurs economic growth. Most available studies have focused on contributory pensions systems and find a negative correlation between contributory pensions and fertility in both developed and developing countries, but the magnitude is generally found to be rather moderate. New results for Brazil and Indonesia find mixed results. While in Brazil, contributory pensions are negatively associated with fertility, in Indonesia, social insurance benefits, including the pension insurance programme, the old-age savings programme, the occupational accident benefit and the death benefit, are insignificant or have at best an effect, either positive or negative, of negligible magnitude on fertility.
Additional evidence points to a possible negative effect of social insurance on skilled emigration. A number of empirical studies suggest that social insurance and migration are negatively correlated in developing countries. Moreover, social insurance benefits may negatively affect the skill composition of migrants in that it favours migration outflows of low-skilled workers. New empirical evidence generated for this report goes a step further and question whether social insurance benefits affect return migration. New findings show that in the case of Brazil and Indonesia, households receiving social insurance benefits are more likely to have members that had a recent experience of migration, but the size of the effect is small. In other words, social insurance could be positively associated with return migration, suggesting that social insurance benefits may act as a substitute for remittances.
…but the positive impact of social insurance on inclusive growth may be moderated by a small negative effect on labour supply, and to a lower extent, on savings.
The existing empirical literature also points to a negative effect, albeit moderate, of social insurance on savings and labour supply. As regards the impact on savings, several studies have looked at contributory pensions and find that pensions tend to partly crowd out private savings, but mostly among the better off. Likewise, there is some empirical evidence suggesting that unemployment insurance negatively affects precautionary savings and leads to a corresponding increase in consumption. In line with previous findings, new empirical evidence for Brazil shows that contributory pensions do not seem to have any effect on household savings, except for some better-off families.
As regards labour supply, several studies find that unemployment benefits, both duration and income replacement rate, tend to have negative but frequently small effects on labour supply, and do not seem very effective in improving the quality of job matching. Yet, more recent studies that are able to control for the fact that a job-seekers’ opportunities and skills deteriorate with unemployment duration find that access to more generous unemployment insurance does indeed tend to help agents to find better jobs. Moreover, activation strategies through the adoption of monitoring and sanction mechanisms – job search requirements conditioning benefits receipt – by public employment services (PES) can overcome the apparent adverse employment effects of unemployment insurance.
Likewise, evidence shows that contributory pensions may have a negative impact on labour supply in developed countries that adequate pensionable ages, limited access to early retirement and actuarially fair benefit formulas could avoid. According to the new empirical evidence produced for this report, contributory pensions in Germany have a negative impact on labour supply for the elderly. The negative impact of pensions on employment gradually increases as household income increases. As regards developing countries, there is some evidence on the negative spillover effect of contributory pensions on the labour force participation of the working age population. New findings for Brazil and Indonesia show that social insurance benefits may lead to a sizeable decrease in employment, as measured by the number of employed household members of working age. This holds true for both men and women, but women could be more negatively affected than men. Profound differences are however found between Brazil and Indonesia when looking at the effect of social insurance across households with different income levels. While in Brazil better-off families may be the most concerned by the decline in employment resulting from contributory pensions, the reverse is true for social insurance benefits in Indonesia.
Making the case for social protection
In many emerging and developing countries, competing claims for scarce government funds make the case for more investments in social protection during budget discussions particularly challenging. Yet, findings from this study suggest that investing in social protection could make sense from a number of perspectives.
Argument 1: Under international human rights law, countries are legally obligated to establish social protection systems.
The most common argument put forward to make the case for social protection is a right-based argument emphasising the moral and legal basis for investing in social protection. This argument flows directly from the right to social security, which is articulated in Article 22 of the Universal Declaration of Human Rights and in Article 9 of the International Covenant on Economic, Social and Cultural Rights (ICESCR). The Sustainable Development Goals (SDGs) and other regional and national commitments embody many elements of a human rights perspective by explicitly mentioning access to social protection as a critical goal.
Argument 2: Social protection is an effective tool to reduce poverty and tackle vulnerability.
A more recent argument that is often made emphasises the effectiveness of social protection vis-à-vis its core objectives, which is to reduce poverty and tackle vulnerability. It relies on overwhelming evidence that social protection schemes can deliver real results in terms of poverty reduction and progress towards decent work, especially when design and implementation issues are carefully taken into account. Numerous evaluations around the world show positive impacts, including a reduction in the poverty gap, greater income security, and better health and education.
Argument 3: Social protection can also make good economic sense.
Another argument, one that can be particularly appealing for policy makers responsible for budget allocations, highlights the broad-based economic potential of social protection investments. Such argument is supported by the findings of this study about the impact of social assistance and social insurance programmes on the micro-level drivers of inclusive growth. It stipulates that a more solid economic case for investing in social protection can be built around two findings discussed in the report: (i) the positive inclusive growth impact of social assistance largely channelled through improved children and youth education outcomes among low income households; and (ii) the pro-growth effect of social insurance driven by increased consumption, and to a lower extent, reduced fertility, that a possible small adverse effect on labour supply and precautionary savings is unlikely to offset.
All in all, this study shows that besides right-based arguments that present the moral and legal basis for directing more resources to social protection, and more recent arguments based on the evidence that social protection can deliver real results in terms of poverty reduction and progress towards decent work, there are also good economic reasons, backed up by micro economic evidence, for investing in social protection.
Yet, there is still much to be learned. As more and better data become available to measure impacts, the quantitative measurement framework presented in this study could be used to undertake new research on the inclusive growth impact of social protection investments and enrich the evidence base discussed in this study. Such quantitative framework may also be enriched through additional qualitative assessments in order to yield important insights as to the role of design issues in the observed outcomes.